Sensing the DNA-mismatch tolerance of catalytically inactive Cas9 via barcoded DNA nanostructures in solid-state nanopores

DNA-nanostructure barcodes

To observe that the CRISPR–dCas9 system is specific enough to detect the single base pair, DNA constructs with different ‘barcodes’ plus an overhang sequence for dCas9 was created. The DNA construct was synthesized from pairing a linearized 7.2 kbp single-stranded (ss) M13mp18 DNA (GuildBiosciences) with 190 complementary oligonucleotides via Watson–Crick base pairing to produce full dsDNA over the period of 1 h in a thermocycler. All oligonucleotides were synthesized by Integrated DNA Technologies and dissolved in IDTE (10 mM Tris–HCl and 0.1 mM ethylenediaminetetraacetic acid, pH 8.0), and the sequences can be found in Supplementary Information. The sample is then filtered using a 100 kDa Amicon filter and measured in a nanodrop spectrophotometer for concentration information. Based on the nanodrop measurement, typical yield is 75–95%. For the nanostructures in Fig. 1, within the 190 oligos are five groups of equally spaced simple dumbbell hairpin motifs to create the spikes that act as a barcode on the DNA nanostructure⁷. Each group consists of 11 DNA dumbbells to create a single spike. The exact sequences with their numbers are shown in Supplementary Table 1 in Supplementary Information following a previous work⁷. The overhang was created by replacing oligos nos. 142 and 143 with a 90 bp oligo made up of 30 bp segments to match the M13 backbone and 50 bp of the specific sequences containing the target sequences we aimed to test. The 50 bp dsDNA overhang is not large enough to generate a current blockade that can be observed. These overhangs are provided in Supplementary Table 2 for the experiments in Fig. 2 and Supplementary Table 3 for the experiments in Fig. 4. For the second nanostructure, shown in Supplementary Fig. 3, oligos no. 44, 45, 81, 82, 118 and 119 were replaced with overhang sequences as found in Supplementary Table 4. The dumbbells were made by replacing oligos no. 23–28, 60–65, 97–102, 134–139, 148–153 and 162–167 with the sequences in Supplementary Table 5. All samples were stored in a storage buffer of 10 mM Tris 0.5 mM MgCl₂.

Design of dCas9 probes and binding

Catalytically deactivated Cas9 D10A/H10A (dCas9) from Streptococcus pyogenes binds with a tracrRNA and a sequence-specific RNA (crRNA), both synthesized by Integrated DNA Technologies and dissolved in IDTE (10 mM Tris–HCl and 0.1 mM ethylenediaminetetraacetic acid, pH 8.0). The target sequences for the crRNA for the probes were designed using online software (http://chopchop.cbu.uib.no/)³⁹ and can be found in Supplementary Information. To assemble the dCas9 RNPs, the tracrRNA (200 nM), crRNA (250 nM) and dCas9 (100 nM) were incubated in a low-salt buffer (25 mM HEPES–NaOH (pH 8.0), 150 mM NaCl and 1 mM MgCl₂) at 25 °C for at least 20 min.

The assembled dCas9 probes were then incubated with the DNA nanostructures for at least 20 min at 25 °C, with the dCas9 probes added in excess of typically 15 dCas9 probes per DNA binding site. The samples containing DNA nanostructures labelled with dCas9 are diluted to 0.1–0.3 nM into a 2 M LiCl, 1× TE buffer solution or 4 M LiCl, 2× TE, depending upon the nanostructure, immediately before the beginning of the measurement in the nanopore system.

Nanopore fabrication and measurement

Nanopores are fabricated from commercially available quartz capillaries (0.2 mm inner diameter/0.5 mm outer diameter Sutter Instruments) using a laser-assisted pipette puller (P-2000, Sutter Instrument) to around 15 nanometres. We produced a polydimethylsiloxane (PDMS) chip with 16 conical nanopores with a communal cis reservoir and individual trans reservoirs. Detailed instructions for production can be found at Bell et al.⁴⁵. Silver/silver-chloride (Ag/AgCl) electrodes are connected to the cis and trans reservoirs in the polydimethylsiloxane chip. The size of each nanopore is estimated before beginning measurements by taking a current–voltage curve in the baseline electrolyte. The central cis reservoir contains the sample and is grounded, while a 500 mV bias voltage is applied to the trans reservoir to drive DNA transport through the nanopore. The measurement is then taken until around 1,000 folded and unfolded events are gathered, with a typical time range of 45 min to 2 h depending upon the concentration used and nanopore. Typically, of these 1,000 events, 300 are unfolded and then analysed. An example of the folded and unfolded events can be seen in Supplementary Fig. 6.

The Axopatch 200B patch-clamp amplifier (Molecular Devices) was used to collect current signals. The set-up is operated in whole-cell mode with the internal filter set to 100 kHz. To reduce noise, an eight-pole analogue low-pass Bessel filter (900CT, Frequency Devices) with a cut-off frequency of 50 kHz is also used. The applied voltage is controlled through an I/O analogue-to-digital converter (DAQ-cards, PCIe-6251, National Instruments), using a program on LabView 2016 to simultaneously record the current signal at a bandwidth of 250 kHz.

Analysis of nanopore data

From the Labview GUI, experimental data are stored as technical data management streaming (TDMS) files. First, a translocation finder Python script (part of the nanopyre package found at is used that identifies the events from the raw traces and stores them in an hdf5 file. After the initial translocation finder analysis, the events from the hdf5 files are read into Python (using the nanopro package and all events with current noise >15 pA are discarded. The parameters to find the spikes are based on manual analysis of the threshold, height, distance and prominence parameters from the Python peakfinder package. This is tested on the first ten and last ten events to ensure the parameters are consistent and will accurately find the peaks. Following this, events are sorted on the basis of the number of spikes. Following the sorting, the events are analysed by eye and events that have folds or knots interfering with the barcode are discarded. Our lab has shown that as few as four events are sufficient for positive detection in the majority of cases, while nine correct events increase the probability of positive detection to more than 90% (ref. ⁴⁶). Percentage of events with dCas9 bound in Figs. 2b and 4 is calculated the following way:

In these equations N_{11111 dCas9} represents the number of events with both the 11111 barcode a dCas9 bound. N_{11111 No dCas9} would represent the number of events with the 11111 barcode and no dCas9 bound.

The calculations of relative concentration have additional parameters because the total number of events without dCas9 bound for both barcoded nanostructures also plays a role. There is always some percentage of error with concentration experiments; thus, to account for that, the normalized bound percentage of accounts for the total number of events with a given barcode as measured in the nanopore. For the events in Fig. 2c,d, the following equations are used:

The first part of this formula just looks at events with dCas9 bound \(\frac{{N}_{11111{\,{\mathrm{dCas}}}9}}{{N}_{11111{\,{\mathrm{dCas}}}9}+{N}_{11001{\,{\mathrm{dCas}}}9}}\), whereas the second part involves multiplying by the total number of events with the barcode being measured (N_{11111 No dCas9} + N_{11111 dCas9}). This normalizes the measurements based on the relative concentrations that are being measured in the nanopore.

For Fig. 3, a different nanostructure is used and values for normalized ratios are changed accordingly. Ratios are calculated the following ways:

This normalization ratio is slightly different and based on the binding efficiency of the target gRNA to its target DNA sequence. Each \({X}_{{{\mathrm{Position}}\; i}}\) represents a different mismatch. X_Control defined as \(\frac{{\Sigma X}_{{{\mathrm{Control}}\; {\mathrm{Position}}}\mathrm{1,2,3}}}{3}\) is the average binding efficiency at each position of the target dCas9 to its target DNA sequence. This treats the measured ratio of the dCas9 RNP to its target as a binding ratio of 1.0 and the other measured ratios relative to that.

Binding efficiency (%) is calculated:

For the samples with multiple measurements, as described in the text, standard deviation of the population was calculated using the following:

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.