Uncharged Morpolino-Based Polymers Having Phosphorous Containing Chiral Intersubunit Linkages

The present invention relates to morpholino-based polymers.

Agarwal, Proc Nat Acad Sci USA, 85:7079 (1988) Balgobin, N., et al., Tetrahedron Lett, 22:3667 (1981). Belikova, Tetrahedron Lett, 37:3557 (1967). Blake et al., Biochem, 24:6132 (1985a). Blake et al., Biochem 24:6139 (1985b). Bower et al., Nucleic Acids Res. 15:4915 (1987). Dikshit et al., Canadian J Chem, 66:2989 (1988). Froehler, et al., Nucleic Acids Res. 16:4831 (1988). Fox, J. J., et al., J Am Chem Soc, 80:1669 (1958). Gait, "Oligonucleotide Synthesis A Practical "Approach," pp. 31-33, IRL Press (Oxford, England) (1984). Goldberg, M. L. et al; Methods in Enzymology 68:206 (1979). Greenlee, J Org Chem, 49 2632 (1984). Grunstein, M. et al; Methods in Enzymology 68:379 (1979). Himmelsbach, F., and W. Pfleiderer, Tetrahedron Lett, 24:3583 (1983). Jayaraman, et al., Proc Natl Acad Sci USA 78:1537 (1981). Kamimura et al., Chem Lett (The Chem. Soc. of Japan) pg. 1051 (1983) LaPlanche et al., Nucleic Acids Res, 14: 9081 (1986). Lerman, L. S., "DNA Probes: Applications in Genetic and Infectious Disease and Cancer," Current Comm in Molec Biol (Cold Spring Harbor Laboratory) (1986). Letsinger and Miller, J Amer Chem Soc, 91:3356 (1969). McBride et al., J Amer Chem Soc 108:2040 (1986). Miller, et al., Biochemistry 18:5134 (1979). Miller, et al., J Biol Chem 255:6959 (1980). Miller, et al., Biochimie 67:769 (1985). Murakami, et al., Biochemistry 24:4041 (1985). Niedballa, U., and H. Vorbruggen, J Org Chem, 39:3668 (1974). Pitha, Biochem Biophys Acta 204:39 (1970a). Pitha, Biopolymers 9:965 (1970b). Reese, C. B., and R. S. Saffhill, J Chem Soc PerkinTrans, 1:2937 (1972). Smith, et al., J.A.C.S. 80:6204 (1958). Smith, et al., Proc Natl Acad Sci USA 83:2787 (1986). Southern, E.; Methods in Enzymology 68:152 (1979) Stirchak E. P. et al., Organic Chem. 52:4202 (1987). Summerton, J., et al., J Molec Biol 122:145 (1978) Summerton, J., et al., J Molec Biol 78:61 (1979a). Summerton, J., J Molec Biol 78:77 (1979b). Szostak, J. W. et al; Methods in Enzymology 68:419 (1979). Thomas, P.; Methods in Enzymology 100:255 (1983). Toulme et al., Proc Nat Acad Sci USA, 83:1227 (1986). Trichtinger et al., Tetrahedron Lett 24:711 (1983).

Polymers which are designed for base-specific binding to polynucleotides have significant potential both for in vitro detection of specific genetic sequences characteristic of pathogens and for in vivo inactivation of genetic sequences causing many diseases--particularly viral diseases.

Standard ribo- and deoxyribonucleotide polymers have been widely used both for detection of complementary genetic sequences, and more recently, for inactivating targeted genetic sequences. However, standard polynucleotides suffer from a number of limitations when used for base-specific binding to target oligonucleotides. These limitations include (i) restricted passage across biological membranes, (ii) nuclease sensitivity, (ii) target binding which is sensitive to ionic concentration, and (iv) susceptibility to cellular strand-separating mechanisms.

In principle, the above limitations can be overcome or minimized by designing polynucleic acid analogs in which the bases are linked along an uncharged backbone. Examples of uncharged nucleic acid analogs have been reported. Pitha et al (1970a, b) have disclosed a variety of homopolymeric polynucleotide analogs in which the normal sugar-phosphate backbone of nucleic acids is replaced by a polyvinyl backbone. These nucleic acid analogs were reported to have the expected Watson/Crick pairing specificities with complementary polynucleotides, but with substantially reduced Tm values (Pitha, 1970a). One serious limitation of this approach is the inability to construct polymers by sequential subunit addition, for producing polymers with a desired base sequence. Thus the polymers cannot be used for base-specific binding to selected target sequences.

Polynucleotide analogs containing uncharged, but stereoisomeric, methylphosphonate linkages between the deoxyribonucleoside subunits have also been reported (Miller, 1979, 1980; Jayaraman; Murakami; Blake, 1985a, 1985b; Smith). More recently a variety of analogous uncharged phosphoramidate-linked oligonucleotide analogs have also been reported (Froehler, 1988). These polymers comprise deoxynucleosides linked by the 3'OH group of one subunit and the 5'OH group of another subunit via an uncharged chiral phosphorous-containing group. These compounds have been shown to bind to and selectively block single-strand polynucleotide target sequences. However, uncharged phosphorous-linked polynucleotide analogs using deoxyribonucleoside subunits are particularly costly and difficult to prepare; the subunit starting material is quite costly and of limited availability.

More recently, deoxyribonucleotide analogs having uncharged and achiral subunit linkages have been constructed (Stirchak 1987). These uncharged, achiral deoxyribonucleoside-derived analogs are, as mentioned above, limited by relatively high cost of starting materials.

It is one general object of the invention to provide a polymer capable of sequence-specific binding to polynucleotides and which overcomes or minimizes many of the problems and limitations associated with polynucleotide analog polymers noted above.

The invention includes a polymer composition containing morpholino ring structures of the form: ##STR1##

The ring structures are linked together by uncharged, chiral linkages, one to three atoms long, joining the morpholino nitrogen of one ring structure to the 5'exocyclic carbon of an adjacent ring structure.

Each ring structure includes a purine or pyrimidine base-pairing moiety P.sub.i which is effective to bind by base-specific hydrogen bonding to a base in a target sequence in a polynucleotide.

These and other objects and features of the invention will become more fully apparent when the following detailed description of the invention is read in conjunction with the accompanying examples and figures.

FIG. 1 shows a basic .beta.-morpholino ring structure which is linked through uncharged, chiral linkages to form the polymer of the present invention;

FIG. 2 shows several exemplary purine and pyrimidine base-pairing moieties (represented as Pi of the ring structures shown in FIG. 1);

FIG. 3 shows several preferred subunits suitable for forming polymers having 5-atom (subunit A), 6-atom (subunit B), and 7-atom (subunits C-E) unit-length backbones;

FIG. 4 shows a repeating subunit segment of exemplary morpholino-based polymers, designated A--A through E--E, constructed using subunits A-E, respectively, of FIG. 3;

FIG. 5 shows the steps in the synthesis of several types of morpholino subunits from a ribonucleoside;

FIG. 6 shows an alternative synthesis of the basic morpholino subunit;

FIG. 7 shows the steps in the synthesis of a morpholino subunit designed for construction of polymers with seven-atom repeating-unit backbones;

FIG. 8 shows the binding mode for 2-amine-containing purines to polar major-groove sites of respective target base-pairs and a representative base sequence of a duplex-binding polymer;

FIG. 9 shows the steps in linking two morpholino subunits through a phosphonamide linkage;

FIG. 10 shows two methods for linking morpholino subunits through a phosphoramidate linkage;

FIG. 11 shows the linking of morpholino subunits through a phosphonoester linkage;

FIG. 12 illustrates a subunit coupling procedure which concurrently generates the morpholino ring structure;

FIG. 13 shows the thermal denaturation plot for a (morpholino-C).sub.6 /poly(G) complex, where the (morpholinoC).sub.6 polymer was constructed according to the present invention; and

FIG. 14 illustrates the use of a morpholino polymer in a probe-diagnostic system.

The present invention includes a morpholino-based polymer which is designed for base-specific binding to a target sequence of a polynucleotide. The polymer is composed of morpholino-based ring structures which are linked together by uncharged, chiral linkages, one to three atoms long, joining the morpholino nitrogen of one structure to the 5'exocyclic carbon of an adjacent structure.

FIG. 1 shows the .beta.-morpholino ring structures on which the polymer subunits are based, where the morpholino carbon atoms are numbered as in the parent ribose. As seen in FIG. 1, the ring structure contains a 5'methylene attached to the 4'carbon in the .beta.-orientation.

Each ring structure includes a purine or pyrimidine or related hydrogen-bonding moiety, P.sub.i, attached to the backbone morpholine moiety through a linkage in the .beta.-orientation.

The purine hydrogen-bonding moieties or bases include purines as well as purine-like planar ring structures having a 5-6 fused ring in which one or more of the atoms, such as N3, N7, or N9 is replaced by a suitable atom, such as carbon. The pyrimidine moieties likewise include pyrimidines as well as pyrimidine-like planar 6-membered rings in which one or more of the atoms, such as N1, is replaced by a suitable atom, such as carbon. Preferred hydrogen-bonding moieties in the invention include the set of purines and pyrimidines shown in FIG. 2. Each base includes at least two hydrogen-bonding sites specific for a polynucleotide base or base-pair. Where the polymers are used for sequence-specific binding to single-stranded polynucleotides, the purine structures 1-3 are designed to bind to thymine or uracil bases; structures 7-8, to guanine bases; structures 4-6, to cytosine bases; and structure 9, to adenine bases.

The polymers of the invention are also effective to bind to hydrogen-bonding sites accessible through the major-groove in duplex polynucleotides having mostly purine bases in one strand and mostly pyrimidine bases in the complementary strand, as discussed below.

Because of the similar type and positioning of the two central polar major-groove sites among the different base-pairs of duplex nucleic acids (i.e., the NH4 and O6 of a CG base-pair present the same H-bonding array as the NH6 and O4 of an AT base-pair), the H-bonding moiety of a duplex-binding polymer must hydrogen-bond to the N7 of its target base-pair in order to uniquely recognize a given base-pair in a target genetic duplex. Thus, in the polymers of the present invention, which are targeted against duplex genetic sequences containing predominantly purines in one strand and predominantly pyrimidines in the other strand, the hydrogen-bonding moieties of the polymer preferably contain purines having an amine at the 2 position since that amine is suitably positioned for H-bonding to the N7 of the target base-pair. More specifically, Structures 2 and 3 of FIG. 2 provide for specific binding to a TA or UA base-pair while Structures 4 and 6 provide for specific binding to a CG base-pair. Two bases which are particularly useful in a duplex-binding polymer are 2,6-diaminopurine (structure 3) and guanine (structure 4). FIG. 8A illustrates the binding of these two bases to the polar major-groove sites of their respective target base-pairs in duplex nucleic acids. FIG. 8B illustrates a representative base sequence of a polymer designed for binding a target genetic sequence in the duplex state.

Polymers comprising predominantly 2-amine-containing purines, thus suitable for high-specificity binding to polar major-groove sites of duplex genetic sequences, can provide effective binding to their targeted genetic duplexes using alternative backbones, in addition to the morpholino-based backbone. Examples of such alternative backbones include phosphodiester-linked deoxyribonucleosides where a pendant group on the phosphorous is one of the following: a negatively charged oxygen (i.e., the natural DNA backbone); a methyl or other alkyl group (referred to as an alkylphosphonate); a methoxy or other alkoxy group (referred to as a phosphotriester); or a mono- or dialkyl amine (referred to as a phosphoramidate).

The morpholino subunits of the instant invention are combined to form polymers by linking the subunits through stable, chiral, uncharged linkages. The linking group of a subunit includes a phosphorous-containing electrophile which is usually reacted with a nucleophile of the subunit to which it is to be linked.

The selection of subunit linking groups for use in polymer synthesis is guided by several considerations. Initial screening of promising intersubunit linkages (i.e., those linkages which are predicted to not be unstable and which allow either free rotation about the linkage or which exist in ,a single conformation) typically involves the use of space-filling CPK or computer molecular models of duplex DNA or RNA. The DNA and RNA duplexes are constructed according to parameters determined by x-ray diffraction of oligodeoxyribonucleotides in the B-form and oligoribonucleotidecontaining duplexes in the A-form.

In each of these constructed duplexes, one of the two sugar phosphate backbones is removed, and the prospective backbone, including the morpholino ring and intersubunit linkage, is replaced, if possible, on the sites of the bases from which the original sugar-phosphate backbone has been removed. Each resulting polynucleotide/polymer duplex is then examined for coplanarity of the Watson/Crick base pairs, torsional and angle strain in the prospective binding polymer backbone, degree of distortion imposed on the nucleic acid strand, and interstrand and intrastrand nonbonded interactions.

Initial studies of this type carried out in support of the invention show that a morpholino-based polymer has a preferred unit backbone length (i.e., the number of atoms in a repeating backbone chain in the polymer) of 6 atoms. However, the modeling studies also show that certain 5-atom and 7-atom repeating-unit morpholino-based backbones meet the requirements for binding to targeted genetic sequences.

Since the morpholino structure itself contributes 4 atoms to each repeating backbone unit, the linkages in the five-atom, six-atom, and seven-atom repeating-unit backbone contributes one, two, and three atoms to the backbone length, respectively. In all cases, the linkage between the ring structures is (a) uncharged, (b) chiral, (c) stable, and (d) must permit adoption of a conformation suitable for binding to the target polynucleotide.

Subunit backbone structures judged acceptable in the above modeling studies are then assessed for feasibility of synthesis. The actual chemical stability of the intersubunit linkage is often assessed with model compounds or dimers.

FIG. 3 shows several preferred .beta.-morpholino subunit types, including linkage groups, which meet the constraints and requirements outlined above. It will be appreciated that a polymer may contain more than one linkage type.

Subunit A in FIG. 3 contains a 1-atom phosphorous-containing linkage which forms the five atom repeating-unit backbone shown at A--A in FIG. 4, where the morpholino rings are linked by a 1-atom phosphonamide linkage. It is noted here that the corresponding chiral thionyl-containing linkage (substituting an S.dbd.O moiety for the phosphorous-containing group) was found to have inadequate stability in aqueous solution.

Subunit B in FIG. 3 is designed for 6-atom repeating-unit backbones, as shown at B--B, in FIG. 4. In structure B, the atom Y linking the 5'morpholino carbon to the phosphorous group may be sulfur, nitrogen, carbon or, preferably, oxygen. The X moiety pendant from the phosphorous may be any of the following: fluorine; an alkyl or substituted alkyl; an alkoxy or substituted alkoxy; a thioalkoxy or substituted thioalkoxy; or, an unsubstituted, monosubstituted, or disubstituted nitrogen, including cyclic structures. Several cyclic disubstituted nitrogen moieties which are suitable for the X moiety are morpholine, pyrrole, and pyrazole.

Subunits C-E in FIG. 3 are designed for 7-atom unit-length backbones as shown for C--C through E--E in FIG. 4. In Structure C, the X moiety is as in Structure B and the moiety Y may be a methylene, sulfur, or preferably oxygen. In Structure D the X and Y moieties are as in Structure B. In Structure E, X is as in Structure B and Y is O, S, or NR.

The most economical starting materials for the synthesis of morpholino-subunits are generally ribonucleosides. Typically, ribonucleosides containing hydrogen-bonding moieties or bases (e.g., A, U, G, C) are synthesized to provide a complete set of subunits for polymer synthesis. Where a suitable ribonucleoside is not available, a 1-haloribose or, preferably, a 1.alpha.-bromoglucose derivative, can be linked to a suitable base and this nucleoside analog then converted to the desired .beta.-morpholino structure via periodate cleavage, and closing the resultant dialdehyde on a suitable amine.

Because of the reactivity of the compounds used for subunit synthesis, activation, and/or coupling, it is generally desirable, and often necessary, to protect the exocyclic ring nitrogens of the bases. Selection of these protective groups is determined by (i) the relative reactivity of the nitrogen to be protected, (ii) the type of reactions involved in subunit synthesis and coupling, and (iii) the stability of the completed polymer prior to base deprotection.

Methods for base protecting a number of the more common ribonucleosides are given in Example 1. The methods detailed in the example are generally applicable for forming nucleosides with amine-protective groups. Standard base-protective groups used for nucleic acid chemistry are often suitable including the following groups: benzoyl for the N4 of C; benzoyl or p-nitrobenzoyl for the N6 of adenine (A); acetyl, phenylacetyl or isobutyryl for the N2 of guanine (G); and N2,N6-bis(isobutyryl) for 2,6-diaminopurine residues. These protective groups can be removed after polymer assembly by treatment with ammonium hydroxide.

It is sometimes desirable to protect the base portion of the morpholino subunit with a group which can be readily removed by other than a nucleophilic base. Suitable base protective groups removable by a strong non-nucleophilic base via a .beta.-elimination mechanism include: 2-(4-nitrophenyl)ethoxy carbonyl or 2-(phenyl sulfonyl)ethoxycarbonyl for both the N4 of C and the N6 of A; and the 9-fluorenyl methoxycarbonyl for the N2 of G and the N2 and N6 of 2,6-diaminopurine. These groups can be removed after polymer assembly by treatment with the strong nonnucleophilic base 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU), under stringently anhydrous conditions.

The syntheses of representative morpholino subunits are described particularly in Examples 2-7. With reference to the synthesis scheme depicted in FIG. 5, a base-protected ribonucleoside is reacted with sodium periodate to form a transient 2', 3'-dialdehyde which then closes upon ammonia to form a morpholino-ring having 2' and 3'hydroxyl groups (numbered as in the parent ribose, see FIG. 1). The compound is then treated with sodium cyanoborohydride to reduce the ring hydroxyl groups. The ring nitrogen is preferably protected by trityl derivatization or by a benzhydraloxycarbonyl group for subsequent subunit coupling. The protective group can be added by reacting the morpholino subunit with trityl chloride or with nitrophenyl benzhydrl carbonate or by reacting the dialdehyde with a primary amine, as illustrated in FIG. 6 and described in Examples 3 and 5. The stereochemistry of the nucleoside starting material is retained as long as the pH of the reaction mixture at the iminium stage is not allowed to go above about 10.

The above synthesis results in a morpholino-ring with an available 5'-hydroxyl. The 5'-hydroxyl can be converted to other active groups including 5' amine and sulfhydral (Example 6) or 5'phosphonate (Example 4).

In the above morpholino synthesis a variety of nitrogen sources can be used--including particularly ammonia, ammonium hydroxide, ammonium carbonate, and ammonium bicarbonate. Best results are obtained when the reaction solution is maintained near neutrality during the oxidation and morpholino ring closure reactions. This can be accomplished by continually titrating the reaction mix or, more conveniently, by using ammonium biborate as the ammonia source. When the solution is too acidic the yield of product is low and when it is too basic, side products (possibly due to epimerization of the 1' and/or 4' carbons) are produced which are difficult to separate from the desired product. It is also noted that the reducing agent can be added before, during, or after the oxidation step with little noticeable effect on product yield.

Ribonucleosides lacking base protection groups can also be successfully oxidized, ring closed, and reduced in aqueous solution to generate the morpholino ring. However, without base protection the number and quantity of undesired side products frequently increases, particularly in the case of cytidine.

The subunits formed by the above methods contain a 5'-OH, SH, or amine which is modified, reacted with, and/or activated, as described below, to be suitable for coupling to a second morpholino subunit. For example, FIG. 5 shows the conversion of a 5'-OH of a morpholino subunit to a phosphonyl linking moiety to form a subunit (Structure 10) which is linked to form a 5-atom unit-length backbone polymer. Details of the subunit synthesis are given in Example 4; modification to the thiophosphonyl linking moiety is also described.

Alternatively, the subunits are designed to include a phosphorous-containing group attached directly or indirectly to the morpholino ring nitrogen, which is coupled to a 5' moiety of a second morpholino subunit (FIG. 7). Subunits of this type are suitable for constructing morpholino polymers with 7-atom repeating-unit backbones.

An example of the synthesis of a subunit suitable for 7-atom unit-length backbones is detailed in Example 5 (with reference to FIG. 7).

Example 7 describes, with reference to Structure E of FIG. 3, the preparation of non-morpholino subunits which are converted into morpholino structures during polymer assembly.

The subunits prepared as above are coupled, in a controlled, sequential manner, often by activating the 5'hydroxyl of one subunit (having a protected morpholino nitrogen) and contacting this activated subunit with another subunit having an unprotected morpholino nitrogen as described in Example 9. It will be recognized that different types of linkages, such as those illustrated below, may be employed in the construction of a single polymer.

The simplest morpholino-type binding polymers are carbamate-linked polymers where the morpholino nitrogen is linked through a carbonyl to the 5' oxygen of another subunit. Experiments conducted in support of the present invention demonstrate that such a polymer effectively binds to a single-stranded DNA target sequence. However, in binding studies with an RNA target, the polymer exhibited unusual binding, as evidenced by a highly atypical hypochromicity profile in the 320 to 230 nm spectral range and lack of a normal thermal denaturation.

Early modeling studies indicated that in a carbamate-linked polymer bound to DNA existing in a B conformation, the backbone of the polymer provides adequate length for binding and the carbamate moieties of the polymer backbone can assume a nearly planar conformation. This modeling result was in good accord with the effective binding of the carbamate-linked polmers to DNA. In contrast, similar modeling studies suggested that binding of the carbamate-linked polymer to an RNA target requires one of the following: (i) the carbamate linkage of the polymer adopt a substantially nonplanar conformation, or (ii) the RNA target sequence adopt a strained conformation in which base-stacking interactions are quite different from that in a normal A conformation. This observation may explain the atypical binding of a carbamate-linked polymer to an RNA target sequence.

The modeling work further indicated that replacing the carbonyl intersubunit linking moiety with either an achiral sulfonyl-containing intersubunit linkage or with a chiral phosphorous-containing linkage would provide added length of about 0.32 angstrom per intersubunit linkage. Such linkages would also provide greater rotational freedom about key bonds, and bond angles of the intersubunit linkage compatible with an oligomer backbone conformation suitable for pairing to both RNA and DNA target sequences in their standard conformations.

The linkage in structure A--A in FIG. 4 (five-atom backbone) can be formed according to the reaction scheme shown in FIG. 9, and detailed in Example 8. Briefly, the 5'-OH of a morpholino subunit is converted to a phosphorous-containing moiety as described in Example 4. This group is activated and coupled to a second subunit having an unprotected ring nitrogen, as shown in FIG. 9 and described in Example 8. The polymer assembly is continued by deprotecting the morpholino ring nitrogen of the dimer, and reacting the dimer with a third activated subunit.

The phosphoramide linkage in Structure B--B of FIG. 4 (6-atom unit-length backbone) can be formed according to the reaction schemes shown in FIG. 10 and detailed in Example 9. The 5'hydroxyl of a protected subunit (Structure 4 of FIG. 5) is reacted with a suitable phosphorous-containing compound, such as dichloro-N,N-dimethylamino phosphate, resulting in an activated subunit. The subunit is then reacted with a second subunit having an unprotected morpholino ring nitrogen. A large number of variations are possible in the pendant X moiety and, as described in Example 9, the identity of the X moiety affects the ease of activation and coupling, the stability of the resulting linkage, and, to some extent, target-binding affinity.

In these syntheses the P.dbd.O group is essentially interchangeable with the P.dbd.S group; reactions with one are generally applicable to the other. An alternative method for forming linkages of type B--B of FIG. 4, as well as phosphonamide and phosphonoester linkages, is to use carbodiimide coupling: an exemplary, carbodiimide is dicyclohexylcarbodiimide (DCC). Carbodiimide coupling is described in Examples 8, 9, and 10. By exploiting an observation of Smith et al. (1958), the carbodiimide reagent can also be used to: (a) add a phosphorous (or thiophosphorous) linking moiety to a subunit; or (b) attach a pendant X moiety to a phosphorous (or thiophosphorous) linking moiety.

Additional linkages of the type B--B (FIG. 4) can be formed by converting the 5'hydroxyl to other functional groups (e.g., SH, CH.sub.2, NR) before activating and coupling the subunits into polymers.

A number of 7-atom unit length backbones prepared from the morpholino subunits (corresponding to Structures C--C and D--D in FIG. 4) allow even more flexibility in the construction of polymers which have specified distances between the base-pairing moieties. Using the 7-atom unit length linkages, distances between the morpholino-subunits, and consequently between the base pairing moieties, can be lengthened. Such lengthening of the intersubunit linkage is particularly useful when targeting duplex genetic sequences in a B conformation.

The 7-atom backbone polymers can be readily synthesized from the subunits C and D constructed as above, employing the general coupling reactions described in Example 10. For example, Structure C--C in FIG. 4 can be produced by (a) reacting the phosphonate (or thiophosphonate) group of subunit C (FIG. 3) with a carbodiimide, and (b) coupling the activated subunit with a second subunit having an unprotected morpholino ring nitrogen.

Similarly, structure D--D in FIG. 4 can be produced by activating the phosphonate (or thiophosphonate) with a carbodiimide, and coupling the activated subunit with a second subunit having an unprotected 5'oxygen, sulfur, or amine, as described in Example 10.

A novel method of forming linkages corresponding to Structure D--D/E--E of FIG. 4 entails oxidizing vicinyl hydroxyls of one subunit (FIG. 3E) and closing the resultant dialdehyde on a primary amine of another subunit followed by reduction with cyanoborohydride. In principle this same scheme could also be used to couple a secondary amine of one subunit and a mono-aldehyde of a second subunit; however, the coupling of a ribose-derived dialdehyde to a primary amine proceeds substantially faster and provides a better yield. Example 7 describes the synthesis of ribonucleosides containing a primary amine at the 5'. Their use in formation of morpholino polymers, as illustrated in FIG. 12, is described in Example 11.

After selecting a desired polymer length and recognition moiety sequence (guidelines for this are presented below), the polymer is assembled using the general procedures described above. One method of polymer assembly involves initial preparation of an appropriate set of dimers, linking selected dimers to form tetramers, linking these to form octamers, and so on. This method is carried out in solution, substantially according to the coupling methods described with reference to Examples 12 and 14. Example 12 outlines such a block assembly synthesis using monomers to form dimers, and dimers to form tetramers. It should be noted that couplings need not involve oligomers of equal size.

A particular merit of this block assembly method is that each coupling product is roughly twice the length of its precursors, so purification of the product of each coupling is simplified. Example 12 details the assembly of a 4-subunit polymer formed by this method.

The polymers may also be synthesized by stepwise subunit addition on a solid support. However, the optimal synthetic approach often uses a combination of the solution and solid support assembly methods where dimers, trimers, or tetramers are synthesized by solution phase and subsequently assembled into the full-length polymer on a solid support, as described in Example 13.

Typically, a solid support, such as glass beads derivatized with acid-stable, long-chain cleavable linkers, are employed as the support material, and prepared for attachment of the first subunit, or block of subunits, as described in Example 13. The glass beads are reacted with a subunit which generally has a readily cleavable protective group on a nitrogen. Whether the morpholino subunit is linked to the support via its morpholino nitrogen or a group at the 5' position depends on the direction of polymer synthesis, i.e. to which group the next subunit will be attached.

After coupling the second subunit (or oligomer which may be assembled in solution) to the support, any unreacted nucleophilic sites can be capped by addition of a suitable capping reagent, such as p-nitrophenyl acetate or acetic anhydride, and thereafter the support is washed. The protecting group on the nitrogen of the terminal subunit is removed, typically by acid treatment, and after neutralization, the support is reacted with an excess of the next-in-sequence subunit (or polymer unit) which is activated by one of the methods outlined above. One feature of the solid support assembly method is the need for high coupling efficiencies at each subunit addition step. This high coupling efficiency is generally achieved by addition of an excess of the activated subunit which maximizes the number of support-bound chains which are chain-elongated.

Chain elongation is continued in this manner, with optional capping of failure sequences after each subunit addition, until the polymer of the desired length and sequence is achieved.

After addition of the final subunit, the terminal backbone moiety may be reacted with a suitable charged or uncharged group, as described in Example 13. The polymer is then cleaved from the support, e.g., by treatment with either ammonium hydroxide or a non-nucleophilic base suitable for effecting .beta.-elimination in the linker joining the polymer to the support. The bases are deprotected and the polymer is purified as described below and in Example 13.

Binding polymers assembled in solution (Examples 12 and 14) are typically base-deprotected by suspending in DMSO or DMF and layering on the suspension an equal volume of concentrated ammonium hydroxide. The preparation is mixed with shaking and incubated at 30.degree. C. for 16 hrs. Workup includes removing the ammonia under reduced pressure. If a protective group (generally trityl or a related acid-labile moiety) is present, this group is cleaved and the crude polymer preparation is suspended in the appropriate buffer for purification (Example 13).

Binding polymers assembled by a solid-phase method (Example 13) wherein they are linked to the support via an ester linkage can be cleaved from the support by suspending the dried support in DMSO, layering on an equal volume of concentrated NH.sub.4 OH, capping tightly, and slowly agitating for 16 hrs at 30.degree. C. The solid support material is removed by filtration and the filtrate is treated as described above.

Alternatively, binding polymers linked to the support via a .beta.-elimination-sensitive linker can be cleaved from the support using a strong nonnucleophilic base 1,8 diazubicyclo(5.4.0.)undec-7-ene (DBU) in DMF. Using this approach one can release the polymer with its bases still protected and thus the polymer is suitable for further modification and/or structural confirmation via fast atom bombardment mass spectroscopy.

Purification of the base-deprotected polymer is preferably carried out at pH 2.5 or pH 11, depending on the pK of the base moieties in the polymer. At pH 2.5 cytosine, adenine, and 2-6-diaminopurine moieties carry a positive charge and guanine carries a partial positive charge. At pH 11 guanine, uracil and hypoxanthine carry a negative charge.

For polymers in which about 50% or more of the base-pairing moieties are ionized at pH 2.5, the purification can be carried out by cation exchange on a column of S-Sepharose fast-flow (Pharmacia) developed with a shallow NaCl gradient buffered at pH 2.5. The effluent is monitored at 254 nm and collected in a fraction collector. The full length polymer, which elutes after the shorter failure sequences, can be further purified and desalted on a column of chromatographic grade polypropylene (Polysciences Inc.), eluted with an aqueous gradient of acetonitrile adjusted to pH 2.5 with formic acid, with the eluant being monitored at 254 nm. The fractions containing the pure product are neutralized and dried under reduced pressure. Salts may be discarded by dissolving the polymer in trifluoroethanol, filtering, and evaporating the trifluoroethanol.

For polymers in which about 50% or more of the base-pairing moieties are ionized at pH 11, the purification may be performed on an anion exchange column of Q Sepharose fast-flow (Pharmacia) developed with an aqueous pH 11 gradient of NaCl. The full-length polymer, which elutes after shorter failure sequences, is further purified and desalted on a polypropylene column eluted with an aqueous pH 11 gradient of acetonitrile. Fractions containing the pure product are processed as above.

The purification methods described above should be carried out so that polymers containing adenine base-pairing moieties are not exposed to pH 11 for more than a few hours at room temperature, to avoid potential base lability problems. The details of the purification methods are outlined in Example 13.

In neutral, aqueous solution, longer morpholino polymers may have solubilities only in the sub-micromolar range. Therefore, it may be advantageous to enhance polymer solubility by addition of one or more hydrophilic moieties, e.g., polyethylene glycol. For most of the polymer types disclosed herein, this can be accomplished by cleaving the terminal backbone protective group from the completed polymer, and reacting the polymer, with the bases still in the protected state, with excess of carbonyldiimidazole-activated polyethylene glycol (PEG). Thereafter the binding polymer is treated with ammonium hydroxide to remove the base-protected groups, and the polymer is purified as above. The level of solubilization is easily adjusted through proper selection of the PEG material. Suitable PEG fractions having average molecular weights of 200, 400, 600, 1,000, 1,540, 3,400, 4,000, 6,000, 7,500, and 18,500 daltons are commercially available (e.g., Polysciences, Inc.) with PEG1000 often providing the best solubilization. The solubilizing moiety may be linked to the polymer through a cleavable linkage, if desired, to allow the polymer to be released from the solubilizing agent, e.g., by esterase or peptidase enzymes.

It will be appreciated that the polymer may be further derivatized or labeled according to known procedures. For example, the polymer may be radiolabeled by preparing the polymer subunits from radiolabeled ribonucleosides or by attaching a radiolabeled amino acid at one terminus. The polymer may be readily derivatized, e.g., employing modifications of the above subunit coupling reactions, with enzymes, chromophoric groups, or the like, where the polymer is to be used as a diagnostic probe. Further, the polymer may be derivatized with biomolecules which serve to target the polymers to specific tissues or cell types.

Fully-protected binding polymers of moderate size (10 to 20 subunits) often give a strong molecular ion in FAB (Fast Atom Bombardment) mass spectroscopy, providing a key confirmation of the polymer length.

Further, COSY-NMR (two-dimensional correlated spectroscopy) of the deprotected and purified polymer provides information on the ratio of the different base-pairing moieties in the polymer as well as quantitative information on the ratio of binding polymer to any solubilizing or other type moiety which may have been linked thereto.

Mobilities on ion exchange columns also provide information on the number of C+A base-pairing moieties in a polymer when purification is carried out at pH 2.5 and information on the number of G+U residues when the purification is run at pH 11. Structural verification is easiest when the polymers have been assembled from oligomer blocks, such as in Examples 12, 13 and 14, since any failure sequences then differ more substantially from the full-length sequences.

The UV profiles of the polymers at pH 1, 7, and 13 can provide information about the relative nucleotide composition of the polymer.

Assessment of a morpholino-based polymer's affinity for its target sequence is carried out by examining the melting curve of the polymer/target duplex, as illustrated in Example 14. Further, comparisons can be made between the melting curve of a regular nucleic acid duplex (such as p(dC).sub.6 /p(dG).sub.6) and the melting curve of a hybrid duplex containing a corresponding morpholino-based polymer (such as (morpholino C).sub.6 /p(dG).sub.6).

The above characterization steps have been applied to a morpholino-based phosphordiamidate-linked poly(C) hexamer where Y is oxygen, X is N(CH.sub.3).sub.2, and Z is oxygen, as described in Example 14. Characterization of the full-length oligomer was achieved by proton NMR and negative ion FAB mass spectroscopy. With these morpholino oligomers, the fragmentation of the oligomers is greatly suppressed so that little sequence information is available. However, the parent ion signal is quite strong and allows confirmation of the composition of the morpholino oligomer (see Example 14).

In order to increase water solubility a polyethylene glycol (PEG) tail was attached to the oligomers. 5 equivalents of PEG 1000 was treated with one equivalent of bis(p-nitrophenyl)carbonate to give monoactivated PEG. Detritylation of the hexamer with 1% acetic acid in trifluoroethanol afforded a free morpholino ring nitrogen. Treatment of the hexamer containing the free amine with activated PEG1000 under standard coupling conditions resulted in attachment of the PEG tail to the hexamer. The bases were deprotected by treatment of the tailed hexamer with concentrated ammonia for 24 hours. The tailed hexamer was taken up in pH 2.5 buffer and purified by cation exchange chromatography on S-Sepharose Fast Flow.TM. eluted with a potassium chloride gradient. After neutralization the eluant was desalted on a polypropylene column eluted with a water/acetonitrile gradient. The tailed hexamer was found to be freely soluble in pH 7.5 buffer.

The stability of complexes of the tailed hexamer with complementary nucleic acids was investigated by thermal denaturation experiments. Difference spectra between mixed and unmixed samples of the tailed hexamer and the selected phosphodiester complement were obtained from 14.degree. C. to 85.degree. C. and over the 320 to 260 nm range (see Example 14). At 60 micromolar in C monomer and 60 micromolar in G monomer the difference UV spectrum of the tailed hexamer, (morphC).sub.6 with poly(dG) gave a T.sub.m value of 79.degree. C. The corresponding (morphC).sub.6 with poly(G) gave a Tm value of 51.5.degree. C. (see Example 14 and FIG. 13).

The target-specific polymers of the invention can be used in a variety of diagnostic assays for detection of RNA or DNA having a given target sequence. In one general application, the polymers are labeled with a suitable radiolabel or other detectable reporter group. Target polynucleotide, typically a single stranded polynucleotide which is bound to a solid support, is reacted with the polymer under hybridization conditions, allowed to anneal, and then the sample is examined for the presence of polymer reporter group.

The diagnostic assay can be carried out according to standard procedures, with suitable adjustment of the hybridization conditions to allow polymer hybridization with the target region. In this regard, it is noted that the polymer can be designed for hybridization with the target at a higher melting temperature than the complementary polynucleotide strand, since polymer binding does not entail backbone charge repulsion effects. Therefore, the polymer can bind to the target at a temperature above the normal polynucleotide melting temperature, an important advantage of the polymer over conventional oligonucleotide probes. This binding at elevated temperature minimizes the problem of competition for binding to the target between the probe and any corresponding single-strand oligonucleotide which may be present in the diagnostic mixture.

In a second general type of diagnostic application, the polymers are linked to a solid support, for capture of target RNA or DNA to the support. The solid support, e.g., polymeric microparticles, can be prepared by linking the polymers to the support according to the methods described above or by conventional derivatization procedures. Alternatively, where the polymers are synthesized on a solid support this support may also serve as the assay support.

According to an important feature of this assay system, the target polynucleotide molecules which are captured on the support by base-specific binding to the polymers can be detected on the basis of their backbone charge, since the support-bound polymers are themselves substantially uncharged. To this end, the assay system may also include polycationic reporter molecules which are designed to bind to the fully charged analyte backbone, but not the uncharged (or substantially uncharged) polymer backbone, under selected binding conditions.

In one embodiment the reporter molecules are composed of a polycationic moiety or tail designed to bind electrostatically to a fully charged polynucleotide, under conditions where the reporter does not bind to the less charged or uncharged binding polymer carried on the diagnostic reagent; one or more reporter groups may be attached to the tail, adapted to produce a signal by which the presence of the reporter can be detected. Methods for forming polycationic molecules and for attaching reporter molecules to cationic compounds are known.

Each reporter molecule carries one or more reporter groups, and each polynucleotide can accommodate binding of typically several thousand or more reporter molecules. Thus the system has an amplification factor, in terms of reporter signal per bound analyte molecule, of several orders of magnitude. In addition, the method has the advantage, noted above, that the polynucleotide binding reaction can be carried out under conditions in which binding competition with complementary nucleotide strands does not occur.

The design considerations applied in preparing a polynucleotide binding polymer for use as a diagnostic reagent are governed by the nature of the target analyte and the reaction conditions under which the analyte is to be assayed. As a first consideration, there is selected a non-homopolymeric target base sequence against which the polymer is directed. This target sequence is generally single-stranded and preferably unique to the analyte being assayed.

The probability of occurrence of a given n-base target sequence is approximately (1/4).sup.n. Accordingly, a given n-base target sequence would be expected to occur approximately once in a polymer containing 4.sup.n bases. Therefore, the probability P that a given n-base sequence will occur in polynucleotides containing a total of N unique-sequence bases is approximately P=N/4.sup.n. To illustrate, the probability P that a 9-base target sequence will be found in a 20 kilobase polynucleotide is about 20.times.10.sup.3 /2.times.10.sup.5 or 0.08, the probability that a 16-base target sequence will be present is about 20.times.10.sup.3 /4.3.times.10.sup.9 or 0.0000047. From these calculations, it can be seen that a polymer having 9-16 recognition moieties specific for a defined 9-16 base target sequence should have high specificity for the target sequence in an assay mixture containing only viral genomes, whose greatest complexities correspond to about 400K of unique-sequence bases.

Similar calculations show that a 12 to 16 subunit polymer can provide adequate specificity for a viral or bacterial target sequence in an assay mixture containing viral and bacterial genomic material only; largest genomic sizes about 5,000 kilobases. A 16 to 22 subunit polymer can provide adequate specificity for a target sequence in a polynucleotide mixture containing mammalian genomic DNA material; genomic sizes of about 5 billion base pairs of unique-sequence DNA.

The polymer/analyte binding affinity, and particularly the temperature at which the polymer just binds with the target sequence (the melting temperature, or Tm) can be selectively varied according to the following criteria: (a) number of subunits in the polymer; (b) the number of hydrogen bonds that can be formed between the base-pairing moieties and the corresponding, complementary bases of the analyte target sequence; (c) unit length of the polymer backbone; (d) the particular intersubunit linkages; and (e) concentration of denaturants, such as formamide, which reduces the temperature of melting.

From a number of studies on model nucleic acid duplexes it is known that the melting temperature of oligonucleotide duplexes in the 10 to 20 bp range increases roughly 3.degree. C. per additional base pair formed by two hydrogen bonds, and about 6.degree. C. per additional base pair formed by three hydrogen bonds. Therefore, the target sequence length originally selected to insure high binding specificity with the polymer may be extended to achieve a desired melting temperature under selected assay conditions.

Also, where the recognition moieties used in constructing the polymer are the standard nucleic acid bases the target sequence may be selected to have a high percentage of guanine plus cytosine bases to achieve a relatively high polymer/analyte melting temperature. On the other hand, to achieve a lower melting temperature a target sequence is selected which contains a relatively high percentage of adenine plus thymine bases.

The binding components in the diagnostic system, as they function in the solid-support diagnostic method just described, are illustrated in FIG. 16. Here "S", the assay reagent, is the solid support having a number of binding polymers attached to its surface through spacer arms indicated by sawtooth lines. In the assay procedure, the target DNA in single strand form is reacted with the support-bound polymers under hybridization conditions, and the solid support is then washed to remove non-hybridized nucleic acid material.

The washed support is then reacted with the reporter, under conditions which favor electrostatic binding of the reporter cationic moiety to the target DNA backbone. The reporter shown in FIG. 16 is a dicationic molecule having a reporter group R.

After reaction with the reporter solution, typically at room temperature for 1-2 minutes, the reagent is washed to remove unbound reporter, and then the assay reagent is assessed for bound reporter. One approach in determining the amount of reporter associated with the reagent, particularly in the case of fluorescent or chromophoric reporter groups, is to elute the reporter from the reagent with a high salt solution and then assess the eluate for reporter.

It is also noted here that the polymer of the invention can undergo sequence-specific binding to duplex nucleic acids, via base-pair-specific hydrogen bonding sites which are accessible through the major groove of the double helix. This bonding can occur in a duplex region in which at least 70% of the bases on one strand are purines and a corresponding percent of the bases on the other strand are pyrimidines. The duplex binding polymer preferably includes 2-aminopurine or 2,6-diaminopurine hydrogen bonding moieties for binding to T-A or U-A base pairs, and guanine or thioguanine hydrogen-bonding moieties for binding to C-G base pairs as illustrated in FIG. 8. Thus, for these special target sequences, the polymer of the invention can be used for diagnostic assays of the types just described, but where the target nucleic acid is in nondenatured, duplex form.

The polymers of the instant invention can be used in place of standard RNA or DNA oligomers for a number of standard laboratory procedures. As mentioned above, morpholino-based polymers can be fixed to a solid support and used to isolate complementary nucleic acid sequences, for example, purification of a specific mRNA from a poly-A fraction (Goldberg et al). The instant polymers are advantageous for such applications since they are inexpensive and straightforward to prepare from activated subunits.

A large number of applications in molecular biology can be found for labeled morpholino-based polymers. Morpholino-based polymers can be easily and efficiently end-labelled by the inclusion in the last step of the polymer synthesis an activated and labelled morpholino-based subunit or, preferably, an .sup.35 S-labelled methionine, as indicated above. The type of label to be used is dependent on the final application of the polymer; such labels include radioactive (.sup.3 H, .sup.14 C, .sup.32 P, or .sup.35 S) nucleosides or biotin. Labelled morpholino-based oligonucleotide analogs can act as efficient probes in, for example, colony hybridization (Grunstein et al), RNA hybridizations (Thomas), DNA hybridizations (Southern), and gene bank screening (Szostak et al).

The polymers of the invention also have important potential use as therapeutic agents. Recently, uncharged anti-sense nucleic acid analogs, which are nearly isostructural with DNA, have been used as anti-viral and anti-tumor agents. The polymers of the present invention provide several advantages over the more conventional anti-sense agents.

First, the morpholino polymers are substantially less expensive to synthesize than oligonucleotides. This is due in part to the fact that the morpholino subunits used in polymer synthesis are derived from ribonucleosides, rather than the much more expensive deoxyribonucleosides. Also, as noted above, the coupling reaction between a phosphorous and an amine of a second subunit occurs under relatively mild conditions, so that protection steps and other precautions needed to avoid unwanted reactions are simplified. This is in contrast to standard ribo- and deoxyribonucleotide polymer synthesis where coupling through a phosphate ester linkage requires that the coupling reagents be highly reactive and that the reaction be carried out under stringent reaction/protection conditions. This advantage in polymer synthesis also applies, of course, to diagnostic uses of the polymer.

Second, polymer binding to its target may give substantially better target inactivation, since the polymer/target duplex is not susceptible to duplex unwinding mechanisms in the cell.

Third, the morpholino-based polymer is also more stable within the cell; the polymer backbone linkage is not susceptible to degradation by cellular nucleases.

Fourth, in therapeutic applications involving cellular uptake of the compound, the uncharged morpholino polymer is much more likely to efficiently enter cells than a charged oligonucleotide.

In the context of therapeutic applications, the morpholino polymers of the present invention may be targeted against double-stranded genetic sequences in which one strand contains predominantly purines and the other strand contains predominantly pyrimidines.

Further, when a messenger RNA is coded by the mostly purine strand of the duplex target sequence, morpholino binding polymers targeted to the duplex have potential for also inactivating the mRNA. Thus such a polymer has the potential for inactivating key genetic sequences of a pathogen in both single-stranded and double-stranded forms.

In 1981 it was reported that short (3 to 7 subunits) methylphosphonate-linked DNA analogs complementary to portions of the Shine-Dalgarano consensus sequence of procaryotic mRNAs were effective in disrupting bacterial protein synthesis in bacterial lysates and in a special permeable strain of bacteria. However, such agents failed to inhibit protein synthesis in normal bacteria (Jayaramon, 1981).

Experiments performed in support of the instant invention show that polymers of 3 to 5 subunits in length can be effective to block protein synthesis in normal bacteria by using a combination of bases which result in a high target-binding affinity. More specifically, the following oligomers and oligomer combinations can perturb protein synthesis in normal intact bacteria (where D is 2,6-Diaminopurine or Adenine; G is Guanine; B is 5-Bromouracil, other 5-Halouracil or Uracil; and sequences are shown with their 5' end to the left): DGG, BDDG, DDGG; DGGD; GGDG; GDGG; DGGB; GGBG; GGAGG; GGDGG; and the combinations BDD+GGDG; DDG+GDGG; DGG+DGGB; GGD+GGBG; BDDG+GDG; DDGG+DGG; DGGD+GGB; GGDG+GBG; BDD+GGDG+GBG.

The use of short binding-enhanced oligomers to disrupt the biological activity of an RNA sequence which plays a key role in the metabolism of a target class of organisms but not a correspondingly important role in higher organisms should be broadly adaptable to a variety of pathogenic organisms (e.g., bacteria and fungi) having a cell wall which excludes the entrance of longer polymers.

The following examples illustrate methods of subunit and polymer synthesis, and uses of the polymer composition of the invention. The examples are in no way intended to limit the scope of the invention.

The following ribonucleosides are obtained from Sigma Chemical Co. (St. Louis, Mo.): uridine, guanosine, 5-methyluridine, adenosine, cytidine, 5-bromouridine, and inosine.

2,6-diamino-9-(B-D-ribofuranosyl)-9H-purine (2,6-diaminopurine riboside) is obtained from Pfaltz and Bauer, Inc., Division of Aceto Chemical Co., Inc. (Waterbury, Conn.).

The following nucleosides are prepared by the literature methods indicated:

1-.beta.-D-ribofuranosyl)-2-pyrimidinone (2-hydroxy-pyrimidine riboside) is prepared by the procedure of Niedballa.

2-amino-9-.beta.-D-ribofuranosyl)-1,6-dihydro-6hpurine-6-thione (thioguanosine) is prepared by the procedure of Fox.

Dimethoxytrityl chloride, N-6-benzoyladenosine, N-4-benzoylcytidine, and N-2-benzoylguanosine are obtained from Sigma Chemicals. 9-fluorenylmethoxycarbonyl chloride (FMOC chloride), trimethylchlorosilane, isobutyric anhydride, 4-nitrobenzoyl chloride, naphthalic anhydride, and all organic solvents for reactions and chromatography were obtained from Aldrich Chemical Co. (Milwaukee, Wis.). Silica Gel is obtained from EM Science (Cherry Hill, N.J.).

When activation of the subunits is achieved using dihalogenated electrophiles (e.g., P(O)Cl.sub.2 N(CH.sub.3).sub.2 or P(S)Cl.sub.2 N(CH.sub.3).sub.2), better yields of activated subunits are often obtained by using protective groups which leave no acidic protons on the purine and pyrimidine exocyclic amines. Examples of such exocyclic amine moieties are as follows: the N6 of adenine, the N4 of cytosine, the N2 of guanine, and the N2 and N6 of diaminopurine. Suitable protective groups for this purpose include the naphthaloyl group (Dikshit) and the amidine groups developed by McBride et al (1986). In addition, use of dihalogenated electrophiles for subunit activation generally requires that the O6 of guanine moieties is protected; this protection is achieved using the diphenylcarboamoyl group (Trichtinger).

In order to minimize side reactions during subunit activations it is often desirable to protect the guanine moity on both the N2 and O6 using the procedure of Trichtinger et al. (1983).

The N-2 9-fluorenylmethoxycarbonyl derivative of guanosine is prepared by the procedure below which is general for the protection of nucleoside amino groups: guanosine (1 mmol) is suspended in pyridine (5 ml) and treated with trimethyl-chlorosilane (5 mmol). After the solution is stirred for 15 minutes, 9-fluorenylmethoxycarbonyl chloride (5 mmol) is added and the solution is maintained at room temperature for 3 hours. The reaction is cooled in an ice bath and water (1 ml) is added. After stirring for 5 minutes conc. ammonia (1 ml) is added, and the reaction is stirred for 15 minutes. The solution is evaporated to near dryness and the residue is dissolved in chloroform (10 ml). This solution is washed with sodium bicarbonate solution (5 ml, 10%), dried over sodium sulfate and evaporated. The residue is coevaporated several times with toluene and the product chromatographed on silica gel using a gradient of methanol in methylene chloride (0-50%).

N-2-Isobutyrylguanosine is prepared by the method of Letsinger. Further protection of the O6 with a nitrophenethyl moiety is often desirable and can be carried out by several methods (Gait, 1984).

N-2-acetylguanosine is obtained by the method of Reese.

N-2-naphthaylguanosine is prepared by the method of Dikshit; this reference provides a general method for the protection of nucleoside amine groups.

The N-6 2-(4-nitrophenyl)-ethoxycarbonyl derivative is prepared by the method of Himmelsbach.

N-6 (4-nitrobenzoyl)adenosine is prepared using the procedure above for FMOC-guanosine except that 4-nitrobenzoyl chloride is substituted for FMOC chloride.

The N-6 2-(phenylsulfonyl)-ethoxycarbonyl derivative is prepared by the procedure for FMOC guanosine except the 2-(phenylsulfonyl)-ethyl chloroformate (Balgobin) is used as the acylating agent and N-methylimidazole or pyridine is used as the solvent.

N-6 naphthoyladenosine is prepared by the method of Dikshit; this reference provides a general method for the protection of nucleoside amine groups.

The N-2,N-6-bis(9-fluorenylmethoxycarbonyl) derivative of 2,6-diaminopurine riboside is prepared by the general procedure described for guanosine.

The N-2,N-6-bis(isobutyryl) derivative is prepared by the general procedure described for guanosine.

The N-2(9-fluorenylmethoxycarbonyl) derivative of thioguanosine is prepared by the general procedure described for guanosine.

To minimize undesired side products during the subunit activation step it is sometimes desirable to protect the N3 of the uracil moiety. 5'O-tritylated uridine-2',3'-acetonide is converted to the N3 anisoyl derivative by the procedure of Kamimura et al. (1983). The product is then treated with hot 80% acetic acid or 0.1N HCl in THF to cleave the protective groups on the ribose moiety.

The steps in the method are illustrated in FIG. 5, with reference to structures shown in FIG. 5.

The base-protected ribonucleoside is oxidized with periodate to a 2'-3' dialdehyde (Structure 1). The dialdehyde is closed on ammonia or primary amine (Structure 2) and the 2' and 3' hydroxyls (numbered as in the parent ribose) are removed by reduction with cyanoborohydride (Structure 3).

An example of this general synthetic scheme is described below with reference to the synthesis of a base-protected cytosine (P.sub.i *) morpholino subunit. To 1.6 l of methanol is added, with stirring, 0.1 mole of N-4-benzoylcytidine and 0.105 mole sodium periodate dissolved in 100 ml of water. After 5 minutes, 0.12 mole of ammonium biborate is added, and the mixture is stirred 1 hour at room temperature, chilled and filtered. To the filtrate is added 0.12 mole sodium cyanoborohydride. After 10 minutes, 0.20 mole of toluenesulfonic acid is added. After another 30 minutes, another 0.20 mole of toluenesulfonic acid is added and the mixture is chilled and filtered. The solid precipitate is washed with two 500 ml portions of water and dried under vacuum to give the tosylate salt of the free amine shown in Structure 3.

The use of a moderately strong (pKa<3) aromatic acid, such as toluenesulfonic acid or 2-naphthalenesulfonic acid, provides ease of handling, significantly improved yields, and a high level of product purity.

The base-protected morpholino subunit is then protected at the annular nitrogen of the morpholino ring using trityl chloride or benzyhydral nitrophenyl carbonate (Structure 4).

Alternatively, the 5' hydroxyl can be protected with a trialkylsilyl group.

As an example of a protection step, to 2 liters of acetonitrile is added, with stirring, 0.1 mole of the tosylate salt from above followed by 0.26 mole of triethylamine and 0.15 mole of trityl chloride. The mixture is covered and stirred for 1 hour at room temperature after which 100 ml methanol is added, followed by stirring for 15 minutes. After drying by rotovaping, 400 ml of methanol is added. After the solid is thoroughly suspended as a slurry, 5 liters of water is added, the mixture is stirred for 30 minutes and filtered. The solid is washed with 1 liter of water, filtered, and dried under vacuum. The solid is resuspended in 500 ml of dichloromethane, filtered, and rotovaped until precipitation just begins, after which 1 liter of hexane is added and stirred for 15 minutes. The solid is removed by filtering, and dried under vacuum.

The above procedure yields the base-protected morpholino subunit tritylated on the morpholino nitrogen and having a free 5' hydroxyl (Structure 4).

This example describes an alternative preparation of a morpholino subunit containing an acid-labile moiety linked to the morpholino ring nitrogen. The steps are described with respect to FIG. 6.

The subunit is prepared by oxidizing a ribonucleoside with periodate, as in Example 2, and closing the resultant dialdehyde (Structure 1) on the primary amine 4,4'-dimethoxybenzhydrylamine (which can be prepared by the method of Greenlee, 1984) buffered with benzotriazole, or p-nitrophenol. Reduction with sodium cyanoborohydride, carried out as in Example 2, gives a morpholino subunit (Structure 2) having a 4,4,-dimethoxybenzhydryl group on the morpholino nitrogen.

It is noteworthy that this procedure is particularly useful for preparing morpholino subunits from ribonucleosides which do not have a protective group on the base (e.g., uridine).

The steps in the method are described with reference to structures in FIG. 5.

The 5' hydroxyl of the doubly-protected morpholino subunit (Structure 4) is converted to a phosphonate as follows. N4-Benzoylcytidine-2',3'-acetonide (1 mmole) is converted to the 5'-Iodo derivative by reaction with methyltriphenoxyphosphonium iodide in DMF (20 ml) under argon at room temperature for 20 hours. Methanol (5 ml) is added to the reaction mixture and after 30 minutes the mixture is evaporated in vacuo. The residue is dissolved in ethyl acetate and the solution washed first with aqueous sodium thiosulfate and then with brine. After drying with sodium sulfate and evaporation of the solvent, the product is purified by chromatography on silica using isopropanol/chloroform solvents.

The iodide derivative prepared above is treated with a large excess of anhydrous phosphine in ethanol for two days at 50.degree. C. is a well-sealed vessel. At the end of this time the vessel is cooled, vented, and the excess phosphine allowed to slowly evaporate. The expelled vapors are bubbled into a solution containing sodium hypochlorite to decompose the toxic gas. The alcoholic solution of the primary phosphine is treated, while cooling the solution, with solid sodium carbonate and an excess of 30% hydrogen perioxide with cooling. The product phosphonic acid is purified by ion exchange chromatography on an anion exchange column.

This dianionic phosphate product, where the counter ions are triethylammonium, is mixed with an excess of reagent which wll result in the carbodiimide-mediated addition of the desired pendant X moiety (FIG. 3a). Examples of suitable reagents are as follows: addition of methanol gives X=methoxy; and addition of dimethylamine gives X=N(CH.sub.3).sub.2. To this mixture a carbodiimide, such as DCC, is added. The resulting subunit is of the form shown in FIG. 3a. A substantially less basic counter ion (e.g., pyridine) should not be used, since it would allow two X moieties to be added to the phosphonate moiety.

This example describes the preparation of a subunit containing a methylphosphonate moiety linked to the morpholino ring nitrogen suitable for preparing polymers (FIG. 3d) with 7-atom unit-length backbones. The steps are described with respect to structures shown in FIG. 7.

A base-protected ribonucleoside is reacted with di(p-methoxy)trityl chloride to give Structure 1. The ribose moiety is then oxidized with periodate in the presence of aminomethanephosphonic acid (AMPA) and N-ethyl morpholine. The oxidation is followed by reduction with sodium cyanoborohydride in the presence of benzotriazole (used to buffer the reaction mix) to give a morpholino subunit having a methanephosphonic acid group on the morpholino nitrogen (Structure 2). Thereafter the product is purified by silica gel chromatography developed with a chloroform/methanol mixture 1% in triethylamine.

This dianionic phosphonate product, where the counter ions are triethylammonium, is mixed with an excess of reagent which will result in the addition of the desired pendant X moiety (FIG. 3d). Examples of suitable reagents are as follows: addition of methanol gives X=methoxy; and addition of dimethylamine gives X=N(CH.sub.3).sub.2. To this mixture a carbodiimide, such as dicyclohexylcarbodiimide (DCC) is added to give Structure 3 of FIG. 7. If a substantially less basic counter ion (e.g., pyridine) is used, then two X moieties may add to the phosphonate moiety.

The steps in the synthesis are described with reference to structures shown in FIG. 5.

The 5'hydroxyl of the doubly-protected morpholino subunit (Structure 4, FIG. 5) can be converted to an amine as follows. To 500 ml of DMSO is added 1.0 mole of pyridine (Pyr), 0.5 mole of triflouroacetic acid (TFA), and 0.1 mole of the morpholino subunit. The mixture is stirred until all components are dissolved and then 0.5 mole of either diisopropylcarbodiimide (DIC) or dicyclohexylcarbodiimide (DCC) is added. After 2 hours the reaction mixture is added to 8 liters of rapidly stirred brine. This solution is stirred for 30 minutes and then filtered. The resultant solid is dried briefly, washed with 1 liter of ice cold hexanes and filtered. The solid is added to 0.2 mole of sodium cyanoborohydride in 1 liter of methanol and stirred for 10 minutes. To this mixture, 0.4 mole of benzotriazole or p-nitrophenol is added, followed the addition of 0.2 mole of methylamine (40% in H.sub.2 O). The preparation is stirred four hours at room temperature. [Note: the benzotriazole or p-nitrophenol buffers the reaction mixture to prevent racemization at the 4' carbon of the subunit at the iminium stage of the reductive alkylation.] Finally, the reaction mixture is poured into 5 liters of water, stirred until a good precipitate forms, and the solid (Structure 6, FIG. 5) is collected and dried.

The 5' hydroxyl of the doubly-protected morpholino subunit is converted to a sulfhydral as follows. One-tenth mole of the 5'-hydroxyl subunit (Structure 4, FIG. 5) is added to 1 liter of pyridine, followed by the addition of 0.12 mole of toluenesulfonylchloride. This solution is stirred for 3 hours at room temperature and the reaction yeilds the product shown as Structure 7 of FIG. 5. The pyridine is removed by rotovapping, and to the solid is added 0.5 mole of fresh sodium hydrosulfide in 1 liter of methanol. This reaction mixture is stirred at room temperature overnight. The mixture is added to 5 liters of water, stirred 20 minutes, and the resulting solid material is collected by filtration and dried to give the product shown as Structure 8 of FIG. 5.

This example describes the preparation of a riboside subunit containing an aminomethylphosphonate moiety linked to the riboside. The structures referred to in this example are shown in FIG. 12.

Aminomethylphosphonic acid (Aldrich Chem. Co.) is reacted with trityl chloride in the presence of triethylamine. The dianionic phosphonate product, where the counter ions are triethylammonium, is mixed with an excess of reagent suitable for adding the desired pendant X moiety (e.g., addition of methanol gives X=methoxy, and addition of dimethylamine gives X=N(CH.sub.3).sub.2) and then a carbodiimide, such as dicyclohexylcarboiimide (DCC), is added. The resultant monoanionic product is shaken with a mixture of water and chloroform containing pyridinium hydrochloride. This procedure gives a monoionic phosphonic acid having a pyridinium counter ion. This product is added to chloroform containing N4-Benzoylcytidine-2',3'-phenylboronate and DCC is added. The product is dried and chromatographed on silica using methanol/chloroform mixtures. The pure product is next treated with 1,3-dihydroxypropane to give Structure 2 of FIG. 12, and a portion is further treated with acetic acid in trifluoroethanol to give Structure 1.

This example describes the coupling of a phosphonate subunit, prepared as in Example 4, with a second subunit having a free morpholino ring nitrogen. The Example is descried with reference to structures in FIG. 9.

The starting material is the base-protected, morpholino nitrogen protected phosphonate subunit prepared in Example 4. A triethylamine salt of this subunit is suspended in chloroform and shaked with water containing pyridinium hydrochloride, resulting in the pyridine salt (Structure 1) of the subunit in the chloroform phase. The organic phase is separated, washed with water, and dried. One portion of the resulting solid is mixed with an escess of 3-hydroxypropionitrile; DCC is then added to the mixture. The product of the resulting reaction is detritylated with 2% acetic acid in trifluoroethanol to give Structure 2. Structures 1 and 2 are then mixed in the presence of DCC resulting in the coupled dimer shown as Structure 3. This dimer may be selectively deprotected on either end for assembly into longer oligomer blocks or polymers. The cyanoethyl group may be removed with DBU in a nonprotic solvent (e.g., DMF) or the trityl moiety may be cleaved as described above.

Example 9A describes the coupling of a 5'hydroxyl subunit, prepared as in Example 2 or 3, to a second subunit having a free morpholino ring nitrogen to give an alkylphosphonamidate intersubunit linkage. The example is described with reference to structures in FIG. 10 where X is an alkyl group.

One mmole of 5'hydroxyl subunit, base-protected and tritylated on the morpholino nitrogen (Structure 4 of FIG. 5), is dissolved in 20 ml of dichloromethane. To this solution 4 mmole of N-ethylmorpholine and 1.1 mmole of methylphosphonic dichloride, for Z.dbd.O (or methylthiophosphonic dichloride, for Z.dbd.S), are added, followed by the addition of 1 mmole of N-methylimidazole. After one hour the reaction solution is washed with aqueous NaH.sub.2 PO.sub.4. The activated subunit is isolated by chromatography on silica gel developed with ethyl acetate (Structure 2 of FIG. 10 where X is methyl). This activated subunit can be directly linked to the morpholino nitrogen of a second subunit (Structure 3) by mixing them in DMF. This coupling reaction yields the dimer shown as Structure 4 (where X is --CH.sub.3).

Alternatively, a small amount of water can be added to the reaction solution instead of washing with NaH.sub.2 PO.sub.4. The resulting solution is stirred for 10 minutes to effect conversion of the activated subunit (Structure 2) to the phosphonate salt (Structure 5, where X is methyl and the counter ion is N-ethylmorpholine). The product is purified by silica gel chromatography developed with a methanol/chloroform mixture 1% in triethylamine. The purified product is shaken with chloroform/water containing 2% pyridinium hydrochloride to change the counter ion to pyridinium. After drying, the product is suitable for coupling to a 5'-protected subunit, having a free morpholino nitrogen (Structure 6), using DCC (dicyclohexylcarbodiimide); preferably, the reaction is carried out in dichloromethane. The carbodiimide coupling reaction yields the dimer shown as Structure 7 (where X is methyl).

The alkylphosphonoamidate intersubunit linkage is very stable to ammonia used for base deprotections. In contrast, the linkage is sensitive to strong acids. For instance, the linkage has a half time of cleavage of about 3 hours in 2% dichloroacetic acid in dichloromethane. However, the linkage showed no detectable cleavage after 18 hours in 2% acetic acid in trifluoroethanol, conditions suitable for detritylation of the morpholino nitrogen.

Example 9B describes the coupling of a 5'hydroxyl subunit, prepared as in Example 2 or 3, to a second subunit having a free morpholino ring nitrogen to give a phosphodiesteramide intersubunit linkage. The example is described with reference to structures in FIG. 10 where X is an alkoxide group.

One mmole of 5'hydroxyl subunit, base-protected and tritylated on the morpholino nitrogen (Structure 4 of FIG. 5), is suspended in 80 ml of benzene and 2.2 mmole of N-methylimidazole is added. After the subunit is dissolved 1.2 mmole of ethyl dichlorophosphate for Z.dbd.O (or ethyldichlorothiophosphate for Z.dbd.S) are added. After an hour the reaction solution is washed with aqueous NaH.sub.2 PO.sub.4. The activated subunit is isolated by chromatography on silica gel developed with ethyl acetate (Structure 2 in FIG. 10, where X is --O--CH.sub.2 CH.sub.3). This activated subunit can be directly linked to the morpholino nitrogen of a second subunit (Structure 3) by mixing in DMF. This coupling reaction yields the dimer shown as Structure 4.

When ethyldichlorothiophosphate (Z.dbd.S) is used for activation of the subunits, improved yields are obtained with the following modifications. One mmole of 5' hydroxyl subunit, base-protected and tritylated on the morpholino nitrogen (Structure 4 of FIG. 5), is suspended in 20 ml of chloroform. To this solution 1 ml of N-methylimidazole is added, followed by the addition of 1.6 ml of ethyldichlorothiophosphate (Aldrich Chem. Co.). After 1 hour the subunit product is purified by silica gel chromatography developed with 20% acetone/80% chloroform. This activated subunit (Structure 2, where X is --O--CH.sub.2 -CH.sub.3 and Z is sulfur) can be coupled to the morpholino nitrogen of a second subunit as described above.

Alternatively, a small amount of water can be added to the reaction solution instead of washing with NaH.sub.2 PO.sub.4. The resulting solution is stirred for 10 minutes to effect conversion of the activated subunit (Structure 2) to the phosphonate salt (Structure 5, where X is --O--CH.sub.2 CH.sub.3 and the counter ion is a protonated N-methylimidazole). The product is purified by silica gel chromatography developed with a methanol/chloroform mixture 1% in triethylamine. The purified product is shaken with chloroform/water containing 2% pyridinium hydrochloride to change the counter ion to pyridinium. After drying the product is suitable for coupling to a 5'-protected subunit having a free morpholino nitrogen (Structure 6) using DCC (dicyclohexylcarbodiimide) in dichloromethane. The carbodiimide coupling reaction yields the dimer shown as Structure 7 (where X is --O--CH.sub.2 CH.sub.3).

Example 9C describes the coupling of a 5'hydroxyl subunit, prepared as in Example 2 or 3, to a second subunit having a free morpholino ring nitrogen to give a fluorophosphoramidate intersubunit linkage. The example is described with reference to structures in FIG. 10 where X is a fluorine.

The starting material is one mmole of 5'hydroxyl subunit, base-protected with groups removable by a beta elimination mechanism and tritylated on the morpholino nitrogen (Structure 4 of FIG. 5). The subunit is dissolved in 20 ml of dichloromethane to which is added 6 mmole of N-methylimidazole, followed by the addition of 2.5 mmole of fluorophosphoric acid. 5 mmole of DCC is added and the solution stirred three hours. The reaction solution is washed with aqueous NaH.sub.2 PO.sub.4 and the organic phase dried under reduced pressure to give the fluorophosphate salt (Structure 5 where X is F and the counter ion is N-methylimidazole). The product is purified by silica gel chromatography developed with a methanol/chloroform mixture 1% in pyridine to give the pyridinium salt. After drying the purified product is suitable for coupling to a 5'-protected subunit, having a free morpholino nitrogen (Structure 6), using DCC in dichloromethane. This carbodiimide coupling reaction yields the dimer shown as Structure 7 (where X is F).

Polymers containing the fluorophosphoramidate intersubunit linkage should not be exposed to strong nucleophiles, such as ammonia. Consequently, bases of the subunits used for assembling such polymers should be protected with groups which can be cleaved without the use of strong nucleophiles. Protective groups cleavable via a beta elimination mechanism, as described in Example 1, are suitable for this purpose.

Example 9D describes the coupling of a 5'hydroxyl subunit, prepared as in Example 2 or 3, to a second subunit having a free morpholino ring nitrogen to give a phosphordiamidate intersubunit linkage. The example is described with reference to structures in FIG. 10, where X is a disubstituted nitrogen.

One mmole of 5'hydroxyl subunit, base-protected and tritylated on the morpholino nitrogen (Structure 4 of FIG. 5) is dissolved in 5 ml of dichloromethane. Six mmole of N-ethylmorpholine and 2 mmole of dimethylaminodichlorophosphate (OP(Cl).sub.2 N(CH.sub.3).sub.2) for Z.dbd.O (or the thiophosphate analog for Z.dbd.S) is added to the solution, followed by the addition of 0.5 mmole of N-methylimidazole. After the reaction is complete (assessed by thin layer chromatography) the reaction solution is washed with aqueous NaH.sub.2 PO.sub.4. The activated subunit is isolated by chromatography on silica gel developed with acetone/chloroform (Structure 2 in FIG. 10, where X is N(CH.sub.3).sub.2). The activated subunit is directly linked to the morpholino nitrogen of a subunit (Structure 3) in DMF containing triehtylamine sufficient to neutralize the HCl produced in the reaction, to give the dimer shown as Structure 4.

The dimethylaminodichlorophosphate (X is --N(CH.sub.3).sub.2 and Z is oxygen) used in the above procedure was prepared as follows: a suspension containing 0.1 mole of dimethylamine hydrochloride in 0.2 mole of phosphorous oxychloride was refluxed for 12 hours and then distilled (boiling point is 36.degree. C. at 0.5 mm Hg). The dimethylaminodichlorothiophosphate (X is --N(CH.sub.3).sub.2 and Z is sulfur) used above was prepared as follows: a suspension containing 0.1 mole of dimethylamine hydrochloride in 0.2 mole of thiophosphoryl chloride was refluxed for 18 hours and then distilled (boiling point 85.degree. C. at 15 mm Hg).

This example describes the coupling of a methylphosphonate subunit, prepared as in Example 5, with a second subunit having a free 5'hydroxyl. The example is described with reference to structures in FIG. 11.

The starting material is the base-protected, morpholino nitrogen-protected methyl phosphonate subunit prepared as in Example 5. A triethylamine salt of this subunit is suspended in chloroform and shaken with water containing 2% pyridinium hydrochloride, resulting in the pyridinium salt (Structure 1) of the subunit in the chloroform phase. The organic phase is separated, washed with water, and dried. One portion of the resulting solid is mixed with an excess of 3-hydroxypropionitrile; and 2 equivalents of DCC is then added to the mixture. The product of the resulting reaction is detritylated with 1% dichloroacetic acid in dichloromethane to give Structure 2. Structures 1 and 2 are then mixed in the presence of DCC resulting in the coupled dimer shown as Structure 3. This dimer may be selectively deprotected on either end for assembly into longer oligomer blocks or polymers; the cyanoethyl group may be removed with DBU in a non-protic solvent (e.g., DMF), or the dimethoxytrityl moiety may be cleaved as above.

This example describes the oxidation of a ribonucleoside containing a protected amine linked through the 5' methylene, such as prepared in Example 7, and coupling to the unprotected amine of another subunit to simultaneously form a morpholino ring structure and join the subunits. The example is described with reference to the structures in FIG. 12.

Ten mmole of ribonucleoside containing a 1.degree. amine linked through the 5' methylene (Structure 1) is reacted with 11 mmole of trityl chloride to protect the amine (Structure 2).

The tritylated subunit (Structure 2), in methanol, is reacted with 11 mmole of NaIO.sub.4 to give the dialdehyde (Structure 3).

If the coupling solution is too acidic the reaction is very slow and if the solution is too basic, epimerization of the Structure 3 component appears to occur. A weak acid is used to neutralize the amine component (Structure 1) and buffer the reaction in this coupling step. Weak acids which have been found suitable for this purpose are: carbonic, ortho and para nitrophenol, and benzotriazole. Accordingly, the dialdehyde (Structure 3) is combined with a suitable salt of Structure 1 in a water/methanol mixture to give the coupled product (Structure 4).

Either during or after the morpholino ring closure step sodium cyanoborohydride is added to reduce the dihydroxymorpholino ring (Structure 4) to the desired morpholino product (Structure 5).

This example describes the assembly of a short oligomer containing a phosphordiamidate-linked backbone (Structure B--B of FIG. 4, where X is N(CH.sub.3).sub.2, Y is oxygen and Z is oxygen) coupled as in Example 9. This solution assembly method is particularly useful for large-scale synthesis of short oligomers suitable for subsequent assembly into longer oligomers using the solid-phase method (Example 13).

5'OH morpholino subunits of C, U, and G tritylated on the morpholino ring nitrogen are prepared as in Example 2 or 3. The U subunit is then activated by conversion to the monochlorophosphoramidate form as in Example 9D. The C subunit and the G subunit are deprotected with trifluoroethanol (TFE) containing 2% acetic acid; the TFE is then removed under reduced pressure. The residue is washed with ether to remove any residual acetic acid. The deprotected C component (1.1 m mole) is dissolved in 5 ml DMF and 0.3 ml TEA, followed by addition of 1.0 m mole of the activated U component. Likewise, the deprotected G component is reacted with the activated U component.

After one hour each of these preparations is added to 100 ml of rapidly stirred brine and the solid collected and washed with water. The GU dimer is dried thoroughly under high vacuum and then activated as in Example 9D. The best tetramer coupling results are obtained when purification of the dimer, via silica gel chromatography, is carried out after, rather than before, this activation step.

The CU dimer is deprotected as above. Better yields of tetramer are obtained when the dimer, after the initial ether precipitation, is thoroughly resuspended in about 2 ml of trifluoroethanol, reprecipitated with 30 ml of ether, and then resuspended in DMF and TEA for subsequent coupling.

Coupling to form the desired tetramer entails simply adding 1 m mole of activated GU dimer to the DMF/TEA solution containing 1.1 mmole of deprotected CU dimer.

Workup of the tetramer entails adding the reaction mixture to brine, washing the solid with water, and drying under vacuum to give the desired tetramer: 5'-CUGU having a hydroxyl at the 5' end and a trityl on the morpholino nitrogen of the U subunit. The structure of this tetramer is most easily confirmed by negative ion fast atom bombardment mass spectroscopy. As a rule the dominant specie in such spectra is the molecular ion.

This example describes the use of tetramer blocks, prepared as in Example 12, for solid-phase assembly of a morpholino polymer containing phosphordiamidate intersubunit linkages. Solid-phase assembly provides a rapid method for assembly of longer binding polymers. The use of short oligomer blocks instead of monomers greatly simplifies separation of the final product from failure sequences.

The following tetramers are synthesized in solution: 5'-CUGU (Example 12); 5'-UCGG; 5'-GCGC; 5'-CACU. These tetramers are converted to their activated monochloro form by the general method described in Example 9D.

Morpholino C subunit containing a trityl moiety on the morpholino ring nitrogen and having a methylamine on the 5' methylene, prepared as in Example 6, is reacted with a 3-fold molar excess of Bis[2-(succinimidooxycarbonyloxy)ethyl]sulfone from Pierce of Rockford, Ill., USA. This product is purified by silica gel chromatography and then added to a suitable solid support containing primary amine functions (e.g., Long Chain Alkyl Amine Controlled Pore Glass, from Pierce of Rockford, Ill.). This procedure links the first tritylated subunit to the synthesis support via a linker which is stable to the acidic conditions used for detritylations, but which can be readily cleaved via a beta elimination mechanism using a strong non-nucleophilic base, such as a 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU).

The coupling cycle for addition of each subunit or oligomer block generally includes deprotection of the terminal backbone moiety, a thorough wash, addition of the next activated subunit or oligomer block, and a thorough wash. The coupling efficiency for each addition can be determined by collecting each detritylation solution and subsequent wash and quantitating the trityl therein.

Detritylation in the present polymer is achieved by slowly passing through the column a solution of 2% acetic acid in trifluoroethanol until the eluant no longer tests positive for trityl (readily determined by adding a drop of eluant to 100 .mu.l methanesulfonic acid and inspecting for the visible yellow color characteristic of the trityl carbonium ion). Thereafter the support is thoroughly washed to remove excess acid and then washed with DMF containing 1% by volume of N-ethylmorpholine (NEM). Coupling of the next subunit or oligomer block in the desired polymer sequence entails addition of a concentrated DMF solution containing the activated monomer or oligomer and a molar equivalent of NEM. Since the rate of coupling is a function of concentration it is desirable to add a substantial molar excess of monomer or oligomer relative to the concentration of support-bound growing chains. A 5-fold molar excess of activated monomer or oligomer over that of the growing chains generally gives acceptable coupling efficiencies. Required coupling times can be determined by removing at specified time intervals small defined quantities of the support material, thoroughly washing, treating the support with methanesulfonic acid, and then spectrophotometrically quantitating the released trityl carbonium ion (molar absorbance at 409 nm is 45,000 in methanesulfonic acid). After coupling is complete the unreacted subunit or oligomer is washed from the support with DMF. The unreacted subunit is generally recovered, purified by chromatography, and reused for later synthesis. The support is thoroughly washed with the solvent trifluoroethanol, without added acid. Washing is complete when addition of a drop of the wash eluant to 100 .mu.l methanesulfonic acid shows no yellow color.

The above coupling cycle is used to add, in order, the four activated tetramers 5'-CUGU, 5'-UCGG, 5'-GCGC, and 5'-CACU. This results in the following polymer: support-linker-CCUGUUCGGGCGCCACU-trityl.

The synthesis support is treated with 20% DBU in DMF for two hours at room temprature in the presence of 2% diethylmalonate, to tie up the vinylsulfone generated during cleavage of the linker. The released morpholino polymer is washed from the support with DMF and precipitated by adding ethylacetate. The precipitate contains full-length polymer having a 5' methylamine, the bases still protected and a trityl moiety on the terminal morpholino nitrogen. In addition, the precipitate contains small amounts of failure sequences. At this stage the polymer size can be confirmed by positive ion fast atom mass spectrometry.

If it is desired to add two solubilizing groups to the morpholino polymer this can be done conveniently by detritylting the N-terminal morpholino nitrogen using 2% acetic acid in trifluoroethanol. Alternatively, if only one solubilizing moiety is to be added, then the 5'methylamine is acetylated with acetic anhydride before the detritylation step.

Polyethylene glycol 1000 (from Polysciences Inc., Warrington, Pa., USA) is thoroughly dried by dissolving in dry DMF and then evaporating the solvent under vacuum. The solid is resuspended in a minimal volume of pure dry DMF and 0.5 mole equivalent (relative to PEG 1000) of bis(p-nitrophenyl)carbonate and 1 mole equivalent of TEA is added and the preparation sealed and incubated overnight at 30.degree. C. to give p-nitrophenyl carbonate-activated PEG 1000.

The full-length morpholino polymer which has been detritylated is added to a substantial molar excess (generally 10- to 20-fold) of activated PEG 1000 and incubated two hours at room temperature. Unreacted PEG 1000 is removed by precipitation of the tailed polymer with ether. The tailed polymer is then air-dried.

The dried polymer is suspended in DMSO, the DMSO solution chilled, and an equal volume of concentrated NH.sub.4 OH is carefully layered on top of the chilled DMSO, and the container tightly capped. The preparation is incubated at 30.degree. C. for eighteen hours. Thereafter, the solution is briefly exposed to aspirator vacuum to remove ammonia.

Purification at pH 2.5 is generally used for binding polymers where about half or more of the base-pairing moieties are of types 1, 2, 3, and 7 of FIG. 2.

Water to be used for chromatography is degassed under aspirator vacuum and phosphoric acid added to give pH 2.5 (solvent A). A corresponding pH 2.5 solution is made 2N in KCl (solvent B). Solvent A is mixed 1:1 by volume with chromatographic-grade acetonitrile to give solvent C.

Load up to about 10 mg of the polymer in 10 ml of solvent A on a chromatography column 1 cm in diameter and 10 to 20 cm in length which is packed with the cation-exchange support S-Sepharose Fast Flow (Pharmacia). Proportionately larger quantities can be loaded on larger columns of the same length, e.g., up to 60 mg can be loaded on a 2.5 cm diameter column and 250 mg on a 5 cm diameter column. After washing the column thoroughly with solvent A elute with a linear gradient ranging from 100% solvent A to 100% solvent B and monitor the eluant to 254 nm. The desired binding polymer is generally the last and the largest peak to elute from the column. When the polymer is prepared by block assembly, base-line separations are often achieved. When peak shapes are unsymmetrical the problem generally has been due to insolubility of the binding polymer rather than a lack of capacity of the chromatographic packing. Such a problem, which is most common when the binding polymers do not contain a PEG solubilizing tail, can often be solved by reducing the quantity of binding polymer loaded in a given run. When peaks are symmetrical but base-line separation is not achieved, substantial improvements are usually attained simply by eluting with a shallower gradient.

The eluant containing the polymer is desalted by loading on an equivalent-sized column packed with 35 micron chromatographic polypropylene (cat. no. 4342 from Polysciences, Inc.) and washing thoroughly with solvent A. If baseline separation was achieved in the foregoing cation exchange chromatography, then pure product is obtained simply by eluting with solvent C; otherwise, the product is eluted with a linear gradient ranging from 100% solvent A to 100% solvent C. When the product is somewhat acid sensitive the eluant is neutralized with dilute NaOH before drying under reduced pressure.

Purification at pH 11 is generally used for binding polymers where about half or more of the base-pairing moieties are of types 4, 5, 6 and 9 of FIG. 2.

N,N-diethylethanolamine (Aldrich) is added to degassed water to adjust the pH to 11.0 (solvent D). A corresponding pH 11 solution 2N in KCl (solvent E) is prepared. A third pH 11 solution is prepared by mixing Solvent D 1:1 by volume with chromatographic grade acetonitrile (solvent F).

The fully-protected binding polymer, prepared as above, is suspended in solvent D at a concentration of about 100 .mu.g/ml. The pH is adjusted, if necessary, to pH 11 with N,N-diethylethanolamine. About 10 ml of this polymer solution is placed on a chromatography column 1 cm in diameter and 10 to 20 cm in length which is packed with anion-exchange support Q-Sepharose Fast Flow (Pharmacia). After washing the column thoroughly with solvent D, the column is eluted with a linear gradient ranging from 100% solvent D to 100% solvent E and the eluant is monitored at 254 nm.

The eluant containing the polymer is desalted by loading on an equivalent-sized column of polypropylene and washing thoroughly with solvent D. If baseline separation is achieved in the foregoing anion exchange chromatography then pure product is obtained simply by eluting with solvent F; otherwise, the product is eluted with a linear gradient ranging from 100% solvent D to 100% solvent F. Fractions containing the product are dried under reduced pressure.

While mass spectral analysis of the full-length polymer in the fully-protected state, as described earlier, does serve to confirm both the polymer length and the base composition, it does not provide information on the subunit sequence. Significant sequence information can be obtained from fragmentation patterns of deoxyribonucleic acids and carbamate-linked deoyribonucleoside-derived polymers (Griffin et al., (1987), Biomed. & Environ. Mass Spectrometry 17 105), however, many of the morpholino polymers of the instant invention are quite resistant to fragmentation and give predominantly the molecular ion with only minimal fragments.

One method for confirming the sequence of the polymer is to take a small portion of the growing polymer after coupling each oligomer block and use mass spectral analysis to follow the elongation of the polymer. This method is applicable except for those rare cases where two blocks used in the synthesis happen to have exactly the same mass.

An indirect method to help verify the correctness of the polymer subunit sequence is to pair the morpholino polymer with its complementary DNA (whose sequence can be confirmed by established methods) and with DNA sequences which might have resulted if the blocks were assembled in the wrong order. Pairing between the polymer and DNA can be evaluated by looking to a hypochromic shift in the 240 to 290 nm wavelength region. Such a shift occurs only between the polymer and its complementary sequence. The polymer/DNA duplex can also be distinguished from any partially-mismatched duplex by slowly raising the temperature while monitoring the absorbance in the 240 nm to 290 nm region. The perfect duplex will have a melting temperature (corresponding to a 50% reduction in the hypochromicity) generally 10 degrees or more above that of any mismatched duplex.

This example describes the preparation, structural confirmation, and assessment of target binding affinity of a simple phosphordiamidate-linked morpholino polymer.

A morpholino hexamer where all Pi moieties are cytosines is assembled from dimer prepared as in Example 9D. One third of that dimer preparation is detritylated (as in Example 12) and the remaining two thirds is activated (as in Example 9D). Half of the activated dimer is reacted with the detritylated dimer to give tetramer, which is purified by silica gel chromatography developed with 6% methanol/94% chloroform. The tetramer is detritylated and reacted with the remaining activated dimer to give hexamer, which is purified by silica gel chromatography developed with 10% methanol/90% chloroform.

This phosphordiamidate-linked 5'OH, base-protected hexamer having a trityl moiety on the morpholino nitrogen is designated pd(mC).sub.6 -trityl. The negative ion Fast Atom Bombardment mass spectrum (3-nitrobenzyl alcohol matrix) shows: M-1=2667.9 (100).

This pd(mC).sub.6 -trityl polymer is next detritylated as in Example 12, and then a polyethylene glycol 1000 tail is added followed by base deprotection, as in Example 13. Purification is by cation exchange chromatography followed by desalting on a column of polypropylene, as described in Example 13.

This purified tailed hexamer, pd(mC).sub.6 -PEG1000, shows an absorption maximum at 267.1 nm in neutral aqueous solution, with a calculated molar absorbance of 42,800. In aqueous solution at pH 1, the same material shows an absorption maximum at 275.7 nm, with a calculated molar absorbance of 77,100.

To assess target-binding affinity of the pd(mC).sub.6 PEG-1000 polymer, 1 mg of polyG RNA (purchased from Sigma Chem Co.) is dissolved in deionized water, and 4 volumes of DMSO (spectrophotometric grade from Aldrich Chem. Co.) is added (stock solution A). Tailed morpholino hexamer, pd(mC).sub.6 -PEG1000, is dissolved in spectrophotometric grade DMSO (stock solution B). Phosphate buffer is prepared by adjusting the pH of 0.05N NaOH to 7.4 using phosphoric acid (Buffer C).

Stock solutions A and B are assayed for the actual concentration of polymer by UV; the absorbance of stock solution A is measured in 0.1N NaOH and stock solution B is measured in 0.1N HCl. Measurements at these pH extremes minimize base stacking and other polymer interactions which can give absorbance values not proportional to the component monomers. Stock solutions A and B are diluted with Buffer C to give solutions of a final concentration of 10 micromolar in polymer. The required dilutions are calculated using molar absorbancies of 65,000 for solution A, poly(G), and 77,100 for solution B, pd(mC).sub.6 -PEG1000.

Assessment of target binding affinity is carried out in a double-beam scanning spectrophotometer having a temperature-controlled cell housing which will accommodate two cells in the reference beam and two in the sample beam.

Using four matched quartz cuvettes, one is filled with 0.5 ml of poly(G), 60 micromolar with respect to G monomer, and 0.5 ml of Buffer C and a second is filled with 0.5 ml of 10 micromolar pd(mC).sub.6 -PEG1000 and 0.5 ml of Buffer C. These two cuvettes are placed in the reference beam of the temperature-controlled cell housing. Next, a third cuvette is filled with 1 ml of Buffer C and a fourth is filled with 0.5 ml of poly(G) 60 micromolar with respect to G monomer and 0.5 ml of 10 micromolar pd(mC).sub.6 -PEG1000. These two cuvettes are placed in the sample beam of the cell housing. The cell housing is then heated to 50.degree. C. and allowed to cool slowly to 14.degree. C. to assure complete pairing between the polymer and its DNA target in the fourth cuvette. A scan is then taken from 320 nm to 240 nm--which shows a substantial absorbance difference due to a hypochromic shift in polymer-target mixture, centered around 273 nm. The temperature of the cell holder is then raised in 2-degree increments to 72.degree. C., with scans taken after each 2-degree rise.

A plot of the absorbance difference as a function of temperature for the morpholino polymer/RNA complex is shown in FIG. 13. The melting temperature, Tm, where the complex is half melted, is seen to be 51.5.degree. C. for this morpholino polymer/RNA.

While specific embodiments, methods, and uses of the invention have been described, it will be appreciated that various changes and modifications of the invention may be made without departing from the invention. In particular, although preferred polymer backbone structures have been described and illustrated, it will be appreciated that other morpholino-based polymers may be constructed according to the backbone constraints and requirements discussed above.