Introduction
DNA doesn't function as a simple linear code. Its biological activity depends on three-dimensional shape: how tightly it supercoils, whether it knots itself during replication, and how separate DNA circles become catenated (linked like chain rings). These features control gene expression and genome stability.
The problem? They're hard to see. Gel electrophoresis sorts DNA by size and crossing number, but it can't reliably tell a knot from a catenane when they share similar geometry.
Electron microscopy shows more detail but demands vacuum conditions and often requires coating DNA with RecA protein just to trace its path.
Atomic Force Microscopy (AFM) fills that gap. It images single DNA molecules in near-native, aqueous buffer, capturing height information that reveals exactly which strand crosses over another.
This guide covers how AFM works, how to prepare DNA samples correctly, where researchers are applying it, how it compares to older techniques, and what to look for in an AFM system built for DNA work.
Key Takeaways
- AFM images single DNA molecules directly in aqueous buffer without staining, fixation, or metal coating
- Height data reveals strand crossing order, enabling classification of knots and catenanes
- Nickel vs. magnesium immobilization buffer controls whether DNA appears compact or open on mica
- Deep-learning tracing pipelines now enable high-throughput quantification of DNA topology across hundreds of molecules
- Reproducible, publication-quality DNA data depends on a BioAFM setup with stable environmental control
Understanding Atomic Force Microscopy: How It Images DNA at the Nanoscale
How AFM Works
An AFM cantilever tip, sharpened to a radius of just a few nanometers, physically rasters across the sample surface. A laser beam reflects off the back of the cantilever and hits a photodetector, tracking every tiny deflection as the tip rises and falls over surface features. That deflection signal builds a topographic height map with sub-nanometer precision.
For DNA, gentle imaging modes matter enormously. Rather than dragging the tip in continuous contact, intermittent-contact (tapping) modes oscillate the cantilever so it touches the surface only briefly on each cycle. This reduces the lateral, dragging forces that would otherwise displace or shred a weakly adsorbed DNA strand.
Molecular Imaging's PicoPlus platforms use MAC Mode (Magnetic AC Mode) for this purpose. Acoustic drive vibrates the whole cantilever holder and can introduce mechanical noise plus spurious resonance peaks. MAC Mode instead drives a magnetically coated cantilever directly with an alternating magnetic field, yielding a cleaner single-peak resonance and finer force control—exactly what fragile DNA strands need.
Why this matters: Bruker's own PeakForce Tapping documentation describes DNA compressing by roughly 20% at applied forces around 70 pN. Keeping imaging force low isn't a nice-to-have. It's the difference between an intact double helix and a smeared artifact.
Why AFM Is Ideal for DNA Structural Research
Three features set AFM apart for DNA work:
- Aqueous imaging. AFM works in liquid buffer near physiological conditions. EM needs vacuum and often RecA coating, which a 2015 AFM review in PMC notes can distort soft specimens.
- Crossing-order determination. Height data shows which duplex passes over another; the over-passing strand reads wider, giving a direct geometric read on knots and catenanes.
- Multi-scale resolution. High-resolution AFM in buffer has resolved the right-handed double helix at 3.4 ± 0.4 nm periodicity and, in the same scan, contour and persistence length (~53 nm on mica).
No other single-molecule technique combines native conditions with this level of structural detail.
Preparing DNA Samples for High-Resolution AFM Imaging
Substrate and Surface Immobilization
Freshly cleaved mica is the standard substrate for DNA-AFM. It cleaves into atomically flat sheets, giving DNA molecules a smooth stage to lie on rather than getting lost in surface texture.
DNA carries a negative charge, so it won't stick to mica (which is also negatively charged) on its own. Divalent cations act as a molecular glue, bridging the two negative surfaces together. The choice of cation changes the outcome dramatically:
| Buffer | Reported Self-Crossings | Conformation |
|---|---|---|
| 25 mM MgCl₂ (10 mM Tris) | 1.7 ± 1.4 | Open, easier to trace |
| 3 mM NiCl₂ (20 mM HEPES) | 4.4 ± 2.3 | Compact, tighter binding |
These figures come from a 2025 study on quantifying DNA structural complexity, which found magnesium immobilization reduced trivial self-crossings and made tracing knotted or catenated molecules considerably easier.
Nickel binds mica more strongly, which compacts the DNA and helps molecules stay put during longer scans. The tradeoff is more crossings, which complicate topological classification.
Molecular Imaging's web store carries freshly prepared mica substrates and gold-on-mica reference substrates for labs that need reliable, atomically flat starting and reference surfaces.
Imaging Parameters and Best Practices
You cannot resolve crossing order reliably at low resolution. The Holmes et al. pipeline achieved contour lengths within 7% of predicted values only at 0.5 to 1 nanometer per pixel. Drop below that resolution and misclassification risk climbs fast.
Environmental stability matters just as much as pixel density:
- Temperature drift during deposition changes ion mobility and can shift DNA conformation mid-experiment
- Humidity fluctuations affect buffer evaporation on an open liquid cell
- Vibration and acoustic noise introduce artifacts that mimic or mask real structural features
This is where temperature-controlled stages (Molecular Imaging's systems span -30°C to 250°C) and environmental isolation chambers earn their keep, holding conditions steady through long imaging sessions.
That same control keeps sample prep simple. Electron microscopy typically needs fixation, staining, or heavy-metal shadowing. AFM skips all of that: deposit DNA in buffer, let it settle, rinse gently, and image. Less handling means less chance to distort the native structure you want to measure.
Key Applications of AFM in DNA Research
Mapping DNA Structure, Supercoiling, and Topology
AFM's crossing-order method turns a flat image into an explicit topological map. Researchers can trace individual plasmids, measure contour length, infer supercoiling and writhe, and classify complex structures with confidence.
The 2025 Holmes et al. study demonstrated this directly:
- Resolved a right-handed 4-node catenane made of a 1,253-bp circle linked to a 398-bp circle
- Correctly identified a five-negative-crossing product as a 5₂* twist knot and a five-positive-crossing product as a 5₁ torus knot
- Used random-forest classification to recover a 56:44 split from a plasmid mixture that was nominally 50:50 nicked versus supercoiled
That last point matters for anyone doing quality control on plasmid preps. Gel electrophoresis would struggle to distinguish some of these species cleanly. AFM's crossing-sign analysis doesn't.
Probing DNA-Protein Interactions
Because AFM images topography directly, it visualizes protein-DNA complexes without needing a protein-specific stain or coating. Transcription factor binding, nucleosome positioning, and recombinase synapses all show up as distinct bumps and bends along the DNA contour.
Time-lapse AFM has captured this in motion:
- Spontaneous nucleosome unwrapping: early studies observed DNA exposing dozens of base pairs from particle ends, occasionally unfolding completely, with no ATP-dependent remodeler involved.
- Asymmetric chromatosome disassembly: high-speed AFM in 2024 showed sequential, asymmetric disassembly where linker histone H1 constrained DNA-arm motion over roughly 50 seconds.
- Stepwise transposase assembly: force-distance-curve AFM mapped how Tn4430 TnpA first binds one transposon end as a compact dimer, then captures the second end to form the paired-end complex.
None of this required RecA decoration or fixation. The protein and DNA sat in buffer, and the cantilever simply traced what was there.
Studying Replication Intermediates and Genome Stability
Replication forks stall for many reasons, and AFM can distinguish how they stall by imaging the resulting intermediates directly.
Holmes et al. compared two replication-blocking systems on different plasmids:
- LacI-blocked plasmid (4,402 bp, 48 LacO sites): theta structures with two forks; unreplicated DNA fell from 313 ± 156 nm at 40 min to 218 ± 153 nm at 80 min, a slowed but incomplete block
- Tus-Ter plasmid (3,574 bp, 24 Ter sites): figure-eight intermediates at 40 min; most molecules fully replicated by 80 min, consistent with a later, milder block
Reversed forks (four-way junctions with one short arm) appeared in 7.7% of all structures across the dataset. That's a level of mechanistic detail gel-based assays simply can't provide.
The same crossing-order approach extends well beyond plasmid biology. AFM combined with simulation has mapped entire kinetoplast DNA networks, revealing minicircle density and valence averaging near 3 links per circle.
Those structural readouts also inform molecular electronics. AFM studies of three-way DNA junctions showed a pyramidal, dynamically shifting structure relevant to charge-transport models through DNA junctions.
AFM in Context: Comparisons and Emerging Advances
Every DNA visualization technique makes trade-offs. Here's how the three main options stack up:
| Technique | Native Conditions | Crossing-Order Detail | Sample Prep |
|---|---|---|---|
| Gel electrophoresis | No | No (relies on average crossing number) | Minimal |
| Electron microscopy | No (vacuum) | Yes, but often requires RecA coating | Extensive |
| AFM | Yes (aqueous buffer) | Yes, directly from height data | Minimal |
Gel electrophoresis is cheap and accessible, but it separates DNA by average crossing number alone. A 4-node knot and a 4-node catenane can co-migrate as if they were the same species. AFM resolves that ambiguity because it measures actual crossing geometry, not just electrophoretic mobility.
Automated tracing is changing the throughput equation
Manually tracing DNA contours from AFM images used to be a bottleneck, one image and one careful eye at a time. Deep-learning pipelines built on tools like TopoStats now segment and classify hundreds to thousands of molecules per experiment, removing much of the observer bias that crept into manual tracing.
That said, these pipelines aren't flawless yet. The 2025 Holmes et al. pipeline achieved 82% agreement with hand-labeled crossing-order classifications. Because whole-knot identification requires every crossing in a molecule to be classified correctly, the paper calculated expected whole-knot correctness at just 56% for 3-node knots and 37% for 5-node knots.
Crossing clustering (where multiple crossings sit close together) and resolution dependence remain the biggest sources of error. Automated tracing is a real advance, but it works best paired with careful imaging parameters, not as a substitute for them.
Choosing the Right AFM System for DNA Research
Not every AFM is built for biomolecular work. For DNA specifically, prioritize:
- Liquid-cell imaging with buffer compatibility across the pH and salt conditions your protocol requires
- Low-noise, sub-nanometer scanning for crossing-order classification, not just general topography
- Gentle-imaging cantilevers, such as liquid-tapping probes rated for biological samples
Molecular Imaging's 5500 ILM PicoPlus BioAFM matches these DNA imaging needs with:
- Sub-angstrom lateral resolution and 0.2 nm vertical resolution
- MAC Mode for magnetically driven, low-force oscillation in fluid
- A patented heating/cooling anti-drift stage for stable extended scans
Pair it with an environmental isolation chamber to control humidity, temperature, and gas composition throughout a session.
To support sample prep and day-to-day DNA runs, the web store also carries:
- Mica and gold-on-mica substrates for atomically flat starting surfaces
- Liquid-tapping cantilevers (silicon and silicon nitride) for extended aqueous imaging
- Low-frequency soft tapping cantilevers for fragile biomolecules
- Calibration samples, including 500 nm grating standards, to verify resolution before a run
If your lab already runs a legacy Agilent or Keysight AFM/STM system, ZoomAFM provides calibration, hardware repair, and preventive maintenance with platform-matched replacement parts so older instruments stay capable of the precision DNA topology work demands.
Frequently Asked Questions
What resolution can AFM achieve when imaging DNA?
High-resolution AFM can resolve the DNA double helix itself, with published measurements reporting helical periodicity of 3.4 ± 0.4 nm. Reliable crossing-order classification typically needs sub-nanometer-per-pixel scan resolution.
Does AFM damage DNA samples during imaging?
Gentle imaging modes like intermittent-contact and magnetically driven tapping (MAC Mode) apply minimal, precisely controlled force compared to contact-mode scanning. This keeps fragile single DNA strands intact throughout imaging.
Can AFM distinguish DNA knots from catenanes?
Yes. AFM height data shows which strand crosses over at each junction, so you can classify knots versus catenanes. Those topologies can look identical when separated by gel electrophoresis alone.
What buffer works best for AFM DNA sample preparation?
It depends on your goal. Magnesium-based buffers produce a more open conformation that's easier to trace for complex or knotted structures. Nickel-based buffers bind DNA more tightly for a more compact, stable layout.
How does AFM compare to electron microscopy for DNA imaging?
AFM works in aqueous buffer with minimal sample prep, without vacuum or staining. Electron microscopy requires vacuum conditions and often RecA protein coating to trace the DNA path, but can add complementary structural detail.
Can AFM image DNA-protein interactions in real time?
Yes. High-speed AFM setups have captured time-lapse imaging of dynamic events like nucleosome unwrapping, chromatosome disassembly, and transposase complex formation directly in liquid buffer.
