Base Pairing in the Purine and Pyrimidine Crystal Structures

Base Pairing in the Purine and Pyrimidine Crystal Structures 

As discussed in Chapter 15, the arrangement of the functional groups on adjacent ring positions in the purine and pyrimidine bases is such as to increase the strength of hydrogen bonding due to 7i-cooperativity. This favors the formation of base-paired dimers and in some cases of trimers and tetramers. These can occur between the same bases (homo- or self-association) and between different bases (hetero-association). 

Both donor functions of amino groups are never involved in base-pair dimers. 

In no case has a base pair been observed, either in crystal structures of monomers or in helical complexes of the polynucleotides, where the amino group utilizes both its hydrogen donor functionalities to form two hydrogen bonds with one associat- 

---

This type of hydrogen bonds includes the N-H 0=C interactions which are the most predominant hydrogen bonds in fibrous and globular proteins. Because they are responsible for the formation of the commonly occurring secondary structure elements a-helix, -pleated sheet and / -turn, a large body of much less accurate data is available from protein crystal structures which will be analyzed in Part III, Chap. 19. The N-H 0=C type hydrogen bond is also the most common in the purine and pyrimidine crystal structures (Thble 7.14), and is one of the two important bonds in the base pairing of the nucleic acids. 

The N-H- -N bonds constitute about a quarter of the hydrogen bonds in the purine and pyrimidine crystal structures (see Thble 7.14). The proportion is much smaller in the nucleosides and nucleotides Thble 7.12, where they compete with the stronger O-H- -O and N-H- -O interactions. In combination with N - H O=C, the N - H N bonds form the Watson-Crick and related base-pair configurations in purine and pyrimidine crystal structures, and in the oligonucleotides and nucleic acids. 

Some of the hydrogen-bonded dimer configurations described in the matrices are observed in the crystal structures of the purines and pyrimidines, but this is by no means the general rule. This may be because the number of crystal structures in which self-base pairing occurs is relatively small compared with the number of possibilities. Certain arrangements appear to be particularly favored for reasons... 

Overlap Geometry A schematic representation of the proposed overlap geometry for proflavine intercalated into a deoxy pyrimidine(3 -5 )purine site is presented below with the (o) symbols representing the location of the phenanthridine ring protons. The mutual overlap of the two base pairs at the intercalation site involves features observed in the crystal structures of a platinum metallointercalator miniature dC-dG duplex complex (55) and the more recent proflavine miniature dC-dG duplex complex (48), as well as features derived in a linked-atom conformational calculation of the intercalation site in the proflavine DNA complex (51). [4]... 

The discovery of these base pairs, which imply the double helical structure of DNA, stimulated a series of crystal structural studies not only of complexes of purines and pyrimidines, but of other complexes involving related molecules and their derivatives. Although we can formulate a large number of possible heterocombinations in matrix form, as shown below these complexes are reluctant to crystallize even when there is spectroscopic evidence of hetero-complex formation in solution. This is presumably because self-(homo)-association is energetically more favorable and only in rare cases were crystals of hetero complexes actually formed. Because of their three hydrogen bonds, G-C complexes form and crystallize more readily. There have been many attempts to crystallize the Watson-Crick A-U base pair, but none was successful and it only formed when the dinucleoside phosphate adenylyl-3,5,-uridine (ApU [536]) or higher oligomers were crystallized (see Part III, Chapter 20). 

Perhaps the prototypical examples of organic heterodimeric co-crystals involve the hydrogen-bond pairing of purine and pyrimidine base pairs, postulated in 1953 by Watson and Crick for adenine thymine (uracil) 60 and guanine cytosine 61 to unravel the structure of DNA [204], and since confirmed by numerous crystal structure analyses, as recently reviewed by Jeffrey and Saenger [10]. 

In the crystal structures, homo base pairs with centrosymmetrical configuration are favored. This general statement is true for all the four different bases of pyrimidine and purine type. The individual bases exhibit considerable dipole moments, as shown in Thble 16.1, and we assume that the preference for base pairs with a center of symmetry is due to the antiparallel orientation of the dipole moments in this particular arrangement. This results in a favorable cancellation of the total electric field over the crystal volume. 

---