DNA is held together by three main types of interactions working in concert: covalent phosphodiester bonds that form the sugar-phosphate backbone, hydrogen bonds that pair bases across the two strands, and base-stacking forces that arise from flat, ring-shaped bases piling on top of one another. Most people learn about the hydrogen bonds first, and they are indeed crucial for the genetic code, but research over the past two decades has revealed that stacking interactions actually contribute more to overall double-helix stability than hydrogen bonding does. Understanding how all these forces cooperate, and sometimes compete, explains everything from why DNA melts at a particular temperature to how cells repair broken strands.
Phosphodiester Bonds Form the Backbone
Each strand of DNA is a long chain of nucleotides linked end to end by phosphodiester bonds. These are true covalent bonds, meaning they involve shared electrons between atoms and are far stronger than the other interactions in the molecule. A phosphate group bridges the 3ʹ carbon of one sugar to the 5ʹ carbon of the next, creating a repeating sugar-phosphate rail that runs the full length of the strand. Because nucleic acids are, at their core, polymeric diesters of phosphoric acid, the chemistry of this linkage has been studied intensively for decades.
1Europe PMC. Phosphodiester models for cleavage of nucleic acidsThe phosphodiester backbone gives DNA its structural continuity. Break it and you get a “nick” in the strand; break both strands close together and you get a double-strand break, one of the most dangerous kinds of DNA damage a cell can experience. Unlike hydrogen bonds and stacking forces, which come and go as the helix breathes and flexes, the covalent backbone holds unless an enzyme deliberately cuts it or a chemical agent damages it. That permanence is what makes the backbone the skeleton of the molecule while the other, weaker forces shape its posture.
Hydrogen Bonds Between the Bases
When people picture the double helix, they usually picture the “rungs” of the ladder: pairs of bases reaching across from one strand to grip the other. Adenine pairs with thymine through two hydrogen bonds, and guanine pairs with cytosine through three. Each hydrogen bond is individually weak compared with a covalent bond, but the sheer number of them along a chromosome adds up to serious collective strength. In a human cell, with roughly six billion base pairs, you are looking at something like 15 billion hydrogen bonds just holding the two strands face to face.
A hydrogen bond forms when a hydrogen atom covalently attached to an electronegative atom (like nitrogen or oxygen) is attracted to another electronegative atom nearby. In DNA, the donor and acceptor atoms sit on the edges of the bases that face into the interior of the helix. The specificity of these bonds is what makes the genetic code work: adenine’s hydrogen-bond donors and acceptors line up perfectly with thymine’s, and guanine’s with cytosine’s, so the correct pairing is always thermodynamically favored.
Recent quantum-chemical work has shown that the aromatic rings in the bases affect how strong these hydrogen bonds are, not through the resonance effects scientists once assumed, but because the ring’s heteroatoms withdraw or donate electrons to the atoms directly involved in bonding. Purines (adenine, guanine) have electron-withdrawing rings, while pyrimidines (thymine, cytosine) have electron-donating rings, and this tunes the strength of the Watson-Crick interaction in subtle ways.
2Royal Society of Chemistry. DNA base pairs: the effect of the aromatic ring on the strength of the Watson–Crick hydrogen bondingThe hydrogen bonds also show cooperative behavior: they do not act in isolation but influence one another. Studies of vibrational properties in Watson-Crick pairs have found that the hydrogen bonds in an A-T pair shift together in ways that amplify their combined effect, meaning the bonds within a pair work as a coordinated unit rather than as independent attachments.
3CrossRef. Cooperative vibrational properties of hydrogen bonds in Watson–Crick DNA base pairsBase Stacking Is the Dominant Stabilizing Force
Here is the part that surprises most people: hydrogen bonding between base pairs is not the main reason the double helix stays together. Base stacking is. When the flat aromatic rings of adjacent bases sit on top of each other like a stack of coins, they interact through a combination of van der Waals forces, hydrophobic effects, and electrostatic interactions between the electron clouds of the rings. These stacking interactions run vertically along each strand, perpendicular to the hydrogen bonds running horizontally between strands.
A study that carefully decomposed the energetic contributions to duplex stability found that base stacking is the main stabilizing factor across all temperatures and salt concentrations tested. More striking, A-T base pairing on its own was found to be destabilizing in thermodynamic terms, and G-C pairing contributed almost no net stabilization. In other words, if you somehow kept the bases paired but prevented them from stacking, the helix would fall apart. Stacking also explains why DNA stability depends so much on sequence: different sequences stack with different strengths.
4Europe PMC. Base-stacking and base-pairing contributions into thermal stability of the DNA double helixThe hydrophobic component of stacking is significant too. The bases are relatively water-repelling compared with the sugar-phosphate backbone, so burying them in the interior of the helix, stacked against each other and away from water, is energetically favorable. Research on how environmental effects induce DNA unstacking has confirmed that hydrophobic base stacking is a major contributor to double-helix stability.
5Europe PMC. Hydrophobic catalysis and a potential biological role of DNA unstacking induced by environment effectsThe practical upshot is that DNA’s famous base-pairing rules (A with T, G with C) are essential for genetic fidelity, but the molecule’s physical integrity relies more on stacking. It is a useful distinction: hydrogen bonds encode the information; stacking holds the structure.
Ions and Water Shape Stability From the Outside
DNA does not exist in a vacuum. In a living cell, it is surrounded by water and dissolved ions, and both of these play active roles in holding the molecule together.
The phosphate groups along the backbone each carry a negative charge at physiological pH. Without something to neutralize that charge, the two strands would repel each other electrostatically. Metal cations, especially magnesium and potassium, cluster around the backbone and screen these negative charges. Infrared spectroscopy studies of DNA thin films exposed to magnesium ions show that at low concentrations, magnesium binds in the backbone region and modifies the local charge environment through long-range electrostatic forces. At higher concentrations, where charge compensation by nearby counterions reaches roughly 92 to 94 percent, DNA can undergo a structural transition into a more compact form.
6Europe PMC. Effect of magnesium ions on the structure of DNA thin films: an infrared spectroscopy studyWater is just as important. In the minor groove of B-form DNA (the common form found in cells), water molecules arrange themselves into an organized network called the “spine of hydration.” First identified in crystal structures, this spine has since been confirmed to form under normal solution conditions as well. The water molecules in the spine are not just sitting there; they form an interconnected, cooperative network that interacts with the base pairs lining the groove.
7Royal Society of Chemistry. The role of water in mediating DNA structures with epigenetic modifications, higher-order conformations and drug–DNA interactionsMolecular dynamics simulations have found that this hydration spine favors structures close to the standard Watson-Crick geometry and makes DNA more rigid in certain sequence contexts, particularly in A-tracts (stretches of consecutive adenines). That rigidity has biological consequences: A-tracts resist wrapping around histone proteins to form nucleosomes, partly because the ordered water network stiffens the helix.
8PubMed Central. Molecular Dynamics Study of the Role of the Spine of Hydration in DNA A-Tracts in Determining Nucleosome OccupancyHow DNA Comes Apart and Gets Sealed Back Together
Because the non-covalent forces holding the two strands together are individually weak, DNA can be “melted” by heating. Raise the temperature enough and the hydrogen bonds and stacking interactions give way, the two strands separate, and you get single-stranded DNA. The temperature at which half the molecules in a sample have separated is called the melting temperature. It depends on the ratio of G-C to A-T pairs (since G-C pairs have three hydrogen bonds and stronger stacking neighbors), the salt concentration, and the overall sequence composition. Models that predict melting temperatures have to account for base pairing, stacking, and the ionic environment simultaneously, because all three contribute to the energy holding the duplex together.
9Europe PMC. A phenomenological model for predicting melting temperatures of DNA sequencesInterestingly, while the melting temperature depends only weakly on heat capacity changes, the way melting temperature shifts with salt concentration depends strongly on DNA composition. AT-rich sequences are stabilized more by increasing ionic strength than GC-rich ones are, which is counterintuitive at first but makes sense when you consider how electrostatic screening interacts differently with the two types of base pairs.
10Biophysical Journal. Heat capacity effects on the melting of DNA. 1. General aspectsInside cells, the backbone’s covalent bonds get broken and resealed all the time by enzymes. DNA ligases are the molecular welders responsible for sealing nicks. They work in three steps: first, the enzyme grabs an energy molecule (ATP or NAD+) and attaches an adenylate group to itself; second, it transfers that group to the 5ʹ phosphate at the nick; third, the 3ʹ hydroxyl on the other side of the nick attacks the activated phosphate, forming a new phosphodiester bond and releasing AMP.
11Oxford Academic. Structural Biology DNA binding with a minimal scaffold: structure–function analysis of Lig E DNA ligasesThis reaction has to be fast and reliable. Real-time single-molecule measurements of DNA ligase at work have clocked the rate of phosphodiester synthesis at roughly 400 per second, comparable to the speed of the polymerases that copy DNA during cell division. If the ligase falls off the nick prematurely after step two but before sealing, the leftover adenylated nick becomes a potentially toxic lesion. That happens in roughly one out of ten attempts, so the enzyme’s ability to clamp tightly around the DNA is a critical part of its design.
12PubMed Central. Dynamics of phosphodiester synthesis by DNA ligaseMeasuring DNA’s Bonds One Molecule at a Time
Scientists can now pull apart individual DNA molecules and watch the bonds give way in real time. Using optical tweezers, researchers attach one strand of a DNA molecule to a bead held by a laser and pull. As the helix unzips, the force required fluctuates depending on the local sequence. GC-rich regions need more force (three hydrogen bonds plus stronger stacking), while AT-rich stretches open more easily. Modern setups achieve sub-piconewton force resolution and millisecond time resolution, and they can detect sequence features as small as ten base pairs. The force signal shows characteristic “flips” between different values at specific positions along the DNA, directly reflecting the sequence-dependent strength of the bonds being broken.
13Biophysical Journal. Unzipping DNA with optical tweezers: high sequence sensitivity and force flipsThese single-molecule experiments have been invaluable for testing the thermodynamic models that predict duplex stability. When the unzipping is done slowly enough, the force measurements match equilibrium predictions well, confirming that our understanding of the balance between hydrogen bonding, stacking, and ionic effects is on solid ground.
When Bonds Go Wrong: Oxidative Damage
DNA’s bonds are not invulnerable. Reactive oxygen species produced during normal metabolism can chemically modify the bases, and one of the most common products is 8-oxoguanine, an oxidized form of guanine. You might expect this to weaken the hydrogen bonds between the damaged G and its partner C, but single-molecule stretching experiments tell a more nuanced story. The base pairing itself is not significantly affected by the 8-oxoG modification. Instead, it is the stacking interactions that take the hit. The oxidized base stacks poorly with its neighbors, destabilizing the local region of the duplex to a degree comparable to having a mismatched base pair.
14Oxford Academic. Quantifying the stability of oxidatively damaged DNA by single-molecule DNA stretchingThis finding reinforces the earlier point about stacking being the dominant structural force. Even when hydrogen bonding is intact, a disruption to stacking alone is enough to seriously weaken the helix locally. It also explains why cells invest heavily in base-excision repair pathways to find and remove oxidized bases: the structural damage they cause goes beyond just misreading the genetic code.
Non-Canonical Structures and Alternative Pairing
Watson-Crick pairing (A-T, G-C) is the default, but DNA can form other hydrogen-bonding arrangements. Hoogsteen base pairs, for instance, use a different face of the purine ring and are essential for building G-quadruplexes, four-stranded structures that form in guanine-rich sequences. In a G-quadruplex, four guanines sit in a plane, each one hydrogen-bonding to two neighbors through Hoogsteen pairing, and the planes stack on top of one another with a monovalent cation (potassium or sodium) sandwiched between them for stabilization.
15ScienceDirect. Effects of deficient of the Hoogsteen base-pairs on the G-quadruplex stabilization and binding mode of a cationic porphyrinDNA can also form triple-stranded structures (triplexes) and i-motifs. In a triplex, a third strand binds in the major groove of a standard duplex through Hoogsteen or reverse-Hoogsteen hydrogen bonds. For example, at slightly acidic pH, cytosine can pick up an extra proton and bind to the guanine of a G-C pair, forming a C+·G-C triad. The i-motif is built from protonated C-C+ pairs and tends to form in cytosine-rich sequences. These structures are not just curiosities: G-quadruplexes and i-motifs have been found at gene promoters and telomeres, where they appear to regulate gene expression and chromosome maintenance.
16Oxford Academic. A unified computational view of DNA duplex, triplex, quadruplex and their donor–acceptor interactionsSynthetic Base Pairs That Bypass Hydrogen Bonding
One of the more striking findings in recent decades is that hydrogen bonding is not the only way to build a functional base pair. Researchers have developed synthetic nucleotides that pair through hydrophobic and packing forces rather than hydrogen bonds, and these unnatural base pairs can be replicated and transcribed with remarkable efficiency. This demonstrates that the information-storage function of DNA does not inherently require hydrogen-bonded base pairs; other intermolecular forces can do the job.
17Wiley Online Library. The expanded genetic alphabetEven more recently, researchers have begun exploring halogen bonding as a pairing force in artificial base pairs. A halogen bond is similar in directionality to a hydrogen bond but uses a halogen atom (like bromine or iodine) as the bridge instead of hydrogen. Early results suggest that halogen-bonded base pairs can maintain proper duplex geometry while being orthogonal to the natural bases, meaning they do not cross-pair with A, T, G, or C.
18ACS Publications. Investigating Halogen Bonds as Pairing Force in an Artificial DNA Base PairThese synthetic systems are not just academic exercises. An expanded genetic alphabet could allow engineered organisms to store more information per base pair, encode novel amino acids, or build biological materials with properties not found in nature.
How Proteins Recognize DNA Through Its Shape and Charge
The bonds within DNA do not just hold the molecule together; they also create the structural features that proteins use to find specific sequences. Transcription factors and other DNA-binding proteins read the genetic sequence in two overlapping ways. In the major groove, proteins make direct hydrogen bonds with the exposed edges of the base pairs, which differ between A-T and G-C pairs. This is sometimes called “base readout.” In the minor groove, the recognition mechanism is different: proteins sense the groove’s width and the electrostatic potential it creates.
A comprehensive analysis of protein-DNA crystal structures revealed that arginine residues binding to narrow minor grooves is one of the most widely used recognition strategies. Narrow minor grooves have a strongly enhanced negative electrostatic potential, and the positive charge on arginine is drawn into these grooves. The nucleosome core particle, the spool around which DNA wraps to pack into chromosomes, is a prominent example of this shape-readout mechanism.
19Europe PMC. The role of DNA shape in protein-DNA recognitionComputational modeling has confirmed that the correlation between minor-groove width and electrostatic potential can be predicted genome-wide, enabling researchers to forecast where proteins are likely to bind based on DNA shape alone.
20Oxford Academic Nucleic Acids Research. Genome-wide prediction of minor-groove electrostatic potential enables biophysical modeling of protein-DNA bindingSimplified protein models have shown that the volume and formal charges of amino acids play the predominant role in binding, with hydrogen bonding and just a handful of key side chains at atomic resolution being enough to reproduce realistic DNA conformations and sequence recognition.
21Oxford Academic. Decomposing protein-DNA binding and recognition using simplified protein modelsDNA Bonds in Extreme Environments
If stacking, hydrogen bonding, and ionic stabilization are all temperature-sensitive, how do organisms that live near boiling water keep their DNA intact? Thermophilic archaea and bacteria, some tolerating temperatures above 100 °C, have evolved several strategies. Their genomes are characterized by the presence of reverse gyrase, a unique enzyme that introduces positive supercoils into DNA. Positive supercoiling winds the helix more tightly, raising the energy barrier for strand separation and effectively counteracting the thermal energy that would otherwise melt the duplex. Reverse gyrase has been suggested to maintain genome integrity both by limiting DNA melting and by participating in DNA repair. Other proteins, including nucleoid-associated proteins, histones, and structural-maintenance-of-chromosomes family proteins, further organize and compact the DNA, adding layers of physical protection.
22Europe PMC. How Do Thermophiles Organize Their Genomes?Polyamines, small positively charged molecules found at high concentrations in thermophiles, also help by neutralizing backbone charge and promoting tighter packing, much like the magnesium ions described earlier but under far harsher conditions. The overall lesson is that the same bonding principles that hold DNA together in your cells hold DNA together at extreme temperatures; thermophiles just add extra mechanical and chemical reinforcement to shift the balance in their favor.
Engineering With DNA Bonds
The predictability of Watson-Crick pairing and stacking interactions has made DNA a building material for nanotechnology. In DNA origami, a long single-stranded “scaffold” is folded into a desired shape by hundreds of short “staple” strands that bind to specific regions through complementary base pairing. The resulting double-helical domains are held at defined lengths, and the structures are stabilized by the same hydrogen bonding and pi-pi stacking interactions that stabilize natural DNA.
23Advanced Science. DNA Origami and Its Applications in Synthetic BiologyResearchers have built two-dimensional tiles, three-dimensional boxes with lids that open in response to molecular signals, drug-delivery vehicles, and nanoscale rulers using DNA origami. The approach works precisely because the bonds are so well understood: you can calculate in advance which sequences will bind to each other, how long the resulting helical segment will be, and how stiff or flexible the joint between segments will be. In a sense, the entire field of structural DNA nanotechnology is an applied test of our knowledge of the bonds that hold DNA together, and the fact that it works so reliably is one of the strongest validations that the bonding picture described here is fundamentally correct.