Z-DNA

Z-DNA is a high-energy, left-handed double-helical conformation of deoxyribonucleic acid in which the sugar-phosphate backbone follows a characteristic zig-zag trajectory around the helical axis, contrasting sharply with the smooth right-handed coiling of canonical B-DNA and A-DNA. First solved at atomic resolution in 1979 by Alexander Rich, Andrew Wang, and colleagues via single-crystal X-ray diffraction of the self-complementary hexanucleotide d(CpGpCpGpCpG), Z-DNA retains antiparallel strand orientation and canonical Watson–Crick base pairing, yet reverses the handedness of the double helix. Rather than the mononucleotide repeat unit of B-DNA, the fundamental repeating structural unit of Z-DNA is a dinucleotide composed of an alternating purine–pyrimidine step (d(GpC) or d(TpG)•d(CpA)).

Under relaxed physiological ionic strength, B-DNA represents the global thermodynamic minimum for duplex nucleotides, whereas the B-to-Z transition carries an unfavorable nucleation free energy of approximately +5.0 kcal mol⁻¹ (two B–Z junctions at ~2.5 kcal mol⁻¹ each) plus a propagation penalty of +0.33 kcal mol⁻¹ per (dG-dC) base pair. Within living cells, Z-DNA is dynamically driven into existence by torsional strain—specifically negative supercoiling generated behind translocating RNA Polymerase II during active transcription—as well as by C5 cytosine DNA methylation (5mC) and specific Z-DNA-binding proteins containing the Zα winged-helix domain, notably the RNA editing enzyme ADAR1 (p150) and the innate immune sensor ZBP1.

Alternating syn-Purine and anti-Pyrimidine Conformations

The structural origin of the left-handed zig-zag backbone lies in the steric behavior of the N-glycosidic bond connecting each nucleobase to its deoxyribose ring. In canonical right-handed B-DNA, every purine and pyrimidine nucleotide adopts the anti glycosidic torsion angle (χ ≈ −102° to −115°), where the bulk of the heterocycle points away from the sugar ring, and every deoxyribose adopts the C2′-endo sugar pucker. Purines (guanine and adenine), however, can rotate around the glycosidic bond into the syn conformation (χ ≈ +60° to +70°), positioning the six-membered pyrimidine ring of the purine directly above the deoxyribose sugar with minimal steric clash. Pyrimidines (cytosine and thymine) resist syn rotation because their exocyclic C2-carbonyl oxygen collides sterically with the C2′ and H3′ atoms of the sugar ring.

When an alternating purine–pyrimidine sequence such as (dG-dC)_n or (dT-dG)_n•(dC-dA)_n transitions into Z-DNA, every deoxyguanosine flips into the syn glycosidic conformation coupled with a C3′-endo sugar pucker, while every deoxycytidine remains in the anti conformation with a C2′-endo sugar pucker. To preserve Watson–Crick hydrogen bonding while alternating between syn and anti glycosidic angles along each strand, the base pairs must flip 180° relative to the B-DNA helical axis. This alternating C2′-endo/anti to C3′-endo/syn geometry forces the phosphodiester linkages to step inward and outward on alternating steps, producing the diagnostic zig-zag phosphate trace and a helical pitch of 44.6 Å spanning 12 base pairs (6 dinucleotide units) per left-handed turn.

Groove Remodeling and Solvent Exposure of the C8/N7 Edge

Flipping the base pairs into the left-handed Z conformation remodels the surface topography of the double helix. Whereas B-DNA displays a wide, accessible major groove and a narrow minor groove, Z-DNA possesses a single deep, narrow groove analogous to the B-DNA minor groove. The former major groove is pushed outward to form a convex cylindrical outer surface along which the C5 position of cytosine and the N7 and C8 atoms of guanine are directly exposed to bulk solvent and protein side chains. Furthermore, the inter-strand phosphate–phosphate distance shrinks to ~7.7 Å across the deep groove in the Z_I sub-conformation (compared to ~11.7 Å in B-DNA), explaining why high salt concentrations (>2.5 M NaCl or millimolar Mg²⁺/spermine) or cytosine C5 methylation screen electrostatic phosphate repulsion and favor the B-to-Z transition.

Negative Supercoiling and Epigenetic Cytosine Methylation

In topologically constrained chromosomal domains, the primary physiological driver of Z-DNA formation is negative supercoiling (−σ). According to twin-supercoiled-domain mechanics described by Leroy Liu and James Wang, as RNA Polymerase II tracks along the helical template during transcription, frictional drag prevents the elongation complex from rotating freely around the DNA axis. This generates positive supercoils ahead of the polymerase and waves of underwound negative supercoils (σ ≈ −0.05 to −0.07) in its wake. Flipping a 12-bp right-handed B-DNA segment (+1.14 turns) into a 12-bp left-handed Z-DNA segment (−1.00 turn) absorbs 2.14 negative superhelical turns, relaxing torsional stress across the surrounding chromatin domain and paying the +5.0 kcal mol⁻¹ nucleation cost of the two B–Z junctions (where one base pair extrudes from the helix at each boundary).

Epigenetic DNA methylation of CpG dinucleotides (5-methylcytosine, 5mC) lowers the torsional threshold required for Z-DNA conversion. In B-DNA, the hydrophobic methyl group at the C5 position of cytosine projects into the major groove, structuring an energetically unfavorable clathrate-like cage of ordered water molecules. When (dG-dm5C)_n flips into Z-DNA, the C5 methyl group packs into a hydrophobic depression formed by the neighbouring syn-guanine imidazole ring and sugar C1′–C2′ patch, releasing structured water and reducing the midpoint NaCl concentration required for the B-to-Z transition from ~2.7 M to <0.7 M.

Human Subtelomeric (TG)n Microsatellites and the AGRN Locus

Computational thermodynamic algorithms such as Z-Hunt and sliding-window alternating purine–pyrimidine scoring reveal that Z-DNA-forming sequences (Z-flipons) cluster near transcription start sites and inside human subtelomere regions. While (dG-dC)_n repeats possess the lowest propagation free energy (+0.33 kcal mol⁻¹/bp), CpG depletion by deamination leaves alternating (dT-dG)_n•(dC-dA)_n microsatellites (+0.66 kcal mol⁻¹/bp) as the most abundant extended Z-DNA reservoirs in the human genome.

A prominent empirical locus occurs within the human Chr1p36.33 subtelomere inside the AGRN (agrin) gene, which encodes the heparan sulfate proteoglycan required for neuromuscular junction postsynaptic clustering. Across AGRN Introns 18–22 (NC_000001.11), dense tandem (TG)_n repeats coexist with flanking subtelomeric G-quadruplex tracts, reaching peak alternating purine–pyrimidine propensities (Z-score > 8.0 SD above background). During high-rate transcription, these intronic Z-DNA tracts act as torsional shock absorbers that buffer supercoiling and modulate co-transcriptional spliceosome assembly across closely spaced exons, but also create fragile B–Z junctions susceptible to ERCC1–XPF and MRE11 endonuclease cleavage and double-strand break formation.

Structural Basis of Zα Winged-Helix Binding

How proteins distinguish left-handed Z-DNA from a million-fold excess of right-handed B-DNA was resolved in 1999 when Alexander Rich and Thomas Schwartz determined the crystal structure of the Zα domain of human ADAR1 bound to d(CGCGCG)₂. The Zα domain folds into a compact α/β winged-helix-turn-helix (α1–α2–α3–β1–β2–β3) architecture that recognizes Z-DNA solely through conformation-specific shape and electrostatic complementarity rather than sequence-specific base reading.

Two Zα monomers bind symmetrically to opposite sides of a 6-bp Z-DNA segment. Residues along the recognition helix α3 (including Lys169, Lys170, Asn173, and Tyr177 in human ADAR1) and the β2–β3 wing (Trp195 and Pro192/Pro193) make direct and water-mediated hydrogen bonds with five consecutive zig-zagging phosphate groups along a single strand. Simultaneously, the aromatic ring of Tyr177 packs Edge-to-Face (CH–π interaction) against the solvent-exposed C8 carbon of a syn-deoxyguanosine residue—a steric contact impossible in B-DNA where the C8 atom is buried inside the major groove floor.

ADAR1 versus ZBP1 in Antiviral Defense and Necroptosis

The Zα domain also binds left-handed double-stranded Z-RNA formed during the replication of viral genomes or transcription of endogenous retroelements (Alu inverted repeats). In mammalian innate immunity, two Zα-containing proteins compete for Z-nucleic acids with opposing physiological outcomes:

1. ADAR1 p150 (Adenosine Deaminase Acting on RNA 1): Induced by type I interferons, ADAR1 p150 uses its N-terminal Zα domain to dock onto endogenous Z-RNA/Z-DNA duplexes and catalyze adenosine-to-inosine (A-to-I) deamination, destabilizing the duplex and preventing false-positive autoimmune activation. Loss-of-function mutation in the ADAR1 Zα domain (Pro193Ala) causes Aicardi–Goutières syndrome, a severe pediatric type I interferonopathy.
2. ZBP1 (Z-DNA Binding Protein 1 / DAI): When viral infection (such as influenza A or herpesviruses) or ADAR1 deficiency allows unedited Z-RNA/Z-DNA to accumulate, ZBP1 binds the left-handed helix via its tandem Zα1 and Zα2 domains. This triggers homotypic RHIM (Receptor-Interacting Protein Homotypic Interaction Motif) oligomerization with the kinase RIPK3, which phosphorylates MLKL to execute inflammatory necroptosis and eliminate the infected cell.

Why does Z-DNA require alternating purine and pyrimidine nucleotides?

A left-handed double helix that preserves Watson–Crick base pairing requires every second nucleotide along a strand to flip its glycosidic bond from anti to syn. Purines (G and A) readily adopt the syn conformation with minimal steric strain, whereas pyrimidines (C and T) experience steric collision between their C2-keto oxygen and the deoxyribose C2′ atom when rotated into syn. Consequently, alternating d(GpC) or d(TpG)•d(CpA) repeats match the natural anti-pyrimidine / syn-purine dinucleotide repeat of Z-DNA.

How does transcription by RNA Polymerase II induce Z-DNA in living cells?

As RNA Polymerase II translocates along chromatin without rotating around the DNA template, it overwinds the DNA ahead of the transcription bubble (positive supercoiling) and underwinds the DNA behind it (negative supercoiling). Because converting one 12-bp turn of right-handed B-DNA into left-handed Z-DNA absorbs more than two negative superhelical turns, negative torsional strain behind the polymerase overcomes the ~5 kcal mol⁻¹ nucleation barrier and drives alternating purine–pyrimidine tracts into Z-DNA.

What is the role of the Tyr177 and Trp195 triad in Zα domain proteins?

In Zα winged-helix proteins (ADAR1 and ZBP1), Trp195 forms the hydrophobic core that positions Tyr177 on the α3 recognition helix perpendicular to the Z-DNA axis. The phenolic ring of Tyr177 makes a diagnostic CH–π van der Waals contact with the exposed C8 atom of syn-guanine while neighboring Asn173 and lysine residues lock onto the zig-zag phosphate backbone, conferring nanomolar specificity for left-handed Z-DNA and Z-RNA.

1. Wang, A. H. J., Quigley, G. J., Kolpak, F. J., Crawford, J. L., van Boom, J. H., van der Marel, G., & Rich, A. (1979). "Molecular structure of a left-handed double helical DNA fragment at atomic resolution." Nature, 282(5740), 680–686. https://doi.org/10.1038/282680a0
2. Schwartz, T., Rould, M. A., Lowenhaupt, K., Herbert, A., & Rich, A. (1999). "Crystal structure of the Zα domain of the human editing enzyme ADAR1 bound to left-handed Z-DNA." Science, 284(5421), 1841–1845. https://doi.org/10.1126/science.284.5421.1841
3. Rich, A., & Zhang, S. (2003). "Timeline: Z-DNA: the long road to biological function." Nature Reviews Genetics, 4(7), 566–572. https://doi.org/10.1038/nrg1115
4. Jiao, H., Wachsmuth, L., Kumari, S., Schwarzer, R., Lin, J., Eren, R. O., Fisher, A., Lane, R., Young, G. R., Kassiotis, G., Kaiser, W. J., & Pasparakis, M. (2020). "Z-nucleic-acid sensing triggers ZBP1-dependent necroptosis and inflammation." Nature, 580(7803), 391–395. https://doi.org/10.1038/s41586-020-2129-8