WO2002078619A2 - Alkynylated single-strand oligonucleotide and uses thereof - Google Patents

Abstract

One aspect of the invention relates to an oligonucleotide which includes : a first nucleotide including at least one alkynyl functional group at a C5 position of a pyrimidine heterocyclic base, and at least one second nucleotide covalently bound to the first nucleotide and including at least one alkynyl functional group at a C5 position of a pyrimidine heterocyclic base; wherein the oligonucleotide comprises a loss in free energy of at least 2 kcal/mol when (a) an alkynyl group of the first nucleotide is removed from the C5 position of the pyrimidine heterocyclic base and (b) the oligonucleotide is covalently or non-covalently bound to a nucleic acid molecule comprising a nucleotide sequence that is substantially Watson-Crick complementary to a sequence of the oligonucleotide.

Description

ALKYNYLATED SINGLE-STRAND OLIGONUCLEOTIDE AND USES THEREOF

This application claims the priority benefit of U.S. Patent Application Serial No. 60/268,429 filed February 13, 2001, which is hereby incorporated by reference in its entirety.

This invention was made, at least in part, with funding available through the National Institutes of Health, grant number GM22939. The U.S. government may have certain rights in this invention.

FIELD OF THE INVENTION

This invention relates to alkynylated single-stranded nucleic acid molecules, duplexes formed therewith, and use of such single-stranded nucleic acid molecules for therapeutic and diagnostic purposes.

BACKGROUND OF THE INVENTION

RNA is a dynamic component of many cellular processes. Consequently, RNA is becoming a target for therapeutics (Pearson et al., 1997) and detection by microarray (Schena et al., 1995) and molecular beacon (Leone et al., 1998) technologies. One powerful approach to targeting RNA is antisense oligonucleotides (Zamecnik et al., 1978). In principle, Watson-Crick base pairing interactions can specifically drive molecular recognition of sense RNA targets by antisense oligonucleotides. Rational design of such therapeutics and probes, however, can be improved by the discovery of new rules for molecular recognition of RNA by antisense compounds.

Studies of the thermodynamic stabilities of nucleic acid duplexes have shown that hydrogen-bonding within Watson-Crick base pairs, and heterocyclic stacking between them, stabilize nucleic acid hybridization (Burkard et al., 1999; Turner, 2000). The magnitudes of these forces vary due to interactions between 'nearest neighbors' (Borer et al., 1974; Xia et al., 1998; SantaLucia, 1998). Extensive studies have elucidated nearest neighbor parameters of Watson-Crick DNA duplexes (SantaLucia, 1998; SantaLucia et al., 1996), RNA duplexes (Xia et al, 1998), and DNA:RNA hybrid duplexes (Sugimoto et al., 1995; Gray, 1997). On the basis of these and other results, models have been constructed to predict the folding of a given sequence (Mathews et al., 1999a) and the optimal unmodified DNA or RNA antisense agents to target a given RNA strand (Mathews et al., 1999b; Walton et al., 1999).

Advances in synthetic nucleic acid chemistry have provided powerful tools to modify the backbone and heterocyclic bases of nucleic acids, including antisense oligonucleotides (Verma et al., 1998). One promising modification that increases antisense target affinity is the substitution of a 1-propynyl functionality at the C5 position of cytosine and uridine, as shown in Figure 1A (De Clercq et al., 1983; Hobbs, 1989; Froehler et al., 1992). For example, C5-(l-propynyl) substitutions on pyrimidines can increase the melting temperature of a DNA:RNA hybrid by 0.9 - 2.6 °C per modification (Froehler et al., 1992; Freier et al., 1997). C5- ( 1-propynyl) substituted pyrimidines are compatible with modifications along the phosphodiester backbone that increase chemical stability, cellular penetration, and therapeutic potency (Wagner et al., 1993).

Previously, an antisense C5-( 1-propynyl) oligodeoxyribonucleotide heptamer with six phosphorothioate backbone modifications was demonstrated to eliminate its SV40 TAg mRNA target with great potency (ICso^ 300 nM) and unexpected specificity (Wagner et al., 1996). The heptamer and its RNA target are shown in Figure IB. What is needed, however, is a systematic study of the effects of propynylation (alkynylation) on the thermodynamics of hybrid duplex formation. This will enable the rationale design of alkynylated antisense molecules for the production of effective therapeutic and diagnostic antisense nucleic acid molecules. The present invention is directed to overcoming these deficiencies in the art.

SUMMARY OF THE INVENTION

A first aspect of the present invention relates to an oligonucleotide which includes: a first nucleotide including at least one alkynyl functional group at a C5 position of a pyrimidine heterocyclic base, and at least one second nucleotide covalentiy bound to the first nucleotide and including at least one alkynyl functional group at a C5 position of a pyrimidine heterocyclic base; wherein the oligonucleotide comprises a loss in free energy of at least about 2 kcal/mol when (a) an alkynyl group of the first nucleotide is removed from the C5 position of the pyrimidine heterocyclic base and (b) the oligonucleotide is covalentiy or non-covalently bound to a nucleic acid molecule comprising a nucleotide sequence that is substantially Watson-Crick complementary to a sequence of the oligonucleotide.

A second aspect of the present invention relates to an oligonucleotide which includes: a first nucleotide including at least one alkynyl functional group at a C5 position of a pyrimidine heterocyclic base, and at least one second nucleotide covalentiy bound to the first nucleotide and including at least one alkynyl functional group at a C5 position of a pyrimidine heterocyclic base; wherein the oligonucleotide comprises a loss in free energy of at least about 2.8 kcal/mol when (a) the oligonucleotide is covalentiy or non-covalently bound to a nucleic acid molecule comprising a nucleotide sequence that is less than substantially Watson-Crick complementary to a sequence of the oligonucleotide and (b) the first nucleotide of the oligonucleotide is covalentiy or non-covalently bound to a nucleotide of the nucleic acid molecule which is not a Watson-Crick base pairing nucleotide for the first nucleotide.

A duplex which includes an oligonucleotide of the present invention is also disclosed.

A third aspect of the present invention relates to a method of designing an oligonucleotide capable of interfering with the function of a target nucleic acid molecule, which method includes: identifying a target sequence of a target nucleic acid molecule and preparing an oligonucleotide including a nucleotide sequence that is substantially Watson-Crick complementary to the target sequence, the oligonucleotide including 6 or more adjacent nucleotide bases that are alkynylated in a manner which more favorably stabilizes the interaction of the oligonucleotide with the target nucleic acid molecule as compared to a second oligonucleotide that includes the same nucleotide sequence but lacks the 6 or more adjacent bases that are alkynylated. A fourth aspect of the present invention relates to a method of interfering with the activity of a target nucleic acid molecule which includes: introducing into an in vitro or in vivo system, which includes a target nucleic acid molecule, an amount of an oligonucleotide of the present invention which is effective to bind to the target nucleic acid molecule in a manner sufficient to interfere with any activity thereof.

A fifth aspect of the present invention relates to a microarray detection device which includes: a substrate and a plurality of oligonucleotides bound to the substrate, each of the oligonucleotides comprising at least 6 nucleotide bases wherein 6 or more adjacent nucleotide bases of each are alkynylated.

A sixth aspect of the present invention relates to a method of identifying an oligonucleotide having binding affinity for a target nucleic acid molecule which includes: introducing a target nucleic acid molecule to a microarray detection device of the present invention under conditions effective for hybridization of substantially complementary sequences between the target nucleic acid molecule and the oligonucleotide; and detecting whether hybridization occurs between the target nucleic acid molecule and one or more of the plurality of oligonucleotides bound to the substrate. A seventh aspect of the present invention relates to a method of detecting the presence of a target nucleic acid molecule in a sample which includes: passing a sample over a microarray detection device according to the present invention under conditions suitable for hybridization to occur between oligonucleotides and target nucleic acid molecules and determining whether any target nucleic acid molecules hybridized to oligonucleotides during said passing.

An eighth aspect of the present invention relates to a method of detecting the localization of a target nucleic acid molecule in an in vitro or in vivo system, said method including: introducing into an in vitro or in vivo system a labeled oligonucleotide including a nucleotide sequence which is substantially complementary and specific to a nucleotide sequence of a target nucleic acid molecule and has 6 or more adjacent nucleotide bases that are alkynylated; allowing sufficient time for the labeled oligonucleotide to hybridize with the target nucleic acid molecule; and determining the location of the labeled oligonucleotide in the system, the location of the labeled oligonucleotide being the same as the location of the target nucleic acid molecule.

A ninth aspect of the present invention relates to a method of making a product, said method comprising: introducing into a reaction medium a first nucleic acid molecule having bound thereto a first molecule or compound and a second nucleic acid molecule having bound thereto a second molecule or compound, the first and second nucleic acid molecules comprising substantially complementary nucleotide sequences that hybridize in the reaction medium and at least one of the first and second nucleic acid molecules comprising at least six adjacent alkynylated bases, wherein hybridization of the first and second nucleic acid molecules brings the first molecule or compound into sufficient proximity to the second molecule or compound for the first and second molecules or compounds to form a product.

A tenth aspect of the present invention relates to a self-assembling system for preparing a product which includes: a first nucleic acid molecule including a first nucleotide sequence, the first nucleic acid molecule having bound thereon a first molecule or compound; and a second nucleic acid molecule including a second nucleotide sequence which is substantially complementary to the first nucleotide sequence, the second nucleic acid molecule having bound thereon a second molecule or compound; wherein at least one of the first and second nucleic acid molecules comprises at least two adjacent alkynylated bases, and wherein upon introduction of the first and second nucleic acid molecules into a reaction medium suitable for hybridization thereof, the first and second molecules or compounds are capable of self-assembly to form a product.

Additional aspects of the present invention will be apparent upon review of the present application, including the Examples and the appended claims. The following abbreviations are used hereinafter: C^p, C5-(l-propynyl) deoxyribocytidine; Cj, total strand concentration; EDTA, ethylenediaminetetraacetic acid; eu, entropy units (i.e. cal K^"1 mol^"1); I, inosine; IC₅₀, antisense oligomer concentration at which 50% of target's expression is inhibited after microinjection; m-DNA, a DNA containing multiple propynyl substitutions but not fully propynylated; NAED, normalized absolute elliptical difference; PODN, C5-(l- propynyl) oligodeoxyribonucleotide; RP-HPLC, reverse phase-high pressure liquid chromatography; s-DNA, a DNA containing a single propynyl substitution; s-PODN, a PODN containing a single propynyl deletion; TBE, 100 mM Tris, 90 mM boric acid and 1 mM ethylenediaminetetraacetic acid; T_m, melting temperature in Celcius; T_M, melting temperature in kelvin; U^p, C5-(l -propynyl) deoxyribouridine; Y^p, C5-(l- propynyl) substituted deoxyribopyrimidine The results described herein reveal non-nearest neighbor interactions of propynylated pyrimidines that are critically dependent on the number and position of consecutive Y^p's within a PODN strand. Without being bound thereby, a preliminary model is proposed to predict the stabilities of Y^p-containing duplexes. Long-range cooperative interactions apparently contribute 2.6 kcal/mol to the stability of the fully propynylated PODN:RNA duplex at 37 °C. Surprisingly, elimination of a single amino group in the target RNA destabilizes the PODN:RNA complex by 3.9 kcal/mol. Circular dichroism spectra indicate that single propynyl deletions within the PODN:RNA duplex affect helix geometry. The physical forces that dictate these observations are addressed herein and the impact of these highly cooperative interactions upon the potency and specificity of PODNs in fields utilizing nucleic acid-based molecular recognition are discussed.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1A illustrates the chemical structures of C5-(l -propynyl) cytosine and C5-( 1-propynyl) uracil. Figure IB illustrates base pairing in the C5-(l- propynyl) oligodeoxyribonucleotide antisense: SV40 TAg mRNA sense complex (GenBank accession no. V01380, which is hereby incoφorated by reference in its entirety). This sequence places the sense RNA target (bold) within its natural 5' and 3' flanking regions (underlined) of the SV40 TAg mRNA (Wagner, 1996).

Figure 2A illustrates representative normalized UV melting curves at 280 nm of the DNA:RNA (grey) and PODN:RNA (black) duplexes at about 10 μM strand concentration. Figure 2B illustrates representative plots of the reciprocal of melting temperature versus log concentration for the DNA:RNA (grey) and

PODN:RNA (black) hybrid duplexes. The concentration range for the PODN:RNA complex is smaller than that of the DNA:RNA complex because high concentrations of the PODN:RNA hybrid have T_M'S that are too high to measure accurately.

Figures 3A-D are plots of ΔH° vs T_M, and ΔS° vs. ln(T_M) for the 5'- dCCUCCUU-3':3'-rGGAGGAA-5' and 5'-dC^pC^pU^pC^pC^pU^pU^p-3':3'-rGGAGGAA-5' duplexes. The R² of these plots are (A) 0J2, (B) 0.73, (C) 0.77, and (D) 0.76.

Figure 4 is a graph illustrating the free energy advantage to the DNA:RNA duplex due to single C5- 1-propynyl additions along the DNA strand (grey) compared to the free energy penalty to the PODN:RNA helix due to single C5- 1 -propynyl deletions along the PODN strand (black). Note that single propynylation at CI and U7 does not affect duplex stability.

Figure 5 is a graph depicting changes in ΔG°₃₇ of duplex formation upon a G to I substitution to give r(3 GAGIAGGAAAU5 in PODN:RNA, DNA:RNA, s-PODN4:RNA, and s-PODN5:RNA duplexes.

Figures 6A-B are CD spectra at 20 °C for the (6A) single strands: 5'-dCCUCCUU-3' (grey), 5'-dC^pC^pU^pC^pC^pU^pU^p-3' (black), 5'-dCC^pU^pC^pC^pU^pU^p-3' (4), 5'-dC^pC^pUC^pC^pU^pU^p-3' (□), and 5'-dC^pC^pU^pC^pC^pU^pU-3' (0), as well as (6B) the duplexes: 5'-rCCUCCUU-3':3'-rGAGGAGGAAAU-5' (— ) 5'-dCCUCCUU- 3':3'-rGAGGAGGAAAU-5' (grey), 5'-dC^pC^pU^pC^pC^pU^pU^p-3':3'- rGAGGAGGAAAU-5 ' (black), 5'-dCC^pU^pC^pC^pU^pU^p-3 ' :3 '-rGAGGAGGAAAU-5 ' (A), S'-dCCUCCU^-S'^'-rGAGGAGGAAAU-S' (D), and 5'- dC^pC^pU^pC^pC^pU^pU-3 ' :3 -rGAGGAGGAAAU-5' ( ). Spectra were smoothed over a 5 nm window with a Savitzky-Golay filter (Press, 1992).

Figure 7 is a graph illustrating the Normalized Absolute Elliptical Differences (NAEDs) calculated between various duplexes formed with 3'- rGAGGAGGAAAU-5 ' . [θi] is for the indicated reference duplex formed with 3'- rGAGG AGGAAAU-5 ' . For example, the bar with a value of 11.9 and labeled DNA quantifies the differences between the CD spectra of the 5'-dCCUCCUU-3':3'- 3'- rGAGGAGGAAAU-5 ' and 5 '-rCCUCCUU-3 ' :3 -rGAGGAGGAAAU-5 ' duplexes.

Figure 8 is a graph depicting the non-nearest neighbor thermodynamics of selected propynyl groups. The free energy increment for propynylation at C_\, U , and U₇, to form the fully propynylated PODN:RNA hybrid (black) is compared to the corresponding increments to form m-DNAl ,2, m-PODNl ,2,3,4,5, and m-DNA6J, respectively (grey).

Figure 9 is a graph depicting representative UV melting curves for DNA:R A duplexes (A:U)-3 (grey-thick) and (G:U)-3 (grey-thin) at A₂₆₀, and PODN:R A duplexes (A:U^p)-3^p (black-thick) and (G:U^p)-3^p black-thin) duplexes at A₂₈₀.

Figure 10 illustrates the chemical structures of cytosine, uridine, C5- (1 -propynyl) cytosine, and C5-(l -propynyl) uridine and backbones discussed in Example 3. In phosphorothioate PODNs, th-PODNs, sulfur substitutes for either a pro-R or pro-S non-bridging oxygen within the phosphodiester backbone of PODNs (i.e. when Y is sulfur, Z is oxygen and vice versa).

Figure 11 is a graph depicting penalties, ΔΔG° ₇(MM)'s, for rG=C and rA=rC substitutions in the duplex,

(5'dC,C₂U₃C₄C₅U₆C₇3'):(3'rG₉A₈G₇G₆A₅G₄G₃A₂A₁A -ιU_-25'), creating dC:rC and dU:rC mismatches in DNA:RNA (white), PODN:RNA (black), and th-PODN:RNA duplexes (striped). On average, propynylation increases ΔΔG°₃ (MM) for dC:rC and dU:rC mismatches by 3.3 and 1.9 kcal/mol, respectively, when the backbone is phosphodiester. Full stereo-random phosphorothioate substitution decreases the average propynylation effect, ΔΔG° ₇(MM), for dC:rC and dU:rC mismatches to 0.7 and 0.5 kcal/mol, respectively.

Figure 12 is a graph depicting average ΔΔG°₃₇(MM)'s for dU:rG, dC:rA, dC:rC, and dU:rC at terminal and internal positions within DNA:RNA (white), PODN :RNA (black), and th-PODN:RNA (striped) duplexes.

Figure 13 is a graph illustrating the effect of full propynylation and full stereo-random phosphorothioate substitutions on ΔΔG°₃₇(MM)'s. Except for terminal dU:rC mismatches, the enhancement of destabilizing ΔΔG°₃₇(MM)'s due to propynylation (black) is greater than the reduction in destabilizing ΔΔG°₃ (MM)'s due to full stereo-random phosphorothioate substitutions in PODN:RNA duplexes (white).

Figure 14 is an illustration of a microarray detection device which includes a substrate and a plurality of oligonucleotides of the present invention that have been printed onto the substrate using standard techniques.

DETAILED DESCRIPTION OF THE INVENTION

The present invention relates generally to the preparation of modified nucleotides and their assembly into oligonucleotides in a manner which affords them to possess greater affinity and higher stability with their target (i.e., substantially Watson-Crick complementary) nucleic acid.

Modified nucleotides can be prepared using known equipment and techniques, including without limitation those techniques described by Matteucci et al. (1981); Usman et al. (1987); and Wincott et al. (1995). Commercially available DNA/RNA synthesizers can be used to carry out such protocols using commercially available reagents. Once prepared, the oligonucleotides can be separated from their solid support and purified according to standard protocols. The modified nucleotides are alkynylated, preferably with a C2 to C6 alkynyl group, most preferably a propynyl group.

For RNA purification, each RNA strand can be deblocked using ethanolic ammonia (Stawinski et al., 1988) followed by incubation in 1 M triethylammonium hydrogen fluoride, and drying. After diethyl ether extraction, oligoribonucleotides can be purified from failure sequences with 20% PAGE. Once separated using the gel matrix, the samples can be eluted with sterile water, spin filtered, and separated from a majority of the urea with a pre-packed PD-10 Sephadex column (Amersham Pharmacia Biotech). Thereafter, an irradiated cellulose ester (MWCO = 1000) Spectra/Por Dispodialyzer (Spectrum Labs Inc.), can be used to dialyze the isolated RNA strands to enhance their purity.

For DNAs and C5-( 1-propynyl) oligodeoxynucleotides (PODNs and th-PODNs), they can be purified from a bulk of the failure sequences on a Poly Pak II cartridge (Glen Research) using the recommended protocol. Oligodeoxynucleotides can be further purified by tic on a Si500F plate (J.T. Baker) with a running buffer of, e.g., n-propanol: ammonium hydroxide:water (55:35:10). Finally, the oligodeoxynucleotides can be desalted on reverse phase C-18 Sep-Pak cartridges (Waters Corp.) and lyophilized.

According to one embodiment of the present invention, an oligonucleotide includes: a first nucleotide including at least one alkynyl functional group at a C5 position of a pyrimidine heterocyclic base, and at least one second nucleotide covalentiy bound to the first nucleotide and including at least one alkynyl functional group at a C5 position of a pyrimidine heterocyclic base; wherein the oligonucleotide comprises a loss in free energy of at least about 2 kcal/mol when (a) an alkynyl group of the first nucleotide is removed from the C5 position of the pyrimidine heterocyclic base and (b) the oligonucleotide is covalentiy or non- covalently bound to a nucleic acid molecule comprising a nucleotide sequence that is substantially Watson-Crick complementary to a sequence of the oligonucleotide. According to another embodiment of the present invention, an oligonucleotide includes: a first nucleotide including at least one alkynyl functional group at a C5 position of a pyrimidine heterocyclic base, and at least one second nucleotide covalentiy bound to the first nucleotide and including at least one alkynyl functional group at a C5 position of a pyrimidine heterocyclic base; wherein the oligonucleotide comprises a loss in free energy of at least about 2.8 kcal/mol when (a) the oligonucleotide is covalentiy or non-covalently bound to a nucleic acid molecule comprising a nucleotide sequence that is less than substantially Watson-Crick complementary to a sequence of the oligonucleotide and (b) the first nucleotide of the oligonucleotide is covalentiy or non-covalently bound to a nucleotide of the nucleic acid molecule which is not a Watson-Crick base pairing nucleotide for the first nucleotide.

These oligonucleotides of the present invention are preferably directed to (i.e., substantially Watson-Crick complementary to) a target nucleic acid molecule which is other than SN40 TAg mRΝA (Wagner et al., 1996). Thus, preferred oligonucleotides of the present invention do not bind the SV40 TAg mRΝA.

The oligonucleotides of the present invention preferably contain a sequence of at least seven nucleotides which includes at least six alkynylated nucleotides of the type described above. The oligonucleotides can also be entirely alkynylated (such that all adjacent bases are alkynylated) within the sequence which is substantially Watson-Crick complementary to the target nucleic acid.

A further aspect of the present invention relates to a duplex formed with an oligonucleotide of the present invention and its target nucleic acid molecule. Duplex formation can be carried out according to known protocols, by varying the temperature and salt concentration of the hybridization medium. Other factors affecting the melting temperature include the GC content of the probe and target. Due to the increased specificity and greater stability between the oligonucleotide of the present invention and its target nucleic acid molecule (i.e., as compared to a similar oligonucleotide containing the same base sequence with unmodified nucleotides), the duplex formed using the oligonucleotides of the present invention will be stable at temperatures which would normally melt a duplex (or inhibit its formation).

Yet another aspect of the present invention relates to the design of such oligonucleotides. Basically, the process of designing the oligonucleotides includes: identifying a target sequence of a target nucleic acid molecule and preparing an oligonucleotide of the present invention (i.e., with modified bases as described above) including a nucleotide sequence that is substantially Watson-Crick complementary to the target sequence. This process can be an iterative process of adjusting the specific sequence which is targeted, as well as the number of modified bases which are adjacent to one another, in order to identify the oligonucleotide which will most favorably stabilize the interaction of the oligonucleotide with the target nucleic acid. The oligonucleotides prepared using iterative process can be analyzed to assess the free energy potential of two or more oligonucleotides of the present invention relative to their target.

Having prepared such oligonucleotides, the oligonucleotides offer a number of therapeutic and diagnostic uses. These include, without limitation, inhibiting the activity of a target nucleic acid (which can be an RNA molecule, a DNA molecule, or a natural or unnatural molecule of related structure), use in microarray detection devices by binding the oligonucleotides to a substrate such that they are available to form a duplex with their target nucleic acid, detection of pathogens or genetic diseases or disorders, and self-assembling micro- or nano-structures.

To inhibit the activity of a target nucleic acid, the inhibition can be performed in vitro for research purposes of identifying viable targets, or the inhibition can be performed in vivo for providing a therapeutic or preventative treatment of a condition which is associated with activity of a particular target nucleic acid molecule. By inhibiting the activity (i.e., expression) of the target nucleic acid molecule, it is possible to lessen the severity or altogether overcome the condition or disorder associated with the activity of the target nucleic acid molecule. For in vivo delivery, one suitable delivery vehicle which has been employed for alkynylated oligonucleotides is a cationic lipid that, when formulated with the fusogenic lipid dioleoylphosphatidyliethanolamine, greatly improves the cellular uptake properties of antisense oligodeoxynucleotides, as well as plasmid DNA. This lipid formulation, termed GS 2888 cytofectin, and its use are described in Lewis et al. (1996). GS 2888 cytofectin was reported to efficiently transfect oligodeoxynucleotides and plasmids into many cell types; use a 4- to 10-fold lower concentration of the agent as compared to the commercially available Lipofectin liposome; and be about 20-fold more effective at eliciting antisense effects in the presence of serum when compared to Lipofectin (Lewis et al., 1996).

Referring to Figure 14, in microarray detection devices, the substrate 12 (i.e., in the form of a microarray chip 10) can be provided with a plurality of oligonucleotides of the present invention, where each is directed to a different target nucleic acid molecule or each is the same. Alternatively, different oligonucleotides can be provided which bind the same target nucleic acid molecule, albeit at a different sequence of the target. The oligonucleotides can be printed onto the substrate 12 in discrete locations 14. Thus, the microarray detection device can include at least one set of oligonucleotides that hybridize to a first target nucleic acid molecule or at least two sets of oligonucleotides that hybridize, respectively, to first and second target nucleic acid molecules.

In use, the microarray of the present invention will be exposed to a target nucleic acid by introducing the target nucleic acid to the array under conditions effective for hybridization of substantially complementary sequences between the target nucleic acid molecule and the oligonucleotide. Thereafter, the array is washed to remove unhybridized nucleic acid molecules, and any hybridization between the oligonucleotide probes of the present invention and target nucleic acids is detected. This process of identifying hybridized target nucleic acid molecules can be carried out using known procedures including, without limitation, fluorescent detection assays. In addition, it is also desirable in some instances to identify the nucleotide sequence and position of alkynylated bases in an oligonucleotide that hybridized to the target nucleic acid molecule.

In accordance with one aspect of the present invention, the microarray detection procedure can be used to identify the presence of a target nucleic acid molecule specific for a particular pathogen and, thus, the presence of that pathogen in a sample. Pathogens which can be detected in accordance with the present invention include, without limitation, bacterial, viral, parasite, and fungal infectious agents. In accordance with another aspect of the present invention, the microarray detection procedure can be used to identify a genetic diseases. This can be carried out by prenatal or post-natal screening for chromosomal and genetic aberrations or for genetic diseases. Examples of detectable genetic diseases include: 21 hydroxylase deficiency, cystic fibrosis, Fragile X Syndrome, Turner Syndrome, Duchenne Muscular Dystrophy, Down Syndrome or other trisomies, heart disease, single gene diseases, HLA typing, phenylketonuria, sickle cell anemia, Tay-Sachs Disease, thalassemia, Klinefelter Syndrome, Huntington Disease, autoimmune diseases, lipidosis, obesity defects, hemophilia, inborn errors of metabolism, and diabetes.

Cancers that can be detected by the method of the present invention generally involve oncogenes, tumor suppressor genes, or genes involved in DNA amplification, replication, recombination, or repair. Examples of these include: BRCA1 gene, p53 gene, APC gene, Her2/Neu amplification, Bcr/Abl, K-ras gene, and human papillomavirus Types 16 and 18. Narious aspects of the present invention can be used to identify amplifications, large deletions, and in particular point mutations and small deletions/insertions of the above genes in the following common human cancers: leukemia, colon cancer, breast cancer, lung cancer, prostate cancer, brain tumors, central nervous system tumors, bladder tumors, melanomas, liver cancer, osteosarcoma and other bone cancers, testicular and ovarian carcinomas, head and neck tumors, and cervical neoplasms.

In the area of environmental monitoring, the present invention can be used for detection, identification, and monitoring of pathogenic and indigenous microorganisms in natural and engineered ecosystems and microcosms such as in municipal waste water purification systems and water reservoirs or in polluted areas undergoing bioremediation. It is also possible to detect plasmids containing genes that can metabolize xenobiotics, to monitor specific target microorganisms in population dynamic studies, or either to detect, identify, or monitor genetically modified microorganisms in the environment and in industrial plants. In the food and feed industry, the present invention has a wide variety of applications. For example, it can be used for identification and characterization of production organisms such as yeast for production of beer, wine, cheese, yogurt, bread, etc. Another area of use is with regard to quality control and certification of products and processes (e.g., livestock, pasteurization, and meat processing) for contaminants. Other uses include the characterization of plants, bulbs, and seeds for breeding purposes, identification of the presence of plant-specific pathogens, and detection and identification of veterinary infections. Another aspect of the present invention involves detecting the localization of a target nucleic acid molecule in an in vitro or in vivo system. Basically, this aspect of the present invention is carried out by using a labeled (i.e., fluorescent, radiolabeled, etc.) oligonucleotide of the present invention, which is introduced into an in vitro or in vivo system , allowed to hybridize with a target nucleic acid molecule, and then the location of the target nucleic acid molecule can be identified by visualizing the labeled duplex formed between the target and the oligonucleotide of the present invention.

The present invention also affords the assembly of a product between two components (i.e., molecules or compounds) which self-assemble when brought into contact with one another. Basically, this process is carried out by binding one of the self-assembling molecules or compounds to an oligonucleotide of the present invention and another of the self-assembling molecules or compounds to a nucleic acid molecule which is a target of the oligonucleotide. These two components are then introduced into a reaction medium (i.e., hybridization medium) under conditions effective for duplex formation, which results in the self-assembling molecules being brought into sufficient proximity with one another such that they self-assemble to form a desired product.

By way of example, Mirkin et al. (1996) used DNA to assemble nanoparticles into macroscopic materials. An alkane dithiol was used as a linker molecule to connect a DNA template to a nanoparticle. The thiol groups at each end of the linker molecule covalentiy attach themselves to colloidal particles to form aggregate structures. Discrete sequences of controlled length and with the appropriate surface binding functionality may be prepared in an automated fashion. In this way, the molecular recognition properties of the oligonucleotides may be used to trigger the colloidal self-assembly process. The interparticle distances and stabilities of the supramolecular structures generated by this can be controlled.

As another example, Mbindyo et al. (2001) have also demonstrated DNA-directed assembly of gold nanowires. DNA were used as templates to provide a skeleton on which gold can be bound to create a wire with increased conductance. EXAMPLES

The following examples are intended to illustrate, but by no means are intended to limit, the scope of the present invention as set forth in the appended claims.

Example 1 - Synthesis of Oligodeoxynucleotides with Multiple C5-(l-Propynyl) Pyrimidines and Analysis of Their Long-Range Cooperativity in Molecular Recognition of RNA

Oligonucleotide Synthesis & Purification:

Riboinosine phosphoramidites were purchased from Chem Genes Corporation. All other phosphoramidites and supports were purchased from Glen Research. All oligonucleotides were synthesized (Matteucci et al., 1981; Usman et al., 1987; Wincott et al., 1995) on an Applied Biosystems 392 DNA/RNA synthesizer using the manufacturer's suggested protocols.

Oligoribonucleotides were deblocked in ethanolic ammonia (1:3 v/v) for 17 h at 55 °C (Stawinski et al., 1988). After filtering away support, RNA oligomers were incubated in 1 M triethylammonium hydrogen fluoride (50 equiv) at 55 °C for 50 h. Solutions of crude products were evaporated, dissolved in water, and extracted against diethyl ether. After removing residual ether by evaporation, 5 mM aqueous ammonium acetate (pH = 7.0) was added and the oligomers were desalted on a reverse phase Sep-Pak C-l 8 cartridge (Waters Corp.). The oligomers were purified by 20% PAGE. The product was UN visualized, cut out, and eluted with sterile water containing 0.5 mM Νa₂EDTA. After separating the sample from a majority of the urea with a pre-packed PD-10 Sephadex column (Amersham Pharmacia Biotech), the samples were dialyzed in an irradiated cellulose ester (MWCO = 1000) Spectra/Por Dispodialyzer (Spectrum Labs Inc.) against 0.1 mM EDTA (pH = 7.0) and subsequently against sterile water. Samples were then lyophilized. To test purity, RNA oligomers were 5'-radiolabeled by incubating 2.4 pmol RNA and 30 pmol [γ-³²P] ATP (New England Nuclear) in 7.5 μL of 50 mM Tris (pH = 7.5), 10 mM MgCl₂, 5 mM DTT, 0.1 mM spermidine, and 0.1 mM Na₂EDTA with 7.5 units of T4 polynucleotide kinase (Gibco-BRL) at 37 °C. After 10 min, the reaction was stopped with 7.5 μL of 2X stop buffer (10 M urea, 10 mM EDTA, pH = 7.0) and applied to a 20% polyacrylamide, 8 M urea denaturing gel for electrophoresis with IX TBE running buffer. Once bromophenol blue dye reached 18 cm, the products were imaged and quantified with a Molecular Dynamics phosphorimager and quantitation software package. Purity was greater than 95%. Product identity was confirmed for all RNA strands by electrospray mass spectroscopy on an Hewlett Packard G1946 LC/MS instrument.

The 5'-trityl oligodeoxyribonucleotides and C5-(l-propynyl)-5'-trityl- oligodeoxyribonucleotides were incubated in concentrated ammonium hydroxide at 55 °C for 2 h. After removing the support by spin filtration, the crude product was applied to, and eluted from, a Poly Pak II cartridge (Glen Research) using the manufacturer's recommended protocol to purify the desired product from most of the failure sequences. The product was further purified by preparative thin layer chromatography (tic) with an n-propanol: ammonium hydroxide:water (55:35:10) running buffer. Reverse phase C-18 Sep-Pak cartridges (Waters Corp.) were used to desalt the products, which were then lyophilized. The purity of all PODNs and DNAs was greater than 95% on the basis of RP-HPLC on a SUPELCOSIL ABZ+ Plus semi- prep column (Supelco) with a gradient from 0 to 30% acetonitrile (by volume) in 100 mM aqueous triethylamine acetate (pH = 7.0) over 1 h at a flow rate of 1 mL/min. Product identity of the DNA and PODN strands was confirmed by electrospray mass spectroscopy. Likewise, the identity of over half of the sequences within the s-DNA, s-PODN, and m-DNA families were tested and confirmed.

UVMelting:

Thermodynamic parameters were measured in 1.0 M NaCl, 0.5 mM Na₂EDTA, 20 mM sodium cacodylate at a pH of 7.0. Single strand oligoribonucleotide concentrations were calculated from high temperature absorbances at 280 nm and predicted single strand extinction coefficients (Borer, 1975; Richards, 1975). Single strand DNA concentrations were determined from high temperature absorbances at 260 nm on the basis of monomer extinction coefficients (Puglisi et al., 1989). Single strand PODN concentrations were calculated from high temperature absorbances at 260 nm on the basis of monomer extinction coefficients of 3200 and 5000 1/M-cm for U^p and C^p, respectively (generously provided by M. D. Matteucci and B. C. Froehler). These were also used in conjunction with DNA monomer extinction coefficients at 260 nm to estimate the concentrations of chimera oligomers containing modified and unmodified pyrimidines.

Appropriate single strands were mixed at 1 : 1 concentration ratios, denatured for 1 min at 90 °C, and re-annealed by slow cooling to 0 °C. UN melting curves were measured at 280 nm with a heating rate of 1 °C/min on a Gilford 250 spectrophotometer. Each melting curve was fit to a non-self-complementary two-state model (Longfellow et al., 1990) with the Meltwin Software package (McDowell et al., 1996). The thermodynamic parameters were averaged over all melts of a given duplex and compared to those generated by plotting the reciprocal of the melting temperature, T_M ^"1, versus Log(Cχ/4), where C_τ is the total concentration of strands (Borer, 1974):

T_M ^"1 = (2.303R/ΔH°)Log(C_τ/4) + ΔS°/ΔH° (1 )

With only two minor exceptions (5'dC^pC^pU^pC^pC^pU^pU^pC^p3':3'rGGAGGAA5' and 5'dCCUCCUU3':3'rGAGIAGGAAAU5'), the enthalpy changes generated from the two methods of data analysis agree within 15%, consistent with two-state melting (Freier, 1986). Errors were calculated as in Xia et al. (1998) and references therein. When comparing the thermodynamic parameters of two different hybrid systems, the T_M ^"1 versus Log[Cχ/4] results were used. The fitting procedure neglects any heat capacity change associated with the reaction, and this introduces additional error in the derived ΔH°.⁴¹ Smaller errors are introduced into ΔG°, however, due to compensating errors in ΔS°.

Multiple Linear Regression and Statistical Analyses:

Using Microsoft Excel software, linear regression analysis (Seber, 1977; Νeter et al., 1985) of the measured ΔG°₃₇'s was used to determine parameters for models to predict the stabilities of Y^p-containing DΝA:RΝA duplexes. The sample data set contains only 22 duplexes. Therefore, the Student's t-test was used to determine whether each individual parameter is statistically significant (Meyer, 1975).

Circular Dichroism (CD) Measurements:

CD spectra of duplexes were measured on a Jasco J-710 spectropolarimeter in a cell with pathlength, L, of 1 cm. Data were collected at 0.1 nm intervals, at a scan speed of 10 nm/min. Sample temperatures were maintained at 20 °C by a waterbath as five scans were collected and averaged. The molar ellipticity, [θ], was calculated from the observed ellipticity, θ, and duplex concentration, c:

[θ] = θ/cL (2)

Results PODN: RNA vs. DNA: RNA Duplex Stability:

Thermodynamic parameters were measured for duplex formation between the DNA, d(5'CCUCCUU3'), or its fully propynylated analogue and an RNA 7-mer, r(3'GGAGGAA5'), that can form the same base pairs as the target sequence within the SV40 TAg mRNA (Figure IB). Figure 2 shows representative melt curves and van't Hoff plots of the DNA:RNA 7-mer and PODN:RNA 7-mer duplexes. Full propynylation has a dramatic effect on duplex stability, ΔG°₃₇(DNA:RNA 7-mer) = -7.6 kcal/mol whereas ΔG°₃₇(PODN:RNA 7-mer) = -15.3 kcal/mol (Table 1). Thus, full propynylation increases the equilibrium constant for duplex formation by more than 200,000-fold. The T_M of the PODN:RNA 7-mer duplex is 31.6 K higher than that of the unpropynylated duplex. If there is a heat capacity change associated with duplex formation, then ΔH° and ΔS° will be temperature dependent (Lewis et al., 1961):

dΔH dT = ΔC°_P (3) dΔS dln(T) = ΔC°_P (4)

This temperature dependence can skew thermodynamic comparisons of systems with very different T_M's. In order to take ΔC°_P into account when comparing the DNA:RNA 7-mer and PODN:RNA 7-mer duplexes at 37 °C, plots of ΔH° vs. T_M and ΔS° vs. ln(T_M) were generated (Figures 3A-D). The slope of each plot provides the ΔC°_P of duplex formation. The average ΔC°_p's of the DNA:RNA 7-mer and

PODN:RNA 7-mer duplexes are -0.86 and -0.44 kcal/mol-K, respectively (Table 2). On a per nucleotide basis, these values are similar to those reported for duplex formation by other nucleic acids (Petersheim et al., 1983; Freier et al., 1985; Chalikian et al., 1999a; Holbrook et al., 1999). The ΔC°_P values allow extrapolation of the ΔH°'s and ΔS°'s to any temperature. The ΔH°'s for the DNA:RNA 7-mer and PODN:RNA 7-mer duplexes at 37 °C are -54.5 and -44.1 kcal/mol, respectively (Table 2). The ΔS°'s for the DNA:RNA 7-mer and PODN:RNA 7-mer duplexes at 37 °C are -151.4 and -90.9 eu, respectively (Table 2). From these results, the ΔG° ₇ of the DNA:RNA 7-mer and PODN:RNA 7-mer duplexes are -7.5 and -15.9 kcal/mol, which correspond well with those derived from T_M ^"' vs. Log[Cχ/4] plots (Table 1). The consideration of heat capacity affects the difference in ΔG°₃₇ only marginally, even though the differences in ΔH° and ΔS° are reduced and change sign (Tables 1 & 2). When the effect of ΔC°_P is considered, the propynylated duplex has a less favorable enthalpy change and a more favorable entropic penalty for formation than the unpropynylated duplex.

The average T_m's of the DNA:RNA and PODN:RNA melts are 41.0 and 72.8 °C, respectively (Figure 2B). Thus, a much shorter extrapolation to 37 °C is required for the DNA:RNA than the PODN:RNA duplex. This difference could skew the quantitative comparison of ΔH°'s and ΔS°'s at 37 °C. To minimize this effect, thermodynamic parameters were also extrapolated to 56.9 °C, halfway between the average T_m's of each data set (Table 2). At 56.9 °C the enthalpy change for PODN:RNA 7-mer formation is also more unfavorable and the entropy change more favorable relative to DNA:RNA 7-mer formation. This suggests that any errors in the ΔH°'s and ΔS°'s, resulting from large extrapolations, are not affecting qualitative comparisons at 37 °C.

Effects of 5 ' and 3 ' Terminal Unpaired Adenosines in the RNA Strand on Hybrid Duplex Stability:

To provide an empirical measure of stacking interactions possible for rA's in a duplex containing an all-purine RNA strand and an all-pyrimidine DNA strand, duplex stabilities were measured for d(5'CCUCCUU3'):r(3'AGGAGGAA5'), d(5'CCUCCUU3'):r(3'AGGAGGAAA5'), and their fully propynylated analogues. The underlined rA's are unpaired. The results in Table 1 allow calculation of free energy increments for the unpaired nucleotides:

ΔΔG°₃₇(3 ' A) = ΔG°₃₇(DNA:RNA 8-mer) - ΔG°₃₇(DNA:RNA 7-mer) (5)

ΔΔG°₃₇(5 ' A) = ΔG°₃₇(DNA:RNA 9-mer) - ΔG°₃₇(DNA:RNA 8-mer) (6) These increments are listed in Table 3. At 37 °C, a 5' dangling rA on a DNA:RNA helix stabilizes the duplex by 0.5 kcal/mol and a 3' dangling rA stabilizes it by 1.3 kcal/mol. On a fully propynylated PODN:RNA duplex, however, they stabilize by 1.9 and 0.9 kcal/mol, respectively. Therefore, a 3' unpaired rA is more stabilizing than a 5' dangling rA in the unmodified duplex, while the reverse is true in the propynylated duplex.

Effects of 5 ' and 3 ' Terminal Unpaired Cytosines in the DNA or PODN Strand on Hybrid Duplex Stability:

To provide an empirical measure of stacking interactions possible for dC's in a duplex containing an all-purine RNA strand and an all-pyrimidine DNA strand, duplex stabilities were measured for d(5'CCCUCCUU3'):r(3'GGAGGAA5'), d(5'CCUCCUUC3'):r(3'GGAGGAA5') and their fully propynylated analogues (Table 1). The underlined dC's are unpaired. Table 3 lists the free energy increments calculated by analogy to eq 5. An unmodified deoxycytidine stacking on the 5' or 3' end of the unmodified DNA:RNA helix stabilizes duplex formation by 0.7 and 0.2 kcal/mol, respectively, while a dC^p stacking on the 5' or 3' end of the fully modified PODN:RNA helix stabilizes by 0.3 and 0.7 kcal/mol, respectively. Thus, effects of propynylation on stacking of dC are relatively small, and the dangling end providing more stabilization appears to switch from 5' to 3'. Effects of Adding More Unpaired Dangling Nucleotides to the 5 ' and 3 ' ends of the RNA Strand in the Hybrid Duplex:

The RNA 11-mer, r(3'GAGGAGGAAAU5'), was selected as the RNA strand for additional experiments because it places the target sequence within its naturally occurring flanking nucleotides of the SV40 TAg mRNA (Figure 1) and because it has been used as the mimic in previous studies (Flanagan, 1999). Addition of the 5' terminal U and the 3' terminal G unpaired nucleotides to the duplex has little effect on stability. The ΔG°₃₇(PODN:RNA 9-mer) and ΔG°₃₇(PODN:RNA 11-mer) are -18.2 and -18.4 kcal/mol, respectively, and ΔG°₃₇(DNA:RNA 9-mer) and ΔG°₃₇(DNA:RNA 11-mer) are -9.4 and -9.6 kcal/mol, respectively (Tables 1 and 4). Evidently, the second unpaired nucleotide on each end provides negligible stacking interactions. Contributions of Single Propynyl Groups to Hybrid Stability in an Otherwise Unmodified DNA Strand:

The thermodynamic advantage of propynyl functionalities was elucidated further by single substitutions in d(5'CCUCCUU3') (Table 4). These singly substituted strands are referred to as s-DNAn oligomers, where n is an integer denoting the site of propynyl substitution. The thermodynamic parameters of the unmodified DNA:RNA 11-mer duplex are subtracted from those of the s-DNAn:RNA 11-mer duplexes to calculate the thermodynamic advantage of each propynyl substitution. For example, ΔΔG°₃₇(C₄C^P ₅U₆) = ΔG°₃₇(s-DNA5 :RNA 11 -mer) -

ΔG°₃₇(DNA:RNA 11 -mer) = - 0.6 kcal/mol (7)

Figure 4 summarizes the changes in free energy for these substitutions at 37 °C. The thermodynamic advantage of these single propynyl substitutions ranges from 0.0 to 1.0 kcal/mol. Substitutions are more stabilizing toward the middle of the helix.

Stability Decrements due to Deletion of Single Propynyl Groups from an Otherwise

Fully Propynylated DNA Strand:

The contributions of individual propynyl groups to the stability of the all-propynylated PODN:RNA 11-mer duplex were measured by systematically removing single propynyl groups. These strands are referred to as s-PODNn strands, where n is an integer denoting the position of propynyl deletion. Table 4 lists these thermodynamic parameters and Figure 4 summarizes these penalties as calculated from equations equivalent to:

ΔΔG°₃₇(C₄ ^PC₅U₆ ^P) = ΔG°₃₇(s-PODN5:RNA 11-mer) -

ΔG°₃₇(PODN:RNA 11-mer) = + 3.7 kcal/mol. (8)

Thus, a single propynyl group at dC5 contributes 3.7 kcal/mol to PODN:RNA 11-mer duplex stability, but only 0.6 kcal/mol to s-DNA5:RNA 11-mer duplex stability. On average, a single internal propynyl group stabilizes the PODN:RNA hybrid by 3.4 kcal/mol, but stabilizes s-DNAn:RNA 11-mer duplexes by only 0.5 kcal/mol. Deletions towards the 3' end of the PODN destabilize the PODN:RNA 11-mer duplex less than those towards the 5' end. Testing Base Pairing and Nearest Neighbor Models for Predicting Stabilities ofΨ- Containing Hybrid Duplexes:

Stabilities of DNA:DNA duplexes have been fit to a nearest neighbor model (SantaLucia et al., 1998), but the dependence of stability on the number of GC pairs suggests stability can be roughly predicted with only a base pairing model such as that used for DNA polymers (Wetmur et al., 1991; Blake et al., 1999). In contrast, stabilities of RNA:RNA duplexes can only be predicted well with a nearest neighbor model (Xia et al., 1998). The results summarized in Figure 4 show that the stability increment from a single propynyl group is dependent on whether a fully propynylated duplex is formed. Thus, a simple base pairing model cannot predict stabilities of Y^P- containing duplexes.

To test the applicability of a nearest neighbor model to propynylated helices, analogues of d(5'CCUCCUU3') were synthesized with multiple propynyl substitutions. These are referred to as m-DNAn,o,p's, where n, o, and p denote the positions of propynyl substitutions. The thermodynamic parameters for binding these m-DNAn,o,p strands to r(3 ' G AGGAGG AAAU5 ') are listed in Table 5. Comparisons of these thermodynamics indicate that a nearest neighbor model is also inadequate for predicting stabilities of propynylated hybrid duplexes. For example, the only difference between d(5'C^pC^pU^pC^pC^pUU3*) and d(5'C^pC^pUC^pC^pUU3') is a propynyl deletion at U₃. Comparing ΔG°₃₇[m-DNAl ,2,3,4,5] with ΔG°₃₇[m-DNAl ,2,4,5] shows that removing the propynyl group at U₃ destabilizes the m-DNAl, 2,3,4,5 :RNA 11-mer duplex by 0.8 kcal/mol. In contrast, removing the propynyl group at U₃ destabilizes the fully propynylated PODN:RNA 11-mer duplex by 4.0 kcal/mol, even though the change in base pairing and nearest neighbors is exactly the same as for the m-DNAl,2,3,4,5:RNA 11-mer duplex. Thus, a nearest neighbor model is not sufficient to explain effects of propynylation on duplex stability.

Effect of Removing an Amino Group from a Single Guanosine:

An inosine substitution for G₆ to give r(3'G₉A₈G G₆A₅G G₃A₂AιA- ιU_-23') replaces the amino group of G with a hydrogen and therefore eliminates any G₆(amino)-C₂(carbonyl) hydrogen bond within a duplex. Thermodynamic parameters of duplexes with the inosine substitution are listed in Table 6. Comparison with Table 4 shows that the single G to I substitution reduces stability at 37 °C by 3.9 kcal/mol for the PODN:RNA duplex, but by only 1.1, 1.0, and 1.2 kcal/mol, respectively, for the DNA:RNA, s-PODN4:RNA, and s-PODN5:RNA duplexes (Figure 5). Thus, the stability of the PODN:RNA duplex is very sensitive to the presence of the amino group on G and presumably its hydrogen bond, but removal of even a single propynyl group two or three base pairs away dramatically reduces this sensitivity. Circular Dichrolsm:

Figure 6 A shows CD spectra of the DNA and PODN single strands. The CD spectrum of the DNA single strand has a large positive band at 270 nm. In contrast, the CD spectrum of the PODN has a large positive band at 245 nm and a larger positive band at 215 nm. Spectral differences can be quantified in a single number, the normalized absolute ellipticity difference (NAED), defined as:

NAED = 100(∑^λ I [θ], - [θ]₂ |_λ)/(∑^λ I [θ]ι| + |[θ]₂ |_λ) (9)

Here, λ is the wavelength at which the molar ellipticities for systems 1 and 2, [θ]ι and [θ]₂, were measured. A large NAED reveals dissimilarity in the CD spectra. The average standard error of NAED comparisons is given by the NAED between two CD spectra collected on the same system at different times. This was performed on each system and averaged. The average error is 5.2, so any NAED > 5.2 is considered significant. The NAED between the DNA and PODN single strands is 81.9, which is very high as expected by inspection of the CD spectra in Figure 6A. The NAED between the DNA:RNA 11 -mer and PODN:RNA 11 -mer duplexes is 30.0, indicating there are also significant differences between the CD spectra of these hybrids. Substantial spectral shifts in the absorbance spectra of the DNA and PODN single strands presumably contribute to the NAED.

A CD spectrum of the RNA:RNA duplex, r(5'CCUCCUU3'):r(3'GAGGAGGAAAU5'), was obtained as a representative of A- form helix geometry. It has a positive and a negative band at 270 and 205 nm, respectively (Figure 6B). These two bands generally distinguish the A-form RNA:RNA helix from the B-form DNA:DNA helix (Gray et al., 1978; Gray et al., 1981). The CD spectrum of the DNA:RNA 11-mer duplex is similar to that of the A- form helix, NAEDRNA = 11.9, consistent with the geometry of pyrimidine-rich

DNA:purine-rich RNA duplexes being similar to A-form (Salazar et al., 1993; Hung et al., 1994; Ratmeyer et al., 1994; Lesnik et al., 1995; Gyi et al., 1996). The CD spectra of d(5'C^pC^pUC^pC^pU^pU^p3') and d(5'C^pC^pU^pC^pC^pU^pU3') are similar to that of the PODN single strand (Figure 6A), NAED_PODN = 1 .6 and 11.4, respectively. The CD spectrum of the d(5'CC^pU^pC^pC^pU^pU^p3') single strand, however, is relatively different, NAED_PODN = 33.0. If the single-stranded PODN is structured, the effects of propynyl deletion on this structure may be position- dependent.

CD spectra were measured for various other duplexes to test the impact of single propynyl deletions on the global helix geometry of the PODN:RNA 11-mer duplex (Figure 6B). CD spectra of the s-PODN3:RNA 11-mer and s-PODN7:RNA 11 -mer duplexes are similar to that of the PODN:RNA 11 -mer duplex, NAED = 17.2 and 1 1.6, respectively. This suggests that at least the s-PODN7:RNA and PODN:RNA duplexes have similar global geometries. The CD spectrum of the s- PODN1 :RNA 11-mer duplex d(5'CC^pU^pC^pC^pU^pU^p3'):r(3^,GAGGAGGAAAU5') is considerably different from that of the PODN:RNA 11-mer duplex, NAED = 31.6 (Figures 6 & 7). This suggests that a single propynyl deletion from a dC^p at the 5' end can have a significant effect on global helix geometry.

A Model For Predicting Stabilities of ^'Ψ '-Containing DNA: RNA Duplexes:

Neither a base pairing nor a nearest neighbor model is sufficient to account for the effects of propynylation on the stabilities of DNA:RNA duplexes. The following model, however, largely accounts for the observed effects of propynylation:

(10) ΔG°,(Y -containing duplex) - ΔG°,(unmodified duplex) = ΔΔG°, =

X_ιΔG°₃₇(int Y^p bonus) + y,ΔG°₃ (5 'dangling end bonus) + z,ΔG°₃ (cooperativity bonus)

where x„ y„ and z, are the number of times that a particular motif occurs within duplex i. The data for 21 Y^p-containing duplexes and the unmodified duplex, d(5'CCUCCUU3'):r(3^,GAGGAGGAAAU5'), were fit to eq 10 by multiple linear regression. Therefore, the model has f = 22 - 3 - 1 = 18 degrees of freedom. Although the y-intercept of eq 10 should be zero because the unmodified DNA:RNA duplex is included in the data set, it was not forced to zero in order to allow for small errors (± 0.5 kcal/mol) in the ΔG°₃₇ of DNA:RNA duplexes. Overall, the model explains the data very well, R² = 0.99 and rmsd = 0.29 kcal/mol. One parameter, ΔG° ₇(int Y^p bonus), accounts for the increased stability of duplexes containing non-terminal propynylated pyrimidines. Multiple linear regression analysis estimates that each internal Y^p stabilizes a DNA:RNA duplex by 0.99 + 0.06 kcal/mol at 37 °C. This parameter has a t-statistic of 3.11 x 10^"12, indicating that it is statistically significant from zero.

The second parameter, ΔG°₃₇(5 'dangling end bonus), accounts for enhanced stacking interactions of a 5' unpaired adenosine on the RNA strand. This enhanced stability is only applied to duplexes containing: (1) A Y^p at the 3' end of the DNA strand, and (2) propynylation of at least five of the remaining six pyrimidines in the DNA strand. Multiple linear regression analysis estimates that this 5' dangling end enhancement stabilizes a PODN:RNA duplex by 1.17 + 0.21 kcal/mol at 37 °C. This parameter has a t-statistic of 1.67 X 10^"5, indicating that it is statistically significant from zero.

The third parameter, ΔG⁰ ₃ (cooperativity bonus), is used to account for the observations that a few duplexes with at least six Y^p's possess unusually enhanced- stability. More specifically, duplexes with the PODN(6-mer), S-PODN7, or S-PODN6 strands are unusually stable. Interestingly, the s-PODNl :RNA 11-mer duplex, which has a very unusual CD spectrum, is not unusually stable. Most propynyl deletions eliminate the long-range cooperative interactions that occur between consecutive Y^p's, but this ability seems dependent upon the number of deletions and the end (5' or 3 ') of the DNA strand at which they occur. Multiple linear regression gives ΔG⁰ ₃ (cooperativity bonus) = -1.94 + 0.23 kcal/mol at 37 °C. This parameter has a t- statistic of 5.09 X 10^" , indicating that it is statistically significant from zero.

Discussion To target RNA in vivo, antisense oligonucleotides must be modified to optimize cellular penetration, half-life, target affinity, target specificity, and other properties (Milligan et al., 1993; Agrawal et al., 1997; Branch, 1998; Crooke, 2000). Design of self-assembling nanostructures based on nucleic acid-like compounds relies on knowledge of sequence specific affinities (Seeman, 1998). Rational optimization of affinity and specificity requires knowledge of the interactions important for nucleic acid associations. Previous work has shown that propynylation of pyrimidines increases duplex stability (Froehler et al., 1992; Freier et al., 1997). Moreover, the propynylated heptamer, d(5'C^pC^pU^pC^pC^pU^pU^p3'), is able to specifically inhibit translation of the SN40 large T antigen in cell culture (Wagner et al., 1996). As shown in Table 1, the propynyl groups on d(5'C^pC^pU^pC^pC^pU^pU^p3') increase the stability of its duplex with r(3'GGAGGAA5') by 7.7 kcal mol. Here, we investigate the sources of this stability enhancement in order to reveal new principles for the design of compounds relying on molecular recognition of nucleic acids.

Propynylation leads to a less favorable enthalpy change and a more favorable entropy change for duplex formation at 37 °C:

After accounting for changes in heat capacity, the enthalpy and entropy changes for duplex formation at 37 °C can be compared (Table 2). The ΔH°₃₇ for the PODΝ:RΝA 7-mer duplex is 10.4 kcal/mol less stabilizing than that of the DNA:RNA 7-mer duplex. The ΔS°₃₇ change is less destabilizing to the PODN:RNA 7-mer duplex by 60.5 eu, corresponding to 310.15(60.5)/l 000 = 18.8 kcal/mol in more favorable free energy at 37 °C. This suggests that the enhanced stability of the fully propynylated 7-mer duplex at 37 °C is due to physical phenomena that affect entropy, such as solvation and/or preorganization within the unpaired DNA single strand.

Stacking Interactions of Terminal Unpaired Adenosines:

Stacking is one sequence dependent interaction that contributes to double helix stability (Turner, 2000). Comparisons of duplex stability in the presence and absence of unpaired terminal nucleotides provide one measure of stacking interactions (Petersheim et al., 1983; Freier et al., 1985; Turner et al., 1988).

Moreover, most RNA target sequences lie within very long RNA strands. Therefore, antisense molecular recognition of an RNA target will involve both 5' and 3' dangling ribonucleotides that can stabilize the double helix.

In the SV40 TAg mRNA antisense:sense duplex in Figure IB (Wagner et al., 1996), d(5'CCUCCUU3^,):r(3'GAGGAGGAAAU5'). rA _Λ can 5' stack upon the rAι-dU₇ base pair and rA₈ can 3' stack upon the rG₇-dCι base pair. The thermodynamic contributions of adenosines stacking on terminal base pairs depends on helix geometry (Table 3). In B-form DNA:DNA helices, stacking interactions for the equivalent sequences favor duplex formation at 37 °C by 0.5 and 0.4 kcal/mol, respectively (Bommarito et al., 2000). In A-form RNA:RNA helices, the corresponding values are 0.3 and 1.1 kcal/mol (Turner, 2000; Freier et al., 1985; Turner et al., 1988). For the unmodified d(5'CCUCCUU3^,):r(3'GAGGAGGAAAU5') duplex studied here, the 5' and 3' dangling rA stacking interactions stabilize the duplex by 0.5 and 1.3 kcal/mol, respectively (Table 3). Evidently, stacking of unpaired adenosines at the ends of this DNA:RNA duplex is similar to stacking at the ends of an A-form RNA:RN A duplex. The 3' rA₈ dangling end stacking on the rG -dCι base pair stabilizes the propynylated duplex, d(5^,C^pC^pU^pC^pC^pU^pU^p3'):r(3^,GAGGAGGAAAU5'), by 0.9 kcal/mol (Table 3). In contrast, the free energy increment of the rA__! 5' dangling end stacking on the rAι-dU^p ₇ base pair stabilizes this PODN:RNA 11-mer duplex by 1.9 kcal/mol (Table 3). Thus, stabilization by the 3' dangling end rA₈ is similar to that observed with A-form helices, but stabilization by the 5' end rA _ι is more favorable than previously observed for equivalent unmodified sequences in either A or B-form helices (Table 3). In fact, similar stabilization of a DNA:DNA duplex requires a 5' unpaired dangling 5-nitroindole or pyrene nucleotide, which stabilize by 1 J kcal/mol (Guckian et al., 2000). These results suggest that a structural change is induced by propynylation and the surface area of face-to-face base stacking at the 5' end may be increased by propynylation.

Stacking Interactions of Terminal Unpaired Cytosines and C5-(l -Propynyl) Cytosines:

An unpaired dC at the 5' end of the rG₇-dCι base pair stabilizes the unmodified DNA:RNA duplex by 0.7 kcal/mol, while an unpaired dC at the 3' end of the rAι-dU₇ base pair stabilizes it by 0.2 kcal/mol. As shown in Table 3, these increments are within experimental error of those for A-form RNA:RNA duplexes (0.3 and 0.1 kcal/mol, respectively) (Turner et al., 1988) and B-form DNA:DNA duplexes (0.5 and 0.2 kcal/mol, respectively) (Bommarito et al., 2000; Guckian et al., 2000). Similarly, unpaired dC^p at the 5' end of the rG₇-dC_! ^P base pair stabilizes the modified PODN:RNA duplex by 0.3 kcal/mol, and an unpaired dC^p at the 3' end of the rAι-dU ^p stabilizes it by 0.7 kcal/mol. Evidently, propynylation does not substantially affect stacking interactions of the cytosines.

The DNA strand within an unmodified DNA: RNA hybrid duplex typically adapts to the RNA strand, progressing toward a predominantly A-form geometry as the purine content within the RNA strand increases (Salazar et al., 1993; Hung et al., 1994; Ratmeyer et al., 1994; Lesnik et al., 1995; Gyi et al., 1996). The stacking increments in Table 3 for the unmodified DNA:RNA hybrid are consistent with such observations. The presence of propynyl groups along the major groove of the PODN: RNA helix, however, changes the relative importance of 5' and 3' stacking on both the RNA and DNA strands, suggesting a change in helix geometry.

Effects of Single Propynyl Substitutions and Deletions: The average free energy advantage of single internal Y^p's within the

DNA:RNA 11-mer hybrid is 0.5 kcal/mol at 37 °C (Figure 4). This is much less stabilizing than the average advantage of 3.4 kcal/mol obtained by adding a single internal propynyl that results in a fully modified PODN:RNA 11-mer duplex (Figure 4). There is no free energy advantage for adding a single propyne at either end of the DNA:RNA 11-mer helix, but adding a propyne at the 5' or 3' end of an otherwise fully propynylated strand provides a free energy advantage of 3.2 or 2.0 kcal/mol, respectively. Thus, the effects of propynylation can not be explained by a simple base pairing model.

Preorganization of the unpaired single strand could explain the 0.5 kcal/mol advantage observed for adding a single internal propyne to an otherwise unmodified strand. Preorganization can arise from steric interactions of a bulky propyne within the single strand. For example, if the number of available conformations is reduced two-fold, then the entropic penalty is more favorable to the free energy of duplex formation by RTln2 = 0.4 kcal/mol at 37 °C. The lack of either a 5' or a 3' neighbor at the ends of the strand could explain the negligible effect of single propynylation observed at Ci and U .

Long-Range Cooperativity of Multiple Propynyls:

The free energy increments associated with propynylating a single pyrimidine are always greater when the fully propynylated PODN:RNA duplex is formed (Figures 4 & 8). For example, propynylation of U₃ to give d(5'C^pC^pU^pC^pC^pUU3') stabilizes the duplex by only 0.8 kcal/mol at 37 °C.

Propynylation of U₃ to give d(5'C^pC^pU^pC^pC^pU^pU^p3'), however, stabilizes the duplex by 4.0 kcal/mol. Similarly, propynylation of C, to gived(5'C^pC^pUCCUU3') or d(5'C^pC^pU^pC^pC^pU^pU^p3') stabilizes the duplex by 0.4 and 3.2 kcal/mol, respectively (Figure 8). On average, insertion of a single propynyl group to give the PODN:RNA

11-mer is 2.3 kcal/mol more stabilizing than insertion at the same position in a m- DNA:RNA 11-mer duplex when the fully propynylated duplex is not formed. This reveals highly cooperative long-range interactions between Y^p's.

A Model for Predicting the Stabilities of Ψ -Containing Duplexes:

The above results, combined with multiple linear regression, provide a preliminary model for approximate prediction of the enhanced stabilities of Y^p- containing hybrid duplexes. Three parameters are apparently able to account largely for the enhanced stability of such duplexes over those with no propynylation (see eq 10). Although the forces underlying these parameters are not well understood, the model predicts stabilities of 22 hybrids within 0.65 kcal/mol at 37 °C, or better than a factor of 3 in the equilibrium constant for duplex formation. In principle, the model can be combined with nearest neighbor parameters for unmodified DNA:RNA sequences (Sugimoto et al., 1985; Gray, 1997) to predict stabilities of other propynylated hybrids.

The first parameter in eq 10, ΔG°₃₇(intY^p bonus), provides about 1.0 kcal/mol in enhanced stability for each internal Y^p. This may be due to preorganization of the PODN single strand and/or enhanced interstrand stacking interactions of ribo-purine nucleotides promoted by propynyl groups (e.g. as observed for a 5' terminal unpaired rA). Bases within the confines of a duplex, however, may not have enough conformational freedom to fully optimize stacking, so the stacking effect may not be as large as observed for dangling ends. Enhanced stacking of Y^p's is unlikely since little or no enhancement is observed for unpaired 5' and 3' terminal unpaired propynylated cytosines (Tables 1 and 3).

The second parameter, ΔG°₃₇(5 'dangling end bonus), provides about 1.2 kcal/mol in enhanced stability if a 5' dangling end rA is adjacent to a terminal propynylated base pair in a hybrid with at least six propynyl groups. This value corresponds well with the measured thermodynamic advantage of 1.9 - 0.5 = 1.4 kcal/mol for the 5' rA _ι stacking interaction on the rAι:dU^p base pair in the PODN:RNA 9-mer duplex (Tables 1 and 3). The magnitude of this parameter will probably be similar for other purines, but different for pyrimidines at the 5' end of the RNA strand, as observed for DNA:DNA (Bommarito et al., 2000; Guckian et al.,

2000) and RNA:RNA (Turner, 2000; Freier et al., 1985; Turner et al., 1988) duplexes. Note that a parameter for 3' dangling end stacking on the PODN:RNA duplex is not included in the model for stability enhancement because such interactions are relatively insensitive to propynyl substitutions (Tables 1 and 3).

The third parameter, ΔG°₃₇(cooperativity bonus), is estimated to provide about 1.9 kcal/mol in enhanced stability. This parameter accounts for the additional enhanced stability of hybrid duplexes formed by the PODN(6-mer), s- PODN6, and S-PODN7 strands. These duplexes apparently have a common feature that is not accounted for by the first and second parameters. Not enough data are available to provide general rules for this parameter. For the data set, however, cooperativity is observed for helices having at least six propynyl groups, with at least five occurring consecutively and no CC^p or C^PC interfaces.

It is also possible that cooperativity is dependent on the side (5' or 3') of the DNA strand containing interruptions, rather than or in addition to the sequence at interfaces. Cooperativity is observed when an unmodified U is the penultimate (s- PODN6) or terminal 3' nucleotide (s-PODN7), but not when an unmodified C is the terminal 5' nucleotide (s-PODNl). Such an effect could be driven by the very favorable 5' interstrand stacking of the rA.i (Figure 1 & Table 3) discussed previously, which could help maintain a duplex geometry that favors cooperative interactions. The observation of an unusual CD spectrum for the s-PODNl :RNA duplex, where only d is not propynylated, is consistent with the hypothesis that cooperativity is dependent upon helix geometry.

With these parameters, the cooperativity model predicts the free energy of the PODN:RNA 11-mer duplex to be approximately -17.7 kcal/mol, which is 0.7 kcal/mol less stable than measured (Table 4). This suggests that interactions responsible for long-range cooperativity may strengthen as the number of consecutive Y^p's increase within a propynylated DNA strand. In this case, the cooperativity increment grows to -2.6 kcal/mol when seven consecutive Y^p's occur in the DNA strand. While considerable effort will be required to fully elucidate the sequence and/or length dependence of cooperativity, it is clearly an important effect within propynylated oligonucleotides. The Enhanced Stability due to Propynylation is Greatly Reduced when the Amino Group on a Single Guanosine is Replaced by Hydrogen:

To test the sensitivity of the cooperative interactions to helix composition, inosine was substituted for G₆ to give d(5'CCUCCUU3'):r(3'GAGIAGGAAAU5'). Inosine substitution for G in a G-C pair typically results in the loss of 0.5 to 1.8 kcal/mol in the free energy of RNA:RNA and DNA:DNA duplexes (Turner et al., 1987; Martin et al., 1985; Aboul-ela et al., 1985; Kawase et al., 1986). This increment has been assigned to hydrogen bonding of the amino group because essentially equivalent free energy increments are provided by stacking of an unpaired G or I at the end of a helix (Turner et al., 1987). Moreover, G and I have similar charge distributions (Burkard et al., 2000). Theoretical support for attributing G to I free energy increments to hydrogen bonding is also provided by molecular modeling of nucleic acids (Stofer et al., 1999). Inosine substitution in the DNA:RNA 11-mer duplex makes hybridization less favorable by 1.7 kcal/mol at 37 °C, consistent with previous values for G to I substitutions. The same substitution, however, destabilizes the PODN:RNA 11-mer duplex by 3.9 kcal/mol. Apparently, this amino group is 2.2 kcal/mol more stabilizing in the PODN:RNA 11-mer duplex than in the DNA:RNA 11-mer duplex. This difference is similar to the calculated stability contribution of 2.6 kcal/mol for long-range cooperative interactions within the PODN:RNA 11-mer duplex, suggesting that minor groove hydrogen bonds play a role in maintaining such interactions. Alternatively, a different conformation for the PODN:RNA hybrid may prevent hydration of the C^p ₂ carbonyl left unpaired by removal of the G₆ amino group. In order to test if the large effect of amino replacement depends on long-range cooperative interactions due to seven consecutive propynyl groups, the propynyls at C₄ and C₅ were separately removed and the thermodynamic effects of inosine substitution for G₆ were measured. When C₄ or C₅ are the only unmodified pyrimidines in the strand, the G to I substitution reduces duplex stability by only 1.0 and 1.2 kcal/mol, respectively (Figure 5). Thus, the increment is made less favorable by about 2.8 kcal/mol when long range cooperativity is eliminated by removing a propynyl group. Again, this difference is similar to the 2.6 kcal/mol assigned to long- range cooperativity in the PODN:RNA 11-mer duplex. Evidently, the contribution of the G₆ amino group to duplex stability is coupled strongly to the cooperative interactions between consecutive propynyl groups on the complementary DNA strand. Apparently, such interactions change the environment of the minor groove, in addition to enhancing the overall stability of the hybrid duplex. Effects of Single Propynyl Deletions on Helical Geometry of the PODNRNA Duplex:

The CD spectra of DNA and PODN single strands are dramatically different, NAED = 81.9 (Figures 6A). This difference is at least partially due to different optical properties of modified and unmodified pyrimidines, but may also indicate that the structures of the unpaired single strands are altered due to propynylation. CD spectra also suggest that the impact of propynyl group deletions on the helical geometry of the PODN:RNA duplex depends on the location of a deletion. The s-PODN7:RNA duplex is missing a propynyl group at the 3' end of the DNA strand and it's CD spectrum is similar to that of the PODN:RNA duplex, NAED_PODN = 11.6 (Figure 7). The s-PODN3:RNA duplex is missing a propynyl group near the middle of the duplex and it's NAEDP_ODN = 17.2 and NAED_S-PODN7 = 18.8 (Figure 7). NAED comparisons show that the CD spectrum of the s- PODN1 :RNA duplex is very dissimilar to that of the PODN:RNA, NAED_PODN = 31.6, and s-PODN7:RNA duplexes, NAED_S-PODN7 = 32.2 (Figure 7). Interestingly, long- range cooperativity is absent in the s-PODNl :RNA duplex, even though it has six consecutive propynyl substitutions. Therefore, long-range cooperative interactions could be dependent on helical structure. These results suggest that thermodynamic interactions governing helical structure at the 5' and 3' ends of the PODN:RNA duplex may not be equal, resulting in an intolerance of propynyl deletions at the 5' end. This could be due to the large difference in unpaired rA stacking interactions at the 5' and 3' ends of the PODN:RNA duplex (Table 3).

Possible Sources ofLong-Range Cooperative Interactions:

There are many possible sources of long-range cooperative effects in duplexes containing multiple consecutive propynyl groups. The enthalpy change for formation of the fully propynylated duplex is less favorable than that of the unmodified duplex (Table 2). This suggests that base stacking is not more favorable in the propynylated duplex, since base stacking is driven by a favorable enthalpy change (Turner, 2000; Petersheim et al., 1983; Holbrook et al., 1999; Freier et al., 1981). Bulky propynyl substitutions may prevent optimal stacking interactions within PODN:RNA duplexes by restricting conformational space available for propeller twisting, roll, etc. It is possible, however, that base stacking is more favorable in the unpaired propynylated strands because of more volume exclusion due to multiple propynyl groups. That is, the preorganization effect described above for single propynyl substitutions can accumulate in a cooperative manner for multiple propynyl substitutions. This would make the enthalpy and entropy changes for duplex formation less and more favorable, respectively. Preliminary ID and 2D NMR spectra of single strands at various temperatures suggest that helical stacking only occurs between propynylated pyrimidines within the PODN, while none occurs within the unmodified DNA.

A classical hydrophobic effect is suggested by the 18.8 kcal/mol more favorable entropy change at 37 °C for PODN: RNA 7-mer duplex hybridization over its unmodified analog (Table 2). Indeed, the dU^p analog is more hydrophobic than dU (Valko et al, 1989). The ΔC°_P of the DNA:RNA duplex is more negative than that of the PODN:RNA duplex, however, arguing against a classical hydrophobic effect (Holbrook et l., 1999).

Chalikian et al. (1999a,b) have suggested that there is a specific 'cooperative patch' of 13 water molecules that lie in the major groove of G-C base pairs and a specific cooperative 'spine' of 8 water molecules in the minor groove of A-U base pairs of duplexes in solution. Egli et al. (1996) found a cooperative patch of water molecules within the major and minor grooves of a crystal of r(5'CCCCGGGG3')₂. They contend that it is 2' hydroxyl groups of RNA, "which lock the sugar-phosphate backbone in a conformation that allows water to bridge adjacent phosphates". Our data suggest that the enhanced stability of the PODN:RNA duplex is due to a less destabilizing entropic penalty than for the DNA:RNA duplex. Therefore, cooperative solvent molecules could be excluded from the helix by geometric distortions induced by multiple consecutive bulky hydrophobic propynyl groups, resulting in a more dehydrated major groove. This could lead to a less favorable enthalpy change and more favorable entropy change for duplex formation. This cumulative bulk could also prevent optimal stacking, which may explain a less favorable enthalpy contribution upon full propynylation (Table 2).

Helix distortion and dehydration could also rationalize the large duplex destabilization due to removal of the amino group from G₆. This effect may be due to strengthening of the G₆(amino)-C₂(carbonyl) hydrogen bond. The length of this hydrogen bond could be shorter due to bulky propynyl groups in the major groove, affecting parameters such as propeller twist and opening, etc. Also, dehydration of the minor groove would reduce the local dielectric constant. The strength of an electrostatic interaction such as a hydrogen bond is inversely proportional to its length and the medium's dielectric constant. A reduction in either or both parameters will lead to an increase in the strength of the G(amino)-C(carbonyl) hydrogen bonds. When one of the propynyl groups is eliminated, as in d(5'C^pC^pU^pCC^pU^pU^p3') or d(5'C^pC^pU^pC^pCU^pU^p3'), the duplex conformation and hydration may change, causing the apparent strength of the minor groove hydrogen bond to revert back to that in the unmodified DNA:RNA helix. Alternatively, the strength of the hydrogen bond in the duplex may not be affected, but the conformation of the fully propynylated duplex may prevent hydration of the unpaired carbonyl on a C^p opposite an inosine. Both possibilities require that propynyls cooperatively induce a global change in helix conformation.

Impact of Enhanced Cooperative Interactions on Affinity and Specificity of Binding:

Propynylation of seven consecutive pyrimidines leads to highly cooperative binding of d(5'C^pC^pU^pC^pC^pU^pU^p3') to r(3 'GAGGAGGAAAU5 '), which drastically increases the potency of the antisense:sense interaction by 8.8 kcal/mol at 37 °C (Table 4). Most of the enhanced stability is focused within the base pairing region since the contribution of stacked unpaired dangling ends is only 1 kcal/mol more favorable for the fully propynylated duplex.

It has been proposed that the sequence, r(3 'G AGGAGGAAAU5 '), exists within a stable 27 nucleotide stem loop structure within the SV40 TAg mRNA (Wagner et al., 1996), and it is therefore surprising that d(5'C^pC^pU^pC^pC^pU^pU^p3') is able to inhibit translation of this mRNA. The results presented here show that the free energy of base pairing for the fully propynylated oligonucleotide is sufficient to allow it to invade RNA targets buried within stable intramolecular secondary structures. Such properties of short PODNs could be used to destroy secondary structures, as well as tertiary contacts, that are crucial for a target RNA's function. Probing for sequences without regard to RNA structure can lead to false-negative results in molecular beacon (Leone et al., 1998; Sokol et al., 1998; Bonnet et al., 1999; Liu et al., 1999) and microarray assays (Schena et al., 1985; Healey et al., 1997; Hacia et al., 1998; Maldonado-Rodriguez et al., 1999; Gerry et al., 1999; Chen et al., 1999; Walt, 2000) if the target sequences are buried within highly stable local secondary structures. The highly cooperative nature of consecutive propynylated pyrimidines could unmask such sequences. The non-nearest neighbor interactions that enhance the cooperativity of PODN:RNA hybridization may also increase an antisense PODN's target specificity. Small changes, such as propyne group deletions and elimination of an amino group greatly destabilize the antisense:sense complex. Thus, larger changes, such as mismatches, may cause even greater destabilization of the complex.

Therefore, the antisense PODN might have high specificity for its intended target due to a loss of long-range cooperative interactions when paired with mismatched bystander targets. This could facilitate applications such as anti sense-based drugs (Milligan et al., 1993; Agrawal et al., 1997; Branch, 1998; Crooke, 2000), microarray screening (Schena et al., 1985; Healey et al., 1997; Hacia et al., 1998; Maldonado- Rodriguez et al., 1999; Gerry et al., 1999; Chen et al., 1999; Walt, 2000) molecular beacon probing (Leone et al., 1998; Sokol et al., 1998; Bonnet et al., 1999; Liu et al., 1999) and design of self-organizing nanostructures that rely on nucleic acid-based molecular recognition (Mirkin et al., 1996; Alivisatos et al., 1996; Seeman et al., 1998).

Example 2 - Long-Range Cooperativity Enhances Specific Recognition of ribo- Adenosine over ribo-Guanosine by C5-(l-PropynyI)-2'-deoxyribo- Uridine

Oligonucleotide Synthesis & Purification:

Phosphoramidites and supports were purchased from Glen Research. All oligonucleotides were synthesized (Matteucci et al., 1981; Usman et al., 1987; Wincott et al., 1995) on an Applied Biosystems 392 DNA/RNA synthesizer using the manufacturer's suggested protocols.

Oligoribonucleotides were deblocked in ethanolic ammonia (1:3 v/v) for 17 h at 55 °C (Stawinski et al., 1988). After filtering away support, RNA oligomers were placed into 1 M triethylammonium hydrogen fluoride (50 equiv) at 55 °C for 50 h. Solutions of crude products were evaporated, dissolved in water, and extracted against diethyl ether. After removing residual ether by evaporation, 5 mM aqueous ammonium acetate (pH = 7.0) was added and the oligomers were desalted on a reverse phase Sep-Pak C-18 cartridge (Waters Corp.). The oligomers were purified by 20%) PAGE. The product was UN visualized, cut out, and eluted with sterile water containing 0.5 mM Νa₂EDTA. After separating the sample from a majority of the urea with a pre-packed PD-10 Sephadex column (Amersham Pharmacia Biotech), the samples were dialyzed in an irradiated cellulose ester (MWCO = 1000) Spectra/Por Dispodialyzer (Spectrum Labs Inc.) against 0.1 mM EDTA (pH = 7.0) and subsequently against sterile water. Samples were then lyophilized. To test purity, RNA oligomers were 5'-radiolabeled by incubating

2.4 pmol RNA and 30 pmol [γ-³²P] ATP (New England Nuclear) in 7.5 uL of 50 mM Tris (pH = 7.5), 10 mM MgCl₂, 5 mM DTT, 0.1 mM spermidine, and 0.1 mM Na₂EDTA with 7.5 units of T4 polynucleotide kinase (Gibco-BRL) at 37 °C. After 10 min, the reaction was stopped with 7.5 uL of 2X stop buffer (10 M urea, 10 mM EDTA, pH = 7.0) and applied to a 20% polyacrylamide, 8 M urea denaturing gel for electrophoresis with IX TBE running buffer. Once bromophenol blue dye reached 18 cm, the products were imaged and quantified with a Molecular Dynamics phosphorimager and quantitation software package. Purity was greater than 95%). The 5'-trityl oligodeoxyribonucleotides and C5-(l-propynyl)-5'- trityl-oligodeoxyribonucleotides were incubated in concentrated ammonium hydroxide at 55 °C for 2 h. After removing the support by spin filtration, the crude product was applied to, and eluted from, a Poly Pak II cartridge (Glen Research) using the manufacturer's recommended protocol to purify the desired product from most of the failure sequences. The product was further purified by preparative thin layer chromatography (tic) with an n-propanol: ammonium hydroxide:water (55:35: 10) running buffer. Reverse phase C-18 Sep-Pak cartridges (Waters Corp.) were used to desalt the products, which were then lyophilized. The purity of all PODNs and DNAs was greater than 95% on the basis of RP-HPLC on a SUPELCOSIL ABZ+ Plus semi- prep column (Supelco) with a gradient from 0 to 30% acetonitrile (by volume) in 100 mM aqueous triethylamine acetate (pH = 7.0) over 1 h at a flow rate of 1 mL/min.

UVMelting:

Thermodynamic parameters were measured in IX melting buffer (1.0 M NaCl, 0.5 mM Na₂EDTA, 20 mM sodium cacodylate at a pH of 7.0). Single strand oligoribonucleotide concentrations were calculated from high temperature absorbances at 280 nm and predicted single strand extinction coefficients (Borer,

1975; Richards, 1975). Single strand DNA concentrations were determined from high temperature absorbances at 260 nm on the basis of monomer extinction coefficients (Puglisi et al., 1989). Single strand PODN concentrations were calculated from high temperature absorbances at 260 nm on the basis of monomer extinction coefficients of 3200 and 5000 M^"'cm^"' for U^p and C^p, respectively (generously provided by M. D. Matteucci and B. C. Froehler). These were also used in conjunction with DNA monomer extinction coefficients at 260 nm to estimate the concentrations of chimera oligomers containing modified and unmodified pyrimidines.

Appropriate single strands were mixed at 1 : 1 concentration ratios, denatured for 1 min at 90 °C, and re-annealed by slow cooling to 0 °C. UN melting curves were measured at 280 nm with a heating rate of 1 °C/min on a Gilford 250 spectrophotometer. Each melting curve was fit to a non-self-complementary two-state model (Longfellow et al., 1990) with the Meltwin Software package (McDowell et al., 1996). The thermodynamic parameters were averaged over all melts of a given duplex and compared to those generated by plotting the reciprocal of the melting temperature, T_M ^"1, versus Log(Cχ/4), where C_τ is the total concentration of strands (Borer et al., 1974):

T_M ^"' = (2.303R/ΔH°)Log(C_τ/4) + ΔS°/ΔH° (11)

With only two minor exceptions, the enthalpy changes generated from the two methods of data analysis agree within 15%, consistent with two-state melting (Freier et al., 1996). Errors were calculated as in Xia et al. (1998) and references therein.

When comparing the thermodynamic parameters of two different hybrid systems, the T_M ^"' versus Log[Cχ/4] results were used and errors were propagated as in Xia et al. (1998) and references therein.

Results & Discussion

C5-(l-Propynyl)ation Enhances Discrimination:

Thermodynamic parameters from UV melting studies are listed in Table 7 for a series of DΝA and PODΝ heptamers hybridized to RΝA strands containing 5' and 3' terminal unpaired nucleotides. The unpaired dangling nucleotides simulate RΝA targets, which are typically longer than DΝA probe strands. These duplexes are denoted as (A:U)-n and (A:U^p)-n, where n is the entry number in Table 7. Single rA-»rG substitutions were made within each RΝA strand, producing single rG:dU or rG:dU^p pairs. These duplexes are denoted (G:U)-n and (G:U^p)-n. Representative melting curves of all four duplex types are shown in Figure 9. Subtracting the free energy of (A:U)-n or (A:U^p)-n duplexes from that of their respective (G:U)-n or (G:U^p)-n duplexes provides the thermodynamic impact, ΔΔG°₃₇, of a single rG:dU wobble on hybrid duplex formation (Table 7).

The ΔΔG° for dU^p ranges from 2.6-4.2 kcal/mol and averages 3.3 kcal/mol compared with a range of 0.1-0.9 and an average of 0.5 kcal/mol for dU. Thus, dU^p provides a roughly 100-fold greater discrimination in the relative binding constants than that observed with dU. Discrimination against G:U formation occurs within all nearest neighbors and positions tested, in contrast with results for 2-thio-rU (Testa et al., 1999). The range of PODN.RNA duplex ΔΔG°₃₇'s suggests that the magnitude of rG:dU^p discrimination is somewhat nearest neighbor dependent. The largest discriminations are seen in the context of d(5'C^pU^pU^p3')/r(3'GGA5').

The average ΔΔG°₃₇ for unmodified dU of 0.5 kcal/mol is the same value as the average ΔΔG°₃₇ expected for rU within RNA:RNA duplexes (Xia et al., 1998; Mathews et al., 1999). Furthermore, the average ΔΔG°₃₇ for dU reported here corresponds well with that expected for dU (0.3 kcal/mol) within DNA:RNA hybrids (Sugimoto et al, 1995; Sugimoto et al., 2000).

Enhanced Discrimination is Coupled to Long-Range Cooperative Interactions Between Ψ 's:

Single propynyl deletions were made within PODN entries 3^P, 4^P, and 5^P from Table 7. These oligonucleotides are referred to as s-PODNs. Their duplexes with RNA sequences having an A:U or G:U pair are denoted (A:U^p)-sn and (G:U^p)-sn, respectively. The thermodynamics of these duplexes are in Table 7. All propynyl deletions within s-PODNs occur at least two base pairs away from the position of an rA:dU^p— »rG:dU^p modification. Thus, nearest neighbor pairs directly adjacent to each position of a rA— >rG modification are not changed by eliminating the propynyl group.

Comparison of the stabilities of (A:U^p)-sn and parent (A:U^p)-n duplexes within Table 7 indicates that the thermodynamic contribution of a single propynyl group to the overall stability of each PODN:RNA duplex ranges from 2.2- 3.3 kcal/mol and averages to 2.6 kcal/mol. This is within experimental error of reported thermodynamic contributions of single propynyl groups (3.1 kcal/mol) to overall PODN:RNA duplex stability (Barnes et al., 2001a). By subtracting the free energy of (A:U^p)-sn duplexes from that of their respective (G:U^p)-sn duplexes in Table 7, the thermodynamic impact, ΔΔG°₃₇, of a single G:U^P pair on s-PODN:RNA duplex formation is found to range from 0.6-1.5 kcal/mol. These values are more similar to those of DNA:RNA than PODN:RNA duplexes with the same sequences in Table 7. The data in Table 7 show that elimination of a single propynyl group reduces discrimination by 2.1-3.6 kcal/mol. The difference in the average ΔΔG°₃₇'s is 2.6 kcal/mol. This is remarkably similar to the 2.6 kcal/mol attributed to highly cooperative long-range interactions between seven consecutive Y^p's in a different PODN:RNA duplex (Barnes et al., 2001a). This suggests that enhanced discrimination against G:U^P is lost due to loss of long-range cooperative interactions between Y^p's when a single propynyl group is removed.

The results in Table 7 show that full propynylation of a DNA strand participating in PODN:RNA hybridization enhances discrimination of rG:dU^p pair formation, thus increasing specificity 100-fold. It is likely that highly cooperative interactions between Y^p's can also enhance discrimination against less stable mismatches within PODN:RNA duplexes. These results show that modifying positions not directly involved in molecular recognition can enhance specificity. It will be interesting to see if modifications can be found to extend specificity to mixed pyrimidine-purine sequences of RNA and DNA, as well as to other backbones suitable for designing therapeutics, diagnostics, and structures for nanotechnology.

Example 3 - C5-(l-Propynyl)-2'-deoxy-Pyrimidines Enhance the Specificity of Hybridization for DNA:RNA Duplexes with Phosphodiester and Phosphorothioate Backbone

Oligonucleotide Synthesis and Purification:

Phosphoramidites, supports, and sulfurizing reagent were purchased from Glen Research. Oligonucleotides were synthesized by standard chemistry (Matteucci et al., 1981; Usman et al, 1987; Wincott et al., 1995) on an Applied Biosystems 392 DNA/RNA synthesizer using the manufacturer's suggested protocol. The sulfurizing reagent, 3H-l,2-benzodithiole-3-one-l,l-dioxide (Iyer et al., 1990), was used for synthesis of C5-(l-propynyl)ated deoxyribophosphorothioate oligonucleotides (th-PODNs). Each RNA strand was deblocked for 17 h at 55 °C in ethanolic ammonia (1 :3 v/v) (Stawinski et al., 1988). After filtering away the solid support, RNA oligomers were incubated in 1 M triethylammonium hydrogen fluoride (50 equivalents) at 55 °C for 50 h and dried. Following diethyl ether extraction, oligoribonucleotides were purified from failure sequences with 20% PAGE. Then, each product was UN visualized and cut out of the gel matrix. The samples were eluted with sterile water, spin filtered, and separated from a majority of the urea with a pre-packed PD-10 Sephadex column (Amersham Pharmacia Biotech). Using an irradiated cellulose ester (MWCO = 1000) Spectra/Por Dispodialyzer (Spectrum Labs Inc.), RΝA strands were dialyzed first against 0.1 mM EDTA (pH = 7.0), then sterile water, ethanol precipitated, and finally lyophilized.

To test purity, 2.4 pmol of RΝA was incubated with 30 pmol [γ-³²P] ATP (New England Nuclear) and 7.5 units of T4 kinase (Gibco-BRL) in 7.5 μL of 50 mM Tris (pH = 7.5), 10 mM MgCl₂, 5 mM DTT, 0.1 mM spermidine and 0.1 mM Na₂EDTA at 37 °C. The reaction was stopped with 7.5 μL of IX stop buffer (10 M urea, 10 mM EDTA, pH = 7.0) after 10 min and applied to a 20% denaturing gel for electrophoresis in IX TBE buffer. Quantification with a Molecular Dynamics phosphorimager showed the products were greater than 95% pure.

DNAs and C5-(l -propynyl) oligodeoxynucleotides (PODNs and th- PODNs) with 5'-trityl attached were cleaved from the support in concentrated ammonium hydroxide at 55 °C for 2 h. The support was removed by spin filtration. Product was purified from a bulk of the failure sequences on a Poly Pak II cartridge (Glen Research) using the recommended protocol. Oligodeoxynucleotides were further purified by tic on a Si500F plate (J.T. Baker) with a running buffer of n- propanol: ammonium hydroxide:water (55:35: 10). They were then desalted on reverse phase C-18 Sep-Pak cartridges (Waters Corp.) and lyophilized. The purities of DNAs and PODNs were greater than 95% as assayed by RP-HPLC on a Supelcosil ABZ+ Plus semi-prep column (Supelco) with a buffer of 100 mM triethylamine acetate (pH = 7.0) in a gradient of 0 to 30% acetonitrile (by volume) over 1 h at a rate of 1 mL/min.

The th-PODN samples are mixtures of 2 = 64 diastereomers possessing pro-R and pro-S linkages. Kination analyses as described above for RNA strands show that > 95% of each racemic mixture migrates as a uniform band on a 20%) denaturing gel in 1 X TBE buffer. UV Melting:

Concentrations of single-stranded RNAs were calculated from high temperature absorbances at 280 nm and predicted single strand extinction coefficients (Borer, 1975; Richards, 1975). Concentrations of single-stranded DNAs were calculated from high temperature absorbances at 260 nm and monomer extinction coefficients (Puglisi et al., 1989). Concentrations of single-stranded PODNs and th- PODNs were determined from high temperature absorbances at 260 nm, assuming monomer extinction coefficients of 3200 and 5000 M^cm^"1 for U^p and C^p, respectively (generously provided by Drs. M. D. Matteucci and B. C. Froehler). These were also used in conjunction with DNA monomer extinction coefficients at 260 nm to estimate the concentrations of chimera oligonucleotides containing C5-(l- propynyl)ated and unmodified pyrimidines. After mixing appropriate strands at a 1 : 1 concentration ratio, duplexes were pre-denatured for 1 min at 90 °C and re-annealed by slow cooling to 0 °C. UV melting studies were performed at 260 nm for DNA:RNA duplexes and at 280 nm for PODN:RNA and th-PODN:RNA duplexes in 1.0 M NaCl, 0.5 mM Na₂EDTA, and 20 mM sodium cacodylate at pH = 7.0. Data were collected at a heating rate of 1 °C/min with a Gilford 250 spectrophotometer. Each melting curve was fit to a non-self-complementary two-state model using Meltwin software (Longfellow et al., 1990; McDowell et al., 1996). Thermodynamic parameters for duplex formation were determined by averaging fits of individual curves and by plotting the reciprocal of the melting temperature, T_M ^"1, versus Log(Cχ/4), where C is the total concentration of strands (Borer et al., 1974):

T_M ^"1 = (2.303R/ΔH°)Log(C_τ/4) + ΔS°/ΔH° (12)

Errors were calculated as in Xia et al., 1998 and references therein. When comparing the thermodynamic parameters of duplexes, the results of the T_M ^"1 versus Log(Cχ/4) plots were used and errors were propagated as in Xia et al., 1998 and references therein. Results

Thermodynamic Parameters of Watson-Crick Paired Duplexes:

Table 8 lists thermodynamic parameters of DNA:RNA, PODN:RNA, and th-PODN:RNA duplexes with only Watson-Crick pairs. In most cases, RNA strands are longer than their deoxyribose counterparts to mimic binding sites of nucleic acid therapeutics and probes. Thus, thermodynamic parameters include contributions from stacking of unpaired ribonucleotides. By comparing duplexes DNA:RNA-1 and PODN:RNA-l^p to DNA:RNA-7 and PODN:RNA-7^p (Table 8), the unpaired ribonucleotides stabilize DNA:RNA-1 and PODN:RNA-l^p by 2.0 and 3.1 kcal/mol, respectively. Subtraction of the free energy of th-PODN:RNA-7^p from that of th-PODN:RNA-l^p (Table 8) indicates that the unpaired ribonucleotides similarly stabilize th-PODN:RNA-l^p by 3.0 kcal/mol.

Thermodynamic Penalties for Duplexes With Single Mismatches:

To determine the thermodynamic penalty for various mismatches in hybrid duplexes, ribopurines were systematically replaced in the following fashions: rA→rC, rG→rC, and rG→rA, to generate single mismatches. The thermodynamic parameters of these duplexes are in Table 9. Thermodynamic parameters for five DNA:RNA and five PODN:RNA duplexes with single dU:rG pairs have been previously reported (Barnes et al., 2001b) and they are also listed in Table 9, along with parameters for five th-PODN:RNA duplexes with similar sequences.

The single mismatches in Table 9 were chosen because they represent three categories, characterized by their impact on the stabilities of DNA:RNA duplexes. First, single dC:rC and dU:rC mismatches destabilize DNA:RNA duplexes within all nearest neighbor contexts studied thus far (Sugimoto et al., 2000). Second, dC:rA mismatches can stabilize or destabilize DNA:RNA duplexes, depending on nearest neighbor pairs (Sugimoto et al., 2000). Thirdly, dU:rG pairs stabilize DNA:RNA duplexes within all nearest neighbor contexts studied thus far (Sugimoto et al., 2000).

Free energy penalties for mismatches, ΔΔG°₃₇(MM)'s, were calculated from:

ΔΔG°₃₇(MM) = ΔG°₃₇(duplex-MM) - ΔG°₃₇(duplex-WC) (13) Here, ΔG°₃₇(duplex-MM) is the free energy of the duplex with a single mismatch and ΔG°₃₇(duplex-WC) is the free energy of the duplex with only Watson-Crick base pairs. Thus, ΔΔG°₃₇(MM) depends on both the free energy of the mismatch and the loss in stabilizing free energy of the substituted base pair. The ΔΔG°₃₇(MM)'s for replacing dU:rA pairs with dU:rG in DNA:RNA, PODN:RNA, and th-PODN:RNA duplexes range from 0.1-0.9, 2.6-4.2, and 1.3-2.4 kcal/mol, respectively (Table 8). The ΔΔG°₃₇(MM)'s for replacing dC:rG base pairs with dC:rA mismatches in DNA:RNA, PODN:RNA, and th-PODN:RNA duplexes range from 1.3-6.0, 4.0-9.1, and 2.1-7.2 kcal/mol, respectively (Table 9). Figure 10 summarizes the position and sequence dependence of

ΔΔG°₃₇(MM)'s for rG→rC and rA→rC substitutions, which yield dC:rC and dU:rC mismatches, respectively. The ΔΔG°₃₇(MM)'s for replacing dC:rG base pairs with dC:rC mismatches in DNA:RNA, PODN:RNA, and th-PODN:RNA duplexes range from 3.6-6.5, 4J-10.2, and 3.6-6.6 kcal/mol, respectively (Table 9). The ΔΔG°₃ (MM)'s for replacing dU:rA base pairs with dU:rC mismatches in DNA:RNA, PODN:RNA, and th-PODN:RNA duplexes range from 1.3-4.7, 2.8-6.9, and 1.3-5.4 kcal/mol, respectively (Table 9). Propynylation of the DNA strand enhances the penalties for all mismatches when the backbone is phosphodiester, and the average enhancement is 2.9 kcal/mol (Table 9). Phosphorothioate backbone substitutions decrease this average enhancement to 1.1 kcal/mol (Table 9).

Free Energy Penalties are Larger for dC.rA, dC.rC, and dU.rC Internal Mismatches than for Terminal Mismatches:

Table 9 and Figure 11 show that not all ΔΔG°₃₇(MM)'s are created equal. One determinant of the magnitude of mismatch penalties is position in the hybrid duplex. For example, ΔΔG°₃₇(dC:rA) at the end of DNA:RNA duplex-1 (C: A) is 3.2 kcal/mol, while that within DNA:RNA duplex-3(C:A) is 6.0 kcal/mol. Figure 12 summarizes the average free energy penalties for terminal and internal mismatches. For all oligonucleotide families investigated, dC:rA, dC:rC, and dU:rC mismatches have substantially larger free energy penalties at internal than at terminal positions. The average ΔΔG°₃ (MM-end) penalties destabilize DNA:RNA,

PODN:RNA, and th-PODN:RNA duplexes by 0.4-3.6, 2.8-5.4 and 1.3-3.6 kcal/mol, respectively (Figure 12). The ΔΔG°₃₇(MM-internal) penalties destabilize DNA:RNA, PODN:RNA, and th-PODN:RNA duplexes by 0.7-5.4, 3.5-9.4, and 2.0-6.7 kcal/mol, respectively (Figure 12). Evidently, full propynylation of DNA strands enhances the penalties of both terminal and internal mismatches. Furthermore, the th-PODN:RNA duplexes retain elevated terminal and internal ΔΔG°₃₇(MM)'s, except for terminal dC^p:rC and dU^p:rC mismatches (Figure 12).

The Effects of Single Propynyl Deletions on Penalties for Mismatches:

The thermodynamic parameters of the Watson-Crick paired duplex, d(5'C^pCU^pC^pC^pU^pU^p3')/r(3'GAGGAGGAAAU5'), which has a single propynyl deletion were measured and compared to those of equivalent duplexes with single dC^p:rC or dC^p:rA mismatches two base pairs away from the single propynyl deletion (Table 10). The d(5'C^pCU^pC^pC^pU^pU^p3')/r(3'GAGGAAGAAAU5') duplex, s- PODN₂(dC^p:rA), which has a dC^p:rA mismatch, is 6.7 kcal/mol less stable than the fully Watson-Crick paired duplex, s-PODN₂ (Table 10). The corresponding penalty is 9.1 kcal/mol for the fully propynylated PODN:RNA duplex (compare entry 3^P (C^P:A) in Table 9 with entry l^p in Table 8).

Similarly, the dC^p:rC mismatch in the d(5'C^pCU^pC^pC^pU^pU^p3')/(3'rGAGGACGAAAU5') duplex, s-PODN₂(dC^p:rC), is destabilizing by 7.2 kcal/mol relative to the fully Watson-Crick paired duplex (Table 10). The corresponding penalty for the fully propynylated PODN:RNA duplex is 10.2 kcal/mol (compare entry 3^p(C^p:C) in Table 9 with entry l^p in Table 8). Differences in ΔΔG°₃₇(MM)'s within s-PODN:RNA and the fully propynylated duplexes are remarkably similar to the 2.6 kcal/mol attributed to long-range cooperative interactions in fully propynylated duplexes with seven Watson-Crick base pairs (Barnes et al, 2000a; 2000b). Apparently, long-range cooperativity between Y^p's enhances the thermodynamic penalties of mismatches within PODN:RNA duplexes.

Discussion Specificity of nucleic acid hybridization is important for designing antisense and probe oligonucleotides. Specificity is typically reduced as oligomer length is increased and free energies per base pair are made more favorable (Herschlag, 1991; Roberts et al., 1991). It is interesting, therefore, that PODNs have more favorable free energies for base pairing (Froehler et al., 1992; Freier et al., 1997; Barnes et al., 2001a; 2001b), but are specific to RNA targets (Wagner et al, 1993; Moulds et al., 1995; Flanagan et al., 1996; Wagner et al., 1996). The results reported here suggest that the specificity at least partially arises from enhanced penalties for mismatches in hybrid duplexes.

Effects of Full Propynylation on Specificity: The order of mismatch penalties in DNA:RNA duplexes is dC:rC- internal ~ dC:rA-internal » dC:rC-end ~ dU:rC-internal > dC:rA-end > dU:rC-end > dU:rG-internal ~ dU:rG-end (Figure 12). The order of mismatch penalties in PODN:RNA duplexes is dC:rC-internal > dC:rA-internal » dC:rA-end ~ dU:rC- internal ~ dC:rC-end > dU^p:rG-internal ~ dU^p:rG-end ~ dU^p:rC-end (Figure 4-3). Thus, the order of mismatch penalties is similar for DNA:RNA and PODN:RNA duplexes, even though the magnitudes of penalties are larger with PODNs.

Full propynylation enhances single mismatch penalties by 1.1-4.0 kcal/mol (Figure 13). The average ΔΔG°₃₇(MM-end) and ΔΔG°₃₇(MM-internal) changes by 2.1 and 3.1 kcal/mol, respectively, upon full propynylation of DNA:RNA duplexes (Figure 13), and the effect enhances the destabilizing penalties of all internal and terminal single mismatches studied. The order of enhanced destabilization is dC^p:rC-internal ~ dC^p:rA-internal > dC^p:rA-end ~ dU^p:rG-internal > dU^p:rG-end ~ dU^p:rC-internal > dU^p:rC-end _dC^p:rC-end (Figure 13). Some mismatches with the least destabilizing ΔΔG°₃₇(MM)'s in DNA:RNA duplexes, such as dU^p:rG-internal and dU^p:rG-end, are highly destabilized upon full propynylation. This provides a way to enhance specificity in nucleic acid hybridization. Other mismatches that are more destabilizing to DNA:RNA duplexes, such as dU^p:rC-internal and dU^p:rC-end, are destabilized the least upon full propynylation. The net result is that PODN:RNA duplexes provide large penalties for all mismatches studied, whereas DNA:RNA duplexes do not.

The magnitude by which mismatch penalties are enhanced is position- dependent. Terminal mismatch penalties are enhanced less than internal mismatch penalties within PODN:RNA duplexes. Figure 13 shows the difference in the magnitudes of ΔΔG° (MM-internal) and ΔΔG°₃₇(MM-end) enhancement upon full propynylation. Penalties for internal dU:rG, dU:rC, and dC:rA mismatches are enhanced by similar magnitudes (0.3 - 0.6 kcal mol) upon full propynylation relative to their equivalent terminal mismatches. In contrast, the penalty for an internal dC:rC mismatch is enhanced 2.9 kcal/mol more than that of a terminal dC:rC mismatch upon full propynylation. The uniqueness of dC:rC mismatches could be related to the fact that C:C mismatches are less likely to hydrogen bond than U:G, U:C, and C:A mismatches, which can form hydrogen bonds at neutral pH (Hare et al., 1986; Tanaka et al., 2000; Pan et al., 1998).

Long-Range Cooperative Interactions of Propynyls Enhance Mismatch Penalties:

Full propynylation of DNA:RNA duplexes increases ΔG°₃ (MM)'s by an average of 2.9 kcal/mol (Table 9). This increment is remarkably similar to the 2.6 kcal/mol of stability provided by a cooperativity bonus in a model for prediction of the stabilities of PODN:RNA duplexes with seven Watson-Crick base pairs (Example 1). These results suggest that long-range cooperative interactions between Y^p's are lost when mismatches are present.

Deleting a single propynyl group from a Y^p in a Watson-Crick pair can eliminate the cooperative interaction of propynyls in an entire 7-mer duplex (Example 1). Deleting a single propynyl group in a Watson-Crick pair two base pairs away from a dU^p:rG pair reduces ΔΔG°₃₇(dU^p:rG) by 2.6 kcal/mol (Example 2). This suggests the enhanced ΔΔG°₃₇(MM) results from long-range cooperative interactions between Y^p's (Examples 1 & 2). Tables 9 and 10 reveal a similar effect for dC^p:rC and dC^p:rA mismatches. In these cases, deleting a single propynyl group two base pairs from the position of the mismatch reduces ΔΔG°₃₇(MM) by 2.4 and 3.0 kcal/mol, respectively (Table 10). Evidently, the enhanced ΔΔG°₃₇(MM) increments in PODN:RNA duplexes depend on the cooperativity between propynylated pyrimidines.

Effects of Full Phosphorothioate Substitution on Specificity: The order of mismatch penalties within th-PODN:RNA duplexes is dC:rA-internal ~ dC:rC-internal » dU:rC-internal ~ dC:rC-end > dC:rA-end > dU^p:rG-internal ~ dU^p:rG-end > dU^p:rC-end, which is similar to those observed with DNA:RNA and PODN:RNA helices (Figure 12). With the exception of ΔΔG°₃₇(dU:rC-end) and ΔΔG°₃₇(dC:rC-end), ΔΔG°₃₇(MM)'s for th-PODN:RNA duplexes are more destabilizing than those for DNA:RNA duplexes.

Full stereo-random phosphorothioate substitution of PODN:RNA duplexes reduces all mismatch penalties, with the average ΔΔG°₃₇(MM-end) and ΔΔG°₃₇(MM-internal) penalties reduced by 1.4 and 1.9 kcal/mol, respectively. The order of reduction of ΔG°₃₇(MM)'s upon stereo-random phosphorothioate substitution of PODN:RNA duplexes is dC^p:rC-internal » dC^p:rA-end ~ dC^p:rA-internal > dU^p:rG-internal ~ dU^p:rC-end ~ dU^p:rC-internal > dC^p:rC-end ~ dU^p:rG-end (Figure 4-4). Interestingly, the order of reduction is similar to the order of enhancement in mismatch penalties due to propynylation.

The magnitude by which mismatch penalties are reduced by phosphorothioate substitution depends on whether the mismatch is terminal or internal (Figure 13). Most dramatically, for a dC^p:rC mismatch, ΔΔG°₃₇(MM-end) is reduced by only 1.1 kcal/mol, while ΔΔG°₃₇(MM-internal) is reduced by 3.1 kcal/mol (Figure 13). The magnitudes of mismatch penalty reduction for all other mismatches are much less position dependent. Interesting, for dC^p:rC and dU^p:rG mismatches, ΔΔG°₃₇(MM-internal) is reduced more than ΔΔG°₃₇(MM-end) upon phosphorothioate substitution of PODN:RNA duplexes. In contrast for dU^p:rC and dC^p:rA mismatches, ΔΔG°₃₇(MM-internal) is reduced less than ΔΔG°₃₇(MM-end).

Effects of Full Phosphorothioate Substitution on PODNRNA Stability:

The T_m's of Watson-Crick PODN:RNA duplexes differ from those of th-PODN:RNA duplexes by an average of only 2.3 °C, and on average the PODN:RNA duplexes are more stable by 1.3 kcal/mol at 37 °C (Table 8). These trends have been observed previously for unpropynylated DNA:RNA and stereo- regular th-DNA:RNA duplexes (Clark et al, 1997; Hashem et al., 1998). In general, the stability change per phosphorothioate substitution is modest.

The two 5' and two 3' unpaired ribonucleotides in the th-PODN:RNA- l^p duplex increase duplex stability by 3.0 kcal/mol over the th-PODN:RNA-7^p duplex. This thermodynamic contribution is similar to that found with the equivalent

PODN:RNA duplexes (3.1 kcal/mol), but more favorable than with the DNA:RNA duplexes (2.0 kcal/mol) (Example 1). Evidently, the helical geometry affording enhanced cross-strand stacking interactions in PODN:RNA duplexes is retained in th- PODN:RNA duplexes. The circular dichroism spectra of th-PODN-l^p (unpublished results) and its phosphodiester analog (Example 1) are very similar, supporting the idea of similar helical geometries. These findings are consistent with those reported for DNA:RNA and stereo-random th-DNA:RNA duplexes (Clark et al., 1997). Comparison of PODN: RNA Duplex Stabilities with Predictions of a Cooperative Model:

A cooperative model has been presented that accounts for the stability enhancements of PODN:RNA duplexes (Example 1). Table 8 shows the free energy changes predicted for the PODN:RNA duplexes studied. The rmsd between measured and predicted ΔG°₃₇ is 0.74 kcal/mol, which is only 4.7 % that of the average measured ΔG°₃₇ for these duplexes. Thus, the cooperativity model (Example 1) is very good at predicting the stabilities of these PODN:RNA duplexes.

Comparison ofDNA:RNA Duplex Stabilities with Predictions: In 1995, Sugimoto et al. reported thermodynamics for many

DNA:RNA duplexes, and derived parameters for the prediction of DNA:RNA duplex stability on the basis of an individual nearest neighbor model (INN) (Sugimoto et al., 1995). In 1997, Gray used the same data to derive parameters for the prediction of DNA:RNA duplex stability on the basis of an independent short sequence model (ISS) (Gray, 1997). After accounting for unpaired terminal nucleotides, the results reported in Table 8 for DNA:RNA duplexes can be compared with predictions if corrections are applied for the expected difference between dU and dT. Studies have shown that the C5-methyl group renders T:A pairs 0.3 kcal/mol more stable, on average, than U:A pairs (Sugimoto et al., 2000; Wang et al, 1995b). As shown in Table 8, after corrections, the predictions correspond very well with experimental values (rmsdiNN = 0.45 kcal/mol and rmsdiss = 0.64 kcal/mol). Evidently, combining the INN or ISS models with the 0.3 kcal/mol penalty for changing dT to dU in DNA results in an accurate prediction of stabilities of Watson-Crick paired DNA:RNA duplexes. The agreement between measured and predicted stabilities of DNA:RNA duplexes and stability enhancements of PODN:RNA duplexes suggests that the stabilities of PODN:RNA duplexes can be predicted by using INN (Sugimoto et al., 1995) or ISS (Gray, 1997) models and the dTVdU correction (Sugimoto et al., 2000; Wang et al., 1995b) to calculate the stability of a DNA:RNA duplex, and then, using the long-range cooperativity model to predict the effects of various degrees of propynylation (Example 1). Comparison of ΔΔCP^MM) 'sfor DNA.-RNA Duplexes with Nearest Neighbor Model Predictions:

The ΔΔG°₃₇(MM) values in Table 9 can be compared with expectations from average trinucleotide parameters (Sugimoto et al., 2000). For example, Sugimoto et al. (2000) determined free energies for single internal dC:rA mismatches within all four permutations of G:C/C:G trinucleotide contexts. The values range from -0.7 to +0.9 kcal/mol and average -0.2 kcal/mol. After correcting for the loss of a dC:rG base pair on the basis of free energies of G:C/C:G nearest neighbor interactions (Sugimoto et al., 1995), the model predicts an average 4.8 kcal/mol mismatch penalty for dC:rA within a DNA:RNA duplex. The average ΔΔG°₃₇(MM- internal) for dC:rA in Table 9 is 5.0 kcal/mol (Figure 12), which is within experimental error of the predicted value.

Similarly, the average ΔΔG°₃₇(MM) for a single internal dU:rG pair within a DNA:RNA duplex is 0.2 kcal/mol as calculated for permutations of G:C/C:G trinucleotide contexts (Sugimoto et al., 1995; 2000). The average ΔΔG°₃₇(MM- internal) in Table 9 for dU:rG is 0.6 kcal/mol (Figure 12), again within experimental error of the predicted value. Evidently, ΔΔG°₃₇(MM)'s for dC:rA and dU,:rG can be predicted roughly from a nearest neighbor model even though the ΔΔG°₃ (MM)'s determined here were generated within different nearest-neighbor motifs. Similar comparisons can not be made for dU:rC and dC:rC mismatches because they were too destabilizing to be measured in a previous study (Sugimoto et al., 2000). Additional stability provided by 5' and 3' unpaired terminal ribonucleotides on the DNA:RNA duplexes studied here provide average values of 3.2 and 5.4 kcal/mol, respectively, for the ΔΔG°₃₇(MM)'s of dU:rC and dC:rC internal mismatches (Figure 12).

Comparing Mismatch Specificity of Linear PODNs to Circular Oligonucleotides:

Circular oligonucleotides provide some of the tightest binding and the highest specificity previously observed for nucleic acid hybridizations (Kool, 1991; Wang et al., 1994; Wang et al., 1995a; Prakash et al., 1991; Wang et al., 1995c). For example, at pH = 7.0 and 25 °C, the DNA 12-mer, (5'dAAGAAAGAAAAG3'), binds to a 34-mer circular DNA to give ΔG°₂₅ = -18.1 kcal/mol with a range in ΔΔG°₃₇(MM-internal) of 7.1-7.5 kcal/mol. At pH = 5.5, binding and specificity are enhanced to give ΔG°₂₅ = -25.5 kcal/mol with a range in ΔΔG⁰ ₃₇(MM-internal) of 8.2 - 10.4 kcal/mol (Wang et al., 1995). The results in Tables 8 and 9 and Figure 12 show that propynylated oligopyrimidine 7-mers provide similar binding and specificity with RNA 11-mers. The results suggest that a variety of approaches can be used to provide the specificity and tight binding that will optimize the many emerging applications of nucleic acid hybridizations.

Table 3: Free energies of 5' and 3' Term inal Unpaired N ucleotides^{3 Δ} G°₃₇ ^Δ G°₃₇

5' stack (kcal/mol) 3' stack (k cal/mol)

5'dAA/dU -0.5" dC/dGA3' -0.4"

5'rAA/rU -0.3^C rC/rGA3' -1 .1 °

5'rAA/dU -0.5 dC/rGA3' -1.3

5'rAA/dU^p -1 .9 dC^p/dGA3' -0.9

5'dCC/dG -0.5" dA/dTC3' -0.2"

5'rCC/rG -0.3^C rA/rUC3' -0.1 °

5'dCC/rG -0.7 rA/dUC-3' -0.2

5'dC^pC^p/rG -0.3 rA/dU^pC^p3' -0J

'Underlined nucleotides are unpaired dangling ends "frc m (65) 'from (1 1 , 46, 64)

Table 5: Thermodynamic Parameters of m -DNA Strands Bound to 3'- ^■GAGGAGGAAAU-5'.^{a b}

1/T_M Plots reference -^ΔG°₃₇ -ΔH° -Δs° τ_m symbol (kcal/mol) (kcal/mol) (eu) (°C)

d(5^,CCUCCU^pU^p3') m-DNA6,7 -10.4 + 0.1 -64.6 + 1.7 -174.6 + 5.3 56.9

d^'C^PCUCCUUS") m-DNA1 ,2 -1 1 .1 + 0.2 -68.3 + 3.1 -184.4 + 9.4 59.3

d(5^,CC^pUCC^pUU3") m-DNA2,5 -11.4 + 0.2 -67.3 + 2.3 -180.4 + 6.9 61.0

d(5'CCUC^pC^pUU3') m-DNA4,5 -11.8 + 0.3 -77.2 + 4.7 -211.1 + 14.2 59.6

d(5'C C^pUC^pC^pUU3') m-DNA1 ,2,4,5 -12.6 + 0.1 -71 .6 + 1.7 -190.2 + 5.1 65.8 d(5"C^pC^pU^pC^pC^pUU3') m-DNA1 , 2,3,4,5 -13.4 + 0.3 -70.1 + 3.0 -183.0 + 87 70.6 ^aBases at which C5-1 -propyne substitutions occur are in bold. ^bSee footnote a from Table 1.

Table 6: Thermodynamic Parameters of DNA, PODN , and s-PODNs Bound to r(3'GAGIAGGAAAU5')^a,b

1/T_M Plots reference -Δ_G°₃₇ -ΔH° -Δs° T_m symbol (kcal/mol) (kcal/mol) (eu) (°C)

d(5'C^pC^pU^pC^pC^pU^pU^p3') PODN -14.5 + 07 -75.5 + 6.2 -1967 + 8.0 73.8 d(5^,CCUCCUU3") DNA (-7.9 + 0.1 ) (-627 + 3.7) (-176.6 + 1 1.0) 44.3

d(5^,C^pC^pU^pCC^pU^pU^p3') S-PODN4 -13.9 + 0.4 -81 .5 + 37 -217.9 + 10.9 67.9

d(5O^pC^pU^pC^pCU^pU^p3') S-PODN5 -13.5 + 0.3 -77.2 + 3.2 -205.5 + 9.5 67.8

Positions of C5-(1 -propyne) c eletions within the PODN strand are in bold. ^bSee footnote a from Table 1

LIST OF REFERENCES CITED

Each of the references listed below or otherwise cited herein is hereby incoφorated by reference in its entirety.

Aboul-ela et al., Nucleic Acids Res. 1985, 13, 4811-4824.

Agrawal et al., Pharmacol. Ther. 1997, 76, 151-160.

Alivisatos et al., Nature 1996, 382, 609-611.

Allawi et al., Biochemistry 1997, 36, 10581-10594. Allawi et al., Biochemistry 1998a, 37, 9435-9444.

Allawi et al., Nucleic Acids Res. 1998b, 26, 2694-2701.

Allawi et al., Biochemistry 1998c, 37, 2170-2179.

Barnes et al., J. Am. Chem. Soc. 2001a, 123, 4107-4118.

Barnes et al, J. Am. Chem. Soc. 2001b, 123, 9186-9187. Bevilacqua et al., Biochemistry 1998, 37, 15877-15884.

Blake et al., Bioinformatics 1999, 15, 370-375.

Bonham et al., Nucleic Acids Res. 1995, 23, 1197-1203.

Bonnet et al., Proc. Natl. Proc. Sci. USA 1999, 4, 6171-6176.

Bommarito, S.; Peyret, N.; SantaLucia, J., Jr. Nucleic Acids Res. 2000, 28, 1929-1934.

Borer et al., J. Mol. Biol. 1914, 241, 246-262.

Borer, In Handbook of Biochemistry and Molecular Bio logy .Nucleic Acids

(Fasman, Ed.) 3^rd ed., CRC Press, Cleveland, OH., 1975, vol. I, p. 589.

Branch, A. D. TIBS 1998, 23, 45-50. Burkard et al., In The RNA World (Gesteland et al., Eds.) 2^nd ed., Cold Spring

Harbor Laboratory Press, Cold Spring Harbor, New York, 1999; pp 233-264.

Burkard et al., Biochemistry 2000, 39, 11748-11762.

Chaires, Biophys. Chem. 1997, 64, 15-23. Chalikian et al., Proc. Natl Acad. Sci. USA 1999a, 96, 7853-7858.

Chalikian et al., Biopolymers 1999b, 50, 459-471.

Chen et al., Nucleic Acids Res. 1999, 27,389-395.

Clark et al., Nucleic Acids Res. 1997, 25, 4098-4105. Crooke, Methods Enzymol 2000, 313, 3-45.

Damberger et al., Nucleic Acids Res. 1994, 22, 3508-3510.

De Clercq et al., J. Med. Chem. 1983, 26, 661-666.

Egli et al., Biochemistry 1996, 35, 8489-8494. Flanagan et al, Nucleic Acids Res. 1996, 24, 2936-2941.

Flanagan et al., Nat. Biotechnol 1999, 17, 48-52.

Freier et al., Biochemistry 1981, 20, 1419-1426.

Freier et al., Biochemistry 1985, 24, 4533-4539.

Freier et al., Proc. Natl Acad. Sci. USA 1986, 83, 9373-9377. Freier et al., Nucleic Acids Res. 1997, 25, 4429-4443.

Froehler et al., Tetrahedron Lett. 1992, 33, 5307-5310.

Gee et al., Antisense Nucleic Acid Drug Dev. 1998, 8, 103-111.

Gerry et al., J. Mol. Biol. 1999, 292, 251-262.

Gralla et al., J. Mol Biol 1973, 78, 310-319. Gray et al., Biopolymers 1978, 17, 85- 106.

Gray et al., Biopolymers 1981, 20, 1337-1382.

Gray, Biopolymers 1997, 42, 795-810.

Guckian et al., J. Am. Chem. Soc. 2000, 122, 2213-2222.

Gutell, Nucleic Acids Res. 1994, 22, 3502-3507. Gyi et al., Biochemistry 1996, 24, 12538-12548.

Hacia et al., Nucleic Acids Res. 1998, 26, 4975-4982.

Hare et al., Biochemistry 1986, 25, 7445-7456.

Hashem et al., Biochemistry 1998, 37, 61-72.

Healey et al., Anal Biochem. 1997, 251, 270-279. Herschlag, Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 6921-6925.

Hobbs, . Org. Chem. 1989, 54, 3420-3422.

Holbrook et al., Biochemistry 1999, 38, 8408-8422.

Hung et al., Nucleic Acids Res. 1994, 22, 4326-4334.

Iyer et al., J. Am. Chem. Soc. 1990, 112, 1253-1254. Kawase et al, Nucleic Acids Res. 1986, 14, 7727-7736.

Kierzek et al. Biochemistry 1999, 38, 14214-14223.

KooL J. Am. Chem. Soc. 1991, 113, 6265-6266.

Leonard et al., Nucleic Acids Res. 1990, 18, 5617-5623. Leone et al., Nucleic Acids Res. 1998, 26, 2150-2155.

Lesnik et al., Biochemistry 1995, 34, 10807-10815.

Lewis et al.,. Thermodynamics, 2ⁿ ed. Revised by Pitzer et al., McGraw-Hill, New York, 1961. Lewis et al., Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 3176-81.

, Liu et al., Anal. Chem. 1999, 71, 5054-5059.

Longfellow et al., Biochemistry 1990, 29, 278-285.

Maldonado-Rodriguez et al., Mol. Biotech. 1999, 11, 13-25.

Martin et al., Nucleic Acids Res. 1985, 13, 8927-8938. Mathews et al., J. Mol Biol. 1999a, 288, 911 -940.

Mathews et al., RNA 1999b, 5, 1458-1469.

Matteucci et al., Biotechnology 1981, 24, 92-98.

Mbindyo et al., Adv. Mater. 2001, 13, 249-254.

McClain et al., Science 1988, 240, 793-796. McDowell, et al, Biochemistry 1996, 35, 14077-14089.

Meroueth et al., Nucleic Acids Res. 1999, 27, 1118-1125.

Meyer, Data Analysis for Scientists and Engineers, Wiley & Sons, Inc., New York, 1975.

Milligan et al., J. Med. Chem. 1993, 36, 1923-1937. Mirkin et al., Nature 1996, 382, 607-609.

Morse et al., Nucleic Acids Res. 1995, 23, 302-306.

Moulds et al., Biochemistry 1995, 34, 5044-5053.

Musier-Forsyth et al., Science 1991, 253, 784-786.

Νeter et al., Applied Statistical Models, 2^nd ed., Richard D. Irwin, Inc., Homewood, IL., 1985.

Otikiti et al., Biochemistry 1997, 36, 11419-11427.

Pan et al., J. Mol. Biol.1998, 283, 977-984.

Pearson et al., Chem. & Biol 1997, 4, 409-414.

Peritz et al., Biochemistry 1991, 30, 6428-6436. Petersheim et al., Biochemistry 1983, 22, 256-263.

Peyret et al., Biochemistry 1999, 38, 3468-3477.

Prakash et al., J. Chem. Soc, Commun. 1991, 1161-1162. Press et al., Numerical Recipes in C 2nd ed., Cambridge Univeristy Press, Cambridge, England, 1992.

Puglisi et al., Methods Enzymol 1989, 180, 304-325.

Pyle et al., Biochemistry 1994, 33, 13856-13863. Ratmeyer et al., Biochemistry 1994, 33, 5298-5304.

Roberts et al., Proc. Natl. Acad. Set. U.S.A. 1991, 88, 9397-9401.

Richards, in Handbook of Biochemistry and Molecular Biology: Nucleic Acids (Fasman, Ed.) 3^rd ed., CRC Press, Cleveland, OH., 1975, vol I, p. 597.

Salazar et al., Biochemistry 1993, 32, 4207-4215. SantaLucia et al., Biochemistry 1991, 30, 8242-8251.

SantaLucia Jr., et al., Biochemistry 1996, 35, 3555-3562.

SantaLucia Jr., Proc. Natl. Acad. Sci. U. S. A. 1998, 95, 1460-1465.

Schena et al., Nucleic Acids Res. 1998, 26, 2150-2155.

Schena et al., Science 1995, 270, 467-470. Seber, Linear Regression Analysis, John Wiley & Sons, Inc., New York, 1977.

Seeman, Anna. Rev. Biomol. Struct. 1998, 27, 225-248.

Sokol et al., Proc. Natl Acad. Sci. U S. A. 1998, 95, 11538-11543.

Stawinski et al., Nucleosides Nucleotides 1988, 7, 779-782.

Stofer et al., J. Am. Chem. Soc. 1999, 121, 9503-9508. Sugimoto et al., Biochemistry 1995, 34, 11211-11216.

Sugimoto et al., Biochemistry 2000, 39, 11270-11281.

Tanaka et al., Biochemistry 2000, 39, 7074-7080

Taton et al., Science 2000, 289, 1757-1760.

Testa et al., Biochemistry 1999, 38, 16655-16662 Turner et al., J. Am. Chem. Soc. 1987, 109, 3783-3785.

Turner et al., Annu. Rev. Biophys. Biophys. Chem. 1988, 17, 167-192.

Turner, In Nucleic Acids: Structures, Properties, and Functions (Bloomfield et al., Eds.) University Science Books, Sausolito, CA., 2000; pp 259-336.

Usman et al., J. Am. Chem. Soc. 1987, 109, 7845-7854. Valko et al, J. Liq. Chromatography 1989, 12, 2103-2116.

Verma et al., F. Anna. Rev. Biochem. 1998, 67, 99-134.

Wagner et al., Science 1993, 260, 1510-1513.

Wagner et al., Nat. Biotechnol 1996, 14, 840-844. Wagner et al., Science 1993, 260, 1510-1513. Wagner et al., Nαt. Biotechnol 1996, 14, 840-844. Walt, Science 2000, 287, 451-452. Walton et al., Biotechnol Bioeng. 1999, 65, 1-9. Wang et al., Nucleic Acids Res. 1994, 22, 2326-2333.

Wang et al., Biochemistry 1995a, 34, 9774-9784. Wang et al., Biochemistry 1995b, 34, 4125-4132. Wang et al., Nucleic Acids Res. 1995c, 23, 1157-1164. Wetmur, Critical Reviews Biochem. Mol. Biol 1991, 26, 227-259. Wincott et al., Nucleic Acids Res. 1995, 23, 2677-2684.

Xia et al., Biochemistry 1998, 42, 14719-14735. Zamecnik et al., Proc. Natl. Acad. Sci. USA 1978, 75, 280-284. Zhong et al., Biochemistry 1997, 36, 2485-2491. Zhu et al., Biochemistry 1997, 36, 15326-15335.

Although preferred embodiments have been depicted and described in detail herein, it will be apparent to those skilled in the relevant art that various modifications, additions, substitutions, and the like can be made without departing from the spirit of the invention and these are therefore considered to be within the scope of the invention as defined in the claims which follow.

Claims

What is Claimed:

1. An oligonucleotide comprising: a first nucleotide comprising at least one alkynl functional group at a C5 position of a pyrimidine heterocyclic base, and at least one second nucleotide covalentiy bound to the first nucleotide and comprising at least one alkynyl functional group at a C5 position of a pyrimidine heterocyclic base; wherein the oligonucleotide comprises a loss in free energy of at least about 2 kcal/mol when

(a) an alkynyl group of the first nucleotide is removed from the C5 position of the pyrimidine heterocyclic base and

(b) the oligonucleotide is covalentiy or non-covalently bound to a nucleic acid molecule comprising a nucleotide sequence that is substantially Watson-Crick complementary to a sequence of the oligonucleotide.

2. The oligonucleotide according to claim 1 wherein the oligonucleotide does not bind to SV40 TAg mRNA.

3. The oligonucleotide according to claim 1 wherein the oligonucleotide comprises a sequence of at least seven nucleotides comprising the first nucleotide, two second nucleotides, and at least three other nucleotides comprising at least one alkynyl functional group at a C5 position of a pyrimidine heterocyclic base.

4. A duplex comprising a nucleic acid molecule and an oligonucleotide according to claim 1 hybridized to the nucleic acid molecule.

5. An oligonucleotide comprising: a first nucleotide comprising at least one alkynyl functional group at a C5 position of a pyrimidine heterocyclic base, and at least one second nucleotide covalentiy bound to the first nucleotide and comprising at least one alkynyl functional group at a C5 position of a pyrimidine heterocyclic base; wherein the oligonucleotide comprises a loss in free energy of at least about 2.8 kcal/mol when (a) the oligonucleotide is covalentiy or non- covalently bound to a nucleic acid molecule comprising a nucleotide sequence that is less than substantially Watson-Crick complementary to a sequence of the oligonucleotide and (b) the first nucleotide of the oligonucleotide is covalentiy or non- covalently bound to a nucleotide of the nucleic acid molecule which is not a Watson- Crick base pairing nucleotide for the first nucleotide.

6. The oligonucleotide according to claim 5 wherein the oligonucleotide does not bind to SV40 TAg mRNA.

7. The oligonucleotide according to claim 5 wherein the oligonucleotide comprises a sequence of at least seven nucleotides comprising the first nucleotide, two second nucleotides, and at least three other nucleotides comprising at least one alkynyl functional group at a C5 position of a pyrimidine heterocyclic base.

8. A duplex comprising a nucleic acid molecule and an oligonucleotide according to claim 5 hybridized to the nucleic acid molecule.

9. A method of designing an oligonucleotide capable of interfering with the function of a target nucleic acid molecule, said method comprising: identifying a target sequence of a target nucleic acid molecule and preparing an oligonucleotide comprising a nucleotide sequence that is substantially Watson-Crick complementary to the target sequence, the oligonucleotide including 6 or more adjacent nucleotide bases that are alkynylated in a manner which more favorably stabilizes the interaction of the oligonucleotide with the target nucleic acid molecule as compared to a second oligonucleotide that includes the same nucleotide sequence but lacks the 6 or more adjacent bases that are alkynylated.

10. The method according to claim 9, further comprising: determining the location of the 6 or more adjacent bases that will most favorably stabilize the interaction of the oligonucleotide with the target nucleic acid.

11. The method according to claim 10 wherein said determining comprises: assessing the free energy potential of two or more sequences, each comprising 6 or more adjacent alkynylated bases for hybridization to the target nucleic acid.

12. A method of interfering with the activity of a target nucleic acid molecule comprising: introducing into an in vitro or in vivo system, which includes a target nucleic acid molecule, an amount of an oligonucleotide according to claim 1 which is effective to bind to the target nucleic acid molecule in a manner sufficient to interfere with any activity thereof.

13. The method according to claim 12, wherein target nucleic acid molecule is an RNA molecule, a DNA molecule, or a natural or unnatural molecule of related structure.

14. The method according to claim 12, wherein the target nucleic acid molecule is not SV40 TAg mRNA.

15. A method of interfering with the activity of a target nucleic acid molecule comprising: introducing into an in vitro or in vivo system, which includes a target nucleic acid molecule, an amount of an oligonucleotide according to claim 5 which is effective to bind to the target nucleic acid molecule in a manner sufficient to interfere with any activity thereof.

16. The method according to claim 15, wherein target nucleic acid molecule is an RNA molecule, a DNA molecule, or a natural or unnatural molecule of related structure.

17. The method according to claim 15, wherein the target nucleic acid molecule is not SV40 TAg mRNA.

18. A microarray detection device comprising: a substrate and a plurality of oligonucleotides bound to the substrate, each of the oligonucleotides comprising at least 6 nucleotide bases wherein 6 or more adjacent nucleotide bases of each are alkynylated.

19. The microarray detection device according to claim 18, wherein each of the plurality of oligonucleotides comprises a nucleotide sequence that is the same.

20. The microarray detection device according to claim 18, wherein the plurality of oligonucleotides comprises two or more oligonucleotides each having a nucleotide sequence that is different.

21. The microarray detection device according to claim 18, wherein the plurality of oligonucleotides comprises at least one set of oligonucleotides that hybridize to a first target nucleic acid molecule.

22. The microarray detection device according to claim 21 , wherein the plurality of oligonucleotides further comprises at least two sets of oligonucleotides that hybridize, respectively, to first and second target nucleic acid molecules.

23. A method of identifying an oligonucleotide having binding affinity for a target nucleic acid molecule comprising: introducing a target nucleic acid molecule to a microarray detection device according to claim 18 under conditions effective for hybridization of substantially complementary sequences between the target nucleic acid molecule and the oligonucleotide; and detecting whether hybridization occurs between the target nucleic acid molecule and one or more of the plurality of oligonucleotides bound to the substrate.

24. The method according to claim 23, further comprising: identifying the nucleotide sequence and position of alkynylated bases in an oligonucleotide that hybridized to the target nucleic acid molecule.

25. A method of detecting the presence of a target nucleic acid molecule in a sample comprising: passing a sample over a microarray according to claim 18 under conditions suitable for hybridization to occur between oligonucleotides and target nucleic acid molecules and determining whether any target nucleic acid molecules hybridized to oligonucleotides during said passing.

26. The method according to claim 25, wherein the target nucleic acid molecule is specific for a pathogen.

27. A method of detecting the localization of a target nucleic acid molecule in an in vitro or in vivo system, said method comprising: introducing into an in vitro or in vivo system a labeled oligonucleotide including a nucleotide sequence which is substantially complementary and specific to a nucleotide sequence of a target nucleic acid molecule and has 6 or more adjacent nucleotide bases that are alkynylated; allowing sufficient time for the labeled oligonucleotide to hybridize with the target nucleic acid molecule; and determining the location of the labeled oligonucleotide in the system, the location of the labeled oligonucleotide being the same as the location of the target nucleic acid molecule.

28. A method of making a product, said method comprising: introducing into a reaction medium a first nucleic acid molecule having bound thereto a first molecule or compound and a second nucleic acid molecule having bound thereto a second molecule or compound, the first and second nucleic acid molecules comprising substantially complementary nucleotide sequences that hybridize in the reaction medium and at least one of the first and second nucleic acid molecules comprising at least six adjacent alkynylated bases, wherein hybridization of the first and second nucleic acid molecules brings the first molecule or compound into sufficient proximity to the second molecule or compound for the first and second molecules or compounds to form a product.

29. The method according to claim 28, wherein both the first and second nucleic acid molecules comprise at least two adjacent alkynylated bases.

30. A self-assembling system for preparing a product, the system comprising: a first nucleic acid molecule comprising a first nucleotide sequence, the first nucleic acid molecule having bound thereon a first molecule or compound; and a second nucleic acid molecule comprising a second nucleotide sequence which is substantially complementary to the first nucleotide sequence, the second nucleic acid molecule having bound thereon a second molecule or compound; wherein at least one of the first and second nucleic acid molecules comprises at least two adjacent alkynylated bases, and wherein upon introduction of the first and second nucleic acid molecules into a reaction medium suitable for hybridization thereof, the first and second molecules or compounds are capable of self-assembly to form a product.

31. The self-assembling system according to claim 30, wherein both the first and second nucleic acid molecules include at least six alkynylated bases.