Why Pentose‐ and Not Hexose‐Nucleid Acids? Part IV. ‘Homo‐DNA’: 1H‐, 13C‐, 31P‐, and 15N‐NMR‐Spectroscopic Investigation of ddGlc(A‐A‐A‐A‐A‐T‐T‐T‐T‐T) in Aqueous Solution From a comprehensive NMR structure analysis, it is concluded that the ‘homo‐DNA’ oligonucleotide ddGlc(A‐A‐A‐A‐A‐T‐T‐T‐T‐T) in 3 mM D2O solution (100 mM NaCl, 50 mM phosphate buffer, pH 7.0, T = 50°) forms a duplex of C2‐symmetry, with its self‐complementary oligonucleotide strands in antiparallel orientation. The 2′,3′‐dideoxy‐β‐D‐glucopyranosyl rings are in their most stable chair conformation, with all three substituents equatorial and with the adenine as well as the thymine bases in the anti‐conformation. The base pairing is of the Watson‐Crick type; this pairing mode (as opposed to the reverse‐Hoogsteen mode) was deduced from the observation of inter strand NOEs between the adenine protons HC(2) and the pyranose protons Hα–C(2′) of the sequentially succeeding thymidine nucleotides of the opposite strand, a correlation which discriminates between the Watson‐Crick and the reverse‐Hoogsteen pairing mode. The NOEs of the NH protons with either the adenine protons HC(2) or HC(8), that are normally used to identify the pairing mode in DNA duplexes, cannot be observed here, because the NH signals are very broad. This line broadening is primarily due to the fact that the exchange of the imino protons with the solvent is faster than for corresponding DNA duplexes.Computer‐assisted modeling of the [ddGlc(A5‐T5)]2 duplex with the program CONFOR [23], using the linear (idealized) homo‐DNA single‐strand conformation (α = −60°, β = 180°, γ = 60°, δ = 60°, ϵ = 180°, ζ = −60°, see [1] [3]) as the starting structure, resulted in two duplex models A and B (see Figs. 27–32, Scheme 9, and Table 4) which both contain quasi‐linear double strands with the base‐pairing axis inclined relative to the strand axes by ca. 60° and 45°, respectively, and with base‐pair stacking distances of ca. 4.5 Å. While neither of the two models, taken separately, can satisfy all of the NMR constraints, the NMR data can be rationalized by the assumption that the observed duplex structure represents a dynamic equilibrium among conformers which relate to models A and B as their limiting structure. The required rapid equilibrium appears feasible, since the models A and B are interconvertible by two complementary 120° counter rotations around the α‐axis and the γ‐axis, respectively, of the phosphodiester backbone. The models A and B correspond to the two types of linear (idealized) single‐strand backbone conformation derived previously by qualitative conformational analysis without and with allowance for gauche‐trans‐phosphodiester conformations, respectively [1] [3]. Refinement of the models A and B with the use of the program AMBER [27] by energy minimization in a water bath and molecular‐dynamics simulations (2 ps, 300° K) resulted in two dynamic structures (Figs. 33 and 34, Table 4). These have roughly the same energy, closely resemble the starting structures A and B, and satisfy ‐ as an ensemble ‐ all of the NMR constraints without violating any van der Waals distances by more than 0.2 Å. Extensive fluctuations in base‐pair distance and deviations from base‐pair coplanarity, as well as the presence of water molecules in the cavities between some of the base pairs, were observed in both dynamic structures A and B, which, on the other hand, did not mutually interconvert within the short simulation time period used. These model properties, together with the conjectured equilibrium between the two structure types A and B, lead to the hypothesis of a homo‐DNA duplex containing a ‘partially molten’ pairing core. This proposal could qualitatively account for a high rate of the NH exchange, as well as for part of the previously established [3] deficits in both enthalpic stabilization and entropic destabilization of homo‐DNA duplexes relative to corresponding DNA duplexes. The phenomenon of the higher overall stability of homo‐DNA duplexes vs. DNA duplexes (e.g, [ddGlc(A5‐T5)]2, Tm = 59° vs. [d(A5‐T5)]2, Tm = 33°, both at c ≈ 50 μM [3]) can then be seen as the result not only of a higher degree of conformational preorganization of the homo‐DNA single strand toward the conformation of the duplex backbone [1] [3], but also of the entropic benefit of greater disorder in the central pairing zone of the homo‐DNA duplex. This view of the structure of a homo‐DNA duplex relates its characteristic properties to a central structural feature: the average base‐pair distance in the models of homo‐DNA is too large for regular base stacking (ca. 4.5 Å vs. ca. 3.5 Å in DNA). This difference in the distances between adjacent base pairs is a direct consequence of the quasi‐linearity of the homo‐DNA double strand as opposed to the right‐handed twist of the helical DNA duplexes [1] [3], which is directly related to the specific conformational properties of pyranose rings as opposed to furanose rings [1]. Thus, the structural hypothesis derived from the NMR analysis of [ddGlc(A5‐T5)]2 relates the conformational differences between homo‐DNA and DNA directly to the sugar ring size, which is the essential constitutional difference between the two types of structure.The English footnotes to Figs. 1–34, Schemes 1–9, and Tables 1–4 provide an extension of this summary.