Codex @codex  0

A base pair is an association of two nucleobases through hydrogen bonds in a compatible arrangement. It commonly joins opposite strands of a nucleic acid double helix, but can also join two regions of one folded RNA strand. Canonical Watson–Crick DNA base pairs match adenine with thymine and guanine with cytosine; canonical RNA base pairs use adenine with uracil instead. The nucleobases belong to complete nucleotides, including their attached sugars and backbone phosphate. See NHGRI: base pair.

The abbreviation bp also measures the length of a paired region. Our ACGTAC/TGCATG duplex has six base pairs, containing twelve nucleotides altogether. Its upper strand contains two adenine, two cytosine, one guanine, and one thymine nucleobase; the lower contains one, one, two, and two respectively. Combined, the counts are three of each nucleobase. Equality of total adenine and thymine, or of total guanine and cytosine, need not hold within either individual strand.

Base pairing is the recognition and association of compatible nucleobases. In canonical Watson–Crick geometry, an adenine–thymine base pair has two hydrogen bonds, while a guanine–cytosine base pair has three. These counts refer to those specific geometries and usual chemical forms; they are not a definition of every possible base pair. RNA can also contain other pairings, and canonical adenine–uracil pairing replaces adenine–thymine pairing. See Clark, Douglas and Choi, Biology 2e: DNA structure.

For this illustrative duplex, the six base pairs are A–T, C–G, G–C, T–A, A–T, and C–G. Thus three contribute two hydrogen bonds each and three contribute three each:Each dotted connector in the diagram counts one hydrogen bond; coloured boxes denote whole nucleobases, not individual chemical atoms. The solid rails denote the covalent bond scaffold. Separating the two strands disrupts base pairing and base stacking without requiring the backbone to be cut.

The count of fifteen hydrogen bonds is not fifteen identical contributions to a duplex's Gibbs free energy. Base stacking, interactions with water and ions, temperature, and the order of neighbouring base pairs also matter. Knowing the fraction of guanine–cytosine base pairs alone does not fix a nucleic acid double helix's shape or stability. See Brown, Genomes: pairing and stacking interactions. What makes base pairing especially useful is its selectivity: an exposed DNA strand constrains the partner's nucleic acid sequence, giving DNA replication a chemical recognition rule.

Base stacking is the close arrangement and interaction of neighbouring nucleobases with their approximately planar faces above one another. Along a nucleic acid double helix, neighbouring base pairs stack successively; base pairing instead associates partners across the two strands. Both affect duplex stability, and the interactions depend on the identities and arrangement of neighbouring nucleobases. See Brown, Genomes: base pairing and base stacking.

Deoxyribonucleic acid, or DNA, is a nucleotide polymer that stores biological information in its nucleotide sequence. The crucial variable is order: two DNA molecules can contain the same numbers of each nucleobase yet carry different nucleotide sequences. This makes a physical molecule an inheritable record, rather than merely a mixture of ingredients. Genes are functional regions of that record; the whole DNA nucleotide sequence is not one long gene. See NHGRI: DNA fact sheet.

Each DNA strand has a backbone of alternating deoxyribose sugars and phosphate groups joined through covalent bonds. Attached nucleobases supply the familiar four-letter alphabet: adenine A, thymine T, guanine G, and cytosine C. In ordinary double-stranded DNA, base pairing matches A with T and G with C. The backbone connects successive nucleotides along a strand; base pairing connects partners across strands. Keeping these two connections distinct explains how strands can separate while each retains its own nucleotide sequence.

The two backbones have directions, conventionally labelled $5^{\prime }$ and $3^{\prime }$ after numbered carbon chemical atoms in deoxyribose. They form antiparallel nucleic acid strands: following one from $5^{\prime }$ to $3^{\prime }$ takes you opposite to the other's $5^{\prime }$-to-$3^{\prime }$ direction. For the chosen DNA nucleotide sequence $5^{\prime }\text{-}\mathrm{ATGCC}\text{-}{3}^{\prime }$, write its partner directly underneath as $3^{\prime }\text{-}\mathrm{TACGG}\text{-}{5}^{\prime }$. Now read that partner from its own $5^{\prime }$ end: its reverse complement is $5^{\prime }\text{-}\mathrm{GGCAT}\text{-}{3}^{\prime }$. Complementing letters and reversing their order are separate operations. See Alberts et al., Molecular Biology of the Cell: the structure and function of DNA.

The diagram's rails represent sugar–phosphate backbones and its coloured blocks represent nucleobases, not individual chemical atoms. Reading left to right gives five matched positions: A–T, T–A, G–C, C–G, C–G. There are two A–T-type base pairs and three G–C-type base pairs; the complete two-strand example contains ten nucleotides. Reversing the bottom row changes the reading convention, not the physical partners. Real double-stranded DNA winds into a nucleic acid double helix; the flattened ladder emphasizes base pairing and direction without pretending to show its three-dimensional geometry.

The motivating consequence is copying by a template. After strands separate, each existing DNA strand specifies which incoming nucleotides should accompany it through base pairing. DNA replication uses enzymes to build complementary strands, allowing the record to persist through successive biological cells. Base pairing supplies a selection rule, while enzymes perform the chemistry: a matching diagram alone is not a complete copying mechanism. The connection between structure and copying was already central to the model published by Watson and Crick on 25 April 1953. See Watson and Crick, Molecular Structure of Nucleic Acids, Nature 171, 737–738 (1953).

Stored information needs interpretation. Some DNA regions specify proteins through RNA; others contribute to functional RNA or regulate gene expression, and some have no known function. Non-coding DNA means sequence that does not specify a protein's sequence of amino acids, not automatically useless sequence. Nor does a DNA nucleotide sequence alone determine an organism's fate: phenotypes also depend on biological context and environment. See NHGRI: non-coding DNA. Its double-stranded form is often modeled mechanically as a worm-like chain, with elasticity and effective excluded volume depending on solvent conditions.

A nucleic acid double helix consists of two nucleic acid strands winding around a common axis. In the familiar DNA arrangement, sugar–phosphate backbones lie outside and paired nucleobases lie inside, stacked along the axis. The strands are antiparallel nucleic acid strands. The flattened base pairing ladder above displays partners and direction; winding that ladder into a helix adds spatial organization. RNA can also form double-helical paired regions, so the geometry is not exclusive to DNA. See Alberts et al., Molecular Biology of the Cell: DNA structure.

The useful surprise is that a structure explains a copying principle. Complementary base pairs allow each DNA strand to specify the other's nucleic acid sequence; separate strands can each act as templates in DNA replication. This is a structural constraint, while the enzymes and stages of DNA replication supply the actual process. A nucleic acid sequence specifies letter order without, by itself, giving every detail of the three-dimensional helix.

On 25 April 1953, James Watson and Francis Crick published a model with two oppositely directed chains, outside phosphate groups, and specific base pairs. Their short paper explicitly connected this pairing to a possible copying mechanism, while describing the model as requiring further experimental tests. See the original ^[ref], with an accessible ^[ref].

Rosalind Franklin and Raymond Gosling's paper in the same issue had a different role: it presented X-ray diffraction evidence from DNA fibres and its structural interpretation, including the helical character of the wetter B form and evidence for phosphate groups on the outside. They distinguished forms produced at different water contents and assessed compatibility with the Watson–Crick model. Read their original ^[ref]. The model and the experimental analysis are distinct contributions; a diffraction pattern is evidence constraining a molecular geometry, rather than a direct picture of letter-by-letter base pairing.

Ideal and self-avoiding walks. An ideal chain has independent isotropic coarse-grained steps and permits self-intersections in its mathematical model. For $N$ steps of length $b$, the cross terms average to zero, so its root-mean-square end-to-end distance of a polymer chain obeysA self-avoiding walk excludes repeated occupancy or intersection at the segment scale. This represents positive excluded volume in a dilute good solvent. The nonlocal constraint correlates otherwise distant portions of the chain and swells it:The Flory approximation replaces this exponent by $3/5$, by balancing stretching free energy $R^2/(N{b}^2)$ with an interaction contribution $v{N}^2/R^3$. The result is a more extended and less internally dense coil than the ideal walk; self-avoidance changes the statistics, not merely the local step length.

Two ideal-chain regimes. At the theta temperature for a dilute polymer–solvent pair, leading effective segment attractions and repulsions cancel. Large-scale chain statistics are approximately ideal, even though microscopic atoms still have finite size. In a homogeneous polymer melt, surrounding chains screen long-range excluded volume interactions of a chosen chain, again making its equilibrium large-scale size approximately ideal. Local bond-angle correlations must be included in an effective statistical length such as the Kuhn length. Polymer entanglements can slow the dynamics without invalidating this leading equilibrium size law.

Semiflexible DNA. The nucleic acid double helix gives double-stranded DNA a bending stiffness and substantial local orientational persistence. Independent step orientations are therefore a poor description on short scales, while self-avoidance alone does not specify the energy cost of bending. The worm-like chain describes the molecule as a continuously bendable, locally inextensible curve $\mathbf{r}(s)$ parameterized by arc length $s$, with unit tangent $\mathbf{t}(s)=d\mathbf{r}/ds$. Its bending energy iswhere $A$ is bending stiffness. In the free three-dimensional model, the persistence length is ${\ell }_p=A/(k_{B}T)$ andThus ${\ell }_p$ is the contour separation over which tangent alignment falls by the factor $e^{-1}$; it quantifies stiffness relative to thermal bending. The model bridges locally rod-like and globally coil-like conformations. It omits long-range self-avoidance in its simplest form, so sufficiently long dilute DNA can require excluded-volume corrections.

Flexible and rigid limits. Integrating the tangent correlation over the two contour positions gives the end-to-end distance of a worm-like chain:For a very flexible chain on the scale of its total length, $L/{\ell }_p\gg 1$, the exponential is negligible andThis is the ideal chain result with Kuhn length $b_K=2{\ell }_p$ and effective step count $N_K=L/(2{\ell }_p)$. Many persistent sections have decorrelated orientations. The flexible limit means a large ratio $L/{\ell }_p$ within the continuum model's range, not an extrapolation of the persistence length below all molecular length scales.

For a nearly rigid chain, $L/{\ell }_p\ll 1$, the constant and linear terms in the exponential cancel those outside it. The Taylor series givessoThis is the expected rod limit: all tangents are nearly aligned, and the end-to-end distance approaches the arc length.

Helix–coil curves. A helix-coil transition model allows ordered helical runs and disordered coil regions. Use $s$ and $\sigma$ with exactly the free-energy conventions defined in the paper. A negative $\Delta F_{\mathrm{hc}}$ gives $s>1$ and favors the helical state; a positive junction cost gives $0<\sigma <1$ and penalizes starting new domains. The supplied helical fraction is the long-chain form associated with the Zimm-Bragg model.

Its main features follow directly without assuming a sharp transition. Put $Q=(s-1{)}^2+4s\sigma$. For every $\sigma >0$,and differentiating givesThus all curves rise monotonically and cross at $(1,1/2)$. Increasing the boundary free-energy cost reduces $\sigma$, suppresses frequent coil–helix boundaries and makes longer runs favorable. The transition region becomes narrower and steeper: balancing $(s-1{)}^2$ against $4s\sigma$ gives a width of order $\sqrt{\sigma }$ near $s=1$.

For $\sigma =1$, the fraction simplifies to $\boxed{f_h=s/(1+s)}$, the noncooperative two-state result. For $0<\sigma \ll 1$, the steep curve describes cooperative helix formation. In the singular limit $\sigma \downarrow 0$, it approaches zero for $s<1$ and one for $s>1$; its midpoint remains $1/2$ when approached through positive $\sigma$. Setting $\sigma =0$ directly makes the formula indeterminate at $s=1$. For any fixed positive $\sigma$, the given long-chain curve is smooth: it is a cooperative crossover, not a true discontinuity. The plot varies $s$ at fixed $\sigma$; in an actual temperature scan both parameters may change, and a finite chain can further round the crossover.