DNA and RNA Preparation
DNA oligonucleotides were acquired from IDT, then carefully purified and quantified for experimental use.
DNA oligonucleotides were initially acquired from IDT and resuspended to a concentration of 200 μM. They then underwent purification using 15% denaturing PAGE, followed by UV shadowing, crushing, soaking, and ethanol precipitation. Quantification of the purified oligonucleotides was performed precisely by spectrophotometry. Some oligos required additional NHS-ester labeling, which was subsequently purified via high-performance liquid chromatography.
sgRNA In Vitro Transcription
Specific sgRNA templates were prepared through a careful annealing process, followed by an overnight transcription reaction.
DNA templates designated for sgRNA synthesis were prepared by heating to 95 °C for 2 minutes and then slowly cooled to 4 °C to facilitate annealing. The in vitro transcription reaction was performed using this annealed double-stranded DNA template. It included dNTPs, MgCl2, inorganic pyrophosphatase, RNase inhibitor, DTT, RNA polymerase buffer, and RNA polymerase. This mixture was incubated overnight at 37 °C to ensure thorough transcription. The resulting product was subsequently purified by 15% denaturing PAGE and ethanol precipitation.
Protein Expression, Purification, and Fluorescence Labeling
Cas9 and dCas9 proteins were expressed, purified, and then specifically dual-labeled using Cy3 and Cy5 maleimide for experimental tracking.
WT and dCas9 proteins were successfully expressed and purified following established protocols. Dual-labeling of Cas9/dCas9 was meticulously achieved using a precise ratio of 1:20:40 (protein to Cy3 maleimide to Cy5 maleimide) in Cas9 storage buffer. The labeling reaction proceeded for 2 hours at room temperature, followed by an overnight incubation at 4 °C. This was succeeded by purification via a Superdex 200 Increase 10/300 GL size-exclusion column. Labeled fractions were carefully pooled, quantified, aliquoted, and then stored at −80 °C.
DNA Minicircle Assembly
A specialized three-step protocol, adapted to include AT-tracts for enhanced bendability, was employed for the precise assembly of DNA minicircles.
Minicircles were assembled utilizing a three-step protocol, which was specifically adapted to incorporate AT-tracts to enhance their bendability.
- First, ssDNA (126 nucleotides) and splint DNA (29 nucleotides) were heat-annealed and then ligated using T4 DNA ligase to form a single-stranded DNA minicircle.
- Second, the ssDNA minicircle underwent a T4 DNA polymerase fill-in reaction, supplied with T4 DNA ligase, dNTPs, and ATP, to complete the circularization.
- Finally, the reaction was treated with Exo V and T5 Exo to meticulously remove any incomplete minicircles. This was followed by proteinase K treatment and final purification using a Monarch PCR & DNA Clean-up Kit.
Negative Supercoiling of DNA Minicircles
DNA minicircles were negatively supercoiled using E. coli gyrase, a crucial step for achieving the desired topological state.
Minicircles were negatively supercoiled through incubation with E. coli gyrase in the presence of EtBr, with the reaction proceeding at 37 °C for at least 1 hour. Subsequent to the supercoiling, reactions were cleaned via phenol–chloroform extraction and ethanol precipitation when prepared for AFM analysis. Alternatively, the Monarch PCR & DNA Clean-up Kit was used for samples intended for Bulk/Cryo-EM.
Bulk Cleavage Assays and Kinetics
Bulk cleavage assays were conducted to monitor Cas9-mediated DNA cleavage kinetics, with reactions halted at specific timepoints and analyzed by native PAGE.
Cas9 was first complexed with sgRNA at a 1:1 stoichiometric ratio for 10 minutes to form the Ribonucleoprotein (RNP). This RNP was then mixed with mcDNA at a 1:10 ratio and incubated at 37 °C in cleavage buffer for 20 minutes. Timepoints were systematically collected at 0s, 20s, 40s, 1min, 2min, 5min, 10min, and 20min. Reactions were swiftly halted by the addition of Cas9 STEB. Samples were subsequently analyzed on 10% native PAGE. Band intensities for supercoiled (SC) and linear (Lin) DNA were precisely measured using Fiji (ImageJ) software.
Cleavage fraction was calculated as Lin / (Lin + Nick + SC), with kinetics fitted to a single exponential using IGOR Pro.
AFM Sample Preparation and Imaging
DNA minicircles and dCas9–sgRNA complexes were precisely adsorbed onto mica for high-resolution AFM imaging.
DNA minicircles were adsorbed onto freshly cleaved mica discs using a divalent cation protocol involving 3 mM NiCl2 and 20 mM HEPES. Unbound DNA was meticulously removed by washing. The dCas9–sgRNA complex was then incubated with DNA minicircles at a 1:1 ratio and subsequently immobilized onto mica using the same methodology. AFM imaging was performed using a FastScan Dimension XR microscope equipped with FastScan-D AFM probes in PeakForce tapping mode, recording force–distance curves at 8 kHz.
Automated Flattening and Analysis of AFM Images
AFM images underwent a rigorous automated processing workflow, utilizing TopoStats software, to ensure accurate and reliable data for analysis.
AFM images were processed using the specialized TopoStats software. This involved applying median row alignment, planar and quadratic tilt removal, and scar removal to correct image distortions. A background mask was specifically employed to refine the flattening process. Images were translated to center the background average around 0, and a 1.1 px Gaussian filter was applied for effective noise reduction.
Molecular Quantification from AFM Images
Structural information from AFM images, including contour length and central mean curvature, was quantified using Python-based analysis.
Python notebooks were utilized for precise grain detection and analysis, extracting crucial structural information such as minimum width (Feret diameter), contour length, and aspect ratio. Grains were carefully vetted for viability based on size, remasked, and meticulously traced to approximate the DNA backbone. Central mean curvature was measured over a 10 nm segment around the trace midpoint, revealing a significant difference (P=0.033) between distributions via a t-test.
Cryo-EM Grid Preparation and Data Collection
Cryo-EM grids were prepared with Cas9/dCas9 RNP complexed with negatively supercoiled mcDNA, enabling high-resolution data collection.
dCas9 or wtCas9 RNP was meticulously complexed with negatively supercoiled mcDNA at a 1:1 stoichiometric ratio. These complexes were then loaded onto glow-discharged copper Quantifoil R2/2 grids with carbon support, blotted using a FEI Vitrobot, and plunged into liquid ethane for vitrification. Grids were screened on a Thermo Fisher Scientific Talos F200i, and high-resolution data were collected on a Thermo Fisher Scientific Titan Krios electron microscope at 300 kV with a K3 direct electron detector, spanning a defocus range of –2.5 μm to –0.5 μm.
Cryo-EM Data Processing
Collected Cryo-EM datasets were rigorously processed using RELION4, incorporating advanced methods for particle picking, classification, and refinement.
Collected datasets were processed using the RELION4 software suite. Micrographs underwent essential CTF- and motion-correction, with those exhibiting resolutions worse than 7 Å meticulously removed. On-the-fly processing involved Topaz for efficient particle picking and subsequent 2D classification. The best classes were then utilized for reference-based picking, followed by multiple rounds of 2D/3D classifications and a final 3D refinement. For focus refinement, soft masks were thoughtfully created in RELION to precisely remove noise originating from the minicircle DNA.
Cryo-EM Modeling
Existing and AlphaFold3 models were leveraged to complete and refine the Cryo-EM structure of Cas9, enhancing its overall accuracy.
An unpublished model (derived from PDB 6O0Z) and an AlphaFold3 model of apo SpCas9 were used to complete unmodeled regions, specifically residues 512–573, 611–678, and 685–730. Modeling was initially performed in Coot, with initial rigid-body fitting to maps achieved using ChimeraX. Minimization was conducted using Namdinator, followed by further refinement in Coot. These initial models subsequently served as a crucial reference for real-space refinement in Phenix, executed with three macro cycles.
Optical Tweezer Correlated Fluorescence Microscopy Experiments and Analysis
Optical tweezers, coupled with correlated fluorescence microscopy, were employed to investigate dCas9–sgRNA binding events on torsionally constrained λ-DNA.
Using a LUMICKS C-trap, torsionally constrained λ-DNA was captured and negatively supercoiled (σ, −0 to −0.5). Channels were passivated with Pluronics PBS, and protein channels with BSA, to minimize non-specific interactions. Streptavidin-coated polystyrene particles were trapped, and λ-DNA was precisely suspended between them. DNA integrity and topology were thoroughly characterized by force–distance curves. 1 nM dCas9 cy3/cy5 RNP, specifically targeting the λ2 site, was flowed in to capture binding via confocal imaging. Data were extracted using LUMICKS Lakeview, force-extension curves were replotted in IGOR Pro, and kymographs were meticulously analyzed for FRET and Cas9 off-target events using LUMICKS scripts.
FRET Analysis
FRET analysis was performed on isolated molecules, calculating FRET efficiency from background-subtracted and smoothed acceptor and donor intensities.
Molecules exhibiting FRET dynamics were carefully isolated in Fiji. Average intensities were obtained from a nine-pixel vertical window and manually background-subtracted. Background-subtracted intensities were plotted and used for subsequent FRET calculations. Green and red intensities were smoothed using an IGOR Pro box algorithm with a five-pixel smoothing window.
FRET was calculated as IA / (IA + ID), where IA is acceptor intensity and ID is donor intensity.