AU679111B2 - Purine oligomers - Google Patents

Description

-r- 1 Backaround Of The Invention Publications and other reference materials referred to herein are incorporated herein by reference and are numerically referenced in the following text and respectively grouped in the appended Bibliography which immediately precedes the claims.

The present invention is directed to novel methods of 10 detecting and locating specific sequences in double stranded DNA using nucleoside oligomers which are capable of specifically complexing with a selected double stranded DNA structure to give a triple helix structure.

Formation of triple helix structures by homopyrimidine oligodeoxyribonucleotides binding to polypurine .:tracts in double stranded DNA by Hoogsteen hydrogen bonding has been reported. (See, e.g. and The homopyrimidine oligonucleotides were found to recognize extended purine sequences in the major groove of double helical DMA via triple helix formation. Specificity was found to be imparted by Hoogsteen base pairing between the homopyrimidine oligonucleotide and the purine strand of the Watson-Crick duplex DNA. Triple helical complexes containing cytosine and thymidine on the Hoogsteen (or third) strand have been found to be stable in acidic to neutral solutions, respectively, but. have been found to dissociate on increasing pH. Incorporation of modified bases of T, such as 5-bromo-uracil and C, such as methylcytosine, into the Hoogsteen strand has been found I: P I to increase stability of the triple helix over a higher pH range. In order for cytosine to participate in the Hoogsteen-type pairing, a hydrogen must be available on the 3-N of the pyrimidine ring for hydrogen bonding.

Accordingly, in some circumstances, C may be protonated at N-3.

DNA exhibits a wide range of polymorphic conformations, such conformations may be essential for biological processes. Modulation of signal transduction by sequence-specific protein-DNA binding and molecular interactions such as transcription, translation, and replication, are believed to be dependent upon DNA conformation. (3) It is exciting to consider the possibility of developing therapeutic agents which bind to critical regions of the genome and selectively inhibit the function, replication, and survival of abnormal cells. (4) The design and development of sequence-specific DNA binding molecules has been pursued by various laboratories 20 and has far-reaching implications for the diagnosis and treatment of diseases involving foreign genetic materials (such as viruses) or alterations in genomic DNA (such as cancer).

Nuclease-resistant nonionic oligodeoxynucleotides (ODN) consisting of a methylphosphonate (MP) backbone have been studied in vitro and in vivo as potential anticancer, antiviral an. ,,~ibacterial agents. The linked internucleotide bonds of these analogs closely approximate the conformation of nucleic acid phosphodiester bonds.

30 The phosphate backbone is rendered neutral by methyl substitution of one anionic phosphoryl oxygen; decreasing inter- and intrastrand repulsion due to the charged phosphate groups. Analogs with MP backbone can penetrate living cells and have been shown to inhibit mRNA translation in globin synthesis and vesicular stomatitis viral protein synthesis, and inhibit splicing of pre-mRNA in inhibition of HSV replication. Mechanisms of action

~C~

for inhibition by the nonionic analogs include formation of stable complexes with complementary RNA and/or DNA.

Nonionic oligonucleoside alkyl- and aryl-phosphonate analogs complementary to a selected single stranded foreign nucleic acid sequence can selectively inhibit the expression or function or expression of that particular nucleic acid without disturbing the function or expression of other nucleic acids present in the cell, by binding to or interfering with that nucleic acid. (See, e.g. U.S.

Patentt- No. 4,469,863 and 4,511,717). The use of complementary nuclease-resistant nonionic oligonucleoside methylphosphonates which are taken up by mammalian cells to inhibit viral protein synthesis in certain contexts, including Herpes simplex virus-1 is disclosed in U.S.

Patent No. 4,757,055.

The use of anti-sense oligonucleotides or phosphorothioate analogs complementary to a part of viral mRNA to interrupt the transcription and translation of viral mRNA into protein has been proposed. The anti-sense constructs 20 can bind to viral mRNA and were thought to obstruct the cells ribosomes from moving along the mRNA and thereby halting the translation of mRNA into protein, a process called "translation arrest" or "ribosomal-hybridization arrest." (6) S 25 The inhibition of infection of cells by HTLV-III by administration of oligonucleotides complementary to highly conserved regions of the HTLV-III genome necessary for HTLV-III replication and/or expression is disclosed in U.S. Patent No. 4,806,463. The oligonucleotides were 30 found to affect viral replication and/or gene expression as assayed by reverse transcriptase activity (replication) and production of viral proteins p15 and p24 (gene expression) The ability of some antisense oligodeoxynucleotides containing internucleoside methylphosphonate linkages to inhibit HIV-induced syncytium formation and expression has been studied.

I

Psoralen-derivatized oligonucleoside methylphosphonates have been reported capable of cross-linking either coding or non-coding single stranded DNA; however, double stranded DNA was not cross-linked. (28) Summary Of The Invention The present invention is directed to methods of detecting or recognizing a specific segment of double stranded DNA or double stranded DNA sequence and to methods of preventing- expression-or function of a specific segment of double stranded DNA having a given sequence.

This is achieved without destroying the intactness of the duplex structure or interrupting the base-pairing of the double stranded DNA. The present invention is also directed to novel modified Oligomers which are useful for preventing expression and/or functioning of a selected double stranded DNA sequence and which optionally include a DNA modifying group. Additionally, the present invention is directed to novel Oligomers which comprise cytoo.o. sine analogs. The present invention is also directed to formation of a triple helix structure by the interaction of a specific segment of double stranded DNA and an Oligomer which is sufficiently complementary to the DNA segment to read it and base pair (or hybridize) thereto.

Accordingly, in one aspect, the present invention is directed to methods of detecting or recognizing a specific segment of double stranded DNA which comprises contacting said segment of DNA with an Oligomer which is sufficiently complementary to the sequence of purine bases in said segment of double stranded DNA or a portion thereof to 30 hydrogen bond (or hybridize) therewith thereby giving a triple helix structure.

In another aspect, the present invention is directed to methods of preventing or inhibiting expression or function of a specific segment of double stranded DNA having a given sequence which comprises contacting said DNA segment with an Oligomer sufficiently complementary to said double stranded DNA segment to form hydrogen bonds therewith, thereby giving a triple helix structure.

The present invention is directed to methods wherein the DNA segment comprises a gene in a living cell and wherein formation of the triple helix structure permanently inhibits or inactivates said gene.

In a preferred aspect, said Oligomer is modified to incorporate a DNA modifying group which, a.ter the Oligomer hydrogen bonds or hybridizes with the selected DNA sequence, is caused to react chemically with the DNA and irreversibly modify it. Such modifications may include cross-linking Oligomer and DNA by forming a covalent bond thereto, alkylating the DNA, cleaving said DNA at a specific location, or by degrading or destroying the DNA.

In another aspect, the present invention provides novel nonionic alkyl- and aryl-phosphonate Oligomers which are sufficiently complementary to the purine sequence of a specific double stranded DNA segment to hydrogen bond 20 and form a triple helix structure. Preferred are nonionic methylphosphonate Oligomers.

One preferred class of methylphosphonate Oligomers comprises only purine bases.

The present invention also provides Oligomers having S 25 nucleosidyl units in which a cytosine analog replaces cytosine and wherein said cytosine analog comprises a heterocycle which has a hydrogen available for hydrogen bonding at the ring position which corresponds to N-3 of cytosine and which is capable of forming two hydrogen bonds with a guanine base at neutral pH.

In another aspect of the present invention, Oligomers which comprise a modified sugar moiety having a lengthening link between the 5'-C and the 5'-OH and/or the 3'-C and 3'-OH are provided. These Oligomers comprising such modified sugar moieties may read purine bases on both strands of a segment of double stranded DNA.

Definitions As used herein, the following terms have the following meanings unless expressly stated to the contrary.

The term "Oligomer" refers to nucleosidyl units which are connected by internucleosidyl phosphorus group linkages and includes oligonucleotides, nonionic oligonucleoside alkyl- and aryl-phosphonate analogs, phosphorothioate analogs of oligonucleotides, phosphoamidate analogs of oligonucleotides, neutral phosphate ester oligonucleotide analogs, and other oligonucleotide analogs and modified oligonucleotides.

The term "phosphonate" refers to the group 0 P-R

I

wherein R is an alkyl or aryl group. Suitable alkyl or aryl groups include those which do not sterically hinder the phosphonate linkage or interact with each other. The phosphonate group may exist in either an or an "S" :"configuration. Phosphonate groups may be used as inter- 20 nucleosidyl phosphorus group linkages (or links) to Sconnect nucleosidyl units.

The term "phosphodiester" refers to the group 0 P- 0e. Phosphodiester groups may be used as internucleosidyl phosphorus group linkages (or links) to connect 25 nucleosidyl units.

The term "alkyl or aryl-phosphonate Oligomer" refers to nucleotide Oligomers having internucleosidyl phosphorus •group linkages wherein at least one alkyl or aryl phosphonate internucleosidyl linkage replaces a 30 phosphodiester internucleosidyl linkage.

The term "methylphosphonate Oligomer" (or "MP- Oligomer") refers to nucleotide Oligomers (or oligonucleotide analogs) having internucleosidyl phosphorus group linkages wherein at least one methylphosphonate internucleosidyl linkage replaces a phosphodiester internucleosidyl linkage.

The term "Third Strand Oligomer" refers to Oligomers which are capable of reading a segment of double stranded DNA and forming a triple helix structure therewith.

The term "nucleoside" includes a nucleosidyl unit and is used interchangeably therewith and includes not only units having A, G, C, T and U as their bases, but also analogs and modified forms of the bases (such as 8substituted purines).

The term "complementary" when referring to an oligomer Third Strand refers to Oligomers having base sequences which hydrogen bond (and base pair or hybridize) with the purine base of a corresponding (Watson-Crick) base pair of a double stranded DNA to form a triple helix structure.

In the various Oligomer sequences listed herein in, as in ApA represents a phosphodiester linkage, and in, as in CEG represents a methylphosphonate linkage. Also, notation such as indicates nucleosidyl groups linked by methyl phosphonate linkages.

The term "purine" or "purine base" includes not only the naturally occurring adenine and guanine bases, but also modifications of those bases such as bases substituted at the 8-position.

The term "read" refers to the ability of a nucleic acid to recognize through hydrogen bond interactions the base sequence of another nucleic acid. Thus, in reading a double stranded DNA sequence, a Third Strand Oligomer is able to recognize through hydrogen bond interactions the base pairs, in particular the purine bases, in the duplex 30 of a segment of double stranded DNA.

The term "triplet" refers to a situation such as that depicted in Figures 1A, 1B and 2A to 2D wherein a base in the Third Strand has hydrogen bonded (and thus base paired) with a (Watson-Crick) base pair of a segment of double stranded DNA.

I

Brief Description Of The Drawings Fig. 1A and 1B depict triplets wherein a pyrimidine base in the Third Strand forms a triplet with the duplex DNA (Watson-Crick) base pair.

Fig. 2A to 2D depict triplets wherein a purine base in the Third Strand forms a triplet with the duplex DNA (Watson-Crick) base pair.

Fig. 3A to 3C depict the base sequences of exemplary polypurine sequence triple helix structures wherein the Third Strand "reads" and, thus, base pairs with purine bases on one strand of the double stranded DNA.

Fig. 4A to 4E depict the base sequences of exemplary mixed sequence triple helix structures wherein the Third Strand "reads" and, thus base pairs with purine bases on both strands of the double stranded DNA.

Fig. 5 depicts a nucleosidyl unit having a modified sugar moiety with an alkyleneoxy link for lengthening internucleoside phosphorus linkages and processes for its preparation.

20 Fig. 6 depicts a nucleosidyl unit having a modified sugar moiety with an alkylene link for lengthening internucleoside phosphorus linkages and processes for its synthesis.

Fig. 7 depicts CD spectra of triple helix structures, depicts a MP-Oligomer Third Strand, depicts an Oligonucleotide Third Strand.

Figs. 8A and 8B depict cross-linking of single stranded DNA and double stranded DNA using psoralenderivatized MP-Oligomers.

Detailed Description Of The Invention The present invention involves the formation of triple helix structures with a selected double stranded DNA sequence by contacting said DNA with an Oligomer which is sufficiently complementary to the purine sequence of the double stranded DNA to form hydrogen bonds (or hybridize) therewith.

General Aspects In addition to other aspects described herein, this invention includes the following aspects.

A first aspect concerns the reading (or recognition) of the base pairs in the double stranded DNA segment without opening the base pair, through the hydrogen bond formation by the bases in the Third Strand with the extra hydrogen bond sites of purines in the double stranded DNA, such as adenine and guanine. In other words, the reading of the base pair sequence in the DNA duplex is always done by reading the purine of the base pair, through hydrogen bond formation with the bases in the Third Strand, with the remaining available hydrogen bonding sites of the purines in the DNA. Either purines (such as adenine (A) or guanine or pyrimidines (thymine or cytosine in the Third Strand can form hydrogen bonds with the purines in the DNA, i.e.: i) Adenine in the base pair of DNA can be read or hydrogen bonded with A or T in the third strand.

20 ii) Guanine in the base pair of DNA can be read or paired with C or G in the third strand.

In a preferred aspect of the present invention, the phosphorus-containing backbone of the Third Strand :comprises methylphosphonate groups as well as naturally 25 occurring phosphodiester groups.

(II) A second aspect of this invention is to read the purine (A or G) in the double stranded DNA totally by the use of purine bases from the Third Strand. In other words, this aspect of the present invention uses only G to 30 read G in the base pairs in the DNA duplex, and similarly to use only A to read A in the base pairs in the DNA duplex. The polarity of the strands, either the antiparallel or parallel direction of the Third Strand in respect to the strand containing the purine to be read in the duplex is important.

The base planes of the purines and pyrimidines are rigid, and the furanose ring only allows a small ripple

~C_

(about 0.5A above or below the plane). Thus, the conformational state of the nucleoside is defined principally by the rotation of these two more or less rigid planes, i.e. the base and the pentose, relative to each other about the axis of the C'-l to N-9 or N-1 bond. The sugar-base torsion angle, has been defined as "the angle formed by the trace of the plane of the base with the projection of the C-'1 to 0-1' bond of the furanose ring when viewed along the C'-1 to a bond. This angle will be taken as zero when the furanose-ring oxygen is antiplanar to C-2 of the pyrimidine or purine ring and positive angles will be taken as those measured in a clockwise direction when viewing C-1' to This angle has also been termed the glycosyl torsion angle. Using the above definition, it was concluded that there were two ranges of c, for the nucleosides, about -300 for the anti conformation and about +1500 for the syn conformation.

The range for each conformation is about ±450. (22, 22a) Other researchers have used or proposed slightly different definitions of this angle. (23,24,25,26,27) Information concerning 0cm has been obtained using procedures such as X-ray diffraction, proton magnetic resonance (PMR) and optical rotatory dispersion-circular dichroism (ORD-CD).

(22a) 25 In order to accommodate the change of location of the purine base to be read from one strand (termed the "Watson strand") to the opposite strand (termed the "Crick strand"), we have recognized that a particular conformation of the nucleoside, defined by the torsion angle of the glycosyl bond, of the purine nucleosidyl unit in the Third Strand is required in order that the purine nucleoside in the Third Strand can be used to read the purine in the duplex. In other words, in order to read the purine bases in the DNA, the conformations of the purine nucleosidyl units in the Third Strand are influenced by the polarity (parallel to or anti-parallel to direction) of the strand containing the purine bases to be read in the DNA in relation to the Third Strand. For the purine nucleosidyl units in the Third Strand, reading the purine in the parallel strand in the duplex, the conformation of the purine nucleosidyl unit in the Third Strand should be in the syn conformation. On the other hand, the conformation of the purine nucleosidyl unit in the Third Strand in reading the corresponding purine in the anti-pox.allel strand in the duplex, should exist in anti conformation. Thus, in reading the purine bases in the duplex distributed-in both strands, one has the choice of using a Third Strand which has the same polarity as (i.e.

is parallel to) either one strand or the opposite strand.

As an example, a.

*a.a *a*a *e e

_M

(these 2 strands are anti-parallel to each other) (Watson strand) (Crick strand) The sequence of the Third Strand in the triplet with the DNA duplex having the same polarity of the Watson strand from 5' to 3' would be as follows:

GS'

I

Aanti 3' (the Third Strand parallel to the Watson strand) The sequence of a Third Strand parallel to the Crick strand would be as follows: Ganti 31

G

5

Y'

ASY"

Anti (the Third Strand parallel to the Crick strand) I According to one aspect of this invention, in constructing the Third Strand forreading the purines in the base pairs of the duplex, the following guidelines apply: Starting from the 5' end toward the 3' end, the purine nucleosidyl units (A or G) of the Third Strand need to be in syn conformation in reading the purines in the base pair (A or G) of the parallel ("Watson") strand of the double stranded DNA. In reading the second purine in the second base pair, the same requirement applies if the purine is located in the same strand as the first purine.

However, if the second purine is located in the opposite anti-parallel ("Crick") strand (now the opposite strand is anti-parallel to the Third Strand), the purine nucleosidyl unit needs to be in anti conformation. In all cases, adenine in the Third Strand is used to read adenine in the duplex and guanine in the Third Strand is used to read guanine in the duplex.

(ii) A third aspect of this invention concerns the length of the linkage of the phosphorus backbone of the Third Strand to allow reading of the purine bases on either strand of the double stranded DNA. In order to be able to "read" (or base pair) with purine bases on either strand, the distance between nucleosidyl units along the 25 phosphorus backbone must be increased. Two examples of types of lengthening link formats for the phosphorus backbone are proposed. One type of link format for the phosphorus backbone would use a universal lengthening link on the individual nucleosidyl units, i.e. all the length- 30 ening links of the Third Strand would be the same. Such a universal link format is particularly suitable for Third Strands comprising only purine bases. Accordingly, the length of the link between the 5' carbon of the "nucleosidyl unit one" to the 3' oxygen of the subsequent "nucleosidyl unit two" may be increased by two atoms (such as -CH 2

CH

2 or by 3 atoms (such as -O-CH 2 thereby lengthening the linkage between individual nucleosidyl ILI I I units by 2 to 6A. In order to allow an appropriate distance between nucleosidyl units, we recommend that separation of the units be increased by a number of atoms ranging from 1 to 6. Figure 6 illustrates an example of a nucleosidyl unit comprising this lengthening link format and proposed synthetic routes.

A second lengthening link format for the phosphorus backbone would comprise non-uniform lengthening links.

Links having internucleosidyl distances on the order of the standard phosphodiester backbone for the Third Strand would be employed when the purines being read were on the same strand, while a lengthened link (15-17A in length) which could comprise lengthening links on the 3' carbon of one nucleosidyl unit and on the carbon of its neighbor, would be employed to read the purine bases located on opposite strands. Such a non-uniform lengthening link format would be particularly suitable for use in Third Strands comprising both pyrimidine (or pyrimidine analog) and purine bases.

20 Formation of Triple Helix Structures Triplet (or Triple-Stranded) base pairing Pyrimidine in Third Strand Forming Triplet In one aspect of the present invention, triplets are formed wherein a pyrimidine base in the Third Strand forms 25 hydrogen bonds and, thus, base pairs with the purine base of a base pair of the double stranded DNA sequence.

Poo. Examples of such triplets where having a pyrimidine base as the base in the Third Strand are shown in Figures 1A and lB.

30 Fig. 1A depicts a triplet having T as the Third Strand base which forms hydrogen bonds and, thus, base pairs with the A of the double stranded DNA. In such circumstances, the T of the Third Strand is aligned parallel to the A-containing strand of the double stranded DNA, and anti-parallel to the T-containing strand of the double stranded DNA (which is also anti-parallel to the Acontaining strand). Accordingly, the sequence for that triplet is written as follows:

TA

Fig. 1B depicts a triplet having a protonated cytosine as the Third Strand base which forms hydrogen bonds and, thus, base pairs with a G of a G-C base pair in the double stranded DNA. In such circumstance, the cytosine base in the Third Strand must be protonated at N3 in order to form hydrogen bonds necessary for a stable triplet. Optionally, a cytosine may be replaced with a cytosine analog having a quaternary nitrogen at a position analogous to N3. In the triplet depicted, the C+ (or its analog) of the Third Strand is aligned parallel to the G-containing strand of the double stranded DNA, and anti-parallel to the C-containing strand of the double 15 stranded DNA (which is also anti-parallel to the G-containing strand). Accordingly, such a triplet sequence is written as follows: C eo\ T i Purine in Third Strand Forming Triplet In another aspect of the present invention, triplets are formed wherein a purine base in the Third Strand forms hydrogen bonds with and, thus, base pairs with the same purine base in one strand of the double stranded DNA.

Accordingly, A will pair with A and G will pair with G.

In contrast to the pyrimidine Third Strand triplets, the purine base in the Third Strand is capable of base pairing with the purine base of double stranded DNA such that the strand polarity of the Third Strand containing the purine base may be aligned either parallel or antiparallel to strand polarity of the strand containing the purine to be read in the double stranded DNA.

For example, Fig. 2A depicts a triplet where the A of the Third Strand is aligned parallel to the A of the double stranded DNA, and anti-parallel to the T of the double stranded DNA. In such a circumstance, the glycosyl torsion angle of the Third Strand A is in the syn conformation, and the glycosyl torsion angles of the A and 20 T bases of the double stranded DNA are both in the anti .1 conformation. Such a triplet sequence is written as follows: om° e II 9

_CL

Alternatively, Fig. 2B depicts a triplet where the A of the Third Strand is aligned anti-parallel to the A of the double stranded DNA and parallel to the T of the double stranded DNA. In such circumstance, the glycosyl torsion angles of all three bases are in the anti conformation. Such a triplet sequence is written as follows: Vd A L-L 0O

C

S

S

C

S S 0

S.

C

S. C t o C C Ira~ II.- Fig. 2C depicts a triplet wherein the G of the Third Strand is aligned parallel to the-G of the double stranded DNA, and anti-parallel to the C of the double stranded DNA. In such circumstance, the glycosyl torsion angle of the Third Strand G is in the syn conformation, and the glycosyl torsion angles of the G-C bases of the double stranded DNA are both in the anti conformation. Such a triplet sequence is written as follows:

C-G

Fig. 2D depicts a triplet wherein the G of the Third Strand is aligned anti-parallel to the G of the double stranded DNA and parallel to the C of the double stranded DNA. In such circumstance, the glycosyl torsion angles for all three bases are in the anti conformation. Such a triplet sequence is written as follows: ol •r '*eeooe ~CIRI~ I -I- Polypurine Sequences Where one strand of a selected double stranded DNA sequence comprises a polypurine sequence, an Oligomer complementary to the polypurine sequence is used and may comprise a complementary sequence of pyrimidines or purines or a mixture thereof. Preferred Oligomers, include nonionic MP-Oligomers which are nuclease resistant and are capable of forming the previously-discussed triplet structures with the double stranded DNA.

Examples of triple stranded helix sequences wherein the Third Strand binds to a polypurine sequence in one strand of a double stranded DNA sequence are schematically depicted in Fig. 3A to 3C. In Figures 3A to 3C, C denotes a protonated C.

Mixed Sequences Mixed sequences are sequences wherein the purine bases of the selected DNA sequence are located on both strands of the double stranded DNA sequence. Accordingly, oe..

the Third Strand Oligomer must be able to hydrogen bond 20 and, thus, base pair with the purine base of each base pair of the double stranded DNA sequence without regard to which strand of the double stranded DNA the purine base is located on. Accordingly, the Third Strand Oligomer must be able to "read" across the double stranded DNA. Ex- :i* amples of mixed sequences including an appropriately complementary Third Strand are depicted in Fig. 4A to 4E.

Fig. 4A depicts a mixed sequence having a pyrimidinerich Third Strand.

Fig. 4B depicts a mixed sequence having a purinerich Third Strand.

Fig. 4C, 4D and 4E depict mixed sequences having Third Strands containing only purine bases. In Figures 4C to 4E, denotes s~Y and denotes anti; no superscript denotes anti.

Where the Third Strand Oligomer must "read" across the double strand from one strand to the opposite strand YIII~ II L -r in order to base pair with the purine base, it may be advantageous to provide a lengthened internucleosidyl phosphorus linkage by incorporation of the previouslydescribed backbone link formats into the phosphorus backbone which connects the sugar moieties of the nucleosidyl units.

For example, an internucleosidyl linkage may be lengthened by the interposition of an appropriate alkylene

(-(CH

2 or alkyleneoxy (-(CH 2 lengthening link between the 5'-carbon and the 5' hydroxyl of the sugar moiety of a nucleosidyl unit or a similar link between the 3'-carbon and the 3'-hydroxyl. (See Figures 5 and 6).

Where indicated, such lengthening links may be interposed at both the and carbons of the sugar moiety.

Where consecutive purines of the selected double stranded DNA sequence occur on the same strand, such lengthening links need not be amployed; internucleosidyl phosphorus linkages such as methylphosphonate linkages allow an appropriate base on the Third Strand to read 20 consecutive purine bases on one strand of the selected double stranded DNA sequence.

Accordingly, by employing such lengthening links where indicated, Oligomers which are capable of reading with the purine bases on both strands of a selected double stranded DNA sequence, may be prepared.

Third Strand Oliqomers Preferred are Oligomers having at least about 7 nucleosides, which is usually a sufficient number to allow for specific binding to a desired purine sequence of a segment of double stranded DNA. More preferred are Oligomers having from about 8 to about 40 nucleotides; especially preferred are Oligomers having from about 10 to about 25 nucleosides. Due to a combination of ease of synthesis, with specificity for a selected sequence, coupled with minimization of intra-Oligomer, in'ernucleoside interactions such as folding and coiling, it is I p I~AQara~ a believed that Oligomers having from about 14 to about 18 nucleosides comprise a particularly preferred group.

Preferred Oligomers These Oligomers may comprise either ribosyl moieties or deoxyribosyl moieties or modifications thereof.

However, due to their easier synthesis and increased stability, Oligomers comprising deoxyribosyl or modified deoxyribosyl moieties are preferred.

Although nucleotide Oligomers having the phosphodiester internucleoside linkages present in natural nucleotide Oligomers, as well as other oligonucleotide analogs) may be used according to the present invention, preferred Oligomers comprise oligonucleoside alkyl and arylphosphonate analogs, phosphorothioate oligonucleoside analogs, phosphoroamidate analogs and neutral phosphate ester oligonucleotide analogs. However, especially preferred are oligonucleoside alkyl- and aryl-phosphonate analogs in which phosphonate linkages replace one or more of the phosphodiester linkages which connect two nucleo- S 20 sidyl units. The preparation of some such oligonucleosidyl alkyl and arylphosphonate analogs and their use to inhibit expression of preselected single stranded nucleic acid sequences is disclosed in U.S. Patent Nos. 4,469,863; 4,511,713; 4,757,055; 4,507,433; and 4,591,614, the 25 disclosures of which are incorporated herein by reference.

A particularly preferred class of these phosphonate analogs are methylphosphonate Oligomers.

Preferred synthetic methods for methylphosphate Oligomers ("MP-Oligomers") are described in Lee, B.L. et al. Biochemistry 27:3197-3203 (1988) and Miller, et Biochemistry 25:5092-5097 (1986), the disclosures of which are incorporated herein by reference.

Preferred are oligonucleosidyl alkyl- and arylphosphonate analogs wherein at least one of the phosphodiester internucleoside linkages is replaced by a 3' linked internucleoside methylphosphonyl (MP) group (or

I,

II

"methylphosphonate"). The methylphosphonate linkage is isosteric with respect to the phosphate groups of oligonucleotides. Thus, these methylphosphonate Oligomers ("MP-oligomers") should present minimal steric restrictio.ns to interaction with the selected DNA sequences.

Also suitable are other alkyl or aryl phosphonate linkages wherein such alkyl or aryl groups do not sterically hinder the phosphonate linkage or interact with each other.

These MP-Oligomers should be more resistant to hydrolysis by various nuclease and esterase activities, since the methylphosphonyl group is not found in naturally occurring nucleic acid molecules. Due to the nonionic nature of the methylphosphonate linkage, these MP-oligomers should be better able to cross cell membranes and thus be taken up by cells. It has been found that certain MP-Oligomers are more resistant to nuclease hydrolysis, are taken up in intact form by mammalian cells in culture and can exert specific inhibitory effects on cellular DNA and protein synthesis (See, U.S. Patent No, 4,469,863).

S: 20 Preferred are MP-Oligomers having at least about seven nucleosidyl units, more preferably at least about 8, which is usually sufficient to allow for specific recognition of the desired segment of double stranded DNA. More preferred are MP-Oligomers having from about 8 to about nucleosides, especially preferred are those having from about 10 to about 25 nucleosides. Due to a combination of ease of preparation, with specificity for a selected sequence and minimization of intra-Oligomer, internucleoside interactions such as folding and coiling, particularly preferred are MP-Oligomers of from about 14 to 18 nucleosides.

Especially preferred are MP-Oligomers where the o 5'-internucleosidyl linkage is a phosphodiester linkage and the remainder of the internucleosidyl linkages are methylphosphonyl linkages. Having a phosphodiester linkage on the 5' end of the MP-Oligomer permits kinase Ir Irl aaec-e~ I I labelling and electrophoresis of the Oligomer and also improves its solubility.

The selected double stranded DNA sequences are sequenced and MP-Oligomers complementary to the purine sequence are prepared by the methods disclosed in the above noted patents and disclosed herein.

These Oligomers are useful in determining the presence or absence of a selected double stranded DNA sequence in a mixture of nucleic acids or in samples including isolated cells, tissue samples or bodily fluids.

These Oligomers are useful as hybridization assay probes and may be used in detection assays. When used as probes, these Oligomers may also be used in diagnostic kits.

If desired, labeling groups such as psoralen, chemiluminescent groups, cross-linking agents, intercalating agents such as acridine, or groups capable of cleaving the targeted portion of the viral nucleic acid such as molecular scissors like o-phenanthrolinecopper or ED- 20 TA-iron may be incorporated in the MP-Oligomers.

These Oligomers may be labelled by any of several well known methods. Useful labels include radioisotopes as well as nonradioactive reporting groups. Isotopic labels include H, S, 32 p, 125

I

Cobalt and C. Most methods of isotopic labelling involve the use of enzymes "and include the known methods of nick translation, end labelling, second strand synthesis, and reverse transcription. When using radio-labelled probes, hybridization can be detected by autoradiography, scintillation S 30 counting, or gamma counting. The detection method selected will depend upon the hybridization conditions and the particular radioisotope used for labelling.

S" Non-isotopic materials can also be used for labelling, and may be introduced by the incorporation of modified nucleosides or nucleoside analogs through the use of enzymes or by chemical modification of the Oligomer, for example, by the use of non-nucleotide linker groups.

I Ir i r a II ~I Ilbs~aa*us ~L Non-isotopic labels include fluorescent molecules, chemiluminescent molecules, enzymes, cofactors, enzyme substrates, haptens or other ligands. One preferred labelling method includes incorporation of acridinium esters.

Such labelled Oligomers are particularly suited as hybridization assay probes and for use in hybridization assays.

When used to prevent function or expression of a double stranded DNA sequence, these Oligomers may be advantageously derivatized or modified to incorporate a DNA modifying group which may be caused to react with said DNA and irreversibly modify its structure thereby rendering it non-functional. 1 rw a e 'a U-.S -S U seziat.-a=A2A3 P, the disclosure of which is incorporated herein by reference, teaches the derivatization of Oligomers which comprise oligonucleoside alkyl and arylphosphonates and the use of such derivatized oligonucleoside alkyl and arylphosphonates to render targeted single stranded nucleic acid sequences non-functional.

A wide variety of DNA modifying groups may be used to derivatize these Oligomers. DNA modifying groups include groups which, after the derivatized Oligomer forms a triple helix structure with the double stranded DNA segment, may be caused to form a covalent linkage, crosslink, alkylate, cleave, degrade, or otherwise inactivate or destroy the DNA segment or a target sequence portion thereof, and thereby irreversibly inhibit the function and/or expression of that DNA segment.

The location of the DNA modifying groups in the Oligomer may be varied and may depend on the particular DNA modifying group employed and the targeted double stranded DNA segment. Accordingy, the DNA modifying group may be positioned at the end of the Oligomer or intermediate the ends. A plurality of DNA modifying groups may be included.

*0t 0 *0.

0* -sP 'I ~I II a ~aau~rl r I In one preferred aspect, the DNA modifying group is photoreactable activated -by a particular wavelength, or range of wavelengths of light), so as to cause reaction and, thus, cross-linking between the Oligomer and the double stranded DNA.

Exemplary of DNA modifying groups which may cause cross-linking are the psoralens, such as an aminomethyltrimethyl psoralen group (AMT). The AMT is advantageously photoreactable, and thus must be activated by exposure to particular wavelength light before crosslinking is effectuated. Other cross-linking groups which may or may not be photoreactable may be used to derivatize these Oligomers.

Alternatively, the DNA modifying groups may comprise an alkylating agent group which, on reaction, separates from the Oligomer and is covalently bonded to the DNA segment to render it inactive. Suitable alkylating agent groups are known in the chemical arts and include groups derived from alkyl halides, haloacetamides, 20 phosphotriesters and the like.

DNA modifying groups which may be caused to cleave the DNA segment include transition metal chelating complexes such as ethylene diamine tetraacetate (EDTA) or o a derivative thereof. Other groups which may be used to effect cleaving include phenanthroline, porphyrin or "bleomycin, and the'like. When EDTA is used, iron may be advantageously tethered to the Oligomer to help generate the cleaving radicals. Although EDTA is a preferred DNA cleaving group, other nitrogen containing materials, such as azo compounds or nitreens or other transition metal chelating complexes may be used.

The nucleosidyl units of Third Strand Oligomers which Sread purine bases on both strands of a double stranded DNA sequence may comprise a mixture of purine and pyrimidine bases or only purine bases.

Where purine bases on both strands of a double stranded DNA sequence are to be read, it is preferred to 9 I' EIIIIIIIl I I I use Oligomers having only purine bases. It is believed that such purine-only Oligomers are advantageous for several reasons: purines have higher stacking properties than pyrimidines, which would tend to increase stability of the triple helix structure; use of purines only eliminates the need for either protonation of cytosine (so it has an available hydrogen for hydrogen bonding at the N-3 position at neutral pH) or use of a cytosine analog having such an available hydrogen at the position which corresponds to N3 on the pyrimidine ring; and allows use of a universal lengthening link.

The purine bases (and pyrimidine bases as well) are normally in the anti conformation; however, the barrier for a base to roll over to the syn conformation is low.

In formation of the third triple helix, the purines on the Third Strand may assume the syn conformation during the hydrogen bonding process. If desired, it is possible to modify the purine so that it is normally in the syn conformation. For example, the purine may be modified at 20 the 8-position with a substitutent such as methyl, bromo, isopropyl or other bulky group so it will assume the syn configuration under normal conditions. Nucleosidyl units "comprising such substituted purines would thus normally assume the syn conformation. Accordingly, where a purine base in the syj conformation is indicated, the present invention contemplates the optional incorporation of such 8-substituted purines in place of unsubstituted A or G.

Studies with our (Kendrew) models indicate that such substitutions should not affect formation of the triple 30 helix structure.

Use of purine nucleosidyl units in the anti and syn conformations, as appropriate (following the rules for reading the double stranded DNA described herein) allows reading of the purines on both strands of the duplex and formation of a triple helix structure by the purine-only third strand with the double stranded DNA.

I L C I If a third strand oligomer comprising both pyrimides and purines is used to read purines on both strands of a double stranded DNA, a non-uniform link format is used as described herein to allow the third strand to read across from one strand of the duplex to the other.

Preferred Purine Oligomers There is also disclosed herein a novel class of purine oligomers which comprise nucleosidyl units selected from: 0 0 N 0 I I I alk alk alk o R 0 R R/ and 0 wherein Bp is a purine base; R is independently alkyl or aryl groups which do not sterically hinder the phosphonate linkage or interact with each other; R' is hydrogen, hydroxy or methoxy; and alk is alkylene of 2 to 6 carbon atoms or alkyleneoxy of 2 to 6 carbon atoms.

Preferred are oligomers which comprise at least about 7 nucleosidyl units.

Preferred are nucleosidyl units where R is methyl. Also preferred are nucleosidyl units wherein R' is hydrogen. Suitable bases Bp include adenine and guanine, either optionally substituted at the 8-position, preferred substitutions include methyl, bromo, isopropyl and the like. Preferred are alk groups having from about 2 to about 3 carbon atoms. Particularly preferred are alk groups having two carbon atoms, and include ethylene.

Oligomers Comprising Cytosine Analogues In another aspect of the present invention, novel oligomers are provided which comprise nucleosidyl units wherein cytosine has been replaced by a cytosine analog comprising a heterocycle which has an available hydrogen at the ring position analogous to the 3-N of the cytosine ring and is capable of forming two hydrogen bonds with a guanine base at neutral pH and thus does not require protonation as does cytosine for Hoogstein-type base pairing, or formation of a triplet.

There is also provided according to the invention an oligomer capable of reading purine bases in a segment of double stranded DNA which comprises at least about seven nucleosidyl units connected by internucleosidyl phosphorus linkages wherein at least one nucleosidyl unit is selected from IG:\WPUSER\LIBFFI00024:JOC I II I -Lh -I 1 1 0 O p 0 P 1 0 P R 0 K 0 R O

R

O

R

,O

R

O R and and wherein Bp is a purine base; R is independently alkyl or aryl groups which do not sterically hinder the phosphonate linkage or interact with each other; R' is hydrogen, hydroxy or methoxy; and alk is alkylene of 2 to 6 carbon atoms or alkyleneoxy of 2 to 6 carbon atoms.

Suitable nucleosidyl units comprise analogues having a six-membered heterocyclic ring which has a hydrogen available for hydrogen bonding at the ring position corresponding to N-3 of cytosine and which is capable of forming two hydrogen bonds with a guanine base at neutral pH and include 2'-deoxy-5,6-dihydro-5-azadeoxycytidine pseudoisocytidine 6-amino-3-(13-D-ribofuranusyl)pyrimidine-2,4-dione (III) and 1-amino-l,2,4-(3-D-deoxyribofuranosyl)triazine-3-[4H]-one the structures of which are set forth in Table Preparation of MP-Oligomers In General As noted previously, the preparation of methylphosphonate oligomers has been :described in U.S. 4 469 863; 4 507 433; 4 511 713; 4 591 614 and 4 757 055.

Preferred synthetic methods for methylphosphonate oligomers are described in Lin, et al., Biochemistry 28:1054-1061 (1989); Lee, et al., Biochemistry 27:3197- 3203 (1988) and Miller, et al., 25:5092-5097 (1986), the disclosures of which are incorporated herein by reference. Oligomers comprising nucleosidyl units which comprise modified sugar moieties having lengthening links (see Figures 5 and 6) may be conveniently prepared by ihese methods.

Oligomers comprising phosphodiester internucleosidyl phosphorus linkages may be synthesised using any of several conventional methods, including automated solid phase chemical synthesis using cyanoethylphosphoroamidite precursors (29).

If desired, the previously-described nucleosidyl units comprising cytosine analogues (see Table 5) may be IG:\WPUSE\LIBFFOO24:JOC I I M I~sll I incorporated into the MP-Oligomer by substituting the appropriate cytidine analog (see Table 5) in the reaction mixture.

Preparation MP-Oligomers Having Lengthening Links in the Phosphorus Backbone 5'-(Ethyleneoxy)-Substituted-Suqar Intermediates MP-Oligomers may be prepared using modified nucleosides where either the bond between the and the 5'-hydroxyl or the 3'-carbon and the 3'-hydroxyl of the sugar moiety has been substituted with a alkyleneoxy group, such as ethyleneoxy group.

Figure 5 shows proposed reaction schemes for preparation of intermediates for modified nucleosides having either a 3'-(ethyleneoxy) or 5'-(ethyleneoxy) link.

In Figure 5, B represents a base, Tr and R represent protecting groups, Tr depicting a protecting group such as dimethoxytrityl and R depicting protecting groups such as t-butyldimethyl silyl or tetrahydropyranyl.

If desired, nucleosidyl units having such lengthening *oo 20 links at both the and positions of the sugar moiety may be prepared.

5'--Hydroxyethyl-Substituted Sugar Intermediate In situations where a double stranded DNA sequence which has purine bases on both strands is to be read, it may be preferred to use MP-Oligomers having a slightly lengthened internucleoside link on the phosphorus backbone.

Such MP-Oligomers may be prepared using nucleosides in which the sugar (deoxyribosyl or ribosyl) moiety has 30 been modified to replace the 5'-hydroxy with a phydroxyethyl (HO-CH 2

-CH

2 group synthetic schemes for the preparation of such a nucleosides is depicted in Figure 6.

Figure 6 depicts a proposed reaction scheme for a 0-hydroxyethyl-substituted sugar analog. In Figure 6, DCC

I

eB~BPsP I I I- denotes dicyclohexylcarbodiimide, DMSO denotes dimethylsulfoxide. B is a base. Suitable protecting groups, R, include t-butyldimethyl silyl and tetrahydropyranyl.

Preparation of MP-Olicomers Having Lenqtheneding Internucleoside Links in the Phosphorus Backbone MP-Oligomers incorporating the above-described modified nucleosidyl units are prepared as described above, substituting the modified nucleosidyl unit.

In the preparation of Oligomers comprising only purine bases, use of nucleosidyl units having the same lenthening links may be employed. However, in the preparation of Oligomers comprising both pyrimidine (or pyrimidine analog) bases and purine bases, a mixture of 15 nucleosidyl units having no lengthening link and lengthening links are used; nucleosidyl units having lengthening links at both the carbon and the 5'-carbon of the *sugar moiety may be advantageous.

Preparation of Derivatized MP-Oliqomers 20 Derivatized Oligomers may be readily prepared by adding the desired DNA modifying groups to the Oligomer.

As noted, the number of nucleosidyl units in the Oligomer and the position of the DNA modifying group(s) in the Oligomer may be varied. The DNA modifying group(s) may be positioned in the Oligomer where it will most effectively modify the target sequence of the DNA. Accordingly, the positioning of the DNA modifying group may depend, in large measure, on the DNA segment involved and its key target site or sites, although such optimum position can be readily determined by conventional techniques known to those skilled in the art.

Preparation of Psoralen-Derivatized MP-Oliqomers The derivatization of MP-Oligomers with psoralens, such as 8-methoxypsoralen and 4'-aminomethyl- I I I I i, I ~BIII~W~RII r~ trimethylpsoralen (AMT), is described in Kean, J.M. et al., Biochemistry 227:9113-9121 (1988) and Lee, B.L. et al., Biochemistry 27:3197-3203 (1988), the disclosures of which are incorporated herein by reference.

Preparation of EDTA-Derivatized MP-Oliqomers The derivatization of MP-Oligomers with EDTA is described in Lin, et al., Biochemistry 28:1054-1061 (1989), the disclosures of which are incorporated herein by reference.

Utility According to the present invention, a specific segment of double stranded DNA may be detected or recognized using an MP-Oligomer Third Strand which reads the purine bases of the duplex of the double-stranded DNA S 15 according to the triplet base pairing guidelines described herein. The MP-Oligomer Third Strand has a sequence selected such that the base of each nucleosidyl unit will form a triplet with a corresponding base pair of the duplex DNA to give a triple helix structure. Detectably S 20 labeled Oligomers may be used as probes for use in hybridization assays, for example, to detect the presence of a particular double-stranded DNA sequence.

S**The present invention also provides a method of preventing expression or function of a selected sequence in double stranded DNA by use of an MP-Oligomer which reads the DNA sequence and forms a triple helix structure.

Formation of the triple helix may prevent expression and/or function by modes such as preventing opening of the duplex for transcription, preventing of binding of effector molecules (such as proteins), etc. Derivatized Oligomers may be used to detect or locate and then irreversibly modify at target site in the DNA duplex by cross-linking (psoralens) or cleaving one or both strands (EDTA). By careful selection of a target site for Irll B8~Pe I cleavage, the Oligomer may be used as a molecular scissors to specifically excise a selected DNA sequence.

The Oligomers may be derivatized to incorporate a DNA reacting or modifying group which can be caused to react with the DNA segment or a target sequence thereof to irreversibly modify, degrade or destroy the DNA and thus irreversibly inhibit its functions.

Thus, these Oligomers may be used to irreversibly inactivate or inhibit a particular gene or target sequence of the game in a living cell, allowing selective inactivation or inhibition. These Oligomers could then be used to permanently inactivate, turn off or destroy genes which produced defective or undesired products or if activated caused undesirable effects.

Another aspect of the present invention is directed to a kit for detecting a particular double stranded DNA sequence which comprises a detectably labeled purine MP- Oligomer Third Strand selected to be able sufficiently complementary to the purine sequence to be able to read the sequence and form a triple helix structure.

To assist in understanding the present invention, the following examples are included which describe the results of a series of experiments, including computer simulations. The following examples relating to this 25 invention should not, of course, be construed in specifically limiting the invention and such variations of the invention, now known or later developed, which would be within the purview of one skilled in the art, are considered to fall within the scope of the present invention as hereinafter claimed.

Examples Example 1 Computer Simulations Of Triple Helix structures The primary purpose of these computer simulations was to determine whether nonionic nucleotide analogs with a methylphosphonate ("MP'")backbone would bind with greater 1- r affinity in comparison with unmodified oligodeoxynucleotides as the Third Strand of the triplestranded helical DNA through Hoogsteen-type base pairing.

Experimental work suggested that ODN binding to duplex DNA and inhibition of transcription could be via triplet formation but no experimental comparison had been made between analogs with MP backbone verses native ODN in formation of the triple-stranded DNA helix. It was previously unknown whether a nonionic analog with MP backbone could be accommodated in the major groove of triple-stranded helical DNA. Furthermore, the conformation of fully solvated triple-stranded helical DNA with native or MP backbone in the third strand had not been determined. Site-specific oligonucleotide binding via double and triple stranded DNA complex formation has recently been shown to suppress transcription of human oncogenes in vitro The goal of this example was to use molecular dynamics simulation to investigate nonionic oligonucleotide analogs with MP backbone in 20 triple-stranded helical complexes, and to gain insight into the molecular mechanism(s) involved in this process.

Simulation Methods Triple-stranded poly (dT 10 )-poly (dAo)-poly (dT 10

[TAT

2 coordinates were obtained from the A-DNA x-ray structure of Arnott and Seising The same coordinates were used for the starting geometry of poly(dT 1 o)-poly -poly (dT 10 methylphosphonate ([TAT 2 MP]. Geometry optimization and partial atomic charge assignments for the dimethyl ester methylphosphonate fragment were calculated by ab initio quantum mechanical methods with QUEST (version 1.1) using 3-21G* and STOG* basis sets, respectively The latter basis set was used to maintain uniform charge assignments with those previously calculated for nucleic acids in the AMBER database. The final monopole atomic charge assignments for the MP fragment were made to obtain a neutral net charge for each base, furanose, and MP backbone of the third DNA strand.

Alternating Rp. and Sp. methyl substitution of the backbone phosphoryl oxygens of the T 2 MP strand was done by stereo computer graphics. The substitution of MP diastereomers was made in this manner to approximate experimental yield, since the synthesis cannot be controlled. Molecular mechanics and molecular dynamics calculations were made with a fully vectorized version of AMBER (version 3.1), using.an all-atom force field (12, 13). All calculations were performed on CRAY X-MP/24 and VAX 8600 computers.

The negative charge of the DNA phosphate backbone was rendered neutral by placement of positive counterions within 4 A of the phosphorus atoms bisecting the phosphate oxygens; counterions were not placed on the MP-substituted strand. The triple helices and counterions were S* surrounded by a 10A shell of TIP3P water (14) molecules with periodic boundary conditions. There are 9,283 and 10,824 atoms in the TiAT 2 and T 1

AT

2 MP systems, respectively. The box dimensions were 101, 686.8 A for TAT 2 20 and 124,321.1 A 3 for TiAT 2 MP. Initially, the DNA and counterion atoms were fully constrained while the surrounding water molecules were energy minimized using an 8.0 A nonbonded cutoff until convergence (root mean-square [rms] of the gradient <0.1 kcal/mole/A). The DNA, 25 counterions, and water were subsequently energy minimized without geometric constraints for an additional 1500 cycles, followed by 220 cycles of minimization with SHAKE activated Molecular dynamics using SHAKE at constant temperature and pressure (300K and 1 bar) was carried out without constraints for 40 psec trajectories for each of the two molecular ensembles.

Results of Simulation Studies The third DNA strand with MP backbone resulted in several changes consistent with enhanced binding of the ODN wit the MP backbone in the triple helix. The average hydrogen bond distances and mean atomic fluctuations are I I I -II BBc~ I r consistently smaller in the T 1 AT2MP triplet (Table 1) The interstrand phosphorus atoms distance was 9.6A for A-T 2 and 8.3A for A-T 2 MP. The reduced interstrand phosphorus atom distance and smaller mean atomic fluctuations between the second and third strands are due to decreased interstrand electrostatic repulsion accompanying MP substitution in the backbone.

Both triple helical DNA systems had strand-specific polymorphic conformational behavior during molecular dynamics. There are significant conformational changes in the furanose relative to the starting geometry in both systems (Table In the TiAT 2 helix the furanose ring populations of the T 1 and T 2 strands remained predominantly in an A-DNA conformation (C3'endo) and the largest proportion of the adenine sugars adopted a B-DNA conformation (C2'endo). A notable percentage of the adenine furanose conformations were in an 01'endo conformation in TAT 2 and T 1 ATMP. The sugar puckering pattern of the MP substituted helix had a greater proportion of 20 01'endo and C2'endo conformations in contrast to the unsubstituted helix. Analysis of other conformational parameters support the hybrid conformational nature of these triple helices. The helical twist angle (between T1 and A strands) averaged 39.4 degrees for the 25 T 1

AT

2 structure and is more consistent with a B-DNA conformation (range 36-45). The TIAT 2 MP helical twist angle averages 32.0 degrees and is closer to that of A-DNA (range 30-32.7). The average helical repeat singles (between the T1 and A strands) for the entire structure are for 10.2 TAT 2 and 11.2 degrees for TAT 2

MP.

The average intrastrand phosphorus atom distances over the psec trajectory are presented in Table 3. In both helices, the intrastrand phosphorus distances of the T, strands are most consistent with an A-DNA conformation (7.0 The interstrand phosphorus distances of the T2MP strand are more consistent with a B-DNA conformation, in I I 36 contrast to values more consistent with A-DNA for the T, strand.

We analyzed the coordination of counterions and water along the backbone of the DNA to determine the changes accompanying MP substitution. The average coordination distance and atomic fluctuations-of the counterions with phosphorus atoms was 3.8A for T 1

AT

2 and 4.6A for the T 1

AT

2 MP helix. The increase in average coordination distance and atomic motion in T 1 ATyMP (by 0.8A) is most likely due to the proximity of Rp (axial projecting) methyl groups to the counterions coordinated to the second strand. The average number of water molecules coordinated to the phosphate groups is not significantly altered in the MP substituted helix.

The DNA backbone and the C1'-N (base) dihedral transitions of the two helical systems are shown in Table 4. Comparisons of the a-dihedral of the adenine strands reveals a slight change in the average position of the dihedral with MP substitution, positioning the dihedral 20 closer to trans, but there is a large overlap in the transitional motions of both dihedrals during the 40 psec trajectory. There was a significant change (by 27.0 degrees) in the average Sp-MP diastereomer 8 dihedral angle from baseline. The fluctuation of the 0 dihedral containing the Sp-MP diastereomer was significantly less than the Rp-MP.

Interpretation of Simulation Results The results of these molecular dynamics calculations predict that an MP-substituted ODN incorporated as a colinear third strand with Hoogsteen pairing will form a more stable triple helical complex than a native ODN as the third strand, as in the case of poly(dT)poly(dA) poly(dT). The enhanced binding of the MP strand is due to reduced interstrand electrostatic repulsion. The MP-substituted helix has reduced hydrogen bond distances, decreased interstrand A-T 2 phosphorus distances, and less 1 I' WO 90/15884 37 fluctuation in atomic position rel?.ive to the native triple helix. The closer fit and reduced atomic motion during molecular dynamics are qualitatively consistent with a greater enthalpy of binding and stability of the MP-substituted triple helical complex. These findings support the MP-substitution of the third strand facilitates formation of a more cohesive triple helical structure by decreased interstrand phosphate repulsion, and will secondarily have closer approximation of Hoogsteen and Watson-Crick hydrogen bond interactions.

One would expect predominant effects on Hoogsteen pairing, but there is an unexpected enhancement of Watson-Crick hydrogen bonding with MP substitution in these calculations. The latter finding is most likely due to decreased electrostatic repulsion and shielding (by the third strand) between the T 1 and TMP strands.

The conformation of these DNA structures differs from experimental data based on the fiber diagram. The S structure of poly(dT)poly(dA)-poly(dT) was determined by 20 x-ray diffraction studies under conditions of 92% humidity, and is a low resolution structure The molecular dynamics simulations are of fully solvated DNA structures under periodic boundary conditions with counterions. DNA in solution is generally believed to predominate in the B-form; A-DNA conformation predominates under conditions of lower humidity Several triple-stranded DNA helical structures have been determined by x-ray diffraction studies and have been uniformly observed in an A-DNA conformation under conditions of low humidity and increased salt concentration (10,17,18). These computer simulations predict that different DNA conformations coexist within the triple helix, that the individual strands of the helices have predominant conformational populations, and that a Hoogsteen-paired, MP-substituted DNA strand is predicted to predominate in the B-form. The large proportion of 01' endo .sugars in both triplexes is of 38 interest since this furanose conformation is 0.6 kcal/mole higher than C2' endo and C3, endo DNA sugar puckers (16).

The DNA dodecamer crystal structure (19) has a notable 01' endo population, and a significant proportion of 01' endo sugar puckers were observed in molecular dynamics simulations of dsDNA by Seibel et al. (20) Both helical structures generally follow the classical observations of purine nucleotides adopting C2' endo geometries and pyrimidines adopting C3' endo geometries.

The large perturbation of the p dihedral and variable conformational fluctuation of the Rp. and Sp. MP diastereoisomers in the triplet are due to nonbonded and hydrophobic interactions. The Sp-MP groups are in close proximity to thymine methyl groups (in the major groove) on the same DNA strand, and interact by Van der Waals forces, which could locally destabilize the helix by "locking" the thymine to the Sp-MP backbone. There are greater deviations in the P dihedral of the Rp-MP groups, since these groups project out into the solvent water and 20 are positioned farther away from the thymine methyl groups. There is a known relationship between the orientation of the methyl substituent on the phosphorus atom and DNA duplex stability, and a proposed mechanism of reduced stability of Sp-MP diastereomers is due to local 25 steric interactions Our calculations suggest that steric interactions contribute very little toward local helix destabilization, and the predominant mechanism is mediated by non-bonded interactions between the methyl groups of the Sp-MP backbone and thymine on the same strand which locally destabilizes the DNA.

Example 2 Detection Of Triple Helix Formation Usinq Circular Dichroism Spectroscopy Circular dichroism spectroscopy studies were performed using Triple Helix Structures formed using a combination following nucleoside oligomers.

I r-q I: d1(CTCTCTCTCTCTCTCT) abbreviated d (CT) 8 E 24= 9.2 X10 4 M1cm, II: d (AGAGAGAGAGAGAGAG) abbreviated d(AG) 8 E 254 1.45 x 10 5M I cm, I III: d (CpTpCpTRCRTpCRTpCRTpCpTpCRTpCpTp) abbreviated d (Cp-T)a E 24= 8.5 x 10 4M, cm, circular dichroism (CD) spectra for the triple helix structures made with 2:1 d(CT) 8 e d(AG), and 1:1:1 d(CT) 8 d(AG 8 9 d(CRTp) 8 were performed using a CD spectropolarimeter in pho~sphate buffer at the indicated pH.

Figure 7 shows the CD spectra for triple helix (a) [2:1 d(CT) 8 9 d(AG)S and [1:1:1 d(CT) 8 a d(AG)a V Example 3 Crosslinking Of Triple Helix Structures Usingr Psoralen-Derivatized MP-Olicqomers **:*.Psoralen derivatized dTp(T-) 6 oligomers were prepared as described in Lee, et Al., Biochemistry 27:3197- 3203 (1988).

The T 7 oligome*rs were allowed to hyh, 'dize with DNA having the following sequence including a 15-mer poly A subsequence: 10 20 30 40 3' d- TAATACGACTCACTATAGGGAGATTTTTTTTTTTTTTTACGAGCT d- ATTATGC2GAGTGATATCCCTCTAAAAAJ~AAATGCTCGA 31 MP-oligomers derivatized with (aminoethyl) aminomethyl-4, 5',8-trianethyl-psoralen 4' -(aminobutyl) aminomethyl-4, 51 8-trimethylpsoralen (ab) AMT1 and, 4 (aminohexyl) aminomethyl-4, 5 -8-trimethylpsoralen L I were allowed to hybridize with single stranded DNA of the above DNA sequences and double stranded DNA of the above sequence at 4 0 C and were irradiated to cause crosslinking as described in Lee, et al.

Results are depicted in Figures 8A and 8B. Figure 8A shows crosslinking of the psoralen derivatized T 7 Oligomers with the single stranded (poly A containing) DNA sequence.

Figure 8B shows crosslinking of the double stranded DNA with the double stranded DNA sequence.

5o Table I Averag~ed Hydrogen Bond Distances (P145) WATSON-CRICK WITHOUT MP WITH MP ADE HN6B THY 04 2.33 1.98 ADE Ni THY H3 2.10 1.95

HOOGSTEEN

ADE HN6A THY 04 2.12 2.09 ADE N7 THY H3 1.94 1.92 Averaged Watson Crick and Hoogsteen hydrogen bond distances (in Angstroms) in T 1 AT, and T 1

AT

2 I4P helices.

These distances are calculated for the triple helical DNA complexes. The fluctuation in atomic position (calculated as the root-mean-square [rms3) are in Y) a Table 2 Averages Of Furanose Pucker Phase, And Conformational Populations

HELICAL

STRAND 0 MFAN (RMS) PHASE (P145) C3'ENDO 01'ENDO C21ENDO 01' EXO

A

0.36 35.70 (±24.7) 0.39 107.44 (±30.8) 0.38 40.99 (±24.1) 86.8% 3.6% 75.3% 6.6% 34.3% 24 .1% 6.6% 0.0% 57.1% 0.6% 0.0% Ti

A

T2-MP 0.38 0.39 0.40 (±0.06) (±0.05) (±0.05) 36.31 109.78 103 .14 36) 59) (±29.50) 79.1% 14.7% 11.2% 18.1% 41.5% 61.8% 2.8% 43.8% 27.0% 0.0% 0.0% .0 .0% Averages of sugar pucker phase, and conformational populations of furanose for T 1

AT

2 and

T

1

AT

2 MP helices (15) Q is in Angstroms, phase is in degrees, and populations are in percent of the total furanose conformations in each of the three DNA strands.

43 Table 3 Average Intrastrand Phosphate Atom Distance (RMS) STRAND WITHOUT MP WITH MP T1 6.5 6.2 A 7.0 7.1 T2 6.3 6.8 Intrastrand phosphate distances of TIAT 2 and T 1

AT

2

MP

helices. The calculated intrastrand phosphate distances (in Angstroms) averaged over the 40 psec trajectory are shown for the entire triple helical systems. Standard interstrand phosphorus distances are 6.0A for A-DNA and for B-DNA V *e d Table 4 Without MP With MP a 03' P 05' T1 280.5 (18.5) 288.4 (11.3) A 252.9 (28.7) 233.2 (28.5) T2 285.1 (11.6) 288.5 (11.3) B P 05' C5' C4' T1 161.2 (11.5) 168.7 (8.7) A 151.5 (10.7) 159.3 (21.5) T2 140.8 POR 158.8 (21.9) POS 167.8 (8.7) Y 05' C5' C4' C3 T1 70.6 (14.6) 62.6 9.3) A 104.8 (27.9) 112.3 (25.5) T2 67.2 (10.5) 64.0 (10.4) 8 C5' C4' C3' 03' T1 90.3 (12.8) 82.5 (11.1) A 112.5 (16.7) 111.6 (17.2) .T2 88.6 (10.4) 104.5 (16.5) 20 E C4' C3' 03' P T1 196.4 (10.1) 199.1 (10.6) A 200.1 (10.9) 198.0 (13.3) .T2 197.5 (10.5) 187.9 7.8) C3' 03' P T1 290.8 (10.6) 290.6 9.6) .A 282.9 (12.0) 282.0 (18.5) T2 285.0 (10.5) 280,5 (11.4) PUR 01' C1' N9 C4 30 ADE 215.3 (14.7) 2J1!.1 (14.2)

PYR

T1 212.3 (11.0) 205.9 (10.0) T2 206.3 (10.7) 213.9 (13.2) Average backbone dihedral angles (rms) for the triple helical DNA structures during the 40 psec trajectory TABLE Cytidine Analoas Structure 1 140 Reference (Preparatioi Goddard, et al., Tetrahedron Letters 29:1767 (1988) Beisler, et al., J. Med.

Chem. 20:806 (1977) 140 Doboszawske, B., Org. Chem (1988); Woodcock, T.M., Cancer Res.

(1980) Burchenal, J.H.

Cancer Res.

(1976) etg',. J.

53: :2777 et al., 40: 423 4 et al. 3 6: 152 0 Chem. Soc. p. 791 (1969) AIN From 1,2,4-triazine-3(4H)one (by reaction with ~4.I ammnonium chloride or by r~'i1(a) nitrosating followed by treatment with j~osodium borohydride) 140 I I i.

47

BIBLIOGRAPHY

1. Mosler, et al., Science 238:645-650 (1987) 2. Povsiz, et al., J. Am. Chem. Soc. 111:3059- 3061 (1989) 3. Wells, et al., FASEB J. 2:2939-2949 (1988) 4. Dervan, Science 232:464-471 (1988) Miller, et al., Anti-Cancer Drug Design, 2:117-128 (1987) 6. Yarchoan, et al., "AIDS Therapies," Scientific American, pp. 110-119 (October 1988) 7. Sarin, et al., Proc. Nat. Acad. Sci. (USA) 85:7448- 7451 (1988) 8. Cooney, et al., Science 241:456-459 (1988) 9. Wickstrom, et al., Proc. Nat. Acad. Sci. (USA) 85:1028-1032 (1988).

Arnoff, et al., J. Mol. Biol. 88:509-521 (1974).

11. Singh, et al., J. Comp. Chem. 5:129-145 (1984).

12. Singh, et al., AMBER (UCSF) Version 3.1 20 (1988).

13.* Weiner, et al., J. Comp. Chem. 7:230-252 (1986).

14. Jorgenson, t al., J. Chem. Phys. 79:926-935 (1983).

15. Berendsen, et al., J. Chem. Phys. 81:3684- 3690 (1984).

16. Sanger, Principles of Nucleic Acid Structure (Springer-Verlag, New York, 1984) 17. Arnott, et al., Nature 244:99-101 (1973) 30 18. Arnott, et al., Science 181:68-69 (1973) 19. Drew, et al., Proc. Nat. Acad. Sci. (USA) 78:2179-2183 (1981).

Seibel, et al., Proc. Nat. Acad. Sci. (USA) 82:6537-6540 (1985) 21. Bower, et al., Nucl. Acids Res. 1:4915-4930 (1987) 22. Donohue, et al., J. Mol. Bio. 2:363 (1960) ImI 48 22a. Ts'o, Basic Principles in Nucleic Acid Chemistry, pp. 453-584 Tso ed., Academic Press, New York, 1974) 23. Sundaralingham, Biopolymers 7:821 (1969).

24. Arnott, Proar. Nuci. Acid Res. Mol. Biol. 10:183 (1970) Sasisekharan, et Al., Conform. Biopolym., Pap.

Int. Sy-mp. 1967, Vol. 2, p. 641 (1967) 26. Lakshminarayanan, et al., Biochem. Biophys.

Acta 204:49 (1970) 27. Lakshminarayanan, et Al., Biopolymers 8:475 and 489 (1970).

28. Lee, gt al., Nucl. Acids Res. 16:10681-10697 (1988) 29. Barone, et al. Nucl. Acids. Res. 12: 4051- 4060 (1984).

I

Claims (3)

1. An oligomer which comprises at least about seven interconnected nucleosidyl units selected from: 0 0 0 alk alk alk R 0 0 PP P 0 0 R' ,O R' 0 R' S R and wherein Bp is a purine base; R is independently alkyl or aryl groups which do not sterically hinder the phosphonate linkage or interact with each other; R' is hydrogen, hydroxy or methoxy; and alk is alkylene of 2 to 6 carbon atoms or alkyleneoxy of 2 to 6 carbon atoms.

2. An oligomer according to claim 1 which comprises a methylphosphonate oligomer.

3. A purine oligomer substantially as hereinbefore described with reference to the Examples. Dated 14 October, 1994 The John Hopkins University 15 Patent Attorneys for the Applicant/Nominated Person SPRUSON FERGUSON 1 [N:\LIBAA]00202:CE Ib L1I~-- Purine Oigomers Abstract An oligomer which comprises at least about seven interconnected nucleosidyl units selected from: 00 0 alk 0 P alk,0B alk 0B 00 51R~ and wherein B~ is a purine, base; R is independently alkyl or aryl groups which do not sterically hinder the phosphonate linkage or interact with each other; R' is hydrogen, hydroxy or methoxy; and alk is alkylene of 2 to 6 carbon atoms or alkyleneoxy of 2 to 6 carbon atoms. 0: 6% [N:\LIBAA]0O202:JOC I