dna-paint
Prepare and acquire DNA-PAINT samples using the full kinetics, buffer, and bench protocol.
- License 许可
- BSD-3-Clause
- 最近更新
- 2026-10-20
- Stars
- ★ 10
- Forks
- 1
01 前言 Front Matter
- name
- dna-paint
- description
- Prepare and acquire DNA-PAINT samples using the full kinetics, buffer, and bench protocol.
03 SKILL.md 正文
DNA-PAINT Experiment Protocol — DNA Origami (in vitro), Singleplex
Returned by load_skill(name="dna-paint"). load_skill(name="smlm") is the
general SMLM reference and covers DNA-PAINT at planning depth; this document is
the depth behind it — kinetics, buffers, strand design, and the full bench
procedure. Where the two differ, this document is the accurate one.
Parameter values below are the published protocol's, measured on the reference instrument in [Schnitzbauer2017]. They are starting points to confirm against the user's own rig, sample and imager stock — never settings to apply unasked.
Sources
Compiled from two reference papers, cited throughout as:
- [Schnitzbauer2017] — Schnitzbauer J, Strauss MT, Schlichthaerle T, Schueder F, Jungmann R.
"Super-resolution microscopy with DNA-PAINT." Nat Protoc. 2017;12(6):1198–1228.
doi:10.1038/nprot.2017.024 — the detailed, step-by-step
DNA-PAINT/Picasso wet-lab protocol;
Step Nbelow refers to this paper. This is also the primary reference setup this document is modeled on — see §0. - [Lelek2021] — Lelek M, Gyparaki MT, Beliu G, Schueder F, Griffié J, Manley S, Jungmann R, Sauer M, Lakadamyali M, Zimmer C. "Single-molecule localization microscopy." Nat Rev Methods Primers. 2021;1:39. doi:10.1038/s43586-021-00038-x — broader SMLM primer covering kinetics and general experimental considerations.
0. This experiment's scope
Decided for this first run:
- Sample: DNA origami nanostructures only (in vitro), no cellular/in situ labeling. This skips immunofixation, antibody-DNA conjugation and cell culture entirely [Schnitzbauer2017, Steps 1–18 apply; Steps 19–33 (in situ) do not].
- Multiplexing: singleplex — one imager species, one acquisition, no Exchange-PAINT fluid exchange and no spectral multiplexing. Simpler chamber, simpler acquisition.
- Oxygen scavenging: the PPT system (PCA/PCD/Trolox) is included from the start — mixed into the imager solution at least 1 h before imaging to improve dye photostability and photon yield [Schnitzbauer2017]. Full ultra-resolution settings (Box 2: 350 ms exposure, 80,000 frames, ~4.5 kW/cm², drift-marker origami) are not adopted here — just the buffer.
- Microscope: any inverted stand with TIRF (or HILO) illumination, a hardware focus
lock, a low-noise camera (EMCCD or sCMOS) and an excitation line matching the imager dye.
The reference setup in [Schnitzbauer2017, EQUIPMENT] is a Nikon Ti-Eclipse + Perfect Focus
System with an Andor iXon Ultra DU-897 EMCCD; its camera/focus recommendations transfer to
a comparably equipped rig — see §5. Ask the user what they have
rather than assuming this one, and call
get_system_stateto see what is configured. - Reagents: all reagents for the experiment are already on hand, so the antibody-labeling and cell-culture reagent lists from the source paper are omitted below.
Assumption to confirm: for a 561 nm excitation line, this protocol defaults to the
paper's 561 nm-compatible reference handle — imager CTAGATGTAT labeled with Cy3B,
docking strand TTATACATCTA (the "P1" handle) [Schnitzbauer2017, MATERIALS]. If the user's origami
structures/imager stock use a different sequence or dye, swap it in — nothing else in this
protocol depends on the specific sequence, only on duplex length and dye/laser match
(see §4).
Assumption to confirm: this protocol includes the full design→fold→purify workflow (§6.A–B) for a start from unfolded staples. With folded, purified origami already on hand, skip straight to §6.C.
1. Background & purpose
In DNA-PAINT, transient hybridization of short, dye-labeled "imager" strands to their complementary "docking" strands (here, staple extensions on a DNA origami structure) creates the stochastic fluorescence "blinking" needed for single-molecule localization microscopy (SMLM). This decouples blinking from dye photophysics, so imaging works with virtually any single-molecule-compatible dye and requires no photoswitching buffer [Schnitzbauer2017, Introduction; Lelek2021, "Temporarily binding" fluorophore class].
Why DNA origami as the first sample: it's a known-geometry, ground-truth test structure — docking sites sit at defined, designed positions, so imager concentration, exposure time and laser power can be validated against a known answer before ever touching a biological sample [Schnitzbauer2017, "Design and preparation of DNA origami structures"]. It's also the field's standard calibration/reference structure for exactly this reason [Schnitzbauer2017, Fig. 1b].
Key properties motivating DNA-PAINT generally:
- No photobleaching as a hard limit — the imager-strand reservoir in solution is effectively infinite, so bleaching of an individual dye has only minimal effect on the overall experiment [Schnitzbauer2017].
- Programmable blinking kinetics — bright time τ_b = 1/k_off is set by duplex stability (length, GC content, salt/temperature); dark time τ_d = 1/(c_i·k_on) is set by imager concentration, with k_on ≈ 10⁶ M⁻¹s⁻¹ typical for DNA-PAINT duplexes [Lelek2021]. A 1-bp increase in duplex length increases τ_b by roughly an order of magnitude [Lelek2021].
- ~1 nm localization precision / sub-5 nm resolution has been demonstrated with this approach under optimized ("ultra-resolution") conditions [Schnitzbauer2017].
Limitations to keep in mind [Schnitzbauer2017, "Advantages and limitations"]: imager strands are non-fluorogenic (always present, always slightly fluorescent in solution), so DNA-PAINT requires optical sectioning (TIRF/HILO) to suppress background from unbound imager.
2. Materials & reagents
Condensed from [Schnitzbauer2017, MATERIALS/REAGENTS/EQUIPMENT], limited to what's needed for a singleplex in vitro DNA origami experiment (antibody-conjugation and cell-culture reagents from the source paper are not needed here and are omitted).
DNA origami folding & purification
- Staple strands, modified (docking-strand extensions) and unmodified core staples.
- M13mp18 ssDNA scaffold (p7249).
- Biotinylated staples (for surface attachment).
- Tris pH 8.0 (1 M), EDTA pH 8.0 (0.5 M), magnesium (1 M), water — for 10× folding buffer (see §3).
- Agarose, 50× TAE buffer, SYBR Safe DNA gel stain, DNA gel loading dye, DNA ladder — for gel purification.
Immobilization & imaging
- BSA-biotin (10 mg/ml stock), streptavidin (10 mg/ml stock).
- Imager strand, dye-labeled (see §4).
- 8 mm-channel custom flow chamber: microscopy slide, high-precision cover glass (No. 1.5H), double-sided adhesive tape, epoxy glue.
- Isopropanol (cleaning slides/coverslips).
- Tween 20 (buffer component).
Oxygen-scavenging system (PPT)
- Protocatechuic acid (PCA), protocatechuate 3,4-dioxygenase (PCD), Trolox, NaOH, methanol — for the 40×/100×/100× stocks mixed 1:1:1 into the imaging buffer (see §3) [Schnitzbauer2017]. ⚠ PCA and Trolox cause skin/respiratory/eye irritation — avoid breathing dust/fumes [Schnitzbauer2017, REAGENTS cautions].
Equipment
- Inverted microscope with TIRF (or HILO) illumination and a hardware focus lock.
- Low-noise camera — EMCCD or back-illuminated sCMOS.
- Excitation laser matching the imager dye (561 nm for Cy3B). A laser power meter + microscopy-slide thermal power sensor to calibrate actual power density at the sample plane, since rated laser output is not the relevant number (see §5) [Schnitzbauer2017, EQUIPMENT SETUP].
- Micro-Manager acquisition software [Schnitzbauer2017] — driven through Microclaw/pycro-manager here.
- Picasso (jungmannlab.org / github.com/jungmannlab/picasso) for origami design, localization, rendering and drift correction [Schnitzbauer2017].
- Thermocycler (for annealing), gel chamber + power supply + blue-light transilluminator + gel imager (for purification/QC).
3. Buffer conditions
Recipes from [Schnitzbauer2017, REAGENT SETUP], limited to the in vitro subset:
| Buffer | Composition | Storage |
|---|---|---|
| Buffer A | 10 mM Tris-HCl, 100 mM NaCl, pH 8.0 | RT, 6 months |
| Buffer A+ | Buffer A + 0.05% (vol/vol) Tween 20 | RT, 6 months |
| Buffer B | 5 mM Tris-HCl, 10 mM MgCl₂, 1 mM EDTA, pH 8.0 | RT, 6 months |
| Buffer B+ | Buffer B + 0.05% (vol/vol) Tween 20 | RT, 6 months |
| 10× folding buffer | 125 mM MgCl₂, 100 mM Tris, 10 mM EDTA, pH 8.0 | RT, 6 months |
| Gel buffer | 1× TAE | RT, 1 year |
| Gel running buffer | 1× TAE + 12.5 mM MgCl₂ | RT, 1 year |
| BSA-biotin stock | 10 mg/ml BSA-biotin in Buffer A; 20 µl aliquots | −20 °C, 6 months |
| BSA-biotin solution | 1 mg/ml BSA-biotin in Buffer A+ | 4 °C, 3 days (fresh) |
| Streptavidin stock | 10 mg/ml streptavidin in Buffer A; 10 µl aliquots | −20 °C, 6 months |
| Streptavidin solution | 0.5 mg/ml streptavidin in Buffer A+ | 4 °C, 3 days (fresh) |
| 100× Trolox | 100 mg Trolox + 430 µl methanol + 345 µl 1 M NaOH in 3.2 ml H₂O; 20 µl aliquots | −20 °C, 6 months |
| 40× PCA | 154 mg PCA in 10 ml water, pH 9.0 (NaOH); 20 µl aliquots | −20 °C, 6 months |
| 100× PCD | 9.3 mg PCD in 13.3 ml (50% glycerol, 50 mM KCl, 1 mM EDTA, 100 mM Tris-HCl pH 8.0); 20 µl aliquots | −20 °C, 6 months |
| PPT oxygen-scavenging mix | 1:1:1 ratio of 1× PCA : 1× PCD : 1× Trolox, mixed into imager solution ≥1 h before imaging | fresh |
| Imager solution (in vitro, with PPT) | 1× Buffer B+ + 1× PCA + 1× PCD + 1× Trolox + fluorophore-labeled DNA strand, 100 pM–10 nM (start ~5 nM for a 12-site structure) | fresh |
PPT improves fluorophore photostability and therefore photon yield per binding event [Schnitzbauer2017, Box 2] — worth having in from the start rather than only for a dedicated ultra-resolution attempt, since it costs little beyond a 1 h pre-mix wait.
Imager concentration is the most consequential tuning knob: too low under-samples the
structure; too high raises unbound-imager background and causes "cross-talk" localizations
where two nearby sites bind simultaneously and get fitted as a single false spot in between
[Schnitzbauer2017, Fig. 7d]. Picasso's Simulate module can screen concentration/exposure
combinations in silico before committing microscope time [Schnitzbauer2017, Box 3].
4. Imager/docking strand design
- Docking and imager strands are short complementary single-stranded DNA oligos, typically 8–10 nucleotides [Schnitzbauer2017]. Here, the docking strand is a staple extension on the origami; the imager strand carries the dye and diffuses freely until it transiently binds.
- Default handle for this protocol (561 nm / Cy3B match — confirm against the user's
actual stock, see §0):
- Docking (staple extension):
TTATACATCTA - Imager:
CTAGATGTATlabeled with Cy3B, excited at 561 nm [Schnitzbauer2017, MATERIALS; this is the same P1 sequence used for the calibration experiments on the paper's own TIRF setup].
- Docking (staple extension):
- Duplex length sets the bright/ON time (τ_b = 1/k_off): a typical 9-bp duplex (like P1 above) gives τ_b ≈ 500 ms. Each added base pair increases τ_b roughly 10×; each removed base pair decreases it roughly 10× [Lelek2021]. Camera exposure time should be matched to τ_b [Schnitzbauer2017, "Data acquisition"].
- Imager concentration sets the dark/OFF time (τ_d = 1/(c_i·k_on)), k_on ≈ 10⁶ M⁻¹s⁻¹ typical [Lelek2021].
- Estimating required imaging time: at 10 nM imager and k_on ≈ 10⁶ M⁻¹s⁻¹, τ_d = (k_on·c)⁻¹ ≈ 100 s. For a ~98% probability that any given site is visited at least once (P = 1 − e^(−t/τ_d)), total imaging time t ≈ 4·τ_d ≈ 400 s; [Schnitzbauer2017] recommends ~33 min total to get multiple binding events per site (needed for decent image quality, not just "visited once").
- Docking-site layout (hexagonal lattice, 5 nm spacing) is designed in
Picasso: Design, which also auto-generates staple order lists and pipetting schemes [Schnitzbauer2017, Steps 1–5].
5. Imaging parameters
The published values, measured on the paper's reference instrument [Schnitzbauer2017, EQUIPMENT]. They transfer to a comparably equipped rig — TIRF, hardware focus lock, low-noise camera — and each is a starting point to confirm, not a setting to apply unasked.
| Parameter | Recommendation | Source |
|---|---|---|
| Exposure time | ~300 ms (paper's in vitro Buffer B+ example, matched to the P1 9-bp duplex τ_b ≈ 500 ms) | [Schnitzbauer2017, Step 37] |
| Frame interval | 0 ms (back-to-back frames) | [Schnitzbauer2017, Step 47] |
| Frame count | Start at 7,500; typical full datasets 10,000–100,000+ depending on desired sampling | [Schnitzbauer2017, Step 46] |
| Pre-focus laser power density | ~0.25 kW/cm² at the sample plane | [Schnitzbauer2017, Step 41] |
| Imaging laser power density | ~2.5–6 kW/cm² at the sample plane (561 nm/Cy3B reference range: 1–6 kW/cm²); raise until bright time (τ_b) starts to decrease — that's imager bleaching while bound, and is the practical ceiling | [Schnitzbauer2017, Step 43; Table 1 troubleshooting] |
| Power density calibration | Power density at the sample plane is what matters, not rated laser output — measure it directly with a microscopy-slide thermal power sensor before picking a % laser power setting. A watt-class source runs at a small fraction of full power for this, so a % setting carried over from another experiment is not evidence of anything. | [Schnitzbauer2017, EQUIPMENT SETUP, "Power density calibration"] |
| Illumination mode | TIRF where available — best background suppression for surface-immobilized origami; HILO otherwise | [Schnitzbauer2017, Step 44] |
| Focus | Engage the hardware focus lock for prefocus and for the run (set_focus_lock) — the paper explicitly recommends a hardware lock over software refocusing |
[Schnitzbauer2017, Step 42, CRITICAL STEP] |
| Camera settings | EMCCD reference values: Output amplifier: Conventional; ROI: Full Image; Frame Transfer: On; PixelType: 16-bit; ReadMode: Image; shutters: Open; readout mode = lowest-noise frequency whose readout time doesn't exceed the exposure time. On an sCMOS, the equivalent choice is the lowest-noise readout mode that sustains the frame rate | [Schnitzbauer2017, Steps 39, 45] |
Photon conversion (for Picasso: Localize) |
EM Gain, Baseline, Sensitivity, Quantum Efficiency set per the camera's actual specs — set EM Gain = 1 when using conventional (non-EM) amplification | [Schnitzbauer2017, Step 54] |
| Equilibration before acquiring | 5–15 min on the microscope before starting, to let thermal/mechanical drift settle | [Schnitzbauer2017, Table 1 troubleshooting] |
6. Step-by-step procedure
A. Design & fold DNA origami (skip if structures are already folded)
[Schnitzbauer2017, Steps 1–9]:
- In
Picasso: Design, lay out docking sites on the hexagonal lattice canvas (5 nm spacing) and assign the extension sequence (§4) (Steps 1–4). - Export the staple order list (
Get plates) and order staples, plus the biotinylated surface-attachment staples (Step 5). (Pause point: 2–10 working days for synthesis.) - Pool staples per the auto-generated pipetting scheme into stock mixes (Steps 6–7).
- Fold via thermal annealing: 80 °C hold, then ramp 60 °C → 4 °C at ~3 min 12 s/°C over ~57 cycles, hold at 4 °C (Step 9). (Pause point: store folded structures at 4 °C up to 1 week, or −20 °C long-term.)
B. Purify
[Schnitzbauer2017, Steps 10–17]:
- Run a 1.5% agarose gel (gel buffer + MgCl₂ + SYBR Safe), 90 V, 90 min, at 4 °C/on ice (Steps 11–15).
- Image the gel; excise the origami band (a single sharp band, distinct from and slightly shifted relative to bare scaffold) (Step 16–17).
- Recover via Freeze 'N Squeeze spin column, 6 min at 1,000g, 4 °C (Step 17). (Pause point: store purified origami at 4 °C for 1 week, or −20 °C for long-term.)
C. Immobilization on glass
[Schnitzbauer2017, Step 18, option A — sealed custom chamber]:
- Clean a microscopy slide and coverslip with isopropanol.
- Build an ~8 mm-wide flow chamber with two strips of double-sided tape between slide and coverslip (~20–30 µl channel volume).
- BSA-biotin solution (1 mg/ml), 2 min → wash with Buffer A+.
- Streptavidin solution (0.5 mg/ml), 2 min → wash with Buffer A+, then Buffer B+.
- Origami solution (~5 nM for a 12-site structure, diluted in Buffer B+), 2 min → wash with Buffer B+.
- Add imager solution with PPT (§3/§4). ⚠ Mix the PCA/PCD/Trolox into the imager solution at least 1 h before this step — prepare it while folding/purifying, not last-minute [Schnitzbauer2017].
- Seal the chamber with epoxy glue; wait ~15 min for it to dry before mounting. ⚠ Wait until fully dry — uncured epoxy can contaminate the objective [Schnitzbauer2017, Step 18A, CRITICAL STEP].
D. Data acquisition
[Schnitzbauer2017, Steps 34–48], with the microclaw call for each step where there is one:
- Mount the sample; raise the objective until immersion oil contacts the coverslip (Step 34).
- With Micro-Manager running, confirm the configured camera and channel with
get_system_state(Step 36). - Set exposure ≈ 300 ms and the camera parameters from §5
(Steps 37, 39). Exposure travels with the acquisition call —
run_timelapse(exposure_ms=300)— rather than being set separately. - Go live, autostretch contrast — confirm you see background noise only (Step 40).
- Open the excitation shutter at low power (~0.25 kW/cm² at the sample — calibrate against
the power meter, not the laser dial) and focus, then engage the hardware focus lock
(
run_autofocus, thenset_focus_lock(enabled=true)) (Steps 41–42). - Raise power to imaging density (~2.5–6 kW/cm²); you should now see individual blinking, diffraction-limited spots (Step 43).
- Adjust the TIRF incident angle for best SNR (Step 44).
- Set EMCCD readout mode for lowest noise consistent with the exposure time (Step 45).
- Acquire with
run_timelapse(n_frames=7500, interval_s=0, exposure_ms=300)—interval_s=0is back-to-back frames, and the dataset is saved to disk by the acquisition itself (Steps 46–47). Per-frame hardware control instead requires distinct consecutiveint(k * interval_s * 1000.0)deadlines throughout the frame count. Equal truncated millisecond deadlines allow hardware-sequencing with no software between exposures; Microclaw refuses such fixed plans before acquisition. Observation-only bursts remain allowed. - Let the sample equilibrate 5–15 min, then acquire (Step 48).
E. Reconstruction (pointer)
Full detail in [Schnitzbauer2017, Steps 50–60]; summary:
- Export the raw stack for the localization software with
export_dataset_as_tiff. Microclaw does not fit localizations; everything below runs outside it. Picasso: Localize— identify and fit single-molecule spots (box side ≈ 6σ_PSF + 1; set a minimum net-gradient threshold; set the camera's photon-conversion parameters from §5) (Steps 50–55).Picasso: Render— render the super-resolution image (Steps 56–58).- Drift correction — redundant cross-correlation (RCC) first; for higher resolution, follow with fiducial/origami-based correction (Steps 59–60).
- Since the docking-site layout is known by design, compare the reconstructed image against the designed pattern as a sanity check [Schnitzbauer2017, "Anticipated results"].