Alternative DNA Forms: A, B, and Z
A crystal of DNA fiber pulled from a humid chamber diffracts X-rays differently than one from a dehydrated chamber. The patterns are visibly different helical shapes. Same chemical sequence, different molecular machine. This is the first clue that the double helix is not a single rigid rod. It is a conformational chameleon instead. It can fold its sugar-phosphate backbone into at least three distinct helical families. The specific fold depends on water activity, counterion identity, and base sequence. The classic B-form dominates textbook cartoons. It is only the most common of these forms. It is stable under near-physiological hydration. A-DNA is a compact, dehydrated form. DNA also adopts A-DNA transiently when paired with RNA. Z-DNA is a left-handed zigzag oddity. It forms preferentially at alternating purine-pyrimidine sequences under high salt or negative supercoiling. Understanding why DNA changes shape matters because conformation is function. RNA polymerase transcribing a gene interacts with a helix. So does a transcription factor reading a promoter. So does a restriction enzyme cleaving a recognition site. In each case, the helix's groove width, base-pair tilt, and backbone torsion angles are locally tunable. The A-form is not a laboratory curiosity. Every DNA-RNA hybrid formed during transcription is forced into an A-like geometry. Every double-stranded RNA molecule is too. This happens because the 2′-OH of ribose sterically forbids the B-form's sugar pucker. Z-DNA transiently appears behind transcribing RNA polymerase. There, torsional strain from DNA Supercoiling and Topology accumulates. Specific Z-DNA-binding proteins recognize this form. These proteins are implicated in immune signaling and RNA editing. The big picture is that a double helix has a small number of internal degrees of freedom. These are the sugar pucker (C2′-endo vs C3′-endo), the glycosidic bond rotation (anti vs syn), and the backbone torsion angles. These couple to macroscopic helical parameters. Those parameters include rise per base pair, twist, base-pair inclination, and major/minor groove dimensions. Different combinations of local conformers are energetically preferred under different solvent conditions. Particular base sequences lock in or destabilize these combinations. This is a direct, mechanistic extension of the ideas developed for The DNA Double Helix. The same Watson-Crick hydrogen-bonding rules persist across A, B, and Z forms. So does antiparallel strand polarity. But the geometric packaging around that chemistry is a variable, not a constant. This article works through several key topics. These include the sugar-pucker pseudorotation cycle that sets backbone geometry. Also the syn/anti glycosidic equilibrium that flips in Z-DNA. Also the hydration and counterion arguments that explain the A/B transition. Also the alternating-sequence and methylation effects that favor Z-DNA. Numerical helical parameters (rise, twist, groove widths) are given for each form. This makes the geometric differences quantitatively, not just qualitatively, established.
- B-DNA (Watson-Crick, ~10.5 bp/turn, right-handed) is the standard form at physiological hydration and salt; deoxyriboses adopt C2′-endo pucker.
- A-DNA is right-handed, more compact (11 bp/turn, ~2.3 Å rise), with C3′-endo sugar pucker; it dominates at low water activity and in RNA/DNA-RNA hybrids where 2′-OH forbids C2′-endo.
- Z-DNA is left-handed, zigzag-backboned, 12 bp/turn, favored by alternating purine-pyrimidine sequences (e.g., (CG)ₙ) because purines flip to the syn glycosidic conformation.
- The interconversions are driven by competing free-energy terms: base-pair stacking, backbone electrostatics (counterion screening), and hydration of the phosphate and minor groove.
- Major and minor groove width/depth differ sharply among the three forms, directly controlling which face of the bases a protein can read without unwinding the helix.
- Negative supercoiling stores torsional free energy that is released by local B→Z conversion at susceptible sequences, linking helix conformation to genome topology.
- Key idea: DNA's double-helical backbone has enough internal conformational freedom (sugar pucker, glycosidic angle) that the same base-pairing chemistry can be packaged into geometrically distinct right- or left-handed helices, and solvent/sequence context selects among them.
Where this lives today
Z-DNA-binding domains (Zα domains) are found in the RNA-editing enzyme ADAR1 (adenosine deaminase acting on RNA), where Z-form recognition is now understood to modulate innate immune signaling — mutations disrupting the Zα domain are linked to Aicardi-Goutières syndrome, a type I interferonopathy, making Z-DNA/Z-RNA biology directly relevant to autoimmune and antiviral disease mechanisms. The related sensor ZBP1 (DAI) uses a Zα domain to detect Z-form nucleic acids generated during viral infection or cell stress, triggering programmed cell death pathways (necroptosis), an area of active pharmacological interest for both infectious disease and cancer immunotherapy. A-form geometry is central to RNA interference and CRISPR biology: guide RNA-target DNA heteroduplexes formed by Cas9 and related nucleases, and the siRNA-mRNA duplexes recognized by Argonaute in RNA interference, are A-form hybrids, and structure-based drug/oligonucleotide design (antisense oligonucleotides, siRNA therapeutics such as patisiran and inclisiran) must account for A-form duplex geometry when engineering chemically modified backbones (2′-O-methyl, locked nucleic acid) that lock the sugar pucker into C3′-endo to enhance hybrid stability and nuclease resistance. In genome biology, alternating CG- and GT-repeat tracts capable of adopting Z-DNA are enriched near transcription start sites and are proposed to act as local relief valves for negative supercoiling generated by RNA polymerase, feeding directly into the broader subject of DNA Supercoiling and Topology and gene-regulatory chromatin dynamics; Z-DNA-forming sequences have also been implicated in genomic instability and are being explored as biomarkers or targets in some cancers. Structural drug design targeting the minor groove (netropsin, distamycin, many DNA-binding chemotherapeutics) is explicitly B-form-selective, since the deep, narrow minor groove of B-DNA — not the shallow A-form or the very differently shaped Z-form groove — is the geometric feature these molecules are designed to fit.
Ask the tutor: How does the B→Z transition free energy change quantitatively when a CpG tract is embedded in a plasmid under a specified superhelical density, and how would you set up that supercoiling free-energy term explicitly?
The physical picture — thorough prose
A nucleotide's conformational freedom is concentrated in two "hinges": the deoxyribose ring pucker and the glycosidic bond torsion angle χ that connects the sugar's C1′ to the base's N9 (purines) or N1 (pyrimidines). The furanose ring is not planar; five ring atoms (C1′, C2′, C3′, C4′, O4′) pucker out of plane in a continuous, one-parameter family called the pseudorotation cycle, parameterized by a phase angle P from 0° to 360°. Two envelope/twist conformations dominate real DNA: C2′-endo (P ≈ 162°, the C2′ atom displaced toward the base side, roughly on the same side as C5′) and C3′-endo (P ≈ 18°, C3′ displaced similarly). These two puckers place the O3′–P–O5′–C5′ backbone torsions in very different relative geometries, which is why they propagate into different helical rise and twist. C2′-endo sugars pack the backbone into an elongated, shallow-helix conformation — this is B-DNA. C3′-endo sugars pull the backbone into a more contracted, steeply-inclined geometry — this is A-DNA (and is obligatory in RNA, whose 2′-OH group sterically clashes with the C2′-endo pucker via a 1,3-interaction with the phosphate, forcing C3′-endo universally). The base pairs in A-DNA are tilted and displaced from the helix axis into the major groove, opening a deep, narrow major groove and leaving a shallow, wide minor groove — almost the geometric inverse of B-DNA's minor/major groove profile. The glycosidic angle χ (defined as O4′–C1′–N9–C4 for purines, O4′–C1′–N1–C2 for pyrimidines) is normally anti in both A- and B-DNA, meaning the base points away from the sugar ring over the O4′ oxygen, with only limited steric contact between base and sugar. In Z-DNA, alternating purine-pyrimidine sequences allow purines (dG in the canonical (dC-dG)ₙ case) to rotate about χ into the syn conformation, in which the base swings back over the sugar ring; simultaneously the purine sugar adopts C3′-endo or C2′-exo pucker while the pyrimidine sugar stays C2′-endo/anti. This alternation of syn-purine/anti-pyrimidine every other residue is what produces the zigzag backbone from which Z-DNA gets its name: tracing the phosphate backbone, you see it kink left-right-left-right rather than sweep smoothly, because each repeat unit is a dinucleotide (not a mononucleotide) and the helix is left-handed.
The core ideas — governing relationships
The helical parameters that distinguish A-, B-, and Z-DNA are not independent; they are geometric consequences of the local backbone torsions and sugar pucker, propagated through the helix by the requirement of a repeating, regular structure. The essential quantitative relationship connects the number of base pairs per helical turn, , to the twist angle per step, :
where:
- is the mean helical twist per base-pair step (degrees),
- is the number of base pairs per complete helical turn (10.5 for B-DNA, 11 for A-DNA, and effectively 12 for the Z-DNA dinucleotide repeat, though Z-DNA twist is more naturally quoted per dinucleotide step).
The helix rise per turn, pitch , and rise per base pair are related by:
where:
- is the pitch (axial length of one complete turn, Å),
- is the rise per base pair along the helix axis (Å),
- is base pairs per turn as above.
For a right-handed helix, successive base pairs rotate through (by convention); for Z-DNA, the average handedness is left, so the net twist per residue is negative even though local dinucleotide twists can have mixed sign — a subtlety that follows directly from having two chemically distinct residues (syn-purine, anti-pyrimidine) alternating in the repeat. A second governing relationship is thermodynamic: which conformer (sugar pucker, glycosidic angle, and hence overall helix form) is populated is set by a Boltzmann-weighted competition among base-stacking energy, backbone conformational strain, and hydration/electrostatic free energy:
where:
- is the equilibrium ratio of two helical (or local conformer) populations,
- is the standard free-energy difference between the two forms (stacking + backbone + solvation terms), kJ mol⁻¹,
- is the gas constant (8.314 J mol⁻¹ K⁻¹), is absolute temperature (K).
Reading: the first two equations are purely geometric bookkeeping — they let you convert a measured X-ray diffraction repeat distance and helical symmetry into rise-per-residue and twist-per-residue, the numbers that experimentally distinguish A, B, and Z fibers. The third equation is the general free-energy framework: it says no helix form is absolutely "correct" — each is the thermodynamically dominant state under its own solvent/sequence/counterion conditions, and shifting those conditions shifts and hence the population ratio, exactly as observed experimentally when fiber humidity or salt is changed.
Goal. Show mechanistically why switching the deoxyribose pucker from C2′-endo (B-DNA) to C3′-endo (A-DNA) necessarily shortens the rise per base pair and increases base-pairs-per-turn, using the backbone torsion geometry rather than fitting parameters.
Step 1 — Identify the torsion chain. The sugar-phosphate backbone repeat unit contains six torsion angles per nucleotide: α (O3′–P–O5′–C5′), β (P–O5′–C5′–C4′), γ (O5′–C5′–C4′–C3′), δ (C5′–C4′–C3′–O3′), ε (C4′–C3′–O3′–P), ζ (C3′–O3′–P–O5′). The sugar pucker directly constrains δ, because δ is defined across the C4′–C3′ bond that is part of the furanose ring. C2′-endo pucker fixes δ near 120–140°; C3′-endo pucker fixes δ near 80–85°. This 40–55° change in a ring-internal torsion is transmitted rigidly to the rest of the chain because the ring is a closed, nearly inextensible five-membered loop — you cannot change δ without simultaneously changing the effective C1′–C4′ virtual bond vector, which sets where the base attached at C1′ points relative to the next phosphate.
Step 2 — Translate ring geometry into base-pair placement. In C2′-endo sugars, the C1′–C4′ vector lies such that the attached base plane sits nearly perpendicular to, and centered on, the helix axis — base pairs stack with rise ≈ 3.4 Å and negligible inclination to the axis (B-DNA). In C3′-endo sugars, the C1′–C4′ vector tilts the base plane, displacing the base pairs radially outward from the axis by roughly 4–5 Å and inclining them by about 20° relative to the helix axis. This inclination is the structural origin of A-DNA's hollow, tube-like core (the base pairs no longer sit on the axis, unlike B-DNA where they nearly do).
Step 3 — Rise per bp is a projection, not a fixed stacking distance. Base-stacking interactions (van der Waals contact plus dipole-dipole stacking energy) still favor a stacking separation of roughly 3.3–3.4 Å measured along the local base-pair normal. But because the base pairs in A-DNA are tilted and displaced off-axis, the projection of that stacking separation onto the helix axis is shorter than the true interplanar spacing. Quantitatively,
where Å is the true base-pair stacking distance and is the inclination angle in A-DNA. Numerically,
which is close to, though not identical to, the experimentally refined A-DNA rise of about 2.3–2.6 Å per bp (the discrepancy reflects that inclination is only one of several coupled parameters — roll and x-displacement also contribute — but the sign and rough magnitude of the compaction are correctly captured by inclination alone).
Step 4 — More rise-compaction per bp forces more base pairs per turn. Because the double helix must complete a fixed total twist of 360° per turn while packing bases with smaller axial rise, and because the phosphate-phosphate through-backbone distance per residue is nearly conserved (the backbone chain length per nucleotide does not change much between forms), a shorter axial rise per base pair is compensated by more base pairs being needed to complete one full turn. This is why A-DNA has 11 bp/turn versus B-DNA's 10.5, while Z-DNA — despite being compacted differently, via the syn/anti alternation, not tilt — ends up with an even larger 12 bp/turn, because its dinucleotide repeat effectively "skips" twist at every other step.
Conclusion. The switch from C2′-endo to C3′-endo pucker is a purely local, ring-geometry event, but because the furanose ring is a rigid mechanical linkage between the backbone and the base, it propagates deterministically into global helical parameters: shorter axial rise, larger base-pair inclination, and more base pairs per turn — precisely the measured differences between A- and B-DNA.
Go deeper · Derivation B · 3 minWhy alternating purine-pyrimidine sequence favors Z-DNA: a stacking and syn-purine argument
Goal. Derive, from base-stacking and steric considerations, why (dC-dG)ₙ (and more generally alternating Pu-Py tracts) are strongly biased toward the left-handed Z conformation relative to random sequence, and why methylation of cytosine at C5 (5-methylcytosine) further stabilizes Z-DNA.
Step 1 — The Z-DNA repeat unit is a dinucleotide, not a mononucleotide. In B-DNA every nucleotide is structurally equivalent: anti glycosidic angle, C2′-endo pucker, uniform twist. In Z-DNA the asymmetric unit is a CpG (or more generally Pu-Py) step in which the pyrimidine (C) stays anti/C2′-endo — essentially B-like — while the purine (G) flips to syn/C3′-endo(or C2′-exo). This alternation requires that every other base along one strand undergoes a conformational change that costs backbone strain energy; a sequence that does not alternate (e.g., all purines on one strand, as in poly(dA)·poly(dT)) cannot adopt this pattern without forcing every purine into the higher-energy syn state, which is thermodynamically unfavorable, especially for adenine relative to guanine.
Step 2 — Guanine's N2 amino group makes syn-G comparatively cheap. In the syn conformation, a purine base is rotated so that its N3/N9 imidazole edge, rather than its Watson-Crick edge, sits over the sugar ring. For guanine, the C8-H and the N2 exocyclic amino group create a relatively favorable steric environment in syn, and moreover the syn guanine's O6/N7 edge can form a stabilizing intramolecular contact with the C3′-endo/C2′-exo sugar-phosphate backbone (a G(syn)–phosphate interaction sometimes described in terms of a stabilizing electrostatic/H-bond-like contact between guanine N2-H and the adjacent phosphate oxygen). Adenine lacks this N2 amino group, so syn-adenine is comparatively destabilized relative to syn-guanine — this is why (dC-dG)ₙ converts to Z-DNA far more readily than (dA-dT)ₙ-type alternating sequences.
Step 3 — Electrostatic penalty of Z-DNA and the role of counterions. Z-DNA's zigzag backbone brings phosphate groups closer together on average than in B-DNA (the minimum P–P distance across the narrow groove is compressed), increasing electrostatic repulsion between backbone phosphates. This repulsion is a real, unfavorable contribution to . High concentrations of monovalent cations (e.g., >2 M NaCl) or the presence of di/trivalent cations (Mg²⁴, spermidine) screen this repulsion via the counterion condensation mechanism, lowering the electrostatic penalty and shifting the B⇌Z equilibrium
toward Z as becomes less positive with increasing ionic screening.
Step 4 — Negative supercoiling substitutes for high salt. Because B-DNA is right-handed and Z-DNA is left-handed, converting a stretch of B-DNA to Z-DNA locally removes right-handed twist from the molecule — exactly what negative supercoiling wants, since negative supercoils represent a deficit of right-handed twist relative to relaxed B-form. Thus in a topologically closed, negatively supercoiled plasmid (a condition generated in vivo behind transcribing RNA polymerase or by DNA Supercoiling and Topology-modifying enzymes), converting an alternating Pu-Py tract to Z-DNA releases stored superhelical free energy, providing an additional, physiologically relevant favorable term:
so that even at physiological (not extreme) salt, a sufficiently negatively supercoiled alternating CG tract can adopt Z-DNA.
Step 5 — Cytosine methylation. 5-methylcytosine adds a hydrophobic methyl group at C5, which in the B-form sits in the (relatively polar, water-accessible) major groove but in Z-DNA relocates to a hydrophobic pocket-like environment on the helix exterior; this differential solvation makes methylation stabilize Z-DNA (lower ) at physiological salt without requiring extreme ionic strength, which is one reason CpG methylation (an epigenetic mark) is linked to potential Z-DNA formation in vivo.
Conclusion. Z-DNA formation is not a generic property of "any" DNA under stress; it requires (i) an alternating Pu-Py sequence so the syn/anti dinucleotide pattern can repeat regularly, (ii) guanine (or another purine with amino-group stabilization) rather than adenine at the syn positions, and (iii) either high ionic screening, negative supercoiling, or cytosine methylation to overcome the intrinsic electrostatic and stacking penalty of the zigzag, phosphate-compressed backbone.
Worked examples
Example 1 (foundational): converting a fiber-diffraction repeat into helical rise and twist for B-DNA. Fiber X-ray diffraction of B-DNA shows a meridional reflection series with a fundamental repeat (pitch) of Å, and independent chemical/diffraction evidence establishes base pairs per turn (the classical Watson-Crick value; refined solution values are closer to 10.5, but the original fiber data reported 10). Rise per base pair:
Twist per base pair:
Limiting-case check: using the modern refined value instead, Å and ; both are close to, but distinguishably different from, the classical numbers, exactly reflecting the known refinement of the original Watson-Crick fiber model — a sanity check that the formula tracks real historical/measured discrepancies rather than hiding an error.
Example 2 (applied): comparing groove-accessible surface for a transcription factor reading A-form vs B-form DNA. A transcription factor must insert an α-helix into the major groove to read a specific base sequence via Watson-Crick edge hydrogen bonds. In B-DNA the major groove is wide (≈ 22 Å across, measuring phosphate-to-phosphate across the groove, minus the sum of van der Waals radii) and moderate in depth (≈ 8.5 Å), comfortably fitting an α-helix (diameter ≈ 12 Å with side chains, or ≈ 10 Å bare backbone). In A-DNA the major groove is deep (≈ 13.5 Å) and very narrow (≈ 2.7 Å) — too narrow for an α-helix to insert without severe clashes — while the minor groove is shallow and wide. Given these approximate groove width values (B major ≈ 11.7 Å if measured as the true groove width parameter rather than P-P distance; different conventions exist, so the qualitative comparison, not the exact absolute number, is what matters here), the mechanistic conclusion is:
so a sequence-specific DNA-binding protein evolved to read B-DNA major-groove chemistry (the vast majority of transcription factors) cannot function on an A-form template — this is one practical reason why cells must actively resolve DNA-RNA hybrids (R-loops), which force local A-form geometry, since persistent R-loops block major-groove-reading protein machinery and are genotoxic if not removed.
Limiting-case check: for double-stranded RNA (obligately A-form, since ribose 2′-OH forbids C2′-endo entirely, i.e., the C3′-endo population is ≈100% rather than the partial preference seen in dehydrated DNA fibers), the same narrow, deep major groove applies universally — consistent with the observation that dsRNA-binding proteins (e.g., PKR, Dicer) use entirely different structural motifs (dsRNA-binding domains contacting the minor groove and phosphate backbone) rather than major-groove-reading helices, exactly as the A-form groove geometry predicts.
Example 3 (exam/competition level): estimating the free-energy bias needed to explain an observed B⇌Z equilibrium shift with salt. Suppose that a (dC-dG)₁₂ plasmid insert is found by circular dichroism to be 5% Z-form at 0.05 M NaCl and 95% Z-form at 2.5 M NaCl, at 25 °C ( K), with all other conditions constant. Estimate the free-energy shift between the two salt conditions.
Step 1. At each salt concentration, the equilibrium constant for Z formation is
where is the fraction Z-form. At low salt: , so
At high salt: , so
Step 2. Convert each to a standard free energy using , with J mol⁻¹K⁻¹, K, so J mol⁻¹ = 2.478 kJ mol⁻¹.
Step 3. The total free-energy shift attributable to the increase in salt from 0.05 M to 2.5 M is
So raising NaCl from 0.05 M to 2.5 M must lower the Z-relative-to-B free energy by about 14.6 kJ mol⁻¹ (roughly 3.5 kcal mol⁻¹) across this dodecamer repeat, consistent in order of magnitude with the known experimental range for CG-repeat B→Z transitions (typically several kcal/mol total, distributed over the many phosphate-phosphate contacts along a 24-bp alternating tract, i.e., of order a few tenths of a kcal/mol per base-pair step) — a check that the numbers are physically reasonable, not just algebraically consistent.
Limiting-case check: if instead were exactly 0.5 at some intermediate salt (the transition midpoint), and , correctly identifying the midpoint salt concentration as the condition where B and Z are isoenergetic — exactly the expected behavior of a two-state equilibrium model.
Limiting cases, scope & common traps
| Regime / question | What actually happens | Common trap |
|---|---|---|
| Pure poly(dA)·poly(dT) tract under high salt | Stays B-form (does not convert to Z); A/T-rich tracts instead show narrowed minor groove, a distinct "B′" variant | Assuming any high-salt condition triggers Z-DNA — sequence alternation is required, not just ionic strength |
| RNA or DNA·RNA hybrid, any sequence | Always A-form; 2′-OH sterically excludes C2′-endo pucker entirely | Thinking A-DNA is just "a dehydrated version of B-DNA" — for RNA it's obligatory, not a humidity artifact |
| Relaxed (nicked) circular DNA with a CG-repeat insert | Alternating tract stays largely B-form even at moderate salt; no supercoiling-derived driving force | Forgetting that in a topologically closed, negatively supercoiled plasmid, B→Z conversion is thermodynamically assisted by superhelical strain relief, which vanishes once the topology is relaxed |
| Single CpG step embedded in otherwise random sequence | Does not spontaneously flip to Z; the syn/anti alternation must propagate over enough dinucleotide repeats (typically ≥ 6 or so) to nucleate a stable Z segment, because junction (B/Z boundary) formation costs free energy | Treating Z-DNA formation as a single-base-pair, purely local event rather than a cooperative, multi-step-nucleated transition |
| Very low hydration fiber (below ~75% relative humidity) of any GC-moderate sequence | Converts to A-form; rehydration reversibly restores B-form | Assuming A-DNA is a "denatured" or non-native structure — it is a fully base-paired, reversible alternative helix, not a melted or damaged state |
The chemist behind it
The discovery of Z-DNA is inseparable from the work of Alexander Rich, a structural biologist whose laboratory at MIT solved the first single-crystal structure of a left-handed double helix in 1979, using a self-complementary hexamer d(CpGpCpGpCpG). Before that, DNA structural knowledge came almost entirely from fiber diffraction — X-ray patterns from stretched, partially oriented fibers of natural or synthetic DNA, which give averaged, lower-resolution information than a true single-crystal structure. Rich's group, collaborating with Andrew Wang, Gary Quigley, and colleagues, obtained genuine single crystals of the hexamer, allowing atomic-resolution refinement that unambiguously revealed the zigzag phosphate backbone, the syn-guanine/anti-cytosine alternation, and the left-handed helical sense — a structure so unexpected relative to prevailing B-form intuition that it initially met considerable skepticism. Rich had earlier been central to establishing the A-form of RNA-containing double helices and to solving the first tRNA crystal structure (with Sung-Hou Kim), giving him deep prior expertise in recognizing non-B helical geometries when fiber and crystallographic data disagreed with simple Watson-Crick model expectations. His career exemplifies a recurring theme in structural biochemistry: anomalies in diffraction data, rather than being discarded as noise, are frequently the signature of a genuinely new molecular conformation.
Origins · 1 minAlexander Rich and how this was discovered
Is Z-DNA biologically real, or only seen in crystals under extreme conditions?
Z-DNA is biologically real. While the original 1979 crystal structure required a self-complementary hexamer under laboratory conditions, subsequent work using Z-DNA-specific antibodies showed transient Z-DNA formation in living cells, especially in transcriptionally active chromatin where negative supercoiling accumulates behind moving RNA polymerase. More recent structural biology has identified dedicated Zα protein domains (in ADAR1 and ZBP1/DAI) that specifically recognize Z-form nucleic acid, strong evidence that cells have evolved to use, not merely tolerate, this conformation as a signal, particularly in innate immune sensing of viral or endogenous nucleic acids.
Why is RNA always A-form and never B-form?
Ribose carries a 2′-hydroxyl group that ribose-free deoxyribose lacks. In the C2′-endo pucker characteristic of B-DNA, the 2′-substituent would clash sterically with the neighboring phosphate group's oxygen atoms across the sugar-phosphate backbone. This 1,3-type steric clash is avoided entirely in the C3′-endo pucker, so RNA (and the RNA strand of any DNA-RNA hybrid) is essentially forced into C3′-endo, and hence into A-form helical geometry, with no accessible B-form alternative. This is a rigid steric constraint, not merely a statistical preference, which is why dsRNA structures are far more uniformly A-form than DNA is uniformly B-form.
Do A-DNA and B-DNA differ in base pairing chemistry?
No. Both use identical Watson-Crick hydrogen-bonding schemes — two hydrogen bonds for A·T (or A·U in RNA) and three for G·C — with the same antiparallel strand polarity. The A/B distinction is entirely about backbone and sugar geometry: pucker, base-pair inclination and displacement from the axis, rise per base pair, and groove dimensions. This is why a single DNA molecule can transition smoothly between A- and B-form as hydration changes, without breaking or reforming any base pair.
What experimental methods distinguish A-, B-, and Z-DNA in the laboratory?
Fiber X-ray diffraction historically distinguished the forms via their characteristic layer-line patterns and helical repeat distances; single-crystal X-ray crystallography (as used for the first Z-DNA structure) gives atomic-resolution confirmation of pucker and glycosidic angle. Circular dichroism (CD) spectroscopy is a fast solution-phase method: B-DNA shows a characteristic positive band near 275 nm and negative band near 245 nm, while Z-DNA shows an inverted CD spectrum (negative band near 290 nm, positive near 260 nm) due to its opposite helical handedness, making CD a convenient assay for monitoring B⇌Z transitions as a function of salt or supercoiling.
Can a single DNA molecule contain both B-form and Z-form segments simultaneously?
Yes. A B-Z junction forms at the boundary between a Z-DNA-forming alternating tract and flanking B-form sequence. These junctions are energetically costly — early theoretical and experimental work estimated a substantial free-energy penalty per junction, meaning short Z-forming tracts are disfavored relative to longer ones, since the junction cost must be paid regardless of tract length while the favorable Z-stabilization energy scales with the number of internal base-pair steps. This cooperative, length-dependent behavior is why Z-DNA formation typically requires a run of at least several alternating purine-pyrimidine repeats to be thermodynamically favorable in a real genomic context.
How does DNA methylation relate to Z-DNA formation, and why does this matter for epigenetics?
5-methylcytosine, the mark deposited by DNA methyltransferases at CpG dinucleotides, stabilizes Z-DNA at physiological salt because its C5 methyl group is more favorably solvated in the Z-form's altered groove geometry than in the B-form major groove. Since CpG methylation is a central epigenetic regulatory mark controlling gene expression, this creates a potential mechanistic link: methylation-promoted Z-DNA formation near promoters could influence transcription factor accessibility or chromatin structure, an active area of ongoing epigenetics research connecting helix conformation directly to gene regulation.
Why does A-DNA have such a different groove structure from B-DNA if the base pairs are chemically identical?
The difference arises because in A-DNA the base pairs are tilted (inclined roughly 20° to the helix axis) and displaced radially outward from the axis, rather than sitting nearly centered and perpendicular to the axis as in B-DNA. This tilt-and-displacement, a direct consequence of the C3′-endo sugar pucker, changes which atoms of the backbone face the major versus minor groove and how far apart phosphates on opposite strands sit when measured across each groove, producing a deep/narrow major groove and shallow/wide minor groove — the near-inverse of B-DNA's shallow major/deep minor groove — without altering a single Watson-Crick hydrogen bond.