Specific Protein Detection Using Designed DNA Carriers and Nanopores

Nanopores are a versatile technique for the detection and characterization of single molecules in solution. An ongoing challenge in the field is to find methods to selectively detect specific biomolecules. In this work we describe a new technique for sensing specific proteins using unmodified solid-state nanopores. We engineered a double strand of DNA by hybridizing nearly two hundred oligonucleotides to a linearized version of the m13mp18 virus genome. This engineered double strand, which we call a DNA carrier, allows positioning of protein binding sites at nanometer accurate intervals along its contour via DNA conjugation chemistry. We measure the ionic current signal of translocating DNA carriers as a function of the number of binding sites and show detection down to the single protein level. Furthermore, we use DNA carriers to develop an assay for identifying a single protein species within a protein mixture.

Followed by heating to 65°C and linearly cooling to 25°C in a thermocycler over 40 minutes. 0.5µL of BamHI-HF and 0.5µL EcoRI-HF (each 100000 units/ml) were then added to the reaction mixture followed by incubation at 37°C for 1 hour. The cut m13mp18 DNA was then immediately purified using a Machery-Nagel NucleoSpin gel and PCR Clean-up kit.
Having cut the scaffold, the double stranded DNA carrier was formed by mixing together the following 4 µL cut m13mp18 DNA (100nM) 20 µL oligonucleotide mix (each oligo 200nM) 4 µL 100mM MgCl 2 1.6 µL 100mM Tris-HCl (pH=8), 10mM EDTA 10.4 µL deionised water And heated to 70°C followed by linearly cooling to 25°C over 1 hour. The sequences of all oligonucleotides used in the various DNA carrier designs is given at the end of the supporting information.
Each oligonucleotide is present at a 10 times excess to the cut m13mp18 DNA in the reaction mixture to ensure a high yield of correctly formed structures. The excess oligonucleotides were removed using Amicon Ultra 100 kDa cut-off centrifugal filters (Millipore) -a procedure commonly used for purification of DNA origami nanostructures from excess oligonucleotides. The protocol was 6 rounds of centrifuging at 3000g and 4°C with each round lasting 10 minutes. The sample was then recovered by turning the filter upside down and centrifuging for 2 minutes at 1000g. S2 Gel analysis revealed that the staple band was removed by this purification procedure (see Supporting Figure 2). The purified DNA carrier was then aliquoted and stored at -20°C before use.

DNA carrier synthesis yield and staple removal
The m13mp18 scaffold produced by New England Biolabs (NEB) is primarily circular and NEB estimate that >90% of molecules are in the circular form with the remaining molecules being in the linear. The linear molecules are inherently due to the purification process (see NEB website for further details).
The base sequence position of the nick which creates a linear DNA is presumably random. Therefore during the scaffold cutting process these linear molecules will be cut at random positions and produce fragments as in Supporting Figure 1. This therefore produces a small background in our experiments.
Supporting Figure 1: Cutting of the m13mp18 results in a small amount of fragments due to the presence of linear molecules in the m13mp18 before cutting. The recognition site is represented in red on the blue DNA strand.
Agarose gel electrophoresis was used to assess the yield of full length double stranded DNA carrier and to check the removal of excess oligonucleotides by Amicon filtration. In Supporting Figure 2 lanes 2 and 3 show there is a slight shift due to the cutting of the m13mp18 which removes 21 bases. Lane 4 shows the double stranded DNA carrier after formation but without excess oligonucleotide purification. The main band runs between the 6kbp and 8kbp markers on the ladder and there is a diffuse band at the bottom of the gel due to the excess oligonucleotides. Lane 5 shows the double stranded DNA carrier after excess oligonucleotide purification. The absence of the staple band in lane 4 confirms that the excess oligonucleotides are removed using the filtration protocol.
Lane 6 shows a sample which was made in order to estimate the amount of linear molecules in the m13mp18 before cutting. For this sample the m13mp18 (uncut) was hybridised with the complete oligonucleotide mix (with the same protocol as the DNA carrier). Therefore the circular scaffolds will form a circular double strand and the already linear scaffolds will form a linear double strand. Indeed we observe that there is a faint band that runs at the same speed as the DNA carrier which can be attributed to linear strands of 7249 bp and a strong band due to circular 7249 bp. Using the ImageJ gel analysis tool the mass percentage of DNA in the linear band was calculated to be 11%. Since the mass of each DNA in the circular and linear bands is the same this means that ~11% of the m13mp18 is present as linear molecules in qualitative agreement with the >90% specified by NEB. Therefore after cutting, hybridising complementary oligos and purifying ~22 % of DNA carriers are not the full length. We decided that this background was acceptable within the experiments since the selection of DNA carrier events based on event charge deficit should preclude most fragments. In future, to create highly pure samples in future the DNA carrier could be purified from an agarose gel. However this typically gives lower yields than Amicon filtration.
Supporting Figure 2: Agarose gel to quantify synthesis quality and yield. The 1% agarose gel was run at 80V for 90 mins in 1xTAE buffer. Lane 1 = New England Biolabs 1kb ladder, Lane 2 = m13mp18 before cutting, Lane 3 = m13mp18 after cutting, Lane 4 = DNA carrier before staple purification, Lane 5 = DNA carrier after staple purification, Lane 6 = Primarily circular DNA carrier made without cutting the m13mp18.
A rough estimation of DNA carrier lengths was also made by atomic force microscopy. Supporting Figure 3 shows a large area scan of 10.5 µm x 10.5 µm of the purified DNA carrier absorbed onto mica. The length of each isolated polymer chain was traced out using ImageJ and is shown in units of microns. A 7228 bp length of double stranded DNA should have a contour length of 2.5 µm (assuming 0.34 nm/bp). We observe most DNA strands have a slightly shorter length of 1.9-2.4 µm which we interpret as due to the large scale image not accurately tracking all the kinks in the DNA and therefore underestimating the true contour. The percentage of lengths less than 1.9 µm is 24% (10/41) which roughly agrees with the estimate of 22% of DNA lengths shorter than the full length made above using gel electrophoresis data.

S4
Supporting Figure 3: Left: Large area AFM of DNA carrier absorbed onto mica. Polymer chain lengths are marked in microns. The dotted area is shown at higher magnification in Figure 1c in the main text. Right: Histogram of lengths measured.

Glass nanopore fabrication
Quartz capillaries with inner diameter 0.2 mm, outer diameter 0.5 mm and containing a 0.05 mm fused filament were purchased from Sutter Instruments. Each capillary was cleaned by sonication in acetone followed by drying under nitrogen. The capillaries were then pulled to a form a glass nanopore using a Sutter P-2000 capillary puller with the following settings: Each glass nanopore was integrated into a microfluidic chip as previously described 1 .

Determination of nanopore geometry and estimation of DNA current change
A sample of 10 glass nanopores, made with the settings in Supporting Table 1, was imaged to determine typical values for the diameter and taper angle at the tip. Imaging was performed using the in-lens detector of a Zeiss FEG Sigma-VP SEM at 2-3 kV acceleration voltage. Supporting Figure 4 shows a view of a glass nanopore at a large scale in (a) and with inner and outer diameters marked in (b). On a few occasions we were able to obtain sufficient resolution and contrast to be able to differentiate between the walls and the inside of the nanopore. The ratio of ID:OD (inner diameter: outer diameter) at the tip is to a good approximation the same as that of the capillary before pulling where ID = 0.2 mm and OD = 0.5 mm. This constant wall thickness ratio has also been observed in larger pores made by glass capillary pulling 2 . Therefore when the inside diameter was not visible, we inferred the inner diameter by measuring the outer diameter and assuming a ID:OD ratio of 0.4. This analysis gives values of d i = 15 ± 3 nm (mean ± s.d., n=10).

S5
Supporting Figure 4: SEM images of a typical glass nanopore prepared using the settings given in Supporting Table 1. As the glass thins out at the tip, contrast is visible between the walls and the inside of the nanopore.
We analysed SEM images and found two taper angles were needed to accurately describe the 3D glass nanopore geometry. The first taper angle is determined from the first 100 nm from the tip. The second taper angle is determined by the next 600 nm as indicated in Supporting Figure 5a and assumed to continue on infinitely.
Supporting Figure 5: (a) Assumed geometry of a glass nanopore. Two cones in series (with two taper angles) are used. L is taken to be 100 nm. The light blue cone is assumed to continue indefinitely to the left. (b) Same as Supporting Supporting Figure 4a but with the outer angles of the two tapers as indicated.
The resistance of the first 100 nm length cone is given by And the infinite cone assumed to form the rest of the nanopore has a resistance given by There is also a contribution from the access resistance at the nanopore tip which can be modelled by a planar disk 3 and is given by Therefore the total resistance is given by summing the contributions from the two cones and the access resistance We can infer the inner taper angles from the measured outer taper angles using the approximation that the ID:OD ratio remains constant at 0.4 along the whole length and therefore tanሺߠ ሻ ൌ ߙtan ሺߠ ሻ Where ߠ is the inner angle, ߠ is the outer angle and ߙ is the ratio ID:OD = 0.4. This SEM model of the 3D geometry can be compared with our observed current change due to the DNA translocation. The resistance of the two taper geometry with an infinite cylindrical rod of radius a running through the centre (Supporting Figure 6) is given by Supporting Figure 6: Estimation of current change due to passage of long DNA through the glass nanopore. The DNA is assumed to pass through the centre of the glass nanopore and to be a cylindrical rod.
Substituting in the mean values from SEM of d i = 15 nm, ߠ ᇱ = 0.092 rad, ߠ ᇱᇱ = 0.046 rad and using our conditions of L = 100 nm, σ = 184 mS/cm and a radius of B-DNA of 1.25 nm gives a predicted current change Of ‫ܫ∆‬ ൌ 105 ‫ܣ‬ at 600 mV. This agrees well with the current change values observed for the DNA carrier in the main text and indicates an accurate geometrical model.

Ionic current recording and analysis
All experiments were performed with a buffer of either 10 mM Tris (pH = 8), 1 mM EDTA, 4M LiCl or 10 mM Tris (pH = 8), 1 mM EDTA, 4M LiCl , 5 mM MgCl 2 in the sample reservoir. The applied voltage was always 600 mV and this typically gave a few DNA carrier S7 translocations per second for a DNA carrier concentration of 1 nM. Ionic current was recorded using Ag/AgCl electrodes and an Axopatch 200B amplifier. The Axopatch 200B output was filtered with an external 8-pole Bessel filter at 49.9 kHz and digitised at 250 kHz.
In an initial step to remove noisy data traces, we automatically remove any section of trace where the standard deviation in 3 seconds was greater than 25 pA. This is typically due to transient, non-specific protein or DNA sticking to the nanopore. For event threshold detection we then zero the baseline ionic current level by fitting a cubic polynomial to every 0.5 seconds of trace. We then count an event when the ionic current passes a threshold of 100 pA from this zeroed baseline level and measure the begin and end of the event as where the current crosses 30 pA (30 pA is chosen rather than 0 pA since DNA translocations through conical glass nanopores give a slow decay as the DNA moves away at the end of the translocation). The one exception to the 100 pA threshold level is Supporting Figure 10 and Supporting Figure 11 where a 60 pA threshold is used for protein only translocations. For  Figure 1f, g in the main text). Each scatter plot and associated ECD histogram shows the first 1000 recorded events.

S9
Translocations of m13mp18 before oligonucleotide hybridisation Supporting Figure 9 shows typical translocation of m13mp18 DNA after cutting at BamHI and EcoRI restriction sites but before hybridizing the 190 oligonucleotides needed to from the DNA carrier. The scatter plot shows data from the first 1000 translocation events. Clearly we observe a significantly different translocation signature compared to the double stranded DNA carrier. The cut m13mp18 DNA shows a short sub-ms event duration and large amplitude current blockage. This reflects the compact globular shape of m13mp18 ssDNA compared to the extended coil-like double stranded DNA carrier as shown by the AFM images in Figure 1 of the main text. The shape of these translocations agrees with previous solid state nanopore measurements 4 .
Supporting Figure 9: Left: Three typical translocations of m13mp18 ssDNA through a glass nanopore with 1xTE, 4M LiCl buffer and 600 mV applied potential. Right: Scatter plot of mean event current against duration for first 1000 translocations.

Translocations of streptavidin and anti-digoxigenin only
Supporting Figure 10 shows typical translocations of streptavidin only through glass nanopores at 600 mV applied voltage in 1x TE, 4M LiCl. The concentration was 500 nM. In common with previous solid state nanopore experiments we measure a scatter plot where the events are cut-off by the detection threshold which was set at 60 pA. Similarly Supporting Figure 11 shows statistics of translocations of anti-digoxigenin antibody only at 500 nM concentration.
Supporting Figure 10: Translocations of streptavidin only through a glass nanopore. The peak current threshold was 60 pA and the first 1000 translocations are shown. Left: event scatter, right: ECD histogram. S10 Supporting Figure 11: Translocations of anti-digoxigenin antibody only through a glass nanopore. The peak current threshold was 60 pA and the first 1000 translocations are shown. Left: event scatter, right: ECD histogram.

Determination of DNA carrier events and selection of events beginning and ending with one DNA strand
For Figures 2-5 in the main text the DNA carrier events we needed to measure only the translocation of full length DNA carriers which begin and end with one DNA strand. Therefore the following protocol was used.

1)
First we recorded all events crossing 100 pA as described in the "Ionic current recording and analysis" section above. We then plotted the event charge deficit (ECD) histogram of all events with a 5 fC bin width. Crucially, the peak due to the DNA carrier typically occurs at ~100 fC to 200 fC (Supporting Figure 8) and is therefore easily discernible from the protein translocation ECD which is mainly <20 fC (Supporting Figure 10 and Supporting Figure 11). We then fit a Gaussian function to the ECD peak and select only the events that have ECD value in the range of µ±2σ where µ is the peak value and σ is the Gaussian standard deviation. Supporting Figure 12 shows an example ECD histogram with the DNA carrier peak labelled.
Supporting Figure 12: In stage 1 the ECD histogram is plotted and only events within µ±2σ are retained. This particular example is an ECD histogram for one nanopore using the 3B+mix2 configuration in Figure 4 in the main text. S11 2) After selecting only full length DNA carrier events in Stage 1, in Stage 2 we select only events beginning and ending with one DNA strand. To do this we plot an all points histogram of the recorded events after stage one selection and fit a Gaussian to the one DNA strand peak (Supporting Figure 13). The centre of this Gaussian determines the one DNA strand level. We then reject events where the current level passes 1.5 times the one DNA strand level in the first 300 µs or last 300 µs of the event see Supporting Figure 14. Supporting Figure 15 shows the scatter plot before and after each selection stage.
Supporting Figure 13: Series of eight concatenated events and histogram from all points of all events showing one DNA strand level peak at 108 pA.
Supporting Figure 14: Examples of selection based on current at start 300 µs and end 300 µs. The event on the left crosses the 1.5 times DNA level in the start 300 µs and is therefore discarded. The event on the right does not cross either the threshold during the start 300 µs or end 300 µs and is therefore not discarded.

S12
Supporting Figure 15: Progression showing scatter plot recorded from all events, events after stage 1 selection and events after stage 2 selection. The number of events decreases left to right from 917 to 517 to 236.
Raw data on detection efficiency (Figure 3b in main text) The following tables show the number of events after stage 1 and stage 2 selection for the different thresholds given in Figure 3.

separated biotin binding sites (Figure 5a in main text)
Oligo no 48 replaced with 5' -biotinTTTAGAAATAAAGAAATTGCGTAGATTTTCAGGTTTAACGT -3' Oligo no 96 replaced with 5' -biotinTTTCTTGAGCCATTTGGGAATTAGAGCCAGCAAAATCACCA -3' Oligo no 143 replaced with 5' -biotinTTTTATGCGATTTTAAGAACTGGCTCATTATACCAGTCAGG -3' The design for binding one digoxigenin was identical as for the biotin design 1B except that a 5' digoxigenin (NHS Ester) was substituted for biotin. The IDT codes for the biotin and digoxigenin modifications are \5Biosg\ and \5DigN\ respectively.