9.2 Reactions of Heterocyclic Bases of Polynucleotides

Nature of Reactive Sites. As can be inferred from previous chapters, the properties of the pyrimidine and purine bases of monomer components are first of all determined by the electronic structure of the atoms and groups of atoms constituting the heterocycle. Before discussing their chemical properties in the context of a polynucleotide (RNA, DNA, or a synthetic polymer), in addition to the concepts outlined in Chapter 2 one must also consider the stereochemistry of the reactions involving the respective bases and, in this connection, the spatial distribution of electron density at each reactive atom of the base. Since the tautomeric equilibrium in heterocyclic bases is normally shifted toward one of the forms, which seems to be responsible for the reactivity of a given base under particular conditions, the electron density distribution will be discussed only in the context of stable forms. Naturally, when we are speaking about the reactivity of a given heterocyclic base at pH values markedly different from 7, it should be remembered that the solution contains other forms that may react in an entirely different fashion.

The following scheme is based on the results of studies into the chemical properties of pyrimidine and purine bases of nucleic acids in aqueous media where the native structure of the latter usually remains intact. In all instances, the reactive sites of similar nature are marked by the same symbol (square, circle, triangle, diamond).

In the past, the influence of steric factors on the reactivity of pyrimidine or purine bases was virtually ignored because in monomer units this influence is minimal and, as a rule, does not determine the course of the reaction. If a base is not hindered sterically, it does not make any difference from which direction its atoms are attacked by the reagent - from the plane of the base or at a right angle thereto. The situation is entirely different when we are speaking about the reactivity of bases within the polymeric nucleic acid molecule. In this case, individual portions of the base may be substantially screened as a result of their involvement in hydrogen bonding or interplanar interactions. This is why the reagent cannot attack a particular atom of the heterocycle from any direction. In this connection we must figure out which direction of attack on each of the reactive sites marked on the generalized scheme will produce any results, and also which of the possible routes is open or blocked in the secondary and tertiary structures of the nucleic acid. Answers to these questions will tell us, firstly, which atoms of a particular pyrimidine or purine base may be modified irrespective of the three-dimensional structure of the corresponding nucleic acid fragment and, secondly, which of these atoms can react only as parts of single-stranded fragments of the polymer chain.

Let us now examine each of the reactive sites separately. The above scheme shows that they may coincide with nitrogen atoms (two species marked by square and triangle), the carbonyl oxygen (marked by square), and carbons (three species marked by circle, diamond, and also by circle and diamond at the same time).

The above-described nitrogen and oxygen atoms apparently may be subject predominantly to electrophilic attack which must be aimed at their lone-pair electrons. Since, as has already been pointed out, the orbital, or orbitals, accommodating these lone-pair electrons lies in the plane of the heterocyclic (or aromatic) nucleus of the base, the reagent's attack must be directed toward this plane to have any results.

A similar electronic structure is also observed in the nitrogen of the exocyclic amino group: the p orbital with a free pair of electrons, conjugated with the aromatic system, is perpendicular to the nuclear plane as well. Nitrogen atoms of this type must also be vulnerable to electrophilic attack, however, depending on the orientation of the orbital accommodating the lone-pair electrons of nitrogen, the attack will be productive only if the reagent approaches from the direction perpendicular to the plane of the heterocyclic base.

Thus, the reactive sites in triangles will be affected by electrophilic reagents when the attack is directed at a right angle to the plane of the heterocyclic base.

Insofar as the reactivity of carbons in pyrimidine and purine bases of nucleic acids is concerned, it should first of all be pointed out that they are in the sp2-hybrid state because each of them forms part of a >C=Y group (where Y = O, N and C). Accordingly, reagents will always attack the carbons (C2 , C4 , C5 , C6, C8) of these bases at a right angle to the plane occupied by the group and, consequently, to the plane of the corresponding base.

As is known, the reaction of electrophilic substitution occurs at C5 in pyrimidine nucleosides and at C8 in purine nucleosides. These bases react similarly within nucleic acids. The corresponding atoms in the generalized scheme are marked by diamonds and circles.

The response of carbons in the >C=Y group to different reagents depends both on the position of this group in the molecule of the base and on the species of atom Y.

In purine bases, for example, the carbons shared by the pyrimidine and imidazole rings are not reactive. In pyrimidine bases, the reactivity of the carbonyl C2 atoms bound to two nitrogens is low. These atoms are left unmarked in the scheme.

Just as expected. the species of atom Y markedly influences the reactivity of the associated carbon.

When Y is nitrogen, we have a >C=N- group in which the nitrogen exerts the same influence on the associated carbon as the carbonyl oxygen does, albeit not so markedly. Consequently, it may sometimes be expected that the corresponding carbons in pyrimidine and purine bases will be sensitive to nucleophilic attack. C4 is such a carbon in the case of cytosine; its reactivity seems to be due to more pronounced electronacceptor behavior of N3 under the influence of the neighboring carbonyl group. In the case of adenine and guanine, nucleophilic attack is aimed at C8.

Thus, only carbon atoms are prone to nucleophilic attack in the pyrimidine and purine bases of nucleic acids. They are shown encircled in the scheme.

Effect of Ionization on Reactivity. In the above discussion of reactivity of pyrimidine and purine bases in nucleic acids it was assumed that the bases are in a neutral form. Yet under the conditions prevailing in reactions of nucleic acids (aqueous solutions, varying pH values), each of the bases may accept or donate a proton, by virtue of its acid-base properties, and thus become charged. This, of course, must affect the course of reactions in which bases are involved to some or other extent.

As has already been mentioned, the proton is attached at the pyridine nitrogens or, in other words, the attachment proceeds as an electrophilic attack in the plane of the heterocyclic base. According to their ability to accept a proton, or be protonated, heterocyclic bases (for their pK_a values, see Table 4.1) can be arranged in the following order even in nucleic acids: cytosine > adenine > guanine. Only uracils, thymines and guanines can give up a proton (see Table 4-1).

A protonated base becomes more immune toward electrophilic agents and more vulnerable toward nucleophiles. When bases give up a proton, their reactivity varies inversely.

An example illustrating the effect of protonation on the chemical properties of a base is the inertness of the pyridine nitrogen N3 in a protonated form of cytosine with respect to electrophilic reagents and simultaneous mitigation of nucleophilic attack at carbon C4 next to the protonated nitrogen.

Thus, the acid-base properties of a heterocyclic base can be used to predict its reactivity, depending on the pH value of the solution in which the reaction is conducted. Since, as has already been mentioned, reactions with nucleic acids are usually carried out in aqueous solutions, the effect of pH on the reactivity of heterocyclic bases must always be taken into account.

Here are some examples of reactions involving constituent bases of nucleic acids in neutral and charged forms.

Example 1. Only the protonated form of the base enters into the reaction with a nucleophile - that is, the reaction rate is almost zero at a pH value when the base is in a free state and maximum at a pH value when practically all bases are in a protonated form. Take the simplest case where the nucleophile Nu is not capable of attaching a proton. Such a reaction may be written as

At pH equal to pK_a of the base, the reaction mixture will contain equal amounts of the protonated and unchanged forms. Apparently, at this point on the reaction rate (v) versus pH curve v will equal 0.5 of v_max.

In this case, the reaction rate changes sharply with varying pH in the region of pK_a of the base. Such reactions, for example, for guanine (pK_a - 3.0), must proceed at a satisfactory rate at pH < 3 and very slowly already at pH 4-5.

Example 2. The protonated form of the base reacts with the nucleophile, but in this case the nucleophile itself can attach a proton: pKa of Nu > pK_a of the base.

In a strongly acidic medium, the reagent and base are protonated, therefore the reaction rate is virtually null. The reaction proceeds at a maximum rate at a pH value when the reagent is already deprotonated, while the base still contains a proton, but there comes a moment at which the base and reagent lose their protons, and the reaction rate drops once more.

Evidently, the reaction rate will be maximum, v_max, at a pH value equal to half the sum of pK_a values of the reactants (pHa):

The relationship between the reaction rate and pH will be represented by a bell-shaped curve.

Example 3. The base reacts with an electrophile only in a deprotonated form. The reaction may take the following general form:

In this case, the reaction rate is zero in the pH range where the base is protonated. It reaches its peak at a pH value when the base loses its proton.

Here, too, the dependence of the reaction rate on pH of the reaction mixture is maximum when pH approaches pKa of the reacting base.

Many more examples could be given showing the relationship between the reactivity of nucleic acid bases and their form - protonated or deprotonated. The relationship between the reaction rate and pH is usually more complex than in the above examples. This depends on the reaction mechanism, ionic strength of the protonating agent. temperature, and many other factors. There is every reason to believe that the rate of reactions between heterocyclic bases of nucleic acids and various reagents is to some extent influenced by negatively charged phosphate groups of the carbohydrate-phosphate backbone.

Effect of the Macromolecular Structure of Nucleic Acids on Reactivity. Different aspects of the ability of heterocyclic bases in nucleic acids to undergo chemical modifications become important when they are regarded as keepers and carriers of genetic information. Since the coding capacity of nucleic acids is determined by the sequence of heterocyclic bases in the polynucleotide chain, the vulnerability of the bases to chemical attack is a measure of stability of the genetic material. Knowledge of the mechanisms underlying the action of mutagens (substances capable of altering the structure of heterocycles) on nucleic acids not only may give an insight into the alterations occurring when the genes become active in the cell, but also provide a tool for their manipulation in an intrusive and directed manner.

It is quite obvious that the reactivity of pyrimidine and purine bases in nucleic acids must be strongly influenced by the three-dimensional structure of the polymer, including its secondary, tertiary and quaternary components. Since the main features of this structure are already a matter of common knowledge for both DNA and RNA, one can predict the effect of conformation of various polynucleotide chain fragments on the reactivity of the heterocyclic bases present in these fragments of located in the neighborhood.

Let us now consider the three basic types of three-dimensional structure of nucleic acids: (a) ideal single-stranded polynucleotide without double-stranded fragments; (b) ideal double-stranded polynucleotide; and (c) polynucleotide with both single- and double-stranded fragments (RNA).

Single-Stranded Polynucleotide. It was already shown in Chapter 7 (Section 3) that the main type of intermolecular interaction responsible for the secondary structure of single-stranded polynucleotides is base stacking.

As can be inferred from the above-described behavior of reactive sites in heterocyclic bases, stacking will hinder a reagent's attack if it is directed at a right angle to the plane of the base. The hindrances will be especially pronounced in reactions where the most fully overlapping reactive sites must be involved. Those sites of bases which are practically not blocked as a result of stacking may react with the corresponding reagents, just as in momomer units but at a slower rate. The reagents attacking a base along the plane of the ring must interact with it at the same rate, irrespective of whether the base forms part of the single-stranded polynueleotide or a monomer unit - a particular nueleotide (with due account, of course, for the polyanionic nature of the polymer). For instance, the rate of reactions at the pyridine nitrogen or carbonyl oxygen, whose orbitals with free electron pairs lie in the plane of the ring, is not affected by stacking.

Ideal Double-Stranded Polynucleotide. Most natural DNAs are double-stranded. Only some viruses have single-stranded DNAs.

In this case, base pairs are stacked to form a helix. The stacking interactions in double-stranded nucleic acids are much stronger than in single-stranded ones. Yet the double-stranded helix is also a dynamic structure. At a point in time, the chains may separate (as a result of breaking hydrogen bonds), and the bases may be turned inside out. How reactive, then, are the bases in double-stranded nucleic acids? Obviously, the reactivity of stacked bases is minimal. Practically no reactions with nucleophilic and electrophilic agents attacking the heterocyclic nucleus at a right angle to the plane of the base are initiated. Only the pyridine nitrogen and carbonyl oxygen atoms are reactive, just as in single-stranded polynucleotides. This applies only to those atoms that are not involved in hydrogen bonding - that is, N7 in A and G as well as C2=0 in T Reactions on the side of the major groove proceed more easily, apparently because of the facilitated approach of the reagent.

However, in spite of their stable secondary structure, high-molecular weight DNAs are partially vulnerable also to reagents attacking bases from a plane perpendicular to that of the heterocyclic nucleus. The primary attack may be aimed at any base which has become vulnerable as a result of fluctuation. The structure of such a segment becomes loose (hydrogen bonding and interplanar interactions are hindered by a new substituent incorporated into the nucleus of the heterocycle). Naturally, looseness of structure renders the neighboring bases vulnerable to attack as well.

Ribonucleic Acids. As has already been mentioned in a previous chapter, the polynucleotide chain of a single-stranded RNA has alternating single- and double-stranded segments. A classical example is tRNA shaped as a cloverleaf folded into an L-shaped tertiary structure. The tRNA portions prone to reagent attack are not involved in its three-dimensional structure.

Consequently, knowing the reactivity of a polynucleotide, one can distinguish those segments of its polymer chain which take part in formation of the secondary or tertiary structures, as well as segments and individual nucleotides of the RNA molecule not involved in intramolecular interactions. This principle underlies studies into the three-dimensional structure of nucleic acids by the chemical modification method.