First generation iSBH-sgRNAs detect short RNA triggers in HEK293T cells. A. Native sgRNA sequences are composed of spacer and scaffold sequences (Jinek et al. (2012)). iSBH-sgRNAs fold into complex secondary structures that interfere with the Cas9 ability to recognise target DNA sequences (OFF-state, Ferry et al. (2017)). iSBH-sgRNAs were designed by extending the 5’ end of the spacer sequence with a 14nt loop and a spacer* sequence partially complementary with the spacer. Bulges were also introduced within the iSBH-sgRNA sequence in order to ensure that the interaction between the spacer* and RNA trigger is more energetically favourable. In the ON-state, iSBH-sgRNAs recognise complementary RNA triggers and become activated, enabling Cas9 to perform its function. Short RNA triggers are complementary with the iSBH-sgRNA loop and spacer* sequence. B. Inside cells, RNA triggers are expected to bind to complementary iSBH-sgRNAs, inducing iSBH-sgRNA activation. Activated iSBH-sgRNAs are recognised by CRISPRa effectors and drive ECFP production from a fluorescent reporter. In this particular example, activated iSBH-sgRNAs interact with dCas9-Vp64 (Maeder et al. (2013)) and drive ECFP production from an 8xCTS-ECFP reporter (Nissim et al. (2014)). Following reporter induction, ECFP production could be monitored by Flow Cytometry. C. Starting from five different sgRNA spacer sequences, we designed 5 different iSBH-sgRNA sequences. For each iSBH-sgRNA, corresponding RNA triggers and 8xCTS-ECFP reporters were also designed. Ability of first-generation iSBH-sgRNA designs to drive expression of the ECFP reporter was assessed in the absence or presence of complementary RNA triggers. Experiments were carried out using dCas9-Vp64 and 8xCTS-ECFP reporters. D. An orthogonality test was performed, in which the 5 iSBH-sgRNA designs were tested against all 5 RNA triggers. Activation is only detected in the presence of matching iSBH-sgRNA and RNA trigger pairs. Figure shows mean +/− standard deviation values measured for 3 biological replicates. Values above bars represent fold turn-on values for iSBH-sgRNA activation (blue) and p-values (black) determined through unpaired t-tests. Figure 1—figure supplement 1. First generation iSBH-sgRNAs detect short RNA triggers in HEK293T cells.

Second generation iSBH-sgRNAs detect longer RNA triggers in HEK293T cells. A. Longer RNA triggers complementary with first-generation iSBH-sgRNAs have a 34nt sequence complementary with the loop and spacer* iSBH-sgRNA sequences. Triggers also have a 100nt flanking sequence immediately downstream from the iSBH-sgRNA complementary region. B. Second-generation designs contain a longer hairpin structure. A 10nt extension region was inserted between the spacer and loop sequences. This enabled increasing the size of the backfold sequence to 30nt. Longer RNA triggers complementary with 2nd generation iSBH-sgRNAs were designed, including triggers with 100nt 3’ flanks, 100nt 5’ flanks as well as 100nt 5’ and 3’ flanks. All trigger designs contain 44nt sequences complementary with the loops and backfold of the second-generation iSBH-sgRNAs. C. Ability of first-generation and second-generation iSBH-sgRNAs to sense 100nt 3’ flank triggers was assessed. D. Ability of second-generation iSBH-sgRNAs to detect different triggers with 100nt 5’ flanks and 100nt 5’ and 3’ flanks was assessed. Figure shows mean +/− standard deviation values measured for 3 biological replicates. Values above bars represent fold turn-on values for iSBH-sgRNA activation (blue) and p-values (black) determined through unpaired t-tests. Figure 2—figure supplement 1. CRISPRa reporters of choice influence ON/OFF ratios of second generation iSBH-sgRNA designs while detecting short RNA triggers.

Modular iSBH-sgRNA designs enable spatial separation of spacer and trigger-sensing sequences. A. In second-generation iSBH-sgRNAs, RNA triggers are complementary with the iSBH-sgRNA backfolds, thus sgRNA spacers influence RNA trigger sequences. In modular iSBH-sgRNAs, design constrains were eliminated as triggers are only complementary with the iSBH-sgRNA loop and first 15nt of the backfold. To increase affinity between iSBH-sgRNAs and RNA triggers, we increased loop sizes. Separation between trigger-sensing and spacer sequences was also achieved by reducing the complementary between the spacer sequence and CTS from 20 to 17nt. B. MODesign enables users to design modular iSBH-sgRNAs starting from input RNA triggers, sgRNA spacers and loop sizes. MODesign calculates the size of trigger-sensing sequences and creates a list of trigger sub-sequences having that size. Script determines the reverse complement of these sequences that could act as trigger-sensing sequences. iSBH-sgRNAs are assembled through adding spacer*, trigger-sensing sequences, extension, spacer and scaffold sequences. Extension sequences are engineered to be partially complementary with trigger-sensing sequences. Before producing a list of output sequences, iSBH-sgRNA folding is checked using NuPACK (Allouche (2012)). Simulations could result in multiple modular iSBH-sgRNA designs. Designs chosen for experimental validation were selected based on the probability of folding into the iSBH-sgRNA structure and lack of trigger secondary structures in the iSBH-sgRNA complementary region. Priority was also given to iSBH-sgRNAs that, by chance, displayed extra complementarity between RNA triggers and the last 15nt of the backfold or more than 17nt complementarity with the CTS. C. MODesign simulations were carried out for designing iSBH-sgRNAs capable of sensing trigger RNA D (146nt eRNA sequence). In each simulation, a different sgRNA sequence was used and a desired loop size of 14nt was kept constant between simulations. Selected designs were transfected to HEK293T cells together with the RNA trigger D sequence (expressed from a U6 promoter). Tests were carried out using dCas9-Vp64 and 8xCTS-ECFP reporters. D. MODesign simulations were run for designing iSBH-sgRNAs capable of sensing trigger RNA A (146nt repetitive RNA sequence), trigger RNA B (267nt repetitive RNA sequence), trigger RNA C (268nt repetitive RNA sequence) and trigger RNA D (146nt eRNA sequence). Tests were performed using different CRISPRa effectors. E. 4 modular iSBH-sgRNAs (A,B,C and D) were co-transfected to HEK293T cells and all iSBH-sgRNA: RNA trigger combinations were tested. Figure shows mean +/− standard deviation values measured for 3 biological replicates. Values above bars represent fold turn-on values for iSBH-sgRNA activation (blue) and p-values (black) determined through unpaired t-tests. Figure 3—figure supplement 1. Modular iSBH-sgRNA designs enable spatial separation of spacer and trigger-sensing sequences. Figure 3—figure supplement 2. Modular iSBH-sgRNAs are specifically activated by complementary RNA triggers.

Insights into the mechanism of iSBH-sgRNA activation. A. Interaction between iSBH-sgRNAs and RNA trigger leads to the formation of long double-stranded RNA structures. A potential activation mechanism might involve RNA strand displacement and formation of stable molecular complexes between the iSBH-sgRNA and the RNA trigger sequence. Supposing this scenario is correct, the size of the iSBH-sgRNA and RNA triggers are expected to remain constant after activation. A second scenario involves double-stranded RNA processing. If this is correct, iSBH-sgRNAs and RNA trigger sequences are expected to be truncated. B. iSBH-sgRNA circularisation assay. Cells were transfected with system components, followed by RNA extraction and ligation. Reverse transcription (RT) was performed on circular RNAs by using RT primers complementary with the sgRNA scaffold. The size of the RT products was determined by two sequential PCR reactions. PCR primers annealed with the scaffold2 and scaffold1 sequences, which are the scaffold sequences found downstream and upstream from the RT primer. For a full-length iSBH-sgRNA sequence, a second PCR product of 137bp is expected, while for a non-engineered native sgRNA, an 81nt product is expected. C. Determining the size of the iSBH-sgRNA after activation. Assays were performed in the presence or absence of complementary 44nt, short RNA triggers and dCas9-Vp64. Non-engineered, native sgRNA controls were also included. D. RNA trigger circularisation assay. After transfection, RNA extraction and RT, RNA trigger size was determined by nested PCR. PCR primers annealed with the 100nt 3’ flank2 and 100nt 3’ flank1 sequences, which are the flank sequences downstream and upstream from the RT primer. For full-length RNA triggers, 220bp PCR bands are expected. E. Determining the size of the RNA triggers after activation. Assays were performed in the presence or absence of a complementary iSBH-sgRNAs and dCas9-Vp64. Figure 4—figure supplement 1. Sequencing results for iSBH-sgRNA circularisation assays.

Testing the ability of second generation iSBH-sgRNA designs to detect short RNA triggers in vivo. A. Transgenic lines encoding dCas9-Vp64 and 8xCTS-ECFP reporters were created. Embryos resulting from in crossing first generation (F1) transgenics were injected with second generation chemically synthesised iSBH-sgRNAs and RNA triggers. B. Second-generation iSBH-sgRNAs were injected into transgenic zebrafish embryos with or without corresponding short RNA triggers. In the absence of RNA triggers (iSBH-sgRNA OFF), embryos are expected to display no ECFP signals, while trigger presence (iSBH-sgRNA ON) should promote ECFP expression. C. Figure presents our second strategy for chemically modifying iSBH-sgRNAs. This strategy involved protecting the iSBH-sgRNA 5’ end as well as the 5’end of the sgRNA spacer. These modifications were used together with sgRNA scaffold modifications. D. In order to quantify the impact of RNA triggers on iSBH-sgRNA activation, we grouped fish according to the intensity of ECFP signals. At 3 days post-fertilisation, embryos displaying no, low or high ECFP expression were counted. E. Embryos injected with iSBH-sgRNAs and non-complementary (iSBH-sgRNA OFF) of complementary RNA triggers (iSBH-sgRNA ON) were scored according to their ECFP intensity. Row number counts determined for 3 experimental replicates are displayed as part of Chi2 contingency tables. P values displayed were determined using Chi2 test. F Figure shows percentage of embryos recovered in each category for the 3 experimental replicates. Percentage of embryos with no ECFP expression varied between the 3 experimental replicates. This was due to the fact that both 8xCTS-ECFP and dCas9-VP64 transgenes are necessary for successfully expressing ECFP. These alleles segregate in a Mendelian fashion and our adult transgenic fish encode variable copy numbers of the transgene. For each individual replicate, we used embryos with identical genetic backgrounds for testing the iSBH-sgRNA (OFF) and iSBH-sgRNA (ON) conditions. Nevertheless, genetic backgrounds were different between the 3 experimental replicates. Figure 5—figure supplement 1. Optimising sgRNA delivery to zebrafish embryos. Figure 5—figure supplement 2. Testing different iSBH-sgRNA chemical modifications in vivo.