9.2.2.3 Reactions at Pyrrole Nitrogens and Exocyclic Amino Groups

When the types of reaction sites in the heterocyclic bases of nucleic acids were discussed, it was pointed out that pyrrole and amine nitrogens must be attacked by reagents in a direction normal to the plane of these bases. Consequently, such nitrogens may be reactive only when their base occupies single-stranded segments of the nucleic acid.

Addition at Active Double Bonds. The most frequently used compounds of this type are water-soluble acrylonitrile and carbodiimide salts. It has already been mentioned that the reaction involves bases containing a -CO-NH-group in the heterocyclic ring, capable of being ionized at alkaline pH values. Thus, the reaction proceeds as nucleophilic addition of the anion emerging from the base to an active multiple bond. In this case, the base reacts as an ambidentate ion at a nitrogen marked by maximum nucleophilic activity, which is N3 in uracils and thymines and N1 in guanines.

The rate of the reaction is maximum at a pH value of about 9. As the medium becomes more basic (at pH > 10), the resulting adduct undergoes almost complete degradation. This allows the investigator to come back to the initial structure, after having studied the modified nucleic acid, and see how well it has been retained.

Guanines react with water-soluble carbodiimides at a much slower rate; some minor bases in tRNAs are modified as easily as uracils and thymines.

Pseudouridylic and inosinic acids provide good examples:

Significantly, internucleotide linkages formed by modified uracils, are not broken by pancreatic RNase.

Acrylonitrile reacts with bases, just as carbodiimides. The reaction proceeds at the highest rate at pH 8-9, the most easily modifiable minor components of tRNA being pseudouridines, 4-thiouridines and inosines. Pseudouridine, for example, reacts at a rate 20 times faster than uridine; similarly to the reaction with carbodiimides, pseudouridine may have two molecules of the reagent added thereto. The following adducts are formed:

Just as expected, the 4-thiouracil ionized in a weakly alkaline medium reacts at a site corresponding to maximum nucleophilic activity - that is, at the sulphur atom:

2-Thiouridines do not react with acrylonitrile under the same conditions.

As regards uracils, they are modified very slowly in a weakly basic medium, whereas cytosines, adenines and guanines are not modified at all.

Reactions with Aldehydes and Ketones. The interaction of nucleotides and nucleic acids with the simplest carbonyl compound, formaldehyde, has been studied in depth. It has been established that uridines, thymidines, pseudouridines and inosines in oligonucleotides and nucleic acids yield the corresponding methylol derivatives as a result of substitution at the imine nitrogen adjacent to one or two carbonyl groups. The reaction proceeds at a rather fast rate in a moderately alkaline medium, slows down in a moderately acidic medium, and is completely inhibited when the acidity of the medium is high. This suggests that in this case, just as in a reaction with compounds having active C=C and C=N bonds, the base reacts as an anion with addition at the C=O bond, for example:

In the case of pseudouridines, mono- and dimethylol derivatives are formed:

For methylol groups to be eliminated from the above heterocyclic bases, it is quite sufficient to gradually remove formaldehyde (e.g., by dialysis) from the reaction mixture. This is another proof of reversibility of the reaction under consideration.

While interacting with the formaldehyde of heterocyclic bases containing exocyclic amino groups, the methylol group becomes incorporated into the latter:

The most thoroughly studied compounds include adenine derivatives. It is the above example that served as demonstration of cross-linking of polynucleotide chains via purine bases in an alkaline medium under the effect of formaldehyde. The methylol derivative formed during the first step of the reaction may have a proton detached, in the presence of an alkali, from the oxygen of the methylol group (in which case it breaks down into the starting substances) and also, to a lesser extent, from the nitrogen with which the group is linked. In the latter case, a rather active Schiff's base is formed to react immediately at the exocyclic amino group with the second purine base:

It has been found that the cross-linking may involve not only two adenines but also an adenine and a guanine:

Such cross-links are quite stable in alkaline media, which has been demonstrated in experiments with certain RNAs. The rate of the reaction between formaldehyde and polynucleotides is strongly dependent on the structure of the latter. Double-stranded nucleic acids react with formaldehyde at a very slow rate (the reaction seems to proceed at fluctuation points). Therefore, formaldehyde treatment is widely used in studies of the secondary structure of nucleic acids, the size and stability of double-stranded segments in RNAs, as well as the structure of denatured and native DNAs (to determine the number of defects in the helical structure). Modification of polynucleotides with formaldehyde has also been used to elucidate the functioning of tRNAs and to determine the template activity of polyadenylic acid.

Fairly recently an interesting observation was made regarding modification of nucleic acids with formaldehyde. It was found that during treatment of denatured nucleic acids with a mixture of formaldehyde and glycine adenines are eliminated from the polynucleotide chain already at neutral pH values, with subsequent cleavage of the chain at depurination sites. This reaction may be instrumental, for example, in controlled cleavage of polynucleotide chains at adenines during investigations of the primary structure.

The fact that glyoxal reacts with 2-N-methylguanine but not with 2-N, N-dimethylguanine and N1-methylguanine suggests that the reaction involves two nitrogen atoms: the heterocyclic N1 atom and amino nitrogen - that is, the reaction calls for a -NH-C(NHR)=N- group.

The structure of products of glyoxal, pyrotartaraldehyde and ketoxal addition to guanines has been established by oxidation with periodate, followed by reduction with lithium aluminium hydride to 2-N-alkylguanine derivatives.

These transformations showing the position of the alkyl radical in the reaction product suggest that the modification begins with reaction at the most active aldehyde group [at R' = CH3 and CH(CH₃)(OC₂H₅)] which is added at the heterocyclic nitrogen, and then the unstable alkylol derivative is converted into a stable cyclic diol. For instance:

Hence, the initial attack of glyoxal, too, proceeds at N1 (pyrrole nitrogen) and, consequently, perpendicularly to the heterocyclic base plane. PMR spectra are indicative of transposition of hydroxyl groups in the cyclic diol which is the glyoxal addition product.

Guanines modified with glyoxal or ketoxal are stable compounds, do not degrade during dialysis, but when the solution gets alkaline they are quickly regenerated, which is used to eliminate the modification agent (the half-life at pH 10 and 20⁰ C is 24 h) if needed. This property of guanines modified with glyoxal and ketoxal makes modification of this kind rather convenient for certain experiments. Since, as has already been mentioned, this modification is carried out as an attack perpendicular to the heterocyclic base plane, glyoxal and ketoxal will modify only those guanines that occupy single-stranded segments in tRNAs, for example. If a tRNA is modified with glyoxal or ketoxal at 70⁰ C (under conditions upsetting the secondary structure), all guanines undergo modification. This property of glyoxal and ketoxal - that is, inability to react with guanines involved in complementary interactions, is put to practical use in studies into the three-dimensional structure of nucleic acids. The modification with glyoxal (ketoxal) is instrumental in studying the primary structure of nucleic acids because guanyl RNase (RNase T₁) does not cleave RNAs at modified units and, therefore, the RNA can be divided into larger fragments (the enzyme exhibits greater specificity). All this renders modification of nucleic acids with glyoxal a rather useful tool for studying the structure and functions of nucleic acids.

This group of modification agents also includes chloroacetaldehyde and similar compounds. It has been found that chloroacetaldehyde with its two reactive groups reacts selectively with adenines and cytosines at pH 4-5, also yielding polycyclic derivatives:

In single-stranded polynucleotides (e.g., polyadenylic and polyeytidylic acids), the bases are modified quantitatively under rather mild conditions (the optimal pH of the reaction medium being 3.5 for modifying cytosines and 4.5 for adenines).

Appropriate conditions have been selected for complete deamination of DNAs, but there are still some side processes taking place, namely, cleavage of N-glycosidic bonds and formation of covalent cross-links between the strands. Most likely, a reaction occurs similar to formation of cyclic nucleosides with the diazonium salt acting as an alkylating agent.

Complete deamination of RNAs can be achieved by using sodium nitrite in glacial acetic acid, however, the polynucleotide chain undergoes marked degradation. Under milder conditions, with side processes with reduced to a minimum, the deamination of both DNAs and RNAs is only partial.

Because of a number of steric reasons (the amino group must be attacked by nitrous acid only at a right angle to the heterocycle plane), the secondary structure (stacking and involvement in hydrogen bonding) must inhibit the reaction with nitrous acid. Yet the reaction does take place in the case of double-stranded nucleic acids. For instance, native calf thymus DNA is deaminated exactly as a denatured one but at different relative rates for guanines, adenines and cytosines. By virtue of the small size of the nitrosonium cation the reaction is possible in double-stranded complexes as well.

The possibility of partial denaturation of double-stranded segments of nucleic acids cannot be ruled out altogether in view of the fact that the reaction proceeds at acid pH values. Table 9-2 lists some reagents commonly used for modifying bases in nucleic acids.

Table 9-2. Some Reagents Used for Modification of Heterocyclic Bases in Nucleic Acids.

Table 9-2. (Continued).