Formation and biological implications of Z-DNA

Abstract

Z-DNA is a left-handed alternative DNA structure that forms at alternating purine-pyrimidine repeats, which are abundant in genomes. It is intrinsically unstable under physiological conditions; however, it can be stabilized by negative supercoiling and specific Z-DNA binding proteins. These stabilizing factors have prompted renewed interest in the biological significance of Z-DNA within the genome. Emerging evidence suggests that Z-DNA plays critical roles in various cellular processes, including transcriptional regulation, genome instability, chromatin remodeling, and the development of human diseases. This review summarizes existing methodologies for local and global identification of Z-DNA, its genomic and epigenetic features, the factors influencing its formation and stability, its biological implications, and future directions to advance our understanding of Z-DNA biology and its potential applications.

Overview of a left-handed DNA helix (Z-DNA)

Z-DNA (see Glossary) is a noncanonical left-handed double-helical nucleic acid structure. In 1972, a circular dichroism (CD) spectrum study showed that high-salt conditions (e.g., 2.5 M NaClO, 1.8 M NaCl, or 0.7 M MgCl ) can induce conformation changes of synthetic poly(dG-dC) DNA, a potential Z-DNA formation sequence. This pioneering study may be the first report of potential Z-DNA formation in vitro, although it did not clearly state the structure as Z-DNA or a left-handed double helix [1]. Z-DNA was formally discovered in double-stranded DNA (dsDNA) fragments containing d(CpGpCpGpCpG) through X-ray diffraction analysis in 1979 [2]. It is inherently unstable under physiological conditions, necessitating additional energy or the assistance of binding proteins to maintain its structure. However, it can become thermodynamically stable under certain nonphysiological conditions, such as high ionic strength or specific chemical modifications. As a result, its genomic role was initially underestimated. Negative supercoiling and Z-DNA binding proteins (ZBPs) have been found to stabilize Z-DNA, leading to reconsideration of its potential biological implications. Accumulating evidence demonstrates that Z-DNA plays important roles in various cellular processes associated with distinct physiological and pathological activities, including transcriptional regulation, genome instability, restriction of nucleosome formation, chromatin remodeling, and human disease etiology. While further studies are still required for a broader understanding of Z-DNA function on a genome-wide scale in organisms, the availability of omics methodologies for the identification of Z-DNA represents a pivotal turning point. This advance is expected to accelerate progress towards a comprehensive understanding of the biological implications of Z-DNA.

Existing methodologies for the identification of Z-DNA

Currently, the detection of Z-DNA conformations, both in vitro and in vivo, faces significant challenges due to its structural heterogeneity, transient nature, and context-dependent stability. Existing methodologies can be broadly categorized into four groups: biophysical methods, biochemical techniques, antibody-based detection and profiling, and computational prediction (Table 1).

Table 1.

Summary of existing methodologies for identification of Z-DNA

Category	Subtype	Methodology	Refs
Biophysical methods	Spectroscopy	CD	[1,3–5]
		IR spectroscopy	[6]
		Raman spectroscopy (including surface-enhanced Raman)	[7,8]
	NMR	Solution-state NMR	[7,9,10]
	X-ray diffraction	Single-crystal X-ray diffraction	[2,11]
	Fluorescence techniques	FRET (Z-to-B transition monitoring)	[12]
		smFRET	[13]
	Hydrogen exchange kinetics	HDX	[14]
	Thermodynamic analysis	ITC	[15]
		DSC	[15]
Biochemical methods	Electrophoresis	Native/denaturing gel electrophoresis	[16–20]
	Torque spectroscopy	Single-molecule torque measurements	[21,22]
	Ultracentrifugation	Sedimentation analysis	[23]
	Membrane-binding assays	Nitrocellulose filtration	[24]
	Ligand-based detection	Ruthenium complex groove binding	[24]
	Nuclease sensitivity	DNase I/restriction enzyme cleavage assays	[25,26]
	Chemical probes	Bromine/OsO4/DEPC modification	[27,28]
Antibody/protein based	High-throughput profiling	ZαADAR1 ChAP-seq	[34]
		Zaa-FLAG ChIP-seq	[30]
		Z22 antibody-based ZIP-seq/ChIP-seq/CUT&TAG	[35,36]
Computational prediction	Algorithmic tools	Z-Hunter	[32]
		DeepZ (machine learning)	[29,32]
		Z-DNABERT/ZSeeker	[33]
		nBMST	[29]
Protein-independent methods	Genome-wide mapping	ssDNA mapping + PZFS overlap	[37]
		ViCAR Hi-C adaptation	[38,39]

Biophysical techniques are commonly used to investigate Z-DNA, including CD), IR spectroscopy, Raman spectroscopy, NMR, X-ray diffraction, fluorescence resonance energy transfer (FRET), single-molecule FRET (smFRET), hydrogen-deuterium exchange (HDX), and thermodynamic techniques such as isothermal titration calorimetry (ITC) and differential scanning calorimetry (DSC) (Figure 1). Each individual method offers unique advantages in characterizing the structural and dynamic features of Z-DNA based on their spectroscopic or thermodynamic signatures. Among them, CD remains a cornerstone for identifying Z-DNA due to its characteristic negative ellipticity at 290 nm; however, its dependence on high concentrations of purified oligonucleotides and salts limits genome-wide scalability [1,3–5]. IR and Raman spectroscopy provide complementary advantages; IR is particularly sensitive to polarity-dependent vibrations, whereas Raman spectroscopy is more suited to analyzing nonpolar bonds with higher spatial resolution, although typically restricted to surface-enhanced formats [6–8]. NMR delivers atomic-resolution dynamics of Z-DNA in solution; however, it necessitates high sample purity and technical expertise [7,9,10]. X-ray diffraction, the gold standard for static structural determination, requires crystalline samples, which limits its applicability to short oligonucleotides [2,11]. For real-time conformational tracking, FRET monitors nanometer-scale B-to-Z transitions, while smFRET resolves conformational heterogeneity at the expense of photo-bleaching artifacts [12,13]. HDX probes Z-DNA stability by assessing the solvent accessibility of hydrogen-bonded regions but requires stringent experimental controls [14]. ITC and DSC assess DNA stability by measuring enthalpy changes and melting profiles, respectively; however, both methods necessitate high DNA concentrations [15].

Figure 1. Summary of existing methodologies for identification of Z-DNA and timeline for important case studies.

Existing methodologies include low- and high-throughput methods. High-throughput ones can be classified into two categories: (i) computer-based prediction of Z-DNA, and (ii) experimental detection of Z-DNA through high-throughput sequencing. Timepoint for the first application of all methodologies and important case studies summarized as a timeline from 1979 to 2025. See also [2,6–30,32–39].

Biochemical techniques primarily comprise gel electrophoresis, torque spectroscopy, sedimentation analysis, nitrocellulose filtration, ligand-based chiral molecule binding, nuclease sensitivity assays, and chemical and enzymatic probe-based detection. In general, gel electrophoresis exploits the altered migration patterns of Z-DNA due to its left-handed helical structure. It also determines certain physical parameters that are important for prediction of Z-DNA formation in native sequences. In these methods, B–Z transition within Z-DNA-forming sequences (ZFSs), which are inserted in substantially longer DNA (usually, a plasmid) with various degrees of negative supercoiling, leads to supercoiling changes in the DNA fragments. This supercoiling change, instead of the Z-DNA structure itself, can result in a change in electrophoretic mobility. Thus, these methods could precisely determine the quantitative parameters of Z-DNA formation (e.g., the energy ‘cost’ of B–Z junctions or a deviation from the perfect ZFS) [16–20]. Torque spectroscopy can also be utilized to study strand separation for DNA containing AT-rich sequences, and study Z-DNA formation of d(pGpC)_n sequences [21,22]. By contrast, ultracentrifugation-based sedimentation analysis enables the separation of Z-DNA primarily due to its unique hydrodynamic properties, making it effective for isolating large DNA fragments, although it requires specialized equipment [23]. Nitrocellulose filtration distinguishes Z-DNA from B-DNA based on their differential binding affinities to the membrane; however, it suffers from suboptimal specificity. Ligand-based chiral molecule binding, such as that involving ruthenium complexes, offers conformation-specific recognition of Z-DNA through groove-binding interactions. However, this approach requires substantial optimization and presents synthetic complexities [24]. Nuclease sensitivity assays, including DNase I and restriction enzyme cleavage, exploit the increased susceptibility of Z-DNA to enzymatic degradation. Nevertheless these methods may be affected by the sequence context, which may limit their specificity [25]. For example, Z-DNA exhibits resistance to the activity of the Hha I restriction enzyme and methylase compared with the surrounding B-DNA [26]. Finally, chemical and enzymatic probe-based detection methods, such as bromine and OsO4 modifications, facilitate structure-specific labeling of Z-DNA regions, balancing specificity and technical complexity [27]. Additionally, diethyl pyrocarbonate (DEPC) has been utilized as a chemical probe to detect Z-DNA by modifying accessible adenine and guanine residues, enabling distinction between Z- and B-DNA according to differential reactivity patterns [28].

In addition to the aforementioned low-throughput methodologies, high-throughput techniques primarily encompass dry-lab (computational algorithms) and wet-lab detection methods, such as ZBP assays, antibody-based ChIP-seq, cleavage under targets and release using nuclease (CUT&RUN)/cleavage under targets and tagmentation (CUT&TAG) assays, and protein-independent methodologies. Computational prediction tools, which are applied for genome-wide identification of potential ZFSs, include Z-Hunter and its enhanced version Z-Hunter II, as well as DeepZ, Z-DNABERT, ZSeeker, and the non-B DNA Motif Search Tool (nBMST) [29]. These tools primarily utilize sequence-based algorithms to identify Z-forming regions (ZFRs) through CG dinucleotide repeats and alternating purine-pyrimidine patterns with or without consideration of additional parameters [30–32]. Z-Hunter tends to systematically overestimate the prevalence of Z-DNA by neglecting essential factors, including epigenetic modifications, chromatin topology, and the presence of ZBPs and/or chemical modifications. To enhance prediction accuracy, machine learning models integrating multiple omics data have facilitated the development of the DeepZ algorithm [32]. Similarly, a novel algorithm called ZSeeker demonstrates significant potential for predicting Z-DNA with enhanced accuracy across the genome, showing high consistency with various experimental data [33]. It is important to conduct experimental validation to ensure the accuracy of sequence-based computational prediction for potential ZFSs. Currently, the reported wet-lab methodologies for global profiling of Z-DNA primarily include ZBP-based ChIP-seq, such as artificially engineered Zaa protein with a FLAG tag [30], Z22 (an anti-Z-DNA antibody)-based ZIP-seq (Z22-based Z-DNA immunoprecipitation coupled with sequencing), and ChIP-seq and CUT&TAG. Zaa-based ChIP-seq has been employed for genome-wide mapping of Z-DNA structures in human cells (n = 391) [30], which is slightly more than the ZαADAR1-based in vitro chromatin affinity precipitation (ChAP) coupled with Sanger sequencing in human cells (n = 186) [34]. However, this number is significantly lower than Z-DNA structures identified using Z22-based ChIP-seq in mice (n = 6022) [35]. Similarly, ZIP-seq (n = 12 189 in vitro) and Z22-based CUT&TAG (n = 3214 in vivo) have recently been applied to the profiling of Z-DNA both in vitro and in vivo in rice [36]. An important caveat to protein and antibody-based approaches is that their binding to DNA could facilitate the formation of Z-DNA, and their specificity to Z-DNA structures versus sequences must be confirmed. Additionally, protein-independent methodologies, such as the mapping of single-stranded DNA (ssDNA) combined with overlapping ZFSs, can also effectively map Z-DNA across the genome [37]. Furthermore, certain techniques such as viewpoint Hi-C on accessible regulatory DNA (ViCAR), which has successfully identified non-B-DNA structures like G-quadruplexes (G4s), can be adapted for genome-wide Z-DNA mapping [38,39]. These two techniques fall under protein-independent methodologies. While each method offers valuable tools for the identification of Z-DNA both locally and globally, they possess distinct advantages and limitations. Therefore, it is crucial to carefully consider essential parameters such as resolution, specificity, and technical feasibility. Ongoing optimization and refinement of these techniques will improve their applicability in research. Undoubtedly, the combination of complementary techniques could provide a comprehensive overview of Z-DNA throughout the genome.

Impact of endogenous and exogenous factors on Z-DNA formation

Unlike the canonical right-handed B-DNA, which represents a thermodynamically favorable conformation under physiological conditions, Z-DNA formation incurs an energetic cost due to its unfavorable geometry and torsional strain. However, the nucleation energy barrier for Z-DNA formation is relatively modest – estimated at approximately 8–10 kcal/mol, which is significantly lower than that required for other non-B-DNA structures such as DNA cruciform or H-DNA structures. As a result, even low levels of negative supercoiling, such as those generated upstream of an elongating RNA polymerase, are often sufficient to induce Z-DNA formation. This property renders Z-DNA a potential sink for absorbing transcription-induced torsional stress, particularly within promoter regions. Once formed, Z-DNA can be further stabilized by sequence composition, base modifications, and specific protein interactions, which collectively distinguish the energy required for its initial formation from the structural persistence of the Z-DNA state [40]. Growing evidence suggests that a variety of endogenous and exogenous factors can influence the formation and stability of Z-DNA, both in vitro and in vivo [41]. These factors encompass intrinsic DNA sequences, chemical modifications of bases [42], mechanical stresses such as negative supercoiling [43], environmental conditions such as ionic strength and type [44], processive enzymes including polymerases and helicases [40], ZBPs [45], and DNA-interactive compounds (e.g., actinomycin D, actinomycin) [46]. At the sequence level, Z-DNA typically contains unique DNA sequence bases and contexts, such as alternating purine-pyrimidine motifs and high GC content. The propensity for Z-DNA formation varies among different DNA sequence contexts, including d(GC/CG), d(TG) /d(GGGC)_n, and d(TA) following a descending order [47,48]. Furthermore, DNA fragments with high GC content or those containing mismatched base pairs, such as G·T, G·Br5U, and 5FU·G wobble pairs, are more likely to adopt a Z-DNA conformation [49,50]. Additionally, the energetic landscape of Z-DNA is influenced not only by nucleation and sequence favorability but also by structural irregularities. Experimental studies have demonstrated that mismatches, bulges, and B–Z junctions impose additional energetic penalties that hinder the stable propagation of the left-handed helix. These irregularities can significantly modulate the stability, length, and genomic distribution of Z-DNA in vivo [51]. In addition to sequence composition, chemical modifications of DNA bases play a pivotal role in promoting and stabilizing the Z-DNA structure. These modifications enhance the thermodynamic stability of the Z-DNA conformation, thereby facilitating its formation under physiological conditions or stress-induced circumstances. For instance, the presence of cytosine (C) modifications, including methylation (5mC), as well as halogenation and bromination at the C5 position, tends to compress the underlying sequences. This compression results in base-pair instability and local alterations in helical conformation, thereby favoring the transition from B-DNA to Z-DNA [3,52,53]. Similarly, guanosine (G) modifications, such as methylation or bromination at the C8 position, 8-oxo-7,8-dihydroguanine (8-oxodG), and the C8 position of G modified with acetylaminofluorene or 4-acetoxyaminoquinoline 1-oxide (Ac-4 HAQO), can promote Z-DNA formation through various chemical reactions [54,55]. These modifications can change the electronic properties of the base or introduce structural distortions that destabilize the traditional Watson–Crick base pairing, thereby reducing free energy and favoring a left-handed helical twist [36,53].

Accumulating evidence indicates that the concentration and types of multivalent cations, dehydrating agents such as ethylene glycol or ethanol, crowding agent such as polyethylene glycol (PEG200), and osmolytes can act individually or synergistically to modulate the formation and stability of Z-DNA in vitro. Cations, including Ru⁴⁺, R h³⁺, Co³⁺ {[Co(NH₃)6]Cl₃}, Co²⁺ (CoCl₂), Mn²⁺, Mg²⁺, Ba²⁺, and Ca²⁺, as well as monovalent cations Li⁺ and Na⁺ at high concentrations (e.g., 3.8 M LiCl and up to 4 M NaCl), can promote Z-DNA formation by shielding the negatively charged DNA backbone and reducing electrostatic repulsion [3,41,56]. Generally, the influence of cations on the conformational transition from B-DNA to Z-DNA is directly related to their valence [57]. This effect is primarily attributed to their electrostatic contributions, the Hofmeister series, and their binding affinity for Z-DNA [41]. The transition from B-DNA to Z-DNA in sequences that have undergone bromination or methylation can occur at significantly lower concentrations of NaCl (150 mM compared with 4 M) or MgCl₂ (≤1.0 mM compared with 0.7 M) compared with the corresponding unmodified sequences. This observation underscores the interplay actions of various factors in facilitating the formation of Z-DNA. Polyamines, such as spermidine and spermine, are naturally occurring positively charged molecules that can promote and stabilize Z-DNA conformations by binding to DNA and neutralizing its negative charges [3]. Additionally, osmolytes, including alcohols (methanol, ethanol, and propanol) and polyols (sucrose and stachyose), can modulate Z-DNA formation by altering the chemical potential of water and osmotic stress, thereby favoring the conformational transition from B-DNA to Z-DNA [58,59].

Mechanical forces, including negative supercoiling, torsional stress, and DNA bending, play a crucial role in modulating the formation of Z-DNA. Negative DNA supercoiling generates torsional strain that under-winds B-DNA, creating topological stress. The formation of left-handed Z-DNA in specific sequences alleviates this underwinding stress without altering the total linking number of the DNA molecule, thereby facilitating an energetically favorable local structural transition. This mechanism serves as a driving force for the B-to-Z transition in negatively supercoiled DNA [23]. Similarly, within cells, the progression of transcriptional machinery induces topological changes as described by the twin supercoiled domain model, in which positive supercoiling accumulates ahead of the RNA polymerase while negative supercoiling builds up behind it. This negative supercoiling introduces torsional strain that promotes the conformational transition from B-DNA to Z-DNA in susceptible sequences [60,61]. In addition, the occurrence of DNA bending alters the helical trajectory, thereby facilitating the conformational transition to form Z-DNA by enhancing thermodynamic stability [62,63]. Consequently, these mechanical stresses collectively induce structural distortions and topological changes, creating conditions that may facilitate the transition from B-DNA to Z-DNA, even under physiological salt conditions that typically inhibit Z-DNA formation.

In addition, ZBPs can modulate the stability of Z-DNA by specifically binding to Z-DNA motifs. These proteins include adenosine deaminase acting on RNA1 (ADAR1), ZBP1 [also known as DLM1 and DNA-dependent activator of IFN-regulatory factors (DAI)], E3L, PKZ, 34L, ORF112, RBP7910, HOP2, DsvD, D2PEW5, feoC, pefI, RPA2, CDC53, CUL1, ANC2, APC2, Rpc34, PBP2, Reut_B4095, and ZBTB43 [64,65]. ZBPs contain the highly conserved Zα domain, which specifically interacts with the left-handed Z-DNA helix. This interaction stabilizes the Z-DNA conformation by preserving the left-handed helical structure of DNA. It has been reported that Zα domains exhibit a diminished ability to induce the B–Z transition and can demonstrate an increased binding affinity for B-DNA [66]. Additionally, negative supercoiling and protein interactions can further enhance the stability of Z-DNA [67]. Notably, several of these ZBPs, such as ADAR1, can also recognize other noncanonical nucleic acid structures. For example, ADAR1 binds to Z-RNA and regulates double-stranded RNA (dsRNA) editing and immune signaling [68]. Similarly, some proteins on the list have been shown to interact with other alternative DNA structures, suggesting that their nucleic acid recognition is not strictly limited to the Z-DNA conformation. This functional overlap implies a broader role for these proteins in sensing and modulating nucleic acid topology in different biological contexts.

In summary, all the aforementioned factors can act either independently or collectively to regulate the conformational transition from B-DNA to Z-DNA (Table 2). These factors subsequently influence the stability of Z-DNA under various experimental and physiological conditions.

Table 2.

Summary of impacts of internal and external factors on Z-DNA formation

Group	Factor	Condition	Potential mechanism	Refs
Internal factors	DNA sequence	Alternating purine-pyrimidine motifs [e.g., d(GC/CG), d(TG), d(GGGC), d(TA) ], d(GC/CG) > d(TG) /d(GGGC) > d(TA)	Higher propensity for Z-DNA formation (descending order for listed motifs)	[42,47,48]
	Base modifications	High GC content, mismatched base pairs (e.g., G·T, G·Br5U, 5FU·G)	Favors Z-DNA conformation via structural destabilization	[49,50]
		Cytosine modifications (e.g., 5mC, C5 halogenation/bromination)	Reduces activation energy, destabilizes B-DNA, promotes B-Z transition	[3,52,53]
		Guanine modifications (e.g., C8 methylation/bromination, 8-oxodG, Ac-4 HAQO)	Distorts Watson-Crick pairing, stabilizes Z-DNA	[53,54]
	ZBPs	ZBPs with Zα domains (e.g., ADAR1, ZBP1, E3L, PKZ, ORF112, RBP7910, HOP2, DsvD, D2PEW5, feoC, pefI, RPA2, CDC53, CUL1 ANC2, APC2, Rpc34, PBP2, Reut_B4095)	Stabilize Z-DNA via specific binding to left-handed helix	[64,65]
	Processive enzymes	Polymerases and helicases	Inducing DNA under-rotation through mechanical stress and triggering B-Z conformational transition	[40]
External factors	Cations	Multivalent ions (Ru4+, Rh3+, Co3+ {[Co(NH3)6]Cl3}, Co2+ {CoCl2}, Mn2+, Mg2+, Ba2+, Ca2+); high NaCl (4 M) or LiCl (3.8 M)	Shielding of DNA backbone, reduces electrostatic repulsion	[3,41,56]
	Polyamines/osmolytes	Spermidine, spermine	Neutralize charges, alter water potential, stabilize Z-DNA	[3]
	Osmolytes	Alcohols (ethanol), polyols (sucrose)	Altering the chemical potentials of water and osmotic stress	[58,59]
	Mechanical stress	Negative supercoiling/torsional stress	Promotes the B-to-Z DNA transition by relieving torsional stress through Z-DNA’s left-handed structure	[21,60,61]
		DNA bending	Changing the spiral trajectory to reduce energy barriers	[62,63]
	Crowding agents	PEG200, ethylene glycol	Dehydration and crowding effects promote B-Z transition	[3,41,56]
	Compounds	Actinomycin D	By synergistically binding Z-DNA and inducing its local reversal to the right-handed B-type conformation	[46]
		Actinomycin	Dependent on positive charge to bind Z-DNA under low-salt/metal ion conditions and partially reverses its structure	[46]

Genomic and epigenetic features of Z-DNA

Computational predictions and experimental profiling of Z-DNA have been conducted across various species, including humans, mice, birds, and several plant species such as Arabidopsis, rice, wheat, and radish, as well as in the bacteriophage phi X174 and the virus SV40 [29,30,32,34,35,47,69–71]. Alternating purine-pyrimidine sequences, such as d(CG)_n, d(TG)_n, and d(TA)_n repeats, can adopt Z-DNA conformations. However, a subset of ZFRs capable of forming Z-DNA contains specific transcription factor (TF) binding sites that promote the recruitment of their corresponding TFs. It has been documented that d(CG)_n repeats exhibit the highest propensity for adopting the left-handed helical structure [40,48]. Additionally, d(TG)_n and d(TA)_n can transit into Z-DNA, although with lower efficiency [1,72]. ZIP-seq has revealed that Z-DNA contains a higher GC content, which aligns with the findings from HUNT predictions of Z-DNA in rice and Arabidopsis [70]. Z-DNA tends to have a higher frequency of CG/GC dinucleotides, while showing a decreased frequency of CC/GG and AA/TT dinucleotides. In terms of genomic distribution, Z-DNA is predominantly enriched in exons, 5′UTRs, and intergenic regions, while it is less frequently found in promoters, introns, and 3′UTRs in the rice genome. This pattern is similar to the ChIP findings reported in humans [30], but distinct from ChAP results that exhibit enrichment of Z-DNA in centromeres and telomeres in humans [34]. Furthermore, read counts of in vitro Z-DNA identified using ZIP-seq or computational predictions are primarily enriched around the transcriptional start site (TSS ) in rice, Arabidopsis, and birds [29,70]. By contrast, read counts of in vivo Z-DNA are enriched from TSS to transcriptional termination site (TTS), reflecting a distinct distribution pattern of in vitro and in vivo Z-DNA reads across genic regions [36]. In vivo Z-DNA is also predominantly enriched in promoter regions with a high frequency of double-strand breaks (DSBs), colocalizing with the Nfe2·Maf-binding motif in medullary thymic epithelial cells (mTECs) [35].

For epigenetic features, in vivo Z-DNA in humans colocalizes with RNA polymerase II and active marks such as H3K4me3 and H3K9ac [30]. In rice, Z-DNA exhibits colocalization with DNA modifications, including CHH methylation, DNA-6mA, and oxidized DNA nucleotides such as 8-oxo-7,8-dihydro-2′-deoxyguanosine (8-oxodG), rather than with 4-acetylcytosine (DNA-4acC). Additionally, certain non-B-DNA structures, including G4s and i-motifs, are associated with rice ZFSs. Furthermore, rice ZFSs are enriched with active epigenetic marks, which include H3K4me3, H3K36me3, H3K9/27ac, and H4K12ac, as opposed to repressive marks like H3K27me3 [36]. Thus, Z-DNA exhibits distinct DNA sequence features in conjunction with specific epigenetic modifications.

Potential functions and biological implications of Z-DNA

It has been documented that Z-DNA plays a significant role in various cellular functions across multiple organisms. These functions include the regulation of gene transcription, the recruitment of ZBPs and other TFs, the maintenance of genome stability, DNA recombination and repair, chromatin remodeling, the restriction of nucleosome formation, biofilm formation, immune response and evasion, evolution, and the development and/or progression of human disease (Figure 2) [38,40,41,73,74]. For example, research has shown that Z-DNA can lead to gene deletions, thereby contributing to the evolution of pelvic fin reduction in stickleback fish [75]. Z-DNA has been implicated in transcriptional regulation, particularly in transcription initiation, through multiple lines of evidence. It has been firmly established that ZFSs are enriched in regulatory regions, including promoters and enhancers [76,77]. Furthermore, genome-wide analyses indicate that promoters containing ZFSs are associated with increased occupancy of RNA polymerase II and an enrichment of active histone marks, including H3K4me3 and H3K9ac [30], supporting its functional role in active transcription. In addition, it has been suggested that Z-DNA may act as a cis-regulatory element (CRE), such as an enhancer or a poised promoter. This could occur by facilitating the recruitment of chromatin remodelers, such as the BAF complex, thereby modulating chromatin accessibility and transcriptional initiation [76,78–80]. However, these mechanistic roles have yet to be fully elucidated and necessitate further experimental validation.

Figure 2. Potential biological relevance of Z-DNA in humans and plants.

They include gene transcription, genome stability, chromatin remodeling in humans, human diseases; and stress responses, potentials in the regulation of translation and epievolution, acting as cis-regulatory elements in plants (Arabidopsis and Oryza sativa). Black arrow on DNA indicates the transcription start site (TSS) for each gene. On and off on the TSS indicates activation and inhibition of the corresponding gene transcription.

Global profiling studies further suggest that Z-DNA may exert genomic region-dependent effects on gene expression. For instance, in rice the formation of Z-DNA at the TSS tends to correlate with transcriptional activation, whereas its presence in gene bodies may be associated with repression [36]. Therefore, while the enrichment of Z-DNA at active regulatory regions is well supported by experimental evidence, its precise mechanistic roles, particularly in transcription initiation and elongation, are only partially understood and are subject to ongoing investigation [81].

Beyond transcriptional regulation, Z-DNA plays a role in homologous recombination and DNA repair, partly by alleviating supercoiling and thereby facilitating DNA repair mechanisms [82]. In bacterial systems, CG-repeat sequences that promote Z-DNA formation can induce small deletions, likely due to misalignment during DNA replication [38]. Z-DNA has been conclusively linked to genomic instability, as supported by extensive experimental evidence across multiple organisms [36,83–85]. The formation of Z-DNA within susceptible DNA sequences introduces local structural distortions and helical transitions, making the DNA more susceptible to strand breaks and erroneous repair mechanisms. Notably, ZFSs can serve as hotspots for DNA DSBs, particularly during periods of transcriptional activity or replication stress. These DSBs may result from topoisomerase trapping, replication fork stalling, or enzymatic cleavage near Z-DNA conformations [79]. In addition, the high torsional strain associated with Z-DNA formation can lead to the accumulation of supercoiling, which further destabilizes the DNA backbone and promotes recombination or deletion events. In mammalian systems, Z-DNA is associated with chromosomal rearrangements and translocation events, which contribute to the onset and progression of oncogenic transformations. Moreover, genome-wide profiling has demonstrated that ZFSs colocalize with regions of recurrent structural variation and mutational clusters, further reinforcing their contribution to genome instability. These findings highlight the necessity of considering Z-DNA as a significant factor in genetic fragility, especially in disease contexts such as cancer.

Z-DNA also plays a significant role in chromatin remodeling and inhibits nucleosome formation. The left-handed helix formed is rigid and structurally incompatible with nucleosome wrapping. Its zigzag backbone and alternating syn-anti base pairs confer a persistence length of approximately 200 nm (compared with ~50 nm for B-DNA) and an estimated free energy of approximately 105 kcal/mol versus approximately 13 kcal/mol. These values reflect the relative energetic stability of Z-DNA and B-DNA conformations in the context of nucleosome wrapping and DNA deformation [65]. Importantly, even relatively short runs of ZFSs (typically 6–12 base pairs), such as (CG)_n repeats, have been shown to influence nucleosome positioning and phasing in vitro [81]. Such short runs may not completely exclude nucleosomes but can act as positioning signals that define nucleosome boundaries or promote partial unwrapping. By contrast, longer uninterrupted ZFSs tend to resist histone wrapping altogether, forming rigid segments that serve as nucleosome-excluding elements. This dual behavior suggests a length-dependent role of Z-DNA in chromatin organization [30,32].

Furthermore, Z-DNA can interact with chromatin remodelers such as SWI/SNF (e.g., SMARCA4), which facilitate nucleosome repositioning and may stabilize Z-DNA conformations in vivo [86–88]. The SWI/SNF complex can remodel nucleosomes to facilitate the formation of Z-DNA [89]. Studies have shown that DNA fragments containing a (CG/GC)_n sequence exhibit a conformation-dependent effect on nucleosome positioning in vitro, favoring nucleosome formation in the B-DNA conformation while disfavoring it in the Z-DNA conformation [88,90]. However, it has been reported that Z-DNA-containing nucleosomes in vitro exhibit enhanced nuclease accessibility compared with those formed from classic B-DNA [89], suggesting structural variations between the two types of nucleosomes. Further investigations are essential to reveal the causal relationship between Z-DNA and nucleosome assembly in vivo.

Importantly, a plethora of evidence shows that Z-DNA is involved in immune responses, cell death or PANoptosis, disease-related signal transduction, and various human diseases. These diseases include neurological disorders, heat stroke, chronic inflammatory conditions, autoimmune disorders such as systemic lupus erythematosus (SLE), acute liver or lung injuries, viral or bacterial infections, and cancer [73,91–93]. This involvement is often mediated by ZBPs, such as ZBP1 [94–108]. For instance, the upregulation of ZBPs is directly associated with tumor growth in various cancer types, such as esophageal cancer (ESCA) and head and neck squamous cell carcinoma (HNSCC) [109,110]. ZBP1-mediated liquid–liquid phase separation functions in the activation of innate immune signaling pathways in response to hepatosplenic fungal infection (HSF) and influenza A virus (IAV) infections [105]. Although (CG)_n repeats are classical Z-DNA-forming motifs, recent structural studies have demonstrated that even non-CG-repeat sequences, such as d(CACGTG), d(CGTACG), and d(CGGCCG), can adopt the Z-DNA conformation when bound to Zα domains of proteins like ADAR1. These studies revealed that Zα recognizes Z-DNA through common conformational features rather than strict sequence specificity [105]. Nevertheless, repeats implicated in neurological disorders such as (CAG)_n in Huntington’s disease and (CGG)_n in fragile X syndrome are not canonical ZFSs and their direct involvement with Z-DNA formation remains to be established [111,112]. The formation of Z-DNA modulates the expression of ADAM-12, a gene involved in cancer metastasis. The loss of Z-DNA-mediated suppression can result in aberrant gene expression in cancer cells, underscoring its potential role in oncogenesis [113]. Immune responses can be mediated by Z-DNA via the binding of ZBP1, a sensor that activates immune signaling and inflammatory pathways, as well as programmed cell death [100,105,114,115]. The functions of ADAR1 in regulating cell death and tumorigenesis are facilitated through its interaction with ZBP1 [116]. Functions of DLM1, a protein associated with the regulation of immune responses, and E3L, a viral protein recognized for its role in immune evasion, are partially mediated by their ability to recognize and stabilize Z-DNA structures [117,118]. The binding of ZBP1 to Z-DNA can trigger RIPK3-dependent necroptosis and inflammation [119]. Additionally, UVB-induced inflammation is closely linked to the UV-induced upregulation of ZBP1, which stabilizes mitochondrial Z-DNA [120].

Additionally, Z-DNA plays a crucial role in various biological processes in plants, including gene regulation, stress responses, and genome evolution. Research has shown that genes with ZFSs in their promoters exhibit functions related to environmental adaptation, such as responses to drought, salinity, and pathogen infection [121]. Moreover, Z-DNA may regulate genes associated with TFs, translation repressors, and metabolic pathways in Arabidopsis, as well as genes involved in vesicle transport, nucleosome organization, stem cell maintenance, and reproductive development in rice [70].

Concluding remarks and future perspectives

A growing body of evidence demonstrates that Z-DNA can function as a regulator or mediator of the genomic code, thereby playing a role in various biological processes in humans and other species. However, compared with the significant progress made in understanding other types of alternative nucleic acid structures (e.g., R-loops, G4s), global profiling and comprehensive characterization of Z-DNA biology, along with underlying mechanisms, remain in their infancy in eukaryotes such as mammals. Furthermore, Z-DNA is even less characterized in plants, where high-throughput experimental studies on Z-DNA are still limited compared with those in animal systems. It is essential to explore the conceptual and theoretical foundations underlying Z-DNA biology. Key areas of focus include: the development of robust and sensitive high-throughput methodologies for global profiling of Z-DNA across the genome under both physiological and pathological conditions; the combined effects of endogenous and exogenous factors on Z-DNA formation in vivo and its dynamics under pathological conditions; and the detailed genetic and epigenetic mechanisms that underlie the functions and biological implications of Z-DNA. Such knowledge not only addresses a significant gap in our current understanding of Z-DNA but also establishes a foundation for further exploration of its biological functions (see Outstanding questions). Such advances will enhance our comprehension of Z-DNA biology, and facilitate the transition of Z-DNA studies from the laboratory to clinical applications, including the development of novel biomarkers and targets for disease screening, diagnosis, treatment, and drug development.

Highlights.

Combined application of complementary methodologies and molecular tools provides a more effective approach to capture both stable and dynamic Z-DNA conformations within the genome.

Utilization of orthogonal high-throughput methodologies offers a comprehensive overview of Z-DNA and/or Z-DNA-forming sequences throughout the genome, thereby advancing our understanding of Z-DNA biology.

Combined effects of intrinsic sequence context, base modifications, and both internal and external environmental factors determine the formation of Z-DNA, both in vitro and in vivo.

Z-DNA plays important roles in multiple cellular processes, including gene transcription, genome stability, DNA recombination and repair, chromatin remodeling, nucleosome and biofilm formation, immune response, and human disease development and/or progression.

Certain Z-DNA loci hold potential for clinical applications, including disease screening, diagnosis, treatment, and drug development.

Outstanding questions.

What strategies can be employed to develop sensitive and robust methodologies for the enhanced visualization or detection of transient Z-DNA conformations under both favorable and unfavorable physiological conditions? Additionally, what approaches can be utilized to identify hybrid Z-flipons that emerge from interactions between DNA and RNA molecules?

What are the combined effects of factors such as sequence composition, base modifications, and both intrinsic and extrinsic environmental factors on the formation of Z-DNA formation in vitro and in vivo?

What are the fundamental mechanisms by which Z-DNA influences biological processes (e.g., transcription, genetic instability) during normal developmental processes or in response to environmental stresses? Additionally, how can we ascertain that the functions of Z-DNA are based on its structural characteristics rather than specific DNA sequences?

Is it possible for Z-DNA to function as a CRE or serve as a platform for the recruitment of trans-factors, thereby influencing the natural or artificial selection of key genes that regulate significant phenotypic traits?

What are the potential biomedical applications of Z-DNA, particularly in disease screening, diagnosis, treatment, and drug development?

Glossary

ChIP-seq

chromatin immunoprecipitation and sequencing; an immunoprecipitation-based method for the global identification of Z-DNA using Z22 followed by high-throughput sequencing.

Cis-regulatory element (CRE)

a DNA sequence that harbors conserved motifs and is essential for the regulation of gene transcription through interaction with trans-acting factors.

Cleavage under targets and release using nuclease (CUT&RUN)

an in situ technique utilized for global profiling of Z-DNA using Z22 in combination with a protein A-MNase fusion protein to release DNA fragments of interest for library preparation and sequencing.

Cleavage under targets and tagmentation (CUT&TAG)

an enzyme-tethering approach for global identification of Z-DNA, utilizing Z22 in combination with a protein A-Tn5 transposase fusion protein to release DNA fragments of interest for library preparation and sequencing.

DeepZ

a deep learning approach for the prediction of Z-flipons, considering multiple parameters such as sequence context, structural features, and experimental data.

G-quadruplexes (G4s)

alternative secondary nucleic acid structures formed by guanine tetrads through Hoogsteen hydrogen bonds.

Negative supercoiling

a type o f D NA topology where the DNA double helix is under-wo und.

Non-B-DNA structures

alternative secondary DNA conformations that distinctly differ from the canonical right-handed B-DNA helix. These include structures such as R-loops, G4s, i-motifs, Z-DNA, hairpins, slipped structures, cruciform, and triplexes (i.e., H-DNA).

R-loop

a triple-stranded nucleic acid structure also known as an RNA-DNA hybrid, comprising an RNA:DNA hybrid plus a ssDNA.

Z22

an antibody specifically designed to bind Z-DNA structures.

Z-DNA

a left-handed double-helical form of DNA featured with a zigzag (Z) backbone conformation. Unlike the more common right-handed B-DNA, Z-DNA typically forms in sequences with alternating purines and pyrimidines and is considered less stable due to its higher-energy state.

Z-DNABERT

a machine learning approach trained on the chemical footprinting data of Z-DNA, utilizing a transformer algorithm that complements DeepZ.

Z-DNA binding proteins (ZBPs)

a class of proteins containing the highly conserved Zα domain enabling specific interactions with the left-handed Z-DNA helix.

Z-DNA-forming sequences (ZFSs)

genomic DNA sequences comprising alternating pyrimidines and purines that have the potential to adopt Z-DNA.

ZIP-seq

a Z22-based immunoprecipitation method coupled with high-throughput sequencing utilized for the global identification of ZDNA across the genome.

Z-Seeker

a novel algorithm applied for the accurate global prediction of ZFSs incorporating experimental data from biophysical, biochemical, molecular biological, and cell-based studies.