www.sciencemag.org/cgi/content/full/335/6070/831/DC1
Supporting Online Material for
A Logic-Gated Nanorobot for Targeted Transport of Molecular Payloads Shawn M. Douglas, Ido Bachelet, George M. Church* *To whom correspondence should be addressed. E-mail: http://arep.med.harvard.edu/gmc/email.html
Published 17 February 2012, Science 335, 831 (2012) DOI: 10.1126/science.1214081
This PDF file includes: Materials and Methods SOM Text Figs. S1 to S27 Table S1
Materials and Methods
Chemicals and Supplies Sigma: EDTA, 2xYT Microbial Medium. Fisher Scientific: magnesium chloride, polyethylene glycol 8000 (PEG8000), sodium chloride (NaCl), Tris base, sodium hydroxide, potassium acetate, lauryl sulfate, glacial acetic acid. BD: LB broth, Bacto agar. New England Biolabs: Nb.BsrDI, T7 exonuclease, lambda exonuclease; Stratagene: XL-10 Gold competent cells. Molecular BioProducts: 8-well PCR strip tubes. Invitrogen: SYBR Safe. Bio-Rad: Freeze ‘N Squeeze DNA gel extraction spin columns. KimbleChase: pellet pestles. Lonza: agarose. EMS: carbon/formvar copper grids, uranyl formate. Nanorobot preparation Nanorobots were produced by mixing 100 nM p7308 scaffold strand (http://www.ncbi.nlm.nih.gov/nuccore/221063642), 400 nM each staple strand, purified by reverse-phase cartridge (Bioneer, Inc.), buffer and salts including 5 mM Tris + 1 mM EDTA (pH 8.0 at 20°C), and 8 mM MgCl2. The mixture was subjected to a thermalannealing ramp according to the following schedule: 80°C to 61°C at 2 min/°C 60°C to 24°C at 150 min/°C Concentration of MgCl2 used for all reported experiments was initially calibrated by a concentration screen followed by agarose-gel analysis (Fig. S2). The value of 8 mM MgCl2 was chosen to maximize yield, though higher concentrations of may slightly improve folding quality. Folded nanorobots were separated from excess staples by centrifugal filtration, loaded with payloads, and then either purified by agarose-gel electrophoresis for imaging by transmission electron microscopy (TEM) or mixed with cells and analyzed by flow cytometry. Gel purification of folded samples Leading monomer bands were visualized with blue light and physically extracted from 2% agarose gels (10 mM MgCl2). Excised bands were crushed with 5–10 strokes using a 1.5-mL pestle, centrifuged in a 1.5 mL tube at for 3 minutes at 3000×g, 15°C. The bottom 5–10 mm of the 1.5 tubes were removed using a microcentrifuge tube cutter and then inverted and placed in a DNA gel extraction spin column, and spun for 1 minute at 3000×g, 15°C. Recovered material was then prepared for imaging as described in Douglas SM, et al. 2009 Nature 459:414–8. 2
Centrifugal filtration of folded samples Excess staples were removed using Amicon Ultra–0.5 mL 100K centrifugal filters (Millipore). Initial filtration was performed by pipetting the folded material into the filter device, adding 1x folding buffer (5 mM Tris, 1 mM EDTA, 8 mM MgCl2) to bring the total volume to 500 μL, and centrifuging for 10 min at 10,000g at 15°C. Flow-through containing excess staples is discarded after each round of filtration. Two wash steps with the same filter were performed by adding 450 μL of 1x folding buffer to the filter and centrifuging for 10 min at 10,000g at 15°C. Retained material was then transferred to a fresh filter device and three additional wash steps were performed. Choice of MgCl2 concentration Our previous work (Douglas et al., Nature 2009) established that the concentration of divalent cations in the DNA origami folding reaction impact both folding rate and degree of aggregation, starting with the formation of multimeric species. These findings were based on extensive analysis by agarose gel electrophoresis and TEM. We found that differences in performance for an identical shape folded under different conditions can be quite subtle. It has been our experience that higher MgCl2 gives better folding results (sharper bands) up to a certain point, after which we get worse results due to aggregation. In this study, we wanted to use the highest concentration of MgCl2 that did not show aggregation in the well, which was the 8 mM MgCl2 condition. Gold nanoparticle payload preparation 4 mg of phosphene powder (4,4′-(Phenylphosphinidene)bis(benzenesulfonic acid) dipotassium salt hydrate) (Sigma) was added to 10 mL of 5-nm collodial gold nanoparticles (AuNPs) (Sigma) and incubated overnight at room temperature on a rotating tube rack. AuNPs were precipitated by addition 1.7 g of sodium chloride powder (Fisher), centrifuged at 1400g for 30 minutes at room temperature, supernatant removed by pipetting, and resuspended in 1 mL of fresh 2.5 mM phosphene buffer. 2 mL of methanol were added to re-precipitate the AuNPs, followed by centrifugation at 1400g for 30 minutes at room temperature, supernatant removal by pipetting, and resuspension in 90 µL of 2.5 mM phosphene buffer. 10 µL of 5’-thiol-modified linker oligo (/5ThioMC6-D/GAACTGGAGTAGCAC, Integrated DNA Technologies) at 100 µM concentration was added to the resuspended AuNP sample and incubated overnight at room temperature on a rotating tube rack. Material was run on a 3% agarose gel (50 µL per well, 30 minutes at 100V). A glass microfibre filter (GF/C, Whatman) was backed by a dialysis membrane (Spectra/Por Cellulose Dialysis Membrane MCWO:10,000, Spectrum Laboratories, Inc.) was inserted to the gel in front of the leading band using a razor blade, and voltage applied an additional 10 minutes. The filter was recovered and placed in a cellulose acetate centrifuge tube filter (Costar 0.45 µM Spin-X filter, Corning), centrifuged at 3000g for 1 minute. For payload loading, recovered AuNP-oligo conjugates are then mixed in 8:1 volume ratio with folded and 0.5 pmol/µL filter-purified nanorobots (e.g. 17.5 µL of AuNP conjugates + 2.5 µL of nanorobots). This step dilutes the 1x folding buffer to 0.125x, giving a final MgCl2 concentration of 1 mM. 3
Antibody payload preparation Mouse antibodies were digested with immobilized ficin in mouse IgG digestion buffer with 25 mM cysteine (Thermo) by shaking at 37 C for 4 h. Antibody Fab’ fragments were purified by centrifugal filtration and evaluated by Nanodrop or SDSPAGE on 4-12% Bis-Tris gels (Invitrogen). Fab’ fragments were conjugated with 5’amine-modified linker oligonucleotide (/5AmMC6/GAACTGGAGTAGCAC, Integrated DNA Technologies) using a commercial kit (Solulink) according to the manufacturer’s instructions. Oligo-conjugated antibodies were cleaned by centrifugal filtration and evaluated by monitoring the absorption at 354 nm on a Nanodrop or by non-reducing PAGE. Loading was performed in folding buffer (8 mM MgCl2 in TE, pH 8.0) at a 2-fold molar excess of payloads to loading sites. Finally, loaded robots were cleaned by centrifugal filtration at 100,000 Da cut-off as described above. Cell lines The NKL cell line, isolated from a patient with aggressive NK leukemia, was previously described [Robertson, 1996] and was a kind gift of Dr. J. Ritz from the Dana Farber Cancer Institute. The cells were maintained in RPMI-1640 supplemented with 10% v/v FBS, 2 mM L-glutamine, 100 U/mL Penicillin-Streptomycin, 1 mM sodium pyruvate and 62.5 U/mL recombinant human IL-2 (Peprotech), and were split twice weekly. Jurkat (ATCC no. TIB-152), Ramos (ATCC no. CRL-1596), CCRF-CEM (ATCC no. CCL-119) and A549 (ATCC no. CCL-185) were purchased from American Type Cell Culture (http://atcc.org) and maintained according to the vendor’s guidelines. Media were purchased from Invitrogen (for NKL cells) or ATCC (Jurkat, Ramos, CCRFCEM, A549). Cell killing and activation assays NKL cells (2×105/sample) were washed once with growth medium (containing rhIL2) and incubated with loaded robots at the designated concentrations and time periods. For activation studies, cells were incubated with loaded robots that either were or were not pre-incubated with S. typhimurium Flagellin (1 ng/mL) and washed in folding buffer. To analyze signaling proteins, cells were fixed with 2% formaldehyde and permeabilized with 100% methanol for at least 10 min. Staining was performed in PBS containing 0.1% w/v BSA and 0.01% w/v NaN3 on ice, with 1 full volume wash prior to analysis. For cell cycle analysis, cells were fixed with frozen 70% ethanol for at least 10 min, and stained with PI solution (5 μg/mL propidium iodide, 0.1 mg/mL RNase A, 0.05% v/v Triton X100 in PBS) and analyzed by flow cytometry. Flow cytometric analysis was performed on Accuri C6 and Becton-Dickinson LSRFortessa Special Order flow cytometers, both equipped with 488 nm and ~650 nm lasers. FCS files were analyzed by FlowPlus or FlowJo and expressed as median fluorescence values ± standard deviation.
4
Flagellin concentration in scavenging experiments The concentration of flagellin in our experiments varied from 100 pg/mL to 10 ng/mL as a range representing concentrations that were shown to be 1) sufficient for induction of significant immune responses (e.g. cell activation and cytokine release, see our ref 28); and 2) realistic in animal models and sepsis patients (see Liaudet et al., 2003 Shock 19, 131–7). We show the 100 pg/mL condition in Figure 3 to demonstrate the low range that was used.
5
SOM Text Open barrel design We chose an open barrel design for the nanorobot because we expected it allow for the production of the highest yield of functional nanorobots for subsequent experimentation. An open design allows for single-step loading of cargo: once the origami is folded and purified from excess staple strands, the DNA-anchor-linked cargo is simply added in excess and allowed to diffuse into the nanorobots and bind. We thus avoid extra preparation steps that may be required of a closed design, such as closing the container after loading it in an open state, or freeing a tethered cargo after loading/activation. This strategy does limit our to payloads that can be linked to DNA, but this is not a problem in our desired application of stimulating cell signaling, because antibody payloads remain active even while attached to the nanorobot. Cargo loading We provide some theoretical basis for the efficiency of antibody cargo loading by passive diffusion, based on the number of cargo particles interact with the front and rear entrances of the nanorobot. We carried out loading in TE buffer (pH 8.0), so the net charge of antibodies and antibody fragments is close to zero. We used a protein charge calculator (http://www.scripps.edu/~cdputnam/protcalc.html) to calculate the net charge at pH 8.0 of various Fab’ fragments (summing all 4 Ig subunits) from published sequences and found that they range between -4.0 and +2.5. (see also Chiodi et al., Electrophoresis 6, 124–8). Thus, we did not expect the negative charge of the DNA origami to significantly affect the diffusion of the antibody cargo near the nanorobots. We used the following Python script to estimate how many cargo particles per second may pass through the ends of the nanorobot. #!/usr/bin/env python nu_cgs = 1.000/100 # viscosity of water at room temp [poise] or [g*s-1*cm-1] nu_kms = nu_cgs/1000*100 # [kg*s-1*m-1] T = 298.15 # [Kelvin] Room Temperature kB = 1.3806488e-23 # [J*K-1] Boltzmann constant r_antibody = 3e-9 # [m] approx. radius of antibody fragment friction_coefficient = (6*3.14*nu_kms*r_antibody) # of a spherical particle # diffusion coefficient estimating the hydrodynamic radius of the antibody D = (kB*T)/friction_coefficient totalParticles = 1.8e13 # [particles] totalVolume = 30 # [uL] C = totalParticles/(totalVolume*1e-6*0.001) # [particles*m-3] # Treat as a diffusion of particles towards a sphere but use # the open area of the ends instead of surface area of the sphere # see Ken A. Dill, Molecular Driving Forces, pp 321-322 a = 45/2.0*1e-9 # [m] spherical radius of nanobot # the nanobot dimensions are 35*45*35*nm^3; assume an entrance is 35nm circle
6
A = 3.14/4.0*(25e-9)**2 # [m2] pi* D2 / 4 entrance area of one end of nanobot J = -D*C/a # flux into the sphere [particles*m-2*s-1] I = J*2*A # current through 2 entrances [particles*s-1] totalTime = 24*60*60 # [seconds] in a day # Output results print "Diffusion coefficient, %e" % (D) print "%.2f, particles/s colliding with the open ends of the nanobot" % (I) print "%.2f, particles colliding with the opening per day" % (totalTime*I) print "%.2f, percent particles interacting with a given nanobot" (totalParticles/(totalTime*I)) print "%f, distance [m] diffuse in 1 hour" % ((6*D*3600)**.5)
%
We estimate that approximately 2000 particles per second diffuse through the open ends of the nanorobot, so the efficiency of cargo loading is unlikely to be diffusion limited on the time scales we used for loading (~24 hrs). One possible concern is steric hindrance in the cavity once payloads bind inside the nanorobot. Currently the loading sites form a single internal ring around the inside of the barrel, but if we decide to have inner and outer positions in future designs, we could perform a multi-step loading procedure with different linker sequences so innermost payloads bind first. Lock optimization and analysis Our nanorobot lock system was inspired by aptamer beacons1, structure-switching aptamers2 and similar aptamer-based biosensors3. These devices undergo a target-induced switching between an aptamer-complement duplex and a aptamer-target complex, resulting in a change in distance between a dye and quencher pair allowing for a fluorescent readout. Many variations on this idea have been implemented, with aptamercomplement duplex lengths ranging from 7 bp2 to 22 bp4 and with activation times of seconds to hours depending on design and reaction conditions5. A novel feature of our system is that the aptamer and complement strands are attached to two separate ~2.4 megadalton domains of a DNA origami nanostructure. Each strand has a 20–22 base extension that anchors it to attachment sites on the top or bottom domain of the nanorobot scaffold (Fig. S4a). Based on effective helix diameter estimates for single-layer origami6 we estimate the distance between the lock attachment sites are at least 3 nm apart. Thus, even a fully complementary lock duplex must remain partially unzipped to bridge this gap between the two domains. Moreover, duplex formation must overcome the repulsion of the negatively charged phosphate backbones of the helices that make up the two domains, as well as the entropic penalty for holding the nanorobot in a closed conformation. To examine the effects of aptamer-complement duplex length, we designed a series of locks with varying mismatches between the aptamer and complement strand. To avoid compromising the aptamer ligand-binding performance, we modified only the complement strand in the region proximal to its attachment sequence, replacing several bases with thymines (Fig. S4b). We designed complement strands with 0, 7, 14, 21, and 28 thymine bases, and then compared their performance in studies of activation speed 7
(Figs. S5–S8), sensitivity (ligand concentration required for activation), and leakiness (i.e. false positives due to spontaneous opening) (Fig. S8). We designed a microbead-based assay coupled with quantitative flow cyotometry (qFCM) to count single nanorobot opening events, which enabled the quantitative calculation of two critical parameters: the rate of opening for a large nanorobot population, and the probability for a given nanorobot to open per unit time. We constructed Cy3-labeled nanorobots, and then loaded them with payload oligos modified with 5’ BioTEG. Nanorobots that have opened in response to the key expose their biotinylated payload and are able to label the microbeads which are then analyzed by qFCM. This assay isolates the open nanorobot fraction from the general population so they can be counted. Acquisition was made at precalibrated photomultiplier tube voltages so that median fluorescence values transform precisely to molecular equivalent of soluble fluorochrome (MESF). We first measured in real time the binding of a population of 100% unlocked nanorobots to determine baseline parameters for the microbead-based assays. We found that open nanorobots saturated the microbeads after 30 minutes, with a binding capacity of approximately 100,000 nanorobots per microbead. To examine test activation rate dependence on lock design, the various nanorobot versions were incubated with 10 nM key analyte (PDGF) for 24 hours (Fig. S5). We observed that nanorobots with weaker locks (more complement thymine mismatches) were able to open faster than those with stronger locks. This behavior was consistent for a lower key concentration of 0.1 nM (Fig. S6). Based on these data we calculated the probability per second (p) that any given nanorobot will open, and plotted the value for each nanorobot version expressed as mean log(p) in Fig. S7. We also performed a functional assay to examine the different lock versions (Fig. S8). We loaded the nanorobots with anti-CD33 and anti-CDw328 Fab’ antibody fragments, and compared the signaling phosphorylation of the p42/44 kinase ERK. Locks with 0 or 7 thymine replacements exhibited similar profiles, which may indicate that the first 7 base pairs are typically unzipped. Nanorobots with weaker locks were able to stimulate cell signaling at lower nanorobot concentrations, which is consistent with the trend of weaker locks opening more readily that we observed in Figs S5 and S6. In designing our lock system, we sought to balance robot sensitivity against robot leakiness. We hypothesized the stability of the aptamer-complement duplex would be inversely proportional to sensitivity and activation rate of the nanorobot5. We also hypothesized that fewer duplex mismatches would reduce unwanted triggering of robot activation that could lead to false positives in downstream experiments. We estimated the percentage of nanorobots that were already open before key ligand addition (caused by initial mis-folding or spontaneous opening), and compared these values to the minimum ligand concentrations needed to activate each nanorobot version (Fig. S9). It was possible to achieve a 100-fold improvement in nanorobot sensitivity by decreasing the duplex 8
strength. Increased sensitivity came at a cost of increased activation in the absence of ligand, but we expect a cargo-affinity-based purification step could be used to mitigate false positives. Since we saw an overall trend of longer duplexes resulting in lower initial leakiness, we wanted to use the longest duplex possible. However, the sensitivity appears to take a significant hit for locks above 23 bp, so we chose this length as the basis for subsequent experimental trials in this study. Optimization of lock sensitivity and strength may be possible on a case-by-case basis by exploring additional lock/complement design strategies, such as elongating the duplex region with extra bases while shortening the complementary region that blocks the aptamer active site, that reported by Nutiu R, et al. JACS 2003. Nanorobot activation by exogenously expressed key on cell surface We tested whether an exogenously expressed cell surface marker in a negative cell line could be used to activate the nanorobot. We chose to use a lock based on an aptamer targeting PTK7, termed sgc8c, has been reported15. To choose an appropriate negative cell line, we screened several (NKL, Jurkat, Ramos-1, CEM-CCRF and A549) for surface expression of PTK7. Only NKL cells were found not to express PTK7 in detectable levels. We expressed PTK7 in NKL cells by transient transfection with a transfectionready human PTK7 (RefSeq NM_002821.3) cDNA clone (Origene), using lipofectamine LTX (Invitrogen) according to the manufacturers’ instructions. Significant expression of PTK7 was expressed as early as 24 h post transfection, peaking at 48 h post transfection (Fig. S22b). PTK7 expression was evaluated by flow cytometry using an APCconjugated antibody (Miltenyi Biotec). After validating PTK7 expression, we found that nanorobots were able to label only the induced expression on NKL cells (Fig. S23). Next, we tested the sensitivity of nanorobot activation to total cell count. Our goal was to see if the nanorobots (sgc8c locks) could detect a target cell population across many cell counts relative to the background. We chose a background of 1e6 key– cells (Ramos/RA1), and mixed that background with serial dilutions (1e6, 1e5, ..., 1e0 cells) of key+ cells (Jurkat). We observed that sensitivity did decrease below the target cell count of 1e4, but was still significant. The capability to bind to targets at the single-cell level could be therapeutically useful, for example in contexts where target cells represent a very small fraction of the total cell population, e.g. detecting cancer metastasis.
9
Fig. S1 Design schematic, produced in cadnano (http://cadnano.org/). 10
Fig. S2 Gel analysis of nanorobot folded with different MgCl2 concentrations, before removal of excess staples. Samples were electrophoresed for 2.5 hours at 70V in a 2% agarose gel containing 10 mM MgCl2, 0.8x SYBR Safe, 45 mM Tris base, 45 mM boric acid, and 1 mM EDTA (pH 8.0). Gel box was submerged in an ice-water bath to prevent overheating. Leading band is nanorobot monomer. Above 12 mM MgCl2, a dimer band starts to appear.
11
Fig. S3 Quantitation of staple flow-through concentration following each wash step. DNA concentration in the flow-through after each wash step was estimated using a Nanodrop spectrophotometer.
12
Fig. S4 Nanorobot lock designs. (a) Schematic orthographic view of the nanorobot, in which helices are represented as circles (top domain in blue, bottom domain orange), along with aptamer-based locks (blue, red, orange, & green lines). (b) Sequence-level detail of five lock variants that attach to the right side of the nanorobot (the left locks are similar, but with the attachment sequences swapped). Each lock consists of two DNA oligonucleotides. The first encodes an aptamer (e.g. anti-PDGF, shown in red) and nanorobot domain attachment sequence that is complementary to the nanorobot scaffold. The second strand encodes a sequence that is complementary to the aptamer (green), and a similar attachment sequence. In order to tune the behavior of the lock, variants of the complement strand were synthesized with mismatching (poly-thymine) regions proximal to the nanorobot.
13
Median Fluorescent Intensity (Cy3)
12000
10000 unlocked 16 bp 8000
23 bp 30 bp
6000
37 bp 44 bp
4000
2000
0 0
5
10
15
20
25
Time (h)
Fig. S5 Nanorobot activation-rate dependence on lock duplex length. Fluorescently labeled nanorobots (10 fmol) with anti-PDGF aptamer-complement locks with varying duplex lengths (n = 16, 23, 30, 37, 44) were loaded with biotinylated cargo and incubated with streptavidin-coated beads in the presence of 10 nM PDGF. Beads were tested for nanorobot labeling by qFCM at 0.5, 1, 3, 6, 12, and 24 hours. Negative control groups consisting of permanently locked nanorobots were measured only at t=24 hrs, and had a median fluorescent intensity of 331 A.U. Error bars represent SEM of two replicates.
14
Median Fluorescent Intensity (Cy3)
7000
6000
5000 16 bp 4000
23 bp 30 bp
3000
37 bp 2000
44 bp
1000
0 0
5
10
15
20
25
Time (h)
Fig. S6 Nanorobot activation rate decreases with lower key concentration. Fluorescently labeled nanorobots (10 fmol) with anti-PDGF aptamer-complement locks with varying duplex lengths (n = 16, 23, 30, 37, 44) were loaded with biotinylated cargo and incubated with streptavidin-coated beads in the presence of 0.1 nM PDGF. Beads were tested for nanorobot labeling by qFCM at 0.5, 1, 3, 6, 12, and 24 hours. Negative control groups consisting of permanently locked nanorobots were measured only at t=24 hrs, and had a median fluorescent intensity of 402 A.U. Error bars represent SEM of two replicates.
15
-4.4
10 nM key
-4.5
0.1 nM key
mean log(p)
-4.6 -4.7 -4.8 -4.9 -5 -5.1 -5.2 -5.3 16
23
30
37
44
lock duplex length (bp)
Fig. S7 Nanorobots activate more readily with weaker locks and in the presence of higher key concentration. Nanorobots with varying lock duplex lengths loaded with biotinylated cargo were incubated with streptavidin-coated beads in the presence of two different concentrations of PDGF key ligand (0.1 and 10 nM). Beads were tested for nanorobot labeling by qFCM at 0.5, 1, 3, 6, 12, and 24 hours. Using baseline parameters for bead binding capacity, we calculated the mean probability for each individual robot type to open per unit time for each lock version. Standard error bars are shown.
16
Median Fluorescence (phospho-p42/44)
350000 300000 Locked 250000
44 bp 37 bp
200000
30 bp 150000
23 bp 16 bp
100000
Unlocked
50000 0 7
7.5
8
8.5
9
9.5
10
Log (nanorobots/! L)
Fig. S8 Nanorobot lock influence on cell signaling suppression. Nanorobots with varying lock duplex lengths (n= 16, 23, 30, 37, 44) were loaded with anti-CD33 and anti-CDw328 Fab’ antibody fragments. Each nanorobot version was tested at varying concentrations (0–10 nM) for its ability to stimulate cell signaling in aggressive NK leukemia cells. After 4 hrs incubation, signaling levels were determined with intracellular flow cytometry by measuring phosphorylation levels of the p42/44 kinase ERK; lower median fluorescence indicates increased cell signaling. Error bars represent SEM of two biological replicates.
17
10000
10
initial % open without ligand
8 1000 7 6 100
5 4 3
10 2
[ligand] (pM) required for activation
9
1 0
1 16
23
30
37
44
lock duplex length (bp)
Fig. S9 Nanorobot lock mismatch influence on leakiness and activation threshold. To examine the percentage of nanorobots that that were open in the absence of activating ligand, fluorescently labeled nanorobots with anti-PDGF aptamer-complement locks (duplex lengths = 16, 23, 30, 37, 44) were loaded with biotinylated cargo and incubated with streptavidin-coated beads, and then analyzed by flow cytometry (plot in red, error bars represent SEM of two replicates). A positive control of 100% open nanorobots was used to establish a baseline fluorescence value, which was used to estimate the percentage of each locked nanorobot that was initially open. The same nanorobot versions were also tested for minimum concentration of ligand was required to achieve a measurable activation response. Each version was incubated with streptavidin-coated beads in the presence of increasing concentrations of PDGF, and then tested for activation by flow cytometry. The ligand concentration was considered to be activating if it resulted in measured fluorescence greater than 50% of the always-open nanorobot control (plot in blue).
18
Fig. S10 TEM images of nanorobots in different lock and loading states. Scale bars: 20 nm. 19
"
!
!
"
#
$! ! ! $ ! ' ! 43! ( &/./)( ' # # $ + /,'
# ! -/. ! /0'15# $ $(" ! ! " '2/' # " !% * ! ! 6! &! (/. '
Fig. S12 AFM micrograph of closed nanorobots. Nanorobots were diluted to ~10 nM into TAE buffer containing 12 mM Mg2+. 5 µL were applied directly on freshly-cleaved grade V mica mounted on a sample plate using optical adhesive no. 61. Samples were visualized in a Veeco/Bruker Multimode Nanoscope V in fluid tapping mode, using SNL-10 C probes.
21
Fig. S13 AFM micrographs showing activated nanorobot-key complexes. Nanorobots bearing anti-PDGF-aptamer locks were incubated overnight at room temperature with 0.25 µM of PDGF-ββ (R&D Systems) in 1x folding buffer, and visualized directly using AFM. For visualization, nanorobots were diluted 1:10 in TAE buffer containing 12 mM Mg2+, and 5 µL were applied directly on freshly-cleaved grade V mica that was mounted on a sample plate using optical adhesive no. 61. Samples were visualized in a Veeco/Bruker Multimode Nanoscope V in fluid tapping mode, in TAE buffer with 12 mM Mg2+, and using sharp silicon nitride (SNL-10) C probe, with a nominal spring constant of 0.24 N/m. Arrows indicate PDGF ligands bound to their keys in three representative images. Scale bars: 50 nm.
22
Fig. S14 Dynamic light scattering analysis of nanorobot activation by keys. Nanorobots bearing anti-PDGF locks (0.5 pmol/sample) were analyzed by dynamic light scattering in three independent trials. Measurements were before activation (black series), 2 hours after keyinduced activation (red series: human PDGF-ββ was added to a final concentration of 25 nM), and following key removal by gel filtration (blue series). Acquisition was carried out on a Malvern Zetasizer Nano ZS instrument at 25°C. Nanorobots appear to become slightly more compact after key removal indicating that some may re-lock in the absence of key. This behavior is consistent with our observation that guide staples are necessary to prepare the nanorobots in a closed state during the initial folding process.
23
Fig. S15 3D Models highlighting guide staples. (a) guide staples (red) bridge top and bottom domains of the nanorobot. 8-base toeholds are included to allow removal by addition of fully-complementary strands after folding and purification. (b) After guide staple removal, only aptamer locks hold the structure in the closed state.
24
Fig. S16 Agarose-gel analysis of nanorobot folded with and without guide staples. Three versions of nanorobots were folded: 1. –locks/–guides, 2. +locks/+guides, 3. +locks/–guides. The +locks/–guides version displayed decreased band sharpness and slower mobility when compared to the version that included guide staples.
25
a
b
26
c
Fig. S17 TEM analysis of nanorobots folded without and with guide staples. (a) Leading band from lane 1 in Fig. S16 was physically extracted from the gel, imaged by negative stain TEM, and manually counted to assess yield. Total closed nanorobots counted = 0. (b) Leading band from lane 2 in Fig. S16 was similarly analyzed. Nanorobots were considered to be in the closed state if they had a uniform rectangular outline. 550 closed and 14 open nanorobots were counted (97.5% closed) (c) Leading band from lane 3 in Fig. S16 was similarly analyzed. 269 closed and 289 open nanorobots were counted (48.2%).
27
Fig. S18 Mixture of NKL cells with whole blood leukocytes. Blood leukocytes were isolated from healthy whole blood and mixed with NKL cells at a 4:1 ratio. Forward- versus sidescatter density plots showing (a) whole blood leukocytes (b) NKL alone and (c) 4:1 mixture of whole blood + NKL cells (reproduced from Figure 2 for comparison). Cell counts: 2e5 NKL, 8e5 leukocytes.
28
Fig. S19 Subpopulation analysis of nanorobot binding to cell-surface CD33. (a) Forward-scatter vs. sidescatter density plots shown on left for 4:1 whole blood + NKL cell mixture, plus locked nanorobots loaded with fluorescently labeled anti-CD33 payload, following . Gated subpopulations are: eosinophils (and possibly basophils), NKL + residual monocytes, monocytes, B- and T- lymphocytes, and granulocytes. Subpopulation gating is not exclusive due to forwardand side-scatter overlaps. Histograms of fluorescent labeling of CD33 in gated subpopulations shown at right. Locked nanorobot does not exhibit significant binding to any population. (b) Unlocked robots label each sub population. Peak segregation of monocyte population suggests developmental stage-specific expression pattern of CD33. (c) Nanorobots gated with antiPDGF/anti-PDGF aptamer locks selectively label NKL cells. In a–c, cell counts: 2e5 NKL cells, 8e5 leukocytes; total nanorobots added: 100 fmol.
29
Fig. S20 Examination of nonselective binding of loaded robots to the surface of NKL cells. Locked nanorobots loaded with DyLight 649-labeled anti-human CD33 were incubated with NKL cells for 30 minutes at 37°C. Cells were analyzed in a BD LSRFortessa flow cytometer. Insets show median values acquired at FL-4 channel following cell staining. (a) NKL cells (count = 10,000) treated with an APC-labeled isotype control IgG showing background staining. (b) NKL cells (count = 10,000) treated with 100 fmol CD33-loaded nanorobots. A slight (but not significant) difference in signal is observed.
30
$" " $# '$) $97<- $-% 66'" %$'$ # "63 %$#$6:>, #'" )* "$##? ')$ $", #$## ' & %#!%"$ -7 ' #$ ,/ 0 #/ % $=43+3330 $"$'$ - # $) $" # ' " % #$ ,/ 0 #/ % $ =43+3330$"$'$433 66- " $#,# $/%$ $# ? $0" # # #"&,
$# % ! #/87.# 0'" $"$ '$ 43 " $# '$ -? % "# $ %# $-% 66 "3+4+ 7"#, #'"$ #$ '$? % "# $ $- %## ") $ ) $" $ #%" /'$ %$%"$" $"$ $0 " " $ $ $ $/$" ?($ " *$ ##" $ +665+9;:-<91534420, #'"'#'$ " "$ )##) ? ')$ $")/ % $ = 433+3330, $ #"& # ? $ " $' #%" ( "## ) $ $ $ $ ) , "" " "# " "# $ $' " $#,
Fig. S22 PTK7 expression on NKL cells 48 h following transfection. (a) APC-isotype control. (b), APC-conjugated anti-human PTK7 antibody.
32
Fig. S23 Nanorobots can be activated by exogenously expressed keys. Nanorobots were prepared with two sgc8c-based locks, loaded with FITC-conjugated anti-human HLA-A/B/C antibody fragments, and mixed with untreated, empty-vector treated, and PTK7-induced cells for 3 h at room temperature. Nanorobots labeled only cells expressing PTK7.
33
Fig. S24 Analysis of nanorobot sensitivity to relative count of target cells. Mixtures of key– (Burkitt’s lymphoma / Ramos RA1) and key+ cells (T-cell leukemia / Jurkat) were incubated with 10 fmol of sgc8c-locked nanorobots loaded with fluorescently labeled anti-HLA A/B/C. Samples containing a fixed concentration of key– cells (count = 10^6) were combined with a dilution series of key+ cells . (A) Representative histogram showing fluorescence counts for control condition (10^6 key– cells) (B) Histogram of fluorescence counts for sample with 10^6 key+ cells (50/50 mixture) (C) 10^5 key+ cells (D) 10^4 key+ cells (E) 10^3 key+ cells (F) 10^2 key+ cells (G) 10^1 key+ cells (H) 10^0 key+ cells. (I) Plot of median fluorescence values for each cell population. Median values decrease for lower cell counts starting below 10^4 key+ target cells. Coefficients of peak variance were ~34% (J) Representation of the nanorobot sensitivity as a function of target to non-target cell ratio. Plot of the key+ values from (I) as a fraction of the max median fluorescence value observed.
34
!
! ! ##% # ' ! ## $!#! @! ## ! #"! # #%#( . ." " 3-8<4- $ ###& ! ##( " ( # "#!# # /("0## ! ($ # ! #"&! # !$ ! ##%# -
1.6 1.4
1.2
1 0.8 0.6 0.4 0.2 0 0.0001
0.001
0.01 0.1 "
1
10
& #( " ( # "! ) .553 !6 8.9;>:; *("# "6.997. 4&! $#&# " %!.' !"" .55 !:7 $#"- ! #"3877 4 &#. #.:: # (! #"&!# # '#$! $# ! # =#! # !#$!- #%. #! # 32 ! #"12 # "4 $ "&"$"# #! " @ $ !" - "#%. #! # 3? ! #"12 # "4&"$"# #! #/#%#0@ $ !" % - ! # #%# &"""""(@ &(# #!(# "$! ! # # "+ " ' !"""!# ##%#@ $ !" % -!! !!"! !" # #& ! #"- 3 " #( # ( #! " # ! ##%# ,# #" & -4
Fig. S26 Validation of antibodies targeting CDw328 and CD33 for growth arrest. (a) An immunosorbent 96-well plate was coated with goat anti-mouse IgG (50 µg/mL, 4h at room temperature), washed 3x with PBS, and coated with mouse anti-human CDw328/AIRM-1 and anti-human CD33 antibodies (1 µg/mL, 4 hours at room temperature) followed by 3 washes with PBS. NKL cells were cultured in the plate at growth medium (RPMI-1640, 10% FBS, 1 mM sodium pyruvate) for 3 days. Cells were then harvested, fixed at 2% formaldehyde (10 min on ice), permeabilized with frozen 100% methanol (10 minutes on ice), and stained with fluorescently-labeled anti-human PCNA and anti-phospho-p42/ERK in FACS buffer (0.1% w/v BSA, 0.05% w/v sodium azide in PBS, pH 7.4, 30 minutes on ice), washed and analyzed in an Accuri C6 flow cytometer. For cell cycle analysis, cells were fixed with frozen 70% ethanol (10 min on ice), stained with cell cycle staining solution (5 µg/mL propidium iodide, 0.1 mg/mL RNase A in PBS containing 0.1% Triton X-100, 30 minutes at room temperature) and immediately analyzed at FL-3 channel. (b) Fraction of cells at G1 stage or sub-G1 (Apo/G0 = apoptotic and G0) are shown. Error bars represent SEM of two replicates. 36
" & ( )#!
' (% ,,)#!! & ! -%
(p)-Akt median fluorescence intensity
12000
10000
8000
6000
PDGF lock unlocked control 4000
2000
0 0
0.01
0.1
1
10
100
[nanorobot] (nM)
! % -% " (! #*$ #+** )., % ! %
Table S1. Nanorobot staple sequences. Description
Sequence
1
core
TTTAGTTAATTTCAATTAATTTTCCCTTTGAGTGA
2
core
AGAAAACTTTTTCATTGAAAACATAGCG
3
core
AATCGCAAGACAAAAGATTAAGACGCTG
4
core
GTTATATTCATAGGTCTGAGACATCAAGAAAACAAATTTCAA
5
core
TGAATTTTACATTTAACAATTTCGCGCA
6
core
ATAACCTCCTTTTACATCGGGTTTCAGGTTTAACGAAAAGTT
7
core
ACAATATATGAGAATCCAATATAT
8
core
ATTCGCCAAATAAAGAAATTGATTTTGC
9
core
TGCATGGAAAATAGCTTGAACGCG
10
core
AAATCATTTGAGAAGAGCAAATCC
11
core
GAGGCGAGGTTAGAACCTACCATCATAT
12
core
TTACCTGTATACTTCTGAATATGATGGC
13
core
TGAGTAAACTCGTATTAAATCCAGAGATACATCGCCATTA
14
core
GGAACAAGACTTTACAAACAACTGAAAGGCGCGAAAGATAAA
15
core
GGAGCGGTTTGAGGATTTAGAGCACAGACAATAATCTCAATC
16
core
TCCTGATGAGCCGTCAATAGACAGTTGGATCAAACAACAGTG
17
core
AATTCATGCACTAACAACTAAAAAGGAATCACCTTAGCAGCA
18
core
TAAAGCATTGAGGATGCAACAGGAAAAATTGC
19
core
AAAATACCGAACGAACCACCAGTGAGAATTAACCGTTGTAATTC
20
core
AGACTGATAGCCCTAAAAGAACCCAGTCACA
21
core
ACAGAGGCCTGAGATTCTTTGATTAGTAATGG
22
core
GCGTATTAGTCTTTAATCGTAAGAATTTACA
23
core
TTAACACACAGGAACACTTGCCTGAGTATTTG
24
core
CCACGCTGGCCGATTCAAACTATCGGCCCGCT
25
core
GCCGCTGAACCTCAAATCAAATCAGGAAATA
26
core
AATGAAACAGAGCGTAATATC
27
core
CGACCAGTCACGCAGCCACCGCTGGCAAAGCGAAAGAAC
28
core
ACCTTCTGACTTCGACACATTATCCGTAGATAGAA
29
core
TTGGCAGGCAATACAGTGTTTCTGCGCGGGCG
30
core
ATTATACGTGAGTATTAAGAAACCAAAACAGTGAT
31
core
GTCTGAAATAACATCGGTACGGCCGCGCACGG
32
core
ACGATCTGGTTAATACAAATTATCATATCAATACA
33
core
CCTACATGAAGAACTAAAGGGCAGGGCGGAGCCCCGGGC
34
core
CATACAGTTGTAGATTATATCAGAATGGAAGATTA
35
core
TGGGGAGCTATTTGACGACTAAATACCATCAGTTT
Id
38
36
core
GGAAGAAGTGTAGCGGTCACGTTATAATCAGC
37
core
AGAGAACGTGAATCAAATGCGTATTTCCAGTCCCC
38
core
CGAACGTTAACCACCACACCCCCAGAATTGAG
39
core
GGAAGGGCGAAAATCGGGTTTTTCGCGTTGCTCGT
40
core
GAGCTTGTTAATGCGCCGCTAATTTTAGCGCCTGCCCTCAAT
41
core
CTAAAGGCGTACTATGGTTGCAACAGGAGAGA
42
core
GCCGTAAAGCAGCACGTATAA
43
core
AAGTAGGGTTAACGCGCTGCCAGCGGCTAGTAGTCCGC
44
core
GATTCCTGTTACGGGCAGTGAGCTTTTCCTGAACGACG
45
core
GCCTTCACCGAAAGCCTCCGCTCATTCCCAG
46
core
GTCCACGCTGCCCAAATCAAG
47
core
GGCGGTTAGAATAGCCCGAGAAGTCCACTATTAAAAAGGAAG
48
core
CAGGGTGCAAAATCCCTTATAGACTCCAACGTCAAAAGCCGG
49
core
CAGTGAGTGATGGTGGTTCCGAAAACCGTCTATCACGATTTA
50
core
ATTGCCCCCAGCAGGCGAAAAGGCCCACTACGTGACGGAACC
51
core
AAATGCCAGTTTGAGGGGGATTGAGTGAGCGAATAGGA
52
core
GGGTAGACCTTTGATAGATTAAATCCGTAAT
53
core
CTCGAATGCTCACTACAGTAT
54
core
AATTGCATGCCTGCAGGACCCGTCGGATTTCAAATCAG
55
core
GCTCATGGTCATAGCTGAACTCACTCGCACT
56
core
TAATGTGAAATTGTTATGGGGTGCGGCACCG
57
core
TCACGACTGTGCTGGCGCAAC
58
core
AACGCCAGGGTTCAATTCCACACAACATACG
59
core
GGGATAGGTGCATCCCTGTCGGGGGAGA
60
core
AAACGGCGACGACGGCCCGCTTGGGCGC
61
core
CGGGCCTAGGAAGAATTAATTTTTTCAC
62
core
TGTTGGGGCTTTCCCTAATGAACAGCTG
63
core
TCGCCATTGCCGGAAAAGTGTCCTGGCC
64
core
CGTAACCGTCACGTCAGCTTTAATTCGC
65
core
CCAGCCAAAGGGCGTGGCGAAAATTCGC
66
core
CTTCTGGTCAGGCTCAAGGCGTAAACGT
67
core
AATCATCAACCGAGGCAACCCGTATAAGGATCGGG
68
core
ACGCCATGAACGGTAATCGTAGAGATCTACAAAGGTAAAAAT
69
core
CTCATTTCATGTCAATCATATGGAGAGGGTAGCTATATATTT
70
core
ATTAAATGGTTGATAATCAGATCTAGCTGATAAATGAGTAAT
71
core
TAATATTCAAAAACAGGAAGAATCAATATGATATTTCAAAAG
72
core
ATATTTAAATTGATTAAGTTGGGT
73
core
ACAAGAGGTCATTGCCTGAGAGCCTTTATTTCAACAATACTT
74
core
TGAAAACTAGTTTAACCAGTAACATCGACTCTACCGAG
75
core
GCCGTACCCCTTTTGTTGCTATTACCAA
76
core
TTTTAGACCAAAAACATTATGCAATAAC
77
core
TAAATGCCATAAAGCTAAATCTTTCATTTGGGGCG
78
core
GGTGAGACAAGGCAAAGAATT
39
79
core
AGCAAAATAAAGATCAACCGTAAAGCCCTTGTTAAAGGGGGAGTTG
80
core
ATTTCGCAAATGGTACCCTGTGCAAGGACTATCAGAATCGATCAAA
81
core
CTGTTTAGCTATATGGTTGTAACCCTCATTTT
82
core
CGAGCTGCTCAGAGAATGCCTTAAT
83
core
TAGATTTAGTTTGAAACCAGAGCGTTTTAGGG
84
core
CCATATATACCTTTCATCAAACTGCGGACCCT
85
core
TTTTAAATGGCTTAGGTCTTTCTTTAAACAAA
86
core
AGGGCCCGAATAGACTGTAAAAACAAATCTATCAT
87
core
TAGAGAGACAGTTGATTCCCAATTCTGCCAAC
88
core
CCTTTTGCTGGAAGTTTCATT
89
core
TCAAAAATCAGAGCTTAATTG
90
core
GGTGCTTTTGCAGTCAGGATTTTAACAG
91
core
AAATGTTAGACTTCAAATATCCCGGAAGCAAACTCGAACGAG
92
core
TGAGTAAGAGCAGGTAGGAGTAGTCAAGAACAATC
93
core
ACGACGAGATAGCGTCCAATAAAGATTAAGAGGAATCAGGAT
94
core
CGTACTAACGATGGTTTCTTCATCTACTTAGGAGG
95
core
ATCATAAATCGTCATAAATATAGCAAAGCGGATTGAATTGCT
96
core
AGGCATAATCCCCCTCAAATGACCCTGACTATTATGTCATTT
97
core
AGGTGAGATTCCTGACGCCAAATCTCGCCTGCGAT
98
core
TACATAACGCCAGTTCAGAAA
99
core
CTTATGCGACGTTGGGAAGAACAAAATAGCGAGAGAATAGTA
100
core
TGTCAGGCGCAGACGGTAGGCACCTGAGGAC
101
core
ACTTTAACGTTAATAAAACGATTACCAG
102
core
GCTTGAGGAACAACATTATTACAACACT
103
core
CCAGAACAAAGATTCATCAGTAATTACG
104
core
GGCTTGCTAGGAATACCA
105
core
ATGCCGAACTAAATACGTGAGGAA
106
core
CGCACACTCAAAGACAGCATCGGAATATGACAACAACC
107
core
AACGAAACCGGAACTTTTTCACGTTGAAGGGA
108
core
CAATAGCAACGGCTACATTTCCAGTGCTAAA
109
core
CACTAAAGACCTGCAAAAAAAAGGCTCCGTTGCGCC
110
core
CCCCCAGATAAATTGCCTTTAATTGTATTTAA
111
core
TATGATCGTCACCCTCAACGCATAGCTTGATACCGATAAAAA
112
core
TCATAATGCCACTACGACAATCATAAAGGAATTGCGAACAAC
113
core
TAAAGACACGATCTTTCAGCGGAGTGAG
114
core
GGCCGCTTCGCTGATCGAGGTGAATTTCCGGTTTATGTATCAAACGTAA
115
core
AGTAAAGTTTTCACCAGTACAAACGGATAAG
116
core
CAACTTTTAATAATGAGGCGC
117
core
GACATTTCTGTATAATCTCCTCCATGT
118
core
ACAACCGATACCACCCTCATTTTCGGAGGTT
119
core
AAAGGAACAACTAAGGGAAAACGGTGTACAGACGAATTAC
120
core
GGAGTGTCGACGGATATTCATTACAGAAACA
121
core
CATTAAATGAACGAGGGAAGAATA
40
122
core
GAACCGCCACCCTCTCAGAAC
123
core
TGCCGTCAGAGGCTGAGACTCCCAGAATGGAAAGCGGTTGAG
124
core
GGGAGTTTCGTGTCGTCGAGGCTTAACCTAA
125
core
AAGTATATTCTGAAACATGAACTGAATTTACCGTTCCGCCGCCAG
126
core
TATCACCTGCCTATTTCGGAAGCGTCATACATGGCCCACCAG
127
core
TCAAGGGATAATCGCCCGCAGCGATCTTTGA
128
core
TAGTACCGTATAAACAGTTAAGATACAGGAGTGTAGAGCCAC
129
core
CCCAGAGCCATATTCGGTTTGCGGACCAAGC
130
core
AAGCTCAAGAACCAGGCTACAACGTAGCGTATTTT
131
core
GCAGGTCAACCGATTTGGGAAACCATTA
132
core
AACCACCCAATAATCAAAATCTATAAAA
133
core
CACCCTCCCAGAGCCCCTTATGACAGAA
134
core
AACCGCCCCTCCCTTCGGCATAGCGTCA
135
core
ATTACCAGAGCCAGTAACCTATTAGCCCGGAAACC
136
core
GCAAGGCGAAACAATAGCCGAACAAAGTTATT
137
core
AACCCGACTTGAGCCATTGAGGGATCACAAT
138
core
ATGAAACTAAGCCCAGGAAACCGAGGAAAAAGACAAATT
139
core
GATAAAGGTGAATTATCTGACGGACCACGGA
140
core
CCGTAATGATAACCAATAACGGAATACCGGCA
141
core
TCAAGTTATTGAGCACTGGCATGATTAAAGAA
142
core
TTTTTTCGGTCATAGCCCACCACCACATACATAAAGGTCAAAAGAGCTA
143
core
CAATAGAAAGCAGATGAAATAAAACGATAGTT
144
core
TTGGGGAAGGGACAGGAGCAGTCTAGTATTAGAGA
145
core
ATAAGTTTACCAGAAATAATACAAAAATCTTT
146
core
GAAACGCACGCAATCACAAGATTACAGACTTACCATTCTAAGCATT
147
core
ACAACCGGAAAGAGCCGTTTTGATTGCCCCCGTAC
148
core
AATGGAACCGACCCTCACTGGTAAAGTGCCCGCCA
149
core
CAAACGTGACTCCTAAAGTCAGGAGAATATTA
150
core
GAAAAGTAAATTCAGACATTCAGACGATATTA
151
core
ATAAACAGGGAGCTACATAGCGAATAATCGGATAGATA
152
core
AAATAAGGCAATAGCACCATTTTAGAGCCAGCAAAAAAGGGCTATGGTT
153
core
ACAGCCCAATGAACAAGCTGTCCACCAGTAACCGACCG
154
core
ATTTGCCTTTTTGTTTAACGTAGAGCAACGGA
155
core
CCAATATAGAACCAAGTACAACATTTAGGCACGTTAAA
156
core
CGAGCGTGAAAATAGCAGCCTATTGAGTCATC
157
core
CCTGAATGAGAATAACATAAATCAGAGACAGTAGCTAGCGTT
158
core
TGCACCCAAGCGCATTAGACGGAGGGTATGCC
159
core
GTTGGAGGTTGAGCATGAAAATAA
160
core
ATCATTACCGCAAATAAACAGC
161
core
TCATCGAAGCAAGCAAATCAGGAGCCTA
162
core
GTATTAAAGGCTTATCCGGTAACGCTAA
163
core
CCAGAACGCGTTAACAAGGAATCA
164
core
TCCTTATAACGCGAGGCGTTTATTTTAT
41
165
core
CAATCAACCTCCCGACTTGCGGCTATTT
166
core
AAGGTAAAGTAATTCAAGCCG
167
core
TTTCGAGGACGACGACAATAAACCGCAC
168
core
ATGTAATGTTCAGCTAATGCAAGAACGG
169
core
GCCATATCCTGTTTATCAACACTGTCTT
170
core
AGTAGGGAGTCCTGAACAAGATAGAAAC
171
core
GGTTTGAAATATAAGAGAATAT
172
core
TGTGATAAATAAGGGAGGCAT
173
core
TAAGAATAAACACCCGCCAAC
174
core
TAATTACTAGAAAAGAGAATC
175
FRET‐site‐A1
TTCAATTAATATCAAAAAACTATA
176
FRET‐site‐A2
TTAGTATCATATGCGCTCAAC
177
FRET‐site‐B1
CTAAAGTACGGTGTATAAGAGAGTCAGATCAT
178
FRET‐site‐B2
GTGTAGGTTAAGCAATAAAGCAAAAGGTGGCATCA
179
internal linker
CAGTACATAAATCAATAACGGTTGTGCTACTCCAGTTC
180
internal linker
AATTACCTTTTTTATTTGAATTTGTGCTACTCCAGTTC
181
internal linker
CATTTTTGAATGGCGTCAGTATTGTGCTACTCCAGTTC
182
internal linker
GCCAGTGCGCCAGCATCGGTGTTGTGCTACTCCAGTTC
183
internal linker
CGGCCTCCTCTCCGTGGGAACTTGTGCTACTCCAGTTC
184
internal linker
ATAGGCTGGCTGACAATTTCATTGTGCTACTCCAGTTC
185
internal linker
AAGAGTAATCTTGAAAATTGGTTGTGCTACTCCAGTTC
186
internal linker
GCAAGCCCAATAGGATAGGTGTTGTGCTACTCCAGTTC
187
internal linker
CATTTAAATATACCGTCAGTCACCATTGTGCTACTCCAGTTC
188
internal linker
TGCCATCTTTTCTTAGCAGCATTGTGCTACTCCAGTTC
189
internal linker
ACCAAGTAATTATTTGCACGTACCAGAATTGTGCTACTCCAGTTC
190
internal linker
TTTACGTTAGGTACCGTAACACTGTTGATATTTGTGCTACTCCAGTTC
191
guide staple 1
CTTAATTAGCCTGTTGTAAATGCTGATGTCAATAGCATCATGG
192
guide staple 2
TTGCGGATATGCAAATTCTACTAATAGTGCTGACGT
193
PDGF aptamer 1
CCAACGTTATACAAATTCTTATACTCAGGGCACTGCAAGCAATTGTGGTCCCAATGGGCTGAGTA
194
PDGF aptamer 2
AGTAGCATTAACATCCAATTACTCAGGGCACTGCAAGCAATTGTGGTCCCAATGGGCTGAGTA
195
PDGF complement 1
196
PDGF complement 2
197
random latch 1
TACTCAGCCCATTGGGACCACAAT ATTTTTTTTTTTTTTTTTTT TT AATGCTGTAGCTCAACATG TACTCAGCCCATTGGGACCACAAT ATTTTTTTTTTTTTTTTTTT TT AACCTCCGGCTTAGGTTGG AGTAGCATTAACATCCAATTAACAGGGTCGCCCATCGGTTCGAATCAGACGGTTTAAGGCAGT
198
random latch 2
CCAACGTTATACAAATTCTTATAACAGGGTCGCCCATCGGTTCGAATCAGACGGTTTAAGGCAGT
199
random complement 1
ACTGCCTTAAACCGTCTGATTCGAACCGATGGGCGACCCTGTTAAATGCTGTAGCTCAACATG
200
random complement 2
ACTGCCTTAAACCGTCTGATTCGAACCGATGGGCGACCCTGTTAAACCTCCGGCTTAGGTTGG
201
TE17 aptamer 1
202
TE17 complement 1
203
TE17 aptamer 2
204
TE17 complement 2
205
sgc8c aptamer 1
206
sgc8c complement 1
CCAACGTTATACAAATTCTTA TT CAGCTACGCAATACAAAACTCCGAACACCTGCTTCTGACTGGGTGCTG CAGCACCCAGTCAGAAGCAGGTGTATTTTTTTTTTTTTTTTTTTTTTT TT AACCTCCGGCTTAGGTTGG CCAACGTTATACAAATTCTTA TT CAGCTACGCAATACAAAACTCCGAACACCTGCTTCTGACTGGGTGCTG CAGCACCCAGTCAGAAGCAGGTGTA TTTTTTTTTTTTTTTTTTTTTTT TT AATGCTGTAGCTCAACATG CCAACGTTATACAAATTCTTA TT ATCTAACTGCTGCGCCGCCGGGAAAATACTGTACGGTTAGA TCTAACCGTACAGTATTTTCCCGG TTTTTTTTTTTTTTTTT TT AACCTCCGGCTTAGGTTGG
42
207
sgc8c aptamer 2
AGTAGCATTAACATCCAAT TT ATCTAACTGCTGCGCCGCCGGGAAAATACTGTACGGTTAGA
208
sgc8c complement 2
TCTAACCGTACAGTATTTTCCCGG TTTTTTTTTTTTTTTTT TT AATGCTGTAGCTCAACATG
209
guide removal strand 1
CCATGATGCTATTGACATCAGCATTTACAACAGGCTAATTAAG
210
guide removal strand 2
ACGTCAGCACTATTAGTAGAATTTGCATATCCGCAA
211
5' BioTEG linker comp.
/5BioTEG/GAACTGGAGTAGCAC
212
5' Amine linker comp.
/5AmMC6/GAACTGGAGTAGCAC
213
5' Thiol linker comp.
/5ThioMC6-D/GAACTGGAGTAGCAC
214
5' Dithiol linker comp.
/5DTPA/GAACTGGAGTAGCAC
43
