DNA Conformations: A, B, and Z Forms

A DNA helix is built from a small number of local geometric choices, repeated at every nucleotide. Each deoxyribose ring can pucker into roughly two families of conformation: C2′-endo (the ring carbon 2′ is displaced out of the sugar-base plane toward the same side as C5′) or C3′-endo (displacement in the other direction). Each glycosidic bond (connecting the base to C1′ of the sugar) can rotate into an anti conformation, where the base points away from the sugar ring and over the backbone, or a syn conformation, where the base swings back over the sugar. Finally, adjacent base pairs are related by a helical twist angle and rise that need not be identical from step to step. B-, A-, and Z-DNA are simply the three conformations in which these local parameters lock into stable, periodic, energetically favorable repeating patterns.

B-DNA is the Watson–Crick helix: right-handed, C2′-endo sugar pucker, all bases anti, roughly 10.5 base pairs per turn, a helical rise of 3.4 Å per base pair, and a helix diameter near 20 Å. Its defining geometric signature is the pronounced asymmetry between a wide, deep major groove and a narrow, deep minor groove — the feature exploited by sequence-reading proteins. A-DNA is also right-handed and all-anti, but the sugar pucker shifts to C3′-endo. This single ring-pucker change tilts the base pairs roughly 20° relative to the helix axis, shortens the rise to about 2.6 Å per base pair, increases the twist to about 11 bp per turn, and — critically — collapses the major groove into a narrow, deep slot while opening the minor groove into a broad, shallow one. Z-DNA is the outlier: left-handed, with a repeating dinucleotide unit in which purines (commonly guanine) sit in the syn conformation with C3′-endo-like (C2′-endo for pyrimidines) sugar puckers, giving the backbone a zigzag path (hence "Z") rather than a smooth spiral, about 12 bp per turn, and a single deep, narrow groove instead of distinct major/minor grooves. The essential conceptual point is that conformation is not imposed by chemistry alone — it is imposed by context. The same phosphodiester backbone and the same four bases can realize any of these three helices depending on water activity, ionic environment, chemical modification (e.g., cytosine methylation), and torsional stress transmitted along the helix. This sequence- and environment-dependence is what makes DNA conformation a biologically meaningful variable rather than a structural curiosity.

Step 1 — the sugar pucker sets the register. The furanose ring of deoxyribose is not planar; it puckers to relieve eclipsing strain between substituents. Two puckers dominate: C2′-endo, which is compatible with a relatively flat, extended backbone geometry, and C3′-endo, which pulls the phosphate backbone into a more contracted, steeper helical path. B-DNA uses C2′-endo throughout; A-DNA uses C3′-endo throughout. This is the single most consequential switch: it changes the P–P (phosphate-to-phosphate) distance along the backbone from about 7.0 Å (B-form) to about 5.9 Å (A-form), which is why A-DNA is a shorter, fatter helix for the same number of base pairs.

Step 2 — hydration selects the pucker. The C2′-endo (B-form) backbone requires a well-organized spine of hydration in the minor groove — a network of ordered water molecules bridging adjacent phosphate groups and base-pair edges. At high relative humidity (roughly 92% and above) and with monovalent counterions like Na⁺, this hydration spine is stable and B-DNA predominates. As water activity drops (fiber humidity below about 75%, or in the presence of dehydrating cosolvents like ethanol, or with certain divalent/trivalent cations that compete for the hydration shell), the minor-groove spine collapses, the sugar pucker shifts to C3′-endo, and the fiber converts to A-form. This is a genuinely reversible, humidity-driven phase transition — the same DNA fiber cycles between A and B diffraction patterns as it is dried and rehydrated.

Step 3 — RNA is structurally forced into A-form. Ribose carries a 2′-OH group that deoxyribose lacks. In the C2′-endo (B-form) pucker, this hydroxyl would clash sterically with the adjacent phosphate backbone; RNA and RNA:DNA hybrid duplexes therefore cannot adopt B-form geometry and default to A-form (or a very close variant). This is why every double-stranded RNA (rRNA stems, siRNA duplexes, retroviral RNA genomes) and every RNA:DNA hybrid (the RNA primer during replication, the R-loop behind RNA polymerase, the reverse-transcription intermediate in retroviruses) is A-form, with a deep major groove that is largely inaccessible to protein side chains and a shallow minor groove — a structural fact with direct consequences for how RNA-binding proteins and reverse transcriptase recognize their substrates.

Step 4 — sequence selects for Z-DNA via the syn/anti alternation. Z-DNA's defining feature is that purines (most readily guanine) flip into the syn glycosidic conformation while the paired pyrimidines (cytosine) remain anti. Because a syn purine and an anti pyrimidine pack together in a geometrically compatible dinucleotide repeat, sequences that alternate purine-pyrimidine along a strand — d(GC)ₙ, d(CA)ₙ, or alternating methylated-CG tracts — can zip into the Z conformation, whereas non-alternating sequences cannot easily accommodate the syn purine without steric clash. High salt (which screens the electrostatic repulsion between the closely juxtaposed phosphates in the zigzag backbone) and negative supercoiling (which directly favors the underwound, higher-twist Z conformation) both push the equilibrium toward Z-DNA. Cytosine methylation at C5, a common epigenetic mark, further stabilizes Z-DNA by increasing the hydrophobic bulk that favors the syn conformation, linking Z-DNA formation to chromatin state.

Step 5 — torsional coupling links transcription to Z-DNA formation in vivo. As RNA polymerase translocates and unwinds the duplex ahead of it, the enzyme cannot itself rotate freely around the DNA axis (a topological constraint reinforced by chromatin attachments), so unwinding ahead of the polymerase is compensated by overwinding ahead of it and underwinding (negative supercoiling) behind it — the twin-supercoiled-domain model. Negative supercoiling behind a moving polymerase transiently favors the Z conformation at susceptible alternating-purine-pyrimidine sequences located just upstream of active promoters. This transient, transcription-coupled Z-DNA is recognized by dedicated Z-DNA-binding domains (Zα domains), found in the RNA-editing enzyme ADAR1 and the innate-immune sensor ZBP1/DAI, connecting a purely structural phenomenon to gene regulation and immune signaling.

Franklin's fiber diffraction distinguished A- from B-DNA before any atomic model existed. In 1952–1953, Rosalind Franklin, working with Raymond Gosling at King's College London, showed that a single DNA fiber gives two distinct X-ray diffraction patterns depending on relative humidity. At high humidity (about 92%), the fiber gave the famous crisp, symmetric X-shaped pattern (Photograph 51) diagnostic of a helical structure with roughly 10-fold screw symmetry and a clear meridional reflection at 3.4 Å corresponding to the base-stacking repeat — this was the B-form. At lower humidity, the same fiber gave a denser, more complex pattern with different layer-line spacing (Franklin's "Structure A"), which she correctly interpreted as a more compact, tilted-base helix. Crucially, Franklin measured that the A-form unit cell was consistent with a helix of shorter pitch and different water content, and she explicitly warned that A- and B-forms were different conformations of the same molecule, interconvertible by hydration — precisely the mechanism later confirmed by structural biology. Watson and Crick's March 1953 model was built using measurements from the B-pattern (helix diameter ~20 Å, rise 3.4 Å, 10 residues per turn) taken from Franklin's data (shown to Watson via Maurice Wilkins) together with Franklin and Gosling's unpublished MRC report data on unit-cell dimensions — a pattern of using her crystallographic measurements to constrain a model built independently, published together in the same 1953 Nature issue as Franklin and Gosling's own paper on the A-form.

Crystallography later confirmed both forms atomically. Because fibers give only relatively low-resolution average structures, unambiguous atomic-resolution proof required single-crystal X-ray diffraction of short, defined oligonucleotides — technically difficult because short DNA duplexes did not crystallize well until synthetic chemistry advanced in the late 1970s. Richard Dickerson and colleagues solved the structure of the self-complementary dodecamer d(CGCGAATTCGCG) in 1980, providing the first atomic-resolution view of B-DNA and confirming, base pair by base pair, the C2′-endo sugar pucker, the anti glycosidic angles, and the characteristic groove-width variation Franklin's diffraction had implied. A-form oligonucleotide crystal structures (e.g., by Olga Kennard's and later others' groups) similarly confirmed the C3′-endo pucker and base-pair tilt predicted from fiber data.

Alexander Rich's 1979 crystal structure proved left-handed Z-DNA exists. Before 1979, a left-handed DNA helix was a theoretical curiosity, not an established structure. Rich, together with Andrew Wang and postdoctoral colleagues at MIT, crystallized the synthetic hexanucleotide d(CGCGCG) — a strictly alternating purine-pyrimidine sequence — under high-salt conditions and solved its structure by X-ray crystallography at 0.9 nm resolution. The result was unambiguous and startling: the phosphate backbone did not spiral smoothly but zigzagged, the helix was left-handed, guanines were in the syn conformation while cytosines remained anti, and there was a single deep groove rather than distinct major and minor grooves. This was the first atomic structure of left-handed DNA and gave Z-DNA its name (from the zigzag backbone path). Subsequent work showed that antibodies raised against Z-DNA could recognize supercoiled plasmid DNA at physiological salt when negatively supercoiled — direct biochemical evidence (from labs including Rich's and David Lilley's) that Z-DNA is not merely a crystallization artifact but a physiologically accessible conformation, later reinforced by chromatin immunoprecipitation-style mapping showing Z-DNA-forming sequences enriched near active transcription start sites and by structures of the Zα domain (from ADAR1) bound to Z-DNA, proving a dedicated cellular reader exists for this conformation.

Example 1 — converting helical parameters into contour length and comparing A- vs. B-form compaction. Suppose a double-stranded DNA molecule is 1,000 base pairs long. Calculate its contour length in B-form versus A-form, and determine how many helical turns each represents. B-form: rise = 3.4 Å/bp. Contour length = 1,000 bp × 3.4 Å/bp = 3,400 Å = 340 nm. Turns = 1,000 bp ÷ 10.5 bp/turn ≈ 95.2 turns. A-form: rise = 2.6 Å/bp. Contour length = 1,000 bp × 2.6 Å/bp = 2,600 Å = 260 nm. Turns = 1,000 bp ÷ 11 bp/turn ≈ 90.9 turns. The same 1,000 bp duplex is 340 nm long in B-form but only 260 nm long in A-form — roughly 24% shorter — while completing almost the same number of helical turns (95 vs. 91). This quantitative contrast explains a subtle but important point: an RNA:DNA hybrid or an RNA duplex of a given base-pair number occupies substantially less contour length than the equivalent B-form DNA, a fact relevant to packaging double-stranded RNA genomes (e.g., in reoviruses) and to estimating footprint sizes of RNA:DNA hybrid R-loops on a gene map.

Example 2 — reasoning through a Z-DNA hotspot at a transcribed promoter. A researcher finds a 24-bp stretch immediately upstream of a heavily transcribed gene's promoter with the sequence (one strand): 5′-CGCGCGCGCGCACACACACACG-3′. Predict whether this region is likely to transition into Z-DNA during active transcription, and explain the expected spatial relationship to the transcription start site. Step through the criteria: 1. Sequence composition. The stretch is dominated by alternating purine-pyrimidine dinucleotide repeats: (CG)ₙ and (CA)ₙ tracts. Both are classic Z-DNA-forming motifs, because C (pyrimidine, anti) can pair stably with G or A (purine, syn) in the required zigzag arrangement. A random or purine-tract sequence (e.g., all-A or all-G stretches) would not satisfy this criterion. 2. Torsional environment. During active transcription, the twin-supercoiled-domain mechanism generates positive supercoiling ahead of RNA polymerase and negative supercoiling behind it. A sequence positioned upstream of (behind, relative to polymerase movement into the gene) an actively transcribed promoter is exactly where negative supercoiling accumulates. 3. Salt/local environment. Physiological ionic strength (~150 mM monovalent salt equivalent, plus polyamines like spermidine in vivo) is generally sufficient to stabilize Z-DNA once negative supercoiling is present, without requiring the artificially high salt needed to form Z-DNA in relaxed, unstressed plasmids in vitro. 4. Epigenetic reinforcement. If the CG dinucleotides are cytosine-methylated (common at CpG-rich promoter-proximal regions), the methyl group further stabilizes the syn-guanine/anti-cytosine Z-conformation, lowering the negative supercoiling threshold needed for the transition. Conclusion: this sequence is a strong candidate for transient Z-DNA formation specifically during, and just after, rounds of active transcription, localized to the region upstream of the promoter where negative supercoiling transiently peaks, and it would be expected to relax back toward B-form once transcription pauses and topoisomerases dissipate the supercoiling. Such regions are, in real genomes, frequently found near actively transcribed promoters and are bound in cells by Zα-domain proteins, consistent with a proposed role in marking or modulating promoter accessibility.

Alexander Rich trained as a physician before turning to structural biology, working first with Linus Pauling at Caltech on the physical chemistry of biological macromolecules before joining MIT, where he built one of the most productive nucleic-acid crystallography programs of the twentieth century. Rich's laboratory was distinctive for pursuing "unusual" or non-canonical nucleic acid structures at a time when most structural biologists focused on refining canonical B-DNA — triple helices, RNA structures like the first crystallographic solution of transfer RNA (with Sung-Hou Kim, 1974), and ultimately left-handed DNA. The 1979 Z-DNA structure emerged from Rich's long-standing interest in why some synthetic polynucleotides gave anomalous circular dichroism spectra, an observation that had puzzled researchers for years before crystallography revealed the left-handed explanation. Rich remained active into the 2000s, extending the Z-DNA story into biology by helping characterize the Zα domain of ADAR1 and its role in recognizing Z-DNA/Z-RNA in cells, connecting a structural discovery to gene regulation and innate immunity.

Conformational polymorphism of DNA and RNA is now a working tool and a diagnostic signal across biomedicine and biotechnology. Because RNA and RNA:DNA hybrids are locked into A-form geometry, antiviral and antisense drug design must account for A-form major-groove inaccessibility: therapeutics targeting double-stranded RNA (e.g., in RNA interference-based drugs like patisiran) or RNA:DNA duplexes (antisense oligonucleotides, and the RNA:DNA hybrid intermediate targeted by RNase H-dependent gapmer drugs) are engineered around A-form, not B-form, groove dimensions. Reverse transcriptase inhibitors used in HIV therapy must dock against the A-form RNA:DNA hybrid produced during reverse transcription, not the B-form product duplex, a structural distinction that matters for structure-based drug design. Z-DNA-forming sequences are used as markers of transcriptional activity and torsional stress in genomic studies: because negative supercoiling behind moving RNA polymerase transiently favors Z-DNA at alternating purine-pyrimidine tracts, mapping Z-DNA-prone sequences (and Zα-domain occupancy) helps identify actively transcribed, torsionally stressed chromatin regions, complementing chromatin accessibility assays. The Zα-domain protein ADAR1 — an RNA-editing enzyme central to distinguishing self from viral nucleic acid — uses Z-DNA/Z-RNA recognition as part of innate immune regulation; mutations disrupting this recognition are implicated in the autoinflammatory disorder Aicardi–Goutières syndrome, linking a purely structural phenomenon directly to human disease. In structural vaccinology and RNA therapeutics, understanding A-form geometry of double-stranded RNA regions in mRNA vaccines and their secondary structures informs formulation and stability engineering. Finally, forensic and diagnostic applications exploit conformation-specific antibodies and conformation-sensitive dyes to distinguish B-form genomic DNA from A-form or Z-form regions in research settings, and structural bioinformatics pipelines predict A/B/Z propensity from sequence to flag regulatory regions genome-wide — a direct descendant of Franklin's and Rich's original diffraction measurements, now applied computationally at genome scale.

Ask the tutor: How does the twin-supercoiled-domain model quantitatively predict where along a gene Z-DNA is most likely to form?