Exquisite Sequence Selectivity with Small Conditional RNAs

Transcription

1 Supplementary Information Exquisite Sequence Selectivity with Small Conditional RNAs Jonathan B. Sternberg and Niles A. Pierce,, Division of Biology & Biological Engineering, Division of Engineering & Applied Science, California Institute of Technology, Pasadena, CA, USA S

2 Contents S Materials and methods S S.Oligonucleotidesequences... S S. Oligonucleotide synthesis and purification S S.Reactionconditions... S S.Polyacrylamidegelelectrophoresis... S S Discrimination of RNA SBS cancer markers and wildtype sequences S S. Annotated single-channel scans for the gels in Figure S S. Further characterization of the s used in Figure S S.QuantificationofON/OFFratios... S S Enhancing selectivity via competitive inhibition with an unstructured scavenger S S. Further characterization of the s and scavenger used in Figure S S. Scavenger studies with additional SNPs S S Genotyping SNPs S S. Annotated single-channel scans for the gel in Figure S S Additional studies S S. Effect of SBS mutation location on selectivity S S. Effect of temperature on selectivity S S

3 S Materials and methods S. Oligonucleotide sequences Each family of -nt RNA targets comprises two or more -nt sequence variants differing by an SBS mutation or an SNP; for each target, the cognate h recognizes an 0-nt marker (a subsequence of the target) containing the mutation or polymorphism. Each HCR has a -nt toehold, a -bp stem, and a -nt hairpin loop. With the exception of the loop of the h and the complementary toehold of the h, the sequences are fully determined by the sequence of the marker (which is identical in sequence to the output domain of h). Hence, the majority of the sequence is designed by setting the register of h relative to the target. We optimized the timing of sequence selectivity for a timescale of choice by adjusting toehold strength for each of the two s relative to the toehold strength of previously tested s. Free energy calculations were performed using the Analysis page of the NUPACK web application. Increasing the energetic driving force for polymerization leads to selectivity at an earlier time point; decreasing the energetic driving force for polymerization leads to selectivity at a later time point. In cases where spontaneous polymerization was too rapid, we increased the stem length of the s to increase the strength of kinetic trapping. It can be useful to purposely include one or more wobble pairs or mismatches between h and the target in order to reduce their affinity, in which case the target no longer fully determines the sequence of the input domain for h. In this study, H U of Figure employs an h that forms a wobble pair with cognate target A (but a Watson Crick pair with h), and H C of Figure Sb employs an h that forms a wobble pair with cognate target G (as well as with h). Each -nt scavenger was designed by adjusting the register and length of the scavenger in order to tune the predicted free energy of hybridization with the target relative to that of previously tested scavengers. Sequences for RNA targets are shown in Table S. Sequences for RNA scavengers are shown in Table S. Sequences for s are shown in Table S. S. Oligonucleotide synthesis and purification Targets, s, and scavengers were synthesized by Integrated DNA Technologies (IDT). Oligonucleotides were resuspended in ultrapure water (Corning, Cat. # -000-CM) in our lab. Fluorescent s were ordered from IDT end-labeled either with a fluorophore or with an amine to permit subsequent coupling of a fluorophore. Dye coupling reactions were performed as follows: to label h of HCR system H C with Cy, we combined µl of sodium tetraborate decahydrate (0. M, ph.),. µl of amine-labeled in ultrapure water (. µm), and. µl of Cy in DMSO ( µg/µl; GE Healthcare, Cat. # PA00), mixing every 0 min for two hours and incubating in the dark at room temperature overnight; to label h of HCR system H U with Alexa, the same procedure was used except that the dye was used in its entirety (Life Technologies, Cat. # A); to label h of HCR system H C with Alexa, we combined. µl of 0 conjugation buffer (0. M Na HPO, 0 mm EDTA, ph.), 0 µl of amine-labeled in ultrapure water (0 µm), and. µl of DMSO containing the full sample of Alexa (Life Technologies, Cat. # A). Oligonucleotides were typically HPLC-purified by IDT. Otherwise, they were desalted by IDT and then purified using a 0% or % denaturing polyacrylamide gel. The band corresponding to the expected mobility was visualized by UV shadowing and excised from the gel. Oligonucleotides were then eluted by soaking in 0. M NaCl overnight and recovered by ethanol precipitation. After dye coupling reactions, s were purified from unincorporated dyes by ethanol precipitation and then resuspended in ultrapure water. Fluophore-labeled s were then purified from unlabeled s using the above strand purification procedure. S. Reaction conditions Oligonucleotides were stored at 0 C, brought to room temperature on the bench top, and quantified in PKR buffer (0 mm HEPES ph., mm MgCl, 00 mm KCl) by measuring absorbance at 0 nm. Prior to use, each oligonucleotide was snap-cooled (heat at C for 0 s followed by incubation on ice for 0 s and incubation at room temperature for a minimum of min) at a convenient concentration ranging from to 0 µm; snap-cooling is helpful in ensuring that h and h predominantly form hairpin monomers prior to the start of each HCR reaction. S

4 For reactions that contained multiple HCR systems, a master mix of the relevant s was created, adding all h s to the test tube, then expelling all h s onto the walls of the test tube, and finally mixing all s together. This procedure enabled the h and h s for each HCR system to be mixed at their diluted master mix concentrations to minimize spontaneous polymerization. In reactions that contained scavengers, targets were first introduced to the bottom of the reaction test tubes and then scavengers and s were expelled onto opposite sides of the test tube walls and mixed in together, providing scavengers and s simultaneous access to the targets. Reactions were typically carried out with targets and s at µm (scavengers at µm, when applicable) in PKR at C for h. Reaction volumes ranged from to 0 µl, with larger volumes favored for long reactions in order to minimize the influence of evaporation effects over the course of the experiment. To move the selectivity window earlier in time, s were used at µm for certain systems (for the BRAF wildtype-detecting system H A of Figures a, Sa, Sa and Sa and the SNP-detecting systems H C and H G of Figures and S). All targets for the SNP studies of Figures and S were also introduced at µm. Reactions for the mutant- and wildtype-detecting BRAF systems of Figures a, Sa, Sa and Sa were carried out for h. Multiple time points were tested for the kinetic discrimination study of Figure, the mutation location study of Figure S, and the temperature study of Figure S (which also tested multiple temperatures). Reactions for the temperature study were performed in the cold room. S. Polyacrylamide gel electrophoresis Following HCR reactions, loading buffer (0% glycerol, PKR) was added directly to the reaction test tubes or to a fraction of the reaction mixture placed in a new test tube. Typically, µl of this mixture was loaded into individual gel lanes. For the fluorophore-labeled mutant- and wildtype-detecting BRAF systems of Figures a, Sa, Sa and Sa, only. µl was loaded into gel lanes. Typically, -lane 0% native TBE gels (Bio-Rad, Cat. # - 0S) were run at 0 V for min at room temperature. Gels for the temperature study of Figure S were run in the cold room. For the quantification studies of Figure S, gels were run at 0 V for min at room temperature. Gels were imaged on an FLA-00 laser scanner (FujiFilm). Gels not containing fluorescently-labeled hairpins were post-stained with SYBR Gold (Invitrogen Cat. # S-) and imaged with a nm laser and a 0 ± 0 nm band-pass filter. Gels containing fluorescently-labeled hairpins were imaged using appropriate excitation lasers and emission filters (Alexa: nm/0 ±0 nm, Cy: nm/0 ±0 nm, Cy and Alexa: nm/ nm long-pass filter). The -channel gel of Figures and S was imaged with the above three channels on the FLA- 00 and then imaged on an Odyssey imaging system (Li-Cor) for the fourth channel (Alexa0: nm/0 nm long-pass filter). Scans from the two instruments were then merged using the TurboReg plugin of ImageJ. S

5 Table S. RNA target sequences. Sequence variants in orange. Name # (nt) Sequence ( 0 to 0 ) Figures GAUUUUGGUCUAGCUACAGAGAAAUCUCGAUGGAGU, S 0 GAUUUUGGUCUAGCUACAGUGAAAUCUCGAUGGAGU, S 00 GAUUUUGGUCUAGCAACAGUGAAAUCUCGAUGGAGU A GAUUUUGGUCUAGCUACAGAGAAAUCUCGAUGGAGU a, Sa, Sa, Sa U GAUUUUGGUCUAGCUACAGUGAAAUCUCGAUGGAGU a, Sa, Sa, Sa U UUGGUUUUAAAUUAUGGAGUAUGUUUCUGUGGAGAC b, Sb, Sb, Sb, Sb G UUGGUUUUAAAUUAUGGAGUAUGUGUCUGUGGAGAC b, Sb, Sb, Sb, Sb G CUGUAAAGCUGGAAAGGGAGGAACUGGUGUAAUGAU c,, Sc, Sc, Sc, S C CUGUAAAGCUGGAAAGGGACGAACUGGUGUAAUGAU c,, Sc, Sc, Sc, S A GAUUUUGGUCUAGCUACAGAGAAAUC, S, Sa G GAUUUUGGUCUAGCUACAGGGAAAUC, S, Sa C GAUUUUGGUCUAGCUACAGCGAAAUC Sa U GAUUUUGGUCUAGCUACAGUGAAAUC Sa A UUGGUUUUAAAUUAUGGAGUAUGUAUCUGUGGAGAC Sb C UUGGUUUUAAAUUAUGGAGUAUGUCUCUGUGGAGAC Sb A CUGUAAAGCUGGAAAGGGAAGAACUG Sc C CUGUAAAGCUGGAAAGGGACGAACUG Sc G CUGUAAAGCUGGAAAGGGAGGAACUG Sc U CUGUAAAGCUGGAAAGGGAUGAACUG Sc A CUGUAAAGCUGGAAAGGGAAGAACUGGUGUAAUGAU, S U CUGUAAAGCUGGAAAGGGAUGAACUGGUGUAAUGAU, S CACACACCACACCACACG Sab M CAGACACCACACCACACG Sa M CACAGACCACACCACACG Sa M CACACACCAGACCACACG Sa M CACACACCACACCAGACG Sa M CAUACACCACACCACACG Sb M CACAUACCACACCACACG Sb M CACACACCAUACCACACG Sb M CACACACCACACCAUACG Sb UGACCUGACUUGACUGAC Scd M UGACCUGACUUGACUUAC Sc M UGACCUGACUUUACUGAC Sc M UGACCUUACUUGACUGAC Sc M UUACCUGACUUGACUGAC Sc M UGACCUGACUUGACUAAC Sd M UGACCUGACUUAACUGAC Sd M UGACCUAACUUGACUGAC Sd M UAACCUGACUUGACUGAC Sd Table S. Scavenger sequences. Nucleotides complementary to target sequence variants in blue. Name # (nt) Sequence ( 0 to 0 ) Figures S C UUUCCCUGUAG, S, Sa S A AUUUCACUGUAG Sa S U AUUUCUCUGUAG Sa S C CAGACACAUAC Sb S U ACAGAUACAUAC Sb S

6 Table S. sequences. Nucleotides complementary to target sequence variants in lavender; toehold nucleotides are underlined. Name # (nt) Sequence ( 0 to 0 ) Figures H fast (h) CAGAGAAAUCUCGACAGAUCGAGAUUUCUCUGUAGC, S H fast (h) UCUGUCGAGAUUUCUCUGGCUACAGAGAAAUCUCGA, S H medium (h) UCUCUGUAGCUAGACCAACAAAUUGGUCUAGCUACA, S H medium (h) UUGGUCUAGCUACAGAGAUGUAGCUAGACCAAUUUG, S H slow (h) GAUUUCUCUGUAGCUAGAGACUUCUAGCUACAGAGA, S H slow (h) UCUAGCUACAGAGAAAUCUCUCUGUAGCUAGAAGUC, S H U (h) BRAF 0 UACAGAGAAAUCUCGAAGAUUCGAGAUUUCUCUGUAGCUA a, Sa, Sa, Sa H U (h) BRAF 0 AUCUUCGAGAUUUCUCUGUAUAGCUACAGAGAAAUCUCGA/iSp//CySp/ a, Sa, Sa, Sa H A (h) BRAF UCACUGUAGCUAGACCAAAUAGUUUUGGUCUAGCUACA a, Sa, Sa, Sa H A (h) BRAF UUUGGUCUAGCUACAGUGAUGUAGCUAGACCAAAACUA/iSp//CySp/ a, Sa, Sa, Sa H A (h) JAK UCCACAGAAACAUACUCCUCUUGGAGUAUGUUUCUG b, Sb, Sb, Sb H A (h) JAK /AlexN//iSp/GGAGUAUGUUUCUGUGGACAGAAACAUACUCCAAGA b, Sb, Sb, Sb H C (h) JAK GUAUGUGUCUGUGGAAGAAUCCACAGACACAUACUCCA b, Sb, Sb, Sb H C (h) JAK UUCUUCCACAGACACAUACUGGAGUAUGUGUCUGUGGA/iSp//AmMO/(Alexa) b, Sb, Sb, Sb H C (h) PTEN GAGGAACUGGUGUAAACAUACACCAGUUCCUCCCUU c,, Sc, Sc, Sc, S H C (h) PTEN UGUUUACACCAGUUCCUCAAGGGAGGAACUGGUGUA/iSp//AmMO/(Cy) c,, Sc, Sc, Sc, S H G (h) PTEN CAGUUCGUCCCUUUCCAGGAAACUGGAAAGGGACGA/iSp//CySp/ c,, Sc, Sc, Sc, S H G (h) PTEN CUGGAAAGGGACGAACUGUCGUCCCUUUCCAGUUUC c,, Sc, Sc, Sc, S H U (h) UCUCUGUAGCUAGACCAACAAAUUGGUCUAGCUACA, S, Sa H U (h) UUGGUCUAGCUACAGAGAUGUAGCUAGACCAAUUUG, S, Sa H A (h) UCACUGUAGCUAGACCAACAAAUUGGUCUAGCUACA Sa H A (h) UUGGUCUAGCUACAGUGAUGUAGCUAGACCAAUUUG Sa H A (h) UCCACAGAAACAUACUCCUCUUGGAGUAUGUUUCUG Sb H A (h) GGAGUAUGUUUCUGUGGACAGAAACAUACUCCAAGA Sb H C (h) GUAUGUGUCUGUGGAAGAAUCCACAGACACAUACUCUA Sb H C (h) UUCUUCCACAGACACAUACUGGAGUAUGUGUCUGUGGA Sb H C (h) CAGUUCCUCCCUUUCCAGGAAACUGGAAAGGGAGGA Sc H C (h) CUGGAAAGGGAGGAACUGUCCUCCCUUUCCAGUUUC Sc H G (h) CAGUUCGUCCCUUUCCAGGAAACUGGAAAGGGACGA Sc H G (h) CUGGAAAGGGACGAACUGUCGUCCCUUUCCAGUUUC Sc H U (h) CACUAGUUCUUCCCUUUCCUCAAGGAAAGGGAAGAACU/iSp//AmMO/(Alexa), S H U (h) GGAAAGGGAAGAACUAGUGAGUUCUUCCCUUUCCUUGA, S H A (h) /Alex0N//iSp/AAGGGAUGAACUGGUCAUCCAGUUCAUCCCUUUCCA, S H A (h) AUGACCAGUUCAUCCCUUUGGAAAGGGAUGAACUGG, S (h) CACCACACCACACGGAGACGUGUGGUGUGGUGUGUG Sab (h) UCUCCGUGUGGUGUGGUGCACACACCACACCACACG Sab (h) GUCAGUCAAGUCAGGUCAAAGAUGACCUGACUUGAC Scd (h) UGACCUGACUUGACUGACGUCAAGUCAGGUCAUCUU Scd S

7 S Discrimination of RNA SBS cancer markers and wildtype sequences S. Annotated single-channel scans for the gels in Figure Single-channel scans for the gels in Figure are shown in Figure S, including an additional lane depicting spontaneous polymerization in the absence of target. Migration of s, s, and is annotated for each of the six HCR systems. The complementarity of the b and b* domains within each h and h (see Figure ) enables formation of off-pathway s in which two h or h monomers hybridize to each other to form two duplex stems separated by an interior loop. Formation of these minor products is kept to a minimum by snap-cooling each species prior to the start of an experiment (Section S.), kinetically trapping most s in the desired monomer state. S. Further characterization of the s used in Figure For each of three SBS cancer markers (BRAF, JAK, PTEN), Figure S characterizes the selectivity of mutantdetecting and wildtype-detecting HCR systems. All HCR systems exhibit high selectivity for their cognate target both in test tubes that contain a single HCR system and in test tubes that contain both mutant-detecting and wildtypedetecting HCR systems. In some instances, competitive inhibition between s enhances selectivity for reactions containing two HCR systems (e.g., for BRAF, H A reduces off-target activation of H U (lane vs of Figure Sa middle panel) and H U reduces off-target activation of H A (lane vs of Figure Sa bottom panel)). S. Quantification of ON/OFF ratios For each of three SBS cancer markers (BRAF, JAK, PTEN), we quantified performance for mutant-detecting and wildtype-detecting HCR systems (H M and H WT ) in test tubes containing both HCR systems labeled with spectrally distinct fluorophores (red channel for H M, green channel for H WT ). Figure S shows both channels for each cancer marker; these data represent one of five replicate sets of experiments. For a given replicate, the normalized polymerization is calculated for each combination of HCR system (H M and H WT ) and target ( M and WT ) : " H M ( M )= r (r + r ) r tot (rtot + r tot) + r (r # + r ) r tot (rtot + r tot " ) H M ( WT )= r (r + r ) r tot (rtot + r tot) + r 0 (r # + r ) r0 tot (rtot + r tot " ) H WT ( M )= g (g + g ) g tot (gtot + g tot) + g (g # + g ) g tot (gtot + g tot " ) H WT ( WT )= g (g + g ) g tot (gtot + g tot) + g 0 (g # + g ) g0 tot (gtot + g tot) Here, r i denotes the integrated red intensity in the upper box labeled i and ri tot denotes the integrated red intensity in the combined upper and lower boxes labeled i (g i and gi tot are the analogous green intensities). The ON/OFF ratio for each target (γ( M ) or γ( WT )) is calculated for each replicate (Table S) as: γ( M )=H M ( M )/H WT ( M ), γ( WT )=H WT ( WT )/H M ( WT ). Here, γ() quantifies the ON state of the cascade with the cognate HCR system relative to the OFF state of the cascade with the off-target HCR system (comparing two channels within a single experiment). For BRAF, H M H U and H WT H A. For JAK, H M H A and H WT H C. For PTEN, M H C and H WT H G. For BRAF, M A and WT U. For JAK, M U and WT G. For PTEN, M G and WT C. S

8 The ON/OFF ratio for each HCR system (γ(h M ) or γ(h WT )) is calculated for each replicate (Table S) as: γ(h M )=H M ( M )/H M ( WT ), γ(h WT )=H WT ( WT )/H WT ( M ). Here, γ(h) quantifies the ON state of the cascade with the cognate target relative to the OFF state of the cascade with the off-target (comparing a single channel between two experiments). To characterize the distribution of measurements for a given ON/OFF ratio, we calculate the sample median over five replicates to provide a robust estimate of location (noting that the sample mean is not robust to outliers) and the sample median absolute deviation (MAD): MAD = median j γ j median j (γ j ) over five replicates to provide a robust estimate of scale (noting that the sample standard deviation is not robust to outliers)., Table S. ON/OFF ratio for each target ( M or WT ) for each of three cancer markers (BRAF, JAK, PTEN). BRAF UA JAK GU PTEN CG γ( M ) γ( WT ) γ( M ) γ( WT ) γ( M ) γ( WT ) Replicate Replicate Replicate Replicate Replicate Median Median absolute deviation Table S. ON/OFF ratio for each HCR system (H M or H WT ) for each of three cancer markers (BRAF, JAK, PTEN). BRAF UA JAK GU PTEN CG γ(h M ) γ(h WT ) γ(h M ) γ(h WT ) γ(h M ) γ(h WT ) Replicate Replicate Replicate Replicate Replicate Median Median absolute deviation S

9 a BRAF UA b JAK GU c PTEN CG H U H U H A A H U H A U H A H C U G H C H G G C H A H A H C H C H A H C H C H G H G Red channel H U Red channel H A Red channel H C Green channel H A Green channel H C Green channel H G Figure S. Annotated single-channel scans for the gels in Figure. Discrimination of RNA SBS cancer markers and wildtype sequences for (a) BRAF UA, (b) JAK GU, and (c) PTEN CG. Each reaction contains two HCR systems labeled with spectrally distinct fluorophores: one targeting the SBS cancer marker (red channel) and one targeting the corresponding wildtype sequence (green channel). For each HCR system, the fluorophore is carried by one of the two s (see Table S). Top: composite image. Middle: red channel. Bottom: green channel. S

10 a BRAF UA HU HU HU HA HA HA A U U A HU HA HU HA A HU HA U Red channel HU HA HA HA HC HC H C U G G U HA HC HA HC U HA HC G Red channel HA Green channel HA JAK GU b PTEN CG HC HC HC HG HG HG G C C G HC HC HC HG HG HG G C Red channel HC Green channel HC c Green channel HG Figure S. Further characterization of the s used in Figure. Discrimination of RNA SBS cancer markers and wildtype sequences for (a) BRAF UA, (b) JAK GU, and (c) PTEN CG. Each experiment contains a mutantdetecting HCR system (lanes ), a wildtype-detecting HCR system (lanes ), or both HCR systems (lanes ). Top: composite image. Middle: red channel (mutant-detecting HCR systems). Bottom: green channel (wildtype-detecting HCR systems). S0

11 a BRAF UA b JAK GU c PTEN CG H U H U H U H A H A H A A A H U H U H A H A U U H A H A H C U H C H A U H C H A H A H C H C G G H C H G H C G H C H G G H G H C H C H G H G C C Red channel H U Red channel H A Red channel H C Green channel H A Green channel H C Green channel H G Figure S. Quantification of ON and OFF states. Discrimination of RNA SBS cancer markers and wildtype sequences for (a) BRAF UA, (b) JAK GU, and (c) PTEN CG. ON and OFF states are quantified using reactions containing both a mutant-detecting HCR system (red channel) and a wildtype-detecting HCR system (green channel) (boxes,,,0). Gel background is quantified using reactions that contain neither HCR system (boxes,,,,,). Spontaneous leakage is depicted in box. S

12 S Enhancing selectivity via competitive inhibition with an unstructured scavenger S. Further characterization of the s and scavenger used in Figure Figure S characterizes the gel migration of reactants and intermediates for the scavenger system used in Figure. h U S C A G h U S C A G S C S C h U A G A G S C A G H U H U H U H U H U S C G 0 Figure S. Enhancing selectivity via competitive inhibition using an unstructured scavenger strand (see Figure ). Without scavenger, HCR system H U is not selective for cognate target A (forming Watson Crick pair U A) over -nt sequence variant G (forming nearly isoenergetic wobble pair U G) (lanes 0 and ). Scavenger S C is selective for G, restoring H U selectivity via competitive inhibition (lanes and ). HCR system H U comprises s h U and h. S. Scavenger studies with additional SNPs To explore the general utility of the scavenger concept, we tested six HCR systems (see Table S) designed to selectively detect the BRAF UA, JAK GU, and PTEN CG mutant and wildtype sequences against all four possible sequence variants at each mutation position. These case studies ( cognate target and off-targets for each of six HCR systems) turned up six -nt sequence variants that challenged the selectivity of HCR cascades; in each case, HCR selectivity was restored via competitive inhibition by the appropriate scavenger (Figure S). a BRAF: H U BRAF: H A b JAK: H A JAK: H C c PTEN: H C PTEN: H G Target: A C G U A G A U A C G U U A U G A C G U U A U G A C G U A C G U A C G U Scavenger: S C S C S A S A S U S U S C S C S U S U S C S C Figure S. Scavenger studies with additional SNPs. For each of three cancer markers, (a) BRAF UA, (b) JAK GU, and (c) PTEN CG, the selectivity of mutant-detecting and wildtype-detecting HCR systems was tested against all four possible sequence variants (A, C, G, U) at the mutation position. Blue lane numbers denote satisfactory OFF states. Red lane numbers denote suboptimal OFF states in the absence of scavenger, which are converted to satisfactory OFF states using the appropriate scavenger (red arrows to blue lane numbers). Strong ON states are observed with or without scavenger (green lane numbers). S

13 S Genotyping SNPs S. Annotated single-channel scans for the gel in Figure Single-channel scans for the gel in Figure are shown in Figure S, including annotation of s, s, and for each of four HCR systems. a target: H U + H G + H C + H A H U H U + H C A C G U A G G b target: H U + H G + H C + H A H U H U + H C A C G U A G G * c target: H U + H G + H C + H A H U H U + H C A C G U A G G d target: H U + H G + H C + H A H U H U + H C A C G U A G G e target: H U + H G + H C + H A H U H U + H C A C G U A G G s s Figure S. Annotated single-channel scans for the gel in Figure. (a) Green channel: H U. (b) Red channel: H G. (c) Blue channel: H C. (d) Yellow channel: H A. (e) Composite image. For each HCR system, the fluorophore is carried by one of the two s (see Table S). S

14 S Additional studies S. Effect of SBS mutation location on selectivity We were curious whether the location of an SBS mutation with respect to the toehold and stem of an h would affect the selectivity of the conditional hybridization cascade. The studies of Figure S did not yield conclusive trends; we include these results to provide a starting point for future work. Based on the evidence obtained to this point, it appears that positioning the SBS mutation so that it hybridizes to the half of the stem near the toehold (M in Figure S) is as good as or better than positioning the mutation elsewhere. a : CACACACCACACCACACG M: CAGACACCACACCACACG M: CACAGACCACACCACACG M: CACACACCAGACCACACG M: CACACACCACACCAGACG b : CACACACCACACCACACG M: CAUACACCACACCACACG M: CACAUACCACACCACACG M: CACACACCAUACCACACG M: CACACACCACACCAUACG M M M M M M M M M M M M M M M M M M M M M M M M min h 0 h min h 0 h c : UGACCUGACUUGACUGAC M: UGACCUGACUUGACUUAC M: UGACCUGACUUUACUGAC M: UGACCUUACUUGACUGAC M: UUACCUGACUUGACUGAC d : UGACCUGACUUGACUGAC M: UGACCUGACUUGACUAAC M: UGACCUGACUUAACUGAC M: UGACCUAACUUGACUGAC M: UAACCUGACUUGACUGAC M M M M M M M M M M M M M M M M M M M M M M M M min h 0 h min h 0 h Figure S. Effect of SBS mutation location on selectivity. For each of two cognate targets, an HCR system was designed to discriminate against two families of -nt mismatches. For each family of mismatches, the mutation and nearest neighbors are constant (panel a: ACA!AGA, panel b: ACA!AUA, panel c: UGA!UUA, panel d: UGA!UAA). All targets are predicted to be relatively free of secondary structure. The hybridization site of the toehold of h is noted with a line above the cognate sequence. S

15 S. Effect of temperature on selectivity To examine the effect of temperature on selectivity, we tested the HCR systems used in Figure (H fast, H medium, and H slow ) with cognate target and -nt sequence variant 0 over a range of temperatures from C to C. Each HCR system exhibits selectivity over a range of temperatures. The timing of the selectivity window shifts to shorter times with increasing temperature and longer times with decreasing temperature. These data demonstrate that -nt selectivity can be engineered at a wide range of timescales and temperatures. a H fast Temperature ( C): min h 0 h b H medium Temperature ( C): min h 0 h c H slow Temperature ( C): min h 0 h Figure S. Effect of temperature on selectivity. The HCR systems (a) H fast, (b) H medium, and (c) H slow were tested with cognate target and -nt sequence variant 0 over a range of temperatures from C to C. S

16 References () Zadeh, J. N.; Steenberg, C. D.; Bois, J. S.; Wolfe, B. R.; Pierce, M. B.; Khan, A. R.; Dirks, R. M.; Pierce, N. A. J. Comput. Chem. 0,, 0. () Choi, H. M. T.; Beck, V. A.; Pierce, N. A. ACS Nano 0,,. () Maronna, R.; Martin, D.; Yohai, V. Robust Statistics: Theory and Methods; Wiley: Chichester, England, 00. () Huber, P.J.; Ronchetti, E.M. Robust Statistics; Wiley: Hoboken, New Jersey, 00. S