Programmable siRNA and its use

Abstract

To provide programmable, conditionally activated siRNA constructs (Cond-siRNAs) in an OFF state.SOLUTION: Provided is a programmable, conditionally activatable siRNA (Cond-siRNA) construct comprising a sensor strand, a core strand, and a guide strand, in which (i) the sensor strand and the core strand form a sensor duplex, (ii) the guide strand and the core strand form a RNAi duplex, and (iii) the sensor duplex is attached to the RNAi duplex to form a single structure.SELECTED DRAWING: None

Description

[Technical Field]

[0001] Claim of priority This application claims priority to U.S. Provisional Patent Application No. 62 / 717,686 filed on 10 August 2018 and U.S. Provisional Patent Application No. 62 / 811,183 filed on 27 February 2019, the contents of which, including the drawings, are incorporated herein by reference in their entirety.

[0002] Government ownership statement This invention was made in part with the support of the Government under grant numbers NSF EFRI-ODISSEI 1332411 and CMMI-SNM 1120890, awarded by the National Science Foundation (NSF). The Government has certain rights in this invention. [Background technology]

[0003] Nucleic acid nanotechnology 1-3 and biomolecular calculations 4 The long-standing goal is the development of conditionally activated oligonucleotide therapeutics that can detect and respond to the cellular expression of specific genes. 3,4 Nucleic acid switches based on toehold-mediated strand displacement. 5,6 It performs logical operations and bacterial cells 7 and mammalian cells 3,8 Although RNA transcripts have been detected, the conditional activation of oligonucleotide drugs by RNA transcripts in mammalian cells has not been convincingly demonstrated. Notable challenges include poorly suppressed background drug activity, weak on-state drug efficacy, input and output sequence duplication, high design complexity, short device lifetime (<24 hours), and required high device concentrations (>10 nM).

[0004] Over the past decade, synthetic RNAi triggers, such as small interfering RNAs (siRNAs), have been used. 10has become a familiar tool in biological research, and extensive preclinical and clinical development efforts have recently culminated in the FDA approval of Onpattro, the first RNAi drug 11 . Despite a rapidly growing drug development pipeline and a broad overview of excipients and delivery technologies that target ligands 9 , the difficulty of delivering RNAi agents to specific populations of disease-related cells continues to limit the potential of RNAi therapies. Repeated attempts over the past 15 years to develop programmable, conditionally activated RNAi agents based on strand displacement switches 12-15 have not convincingly demonstrated the intended effect, despite significant progress 3,8,16-19 . Accordingly, novel conditionally activated siRNAs (Cond-siRNAs) are provided herein to overcome the problems in the art

SUMMARY OF THE INVENTION

[0005] In one aspect, the disclosure relates to an off-state, programmable, conditionally activated siRNA construct, the construct comprising a sensor strand, a core strand, and a guide strand, wherein the sensor strand and the core strand complementarily bind to form a sensor double helix, the guide strand and the core strand complementarily bind to form an RNAi double helix, and the sensor double helix and RNAi double helix are linked to each other to form a single structure. In some embodiments, the sensor double helix and RNAi double helix are linked to each other via the core strand. In some embodiments, the sensor strand complementarily binds to 5' and 3' fragments of the core strand to form a sensor double helix, and the guide strand complementarily binds to a central fragment of the core strand to form an RNAi double helix, the central fragment not comprising the 5' or 3' sequence of the core strand, such that the sensor double helix and RNAi double helix are linked to each other via two different fragments of the core strand. In some embodiments, the sensor domain of the Cond-siRNA construct includes a sensor duplex formed by complementary binding of the 3' and 5' fragments of the sensor strand and the core strand, as well as a sensor overhang that is not paired with the core strand. The sensor overhang is located at either the 3' or 5' end of the sensor strand. In some embodiments, the sensor strand, core strand, and guide strand form a single construct through self-assembly upon contact with each other. In some embodiments, the sensor duplex contains 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, or 30 base pairs. In some embodiments, the RNAi duplex contains 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, or 30 base pairs.

[0006] In some embodiments, the Cond-siRNA is chemically modified to further improve its off-state stability and / or its dissociation efficiency when turned on upon contact with the input strand. For example, the bases of the sensor strand in the double-stranded region are modified by LNA modification, 2'-O-methyl modification, or both, but not by phosphorothioate (PS) modification; one or both ends of the core strand are modified by PS modification, 2'-O-methyl modification, or their ends; or the single-stranded overhang of the sensor strand is modified by LNA modification, 2'-O-methyl modification, PS modification, or any combination thereof. In some embodiments, the sensor domain and RNAi domain are modified by different chemical modifications.

[0007] The off-state Cond-siRNA is activated or turned on by contacting the input strand, where one end of the input strand, along with the sensor strand overhang, forms a foothold, inducing the replacement of the sensor strand from the core strand via complementary binding between the input strand and the sensor strand, thereby forming a discarded duplex, and the RNAi duplex completely dissociates from this construct.

[0008] This application includes at least one drawing created in color. [Brief explanation of the drawing]

[0009] [Figure 1a] Figure 1 shows the conceptual design, operation, and molecular dynamics simulation of Cond-siRNA. Figure 1a shows the conceptual secondary and tertiary structures of Cond-siRNA. The docking of the RNAi double helix to the (X-ray analyzed) crystal structure of Dicer from the genus Giardia shows a large steric clash between the sensor double helix and Dicer. [Figure 1b-d]Figure 1 shows the conceptual design, operation, and molecular dynamics simulation of Cond-siRNA. Figure 1b shows RNAi activation by strand substitution. When complementary input RNA encounters Cond-siRNA (I), the input strand forms a foothold with the 3' or 5' single-strand overhang on the sensor strand (II), resulting in strand substitution (III). The substitution separates the sensor strand from the RNAi double helix (IV). A cellular nuclease removes the core strand overhang on the RNAi double helix (V), leaving an active RNAi trigger for processing by Dicer (VI). Figures 1c and 1d show molecular dynamics (MD) optimized models of constructs I.1 and III.1. Green arrows indicate the direction of strand substitution from the 3' and 5' ends of the sensor strand. [Figure 2] Figure 2 (SEQ ID NOs. 77 and 113) shows the intracellular degradation of chemically modified single-stranded overhangs. Northern blots show the stepwise degradation of phosphorothioate (PS)-protected 5' overhangs on the test constructs. HCT116 cells were transfected with the test constructs (Dicer substrates with segmented passenger strands and various 5' overhangs) over 24 hours. All RNA was extracted and analyzed by Northern blot. Analysis of two sets of similar samples showed similar results at different loading concentrations. [Figure 3a-1] Figure 3a-1 (Sequence IDs 1-12) shows the array diagram and map of the optimization regions of the prototype construct. [Figure 3a-2] Figure 3a-2 (Sequence IDs 1-12) shows the array diagram and map of the optimization regions of the prototype construct. [Figure 3b-1] Figure 3b-1 (Sequence IDs 13-36) shows the array diagram and map of the optimization regions for the construct II variant. [Figure 3b-2] Figure 3b-2 (Sequence IDs 13-36) shows the array diagram and map of the optimization regions for the construct II variant. [Figure 3c] Figure 3c (Sequence IDs 13-36) shows the array diagram and map of the optimization regions for the construct II variant. [Figure 3d]Figure 3d (Sequence IDs 37-60) shows the array diagram and map of the optimization regions for the construct III variant. [Figure 3e-1] Figure 3e-1 (Sequence IDs 37-60) shows the array diagram and map of the optimization regions for the construct III variant. [Figure 3e-2] Figure 3e-2 (Sequence IDs 37-60) shows the array diagram and map of the optimization regions for the construct III variant. [Figure 3f-1] Figure 3f-1 (Sequence IDs 61-69) shows the array diagram and map of the optimization regions for the construct IV variant. [Figure 3f-2] Figure 3f-2 (Sequence IDs 61-69) shows the array diagram and map of the optimization regions for the construct IV variant. [Figure 4a-b] Figures 4a and 4b show Northern blots of input RNA in HCT 116 cells. Northern blot assays probe tat / rev and AML input RNA recovered from HCT 116 cells after 48 hours. Figure 4a shows tat / rev RNA transcripts probed with mutant tRNALys3 that matches their common leader sequence. Lanes: (L) Ambion decade marker; (0) Negative control using RNA from mock transfection; (1) Perfectly matched input RNA; (2) 5' mismatched input RNA; (3) Perfectly mismatched input RNA; (4) Double-strand mismatched activator (not used); (5) 3' mismatched activator. The expected size of the input RNA was 145–150 nt. Figure 4b shows CBFβ-MYH11 RNA transcripts. Lanes: (0) Mock transfection; (1) Tat / rev fully matched input RNA (for comparison); (2) CBFβ-MYH11 fusion; (3) MYH11 parent; (4) MYH11 parent. Consecutive panels show the same samples probed with mutant tRNALys3 probe, MYH11 probe, and CBFβ probe. Input RNA expression levels were comparable across all cohorts. [Figure 5a-f]Figures 5a-5f show the assembly, strand substitution, RNAi activity, and RNAi activation of constructs I, II, and III variants. Figures 5a and 5d show non-denaturing PAGE and isothermal strand substitution at 37°C in 1×PBS buffer for constructs I and II assemblies. Constructs I and II are disassembled by their respective input strands and are unaffected by mismatched input strands. The control lane is I=input strand, C=construct, P=RNAi double helix, and W=discarded double helix. Figures 5b and 5e show the RNAi activity of the constructs in the presence or absence of mismatched input strands, and the on-state construct in the presence of mismatched input strands. Figures 5c and 5f show the RNAi activity of the off-state construct in cells expressing the indicated input strands. [Figure 5g-j] Figures 5g and 5h show the off- and on-state RNAi activity of construct II variants in cells expressing mismatched input strands. Figures 5i and 5j show the off-state RNAi activity of off-state construct II and III variants with different core strand modifications in cells expressing either mismatched or matched input strands. [Figure 5k] Figure 5k shows the RNAi activity of construct III variants with different sensor strand modifications in cells with matched or mismatched input strands. [Figure 6]Figure 6 shows the processing of pre-activated Cond-siRNA constructs by Dicer. Northern blot assay probes Cond-siRNA guide strands recovered from HCT 116 cells after 48 hours. The lanes are as follows: (L) Ambion decade marker; (0) RNA from mock-transfected cells; (1) and (2) guide strands from a third prototype Cond-siRNA not disclosed in this specification; (3) and (4) prototype HIV constructs in off and on states; (5) and (6) AML constructs in off and on states. Arrows indicate the location of Dicer-cleaved guide strands. The product of Dicer cleavage (guide strand fragments of approximately 21 nt) was detected in RNA material extracted from cells transfected with on-state Cond-siRNA, but not in RNA material extracted from cells transfected with off-state Cond-siRNA. [Figure 7] Figure 7 shows the base pair parameters for 23 base pairs in each helix, measured every 5 ps using x3DNA43 and plotted as a heatmap. Base pairs are numbered from top (1) to bottom (23). The definition of base pair parameters is shown on the website x3dna.org / articles / seeing-is-understanding-as-well-as-believing. [Figure 8a] Figure 8a shows the base pair parameters obtained from 5-nanosecond molecular dynamics simulations of the sensor and siRNA double helix. The twist parameters of base pairs 1-22 relative to subsequent base pairs are measured. The heat scale for each measurement is centered on the mean of the control MD trajectory measured for 5-nanoseconds for four unmodified RNA double helixes having the same base sequence as the siRNA and sensor double helixes of the HIV and AML constructs. The heat scale extends to three standard deviations measured on the same control trajectory. The inventors observed significant deviations from normal RNA values. In these plots, construct 1 is the HIV construct and construct 2 is the AML construct. [Figure 8b]Figure 8b shows the base pair parameters obtained from 5-nanosecond molecular dynamics simulations of the sensor and siRNA double helix. The rise parameters of base pairs 1-22 relative to subsequent base pairs are measured. The heat scale for each measurement is centered on the mean of control MD trajectories measured for 5-nanoseconds using four unmodified RNA double helixes with identical base sequences to the siRNA and sensor double helixes of the HIV and AML constructs. The heat scale extends to three standard deviations measured using the same control trajectory. The inventors observed significant deviations from normal RNA values. In these plots, construct 1 is the HIV construct and construct 2 is the AML construct. [Figure 8c] Figure 8c shows the base pair parameters obtained from 5-nanosecond molecular dynamics simulations of the sensor and siRNA double helix. The roll parameters of base pairs 1-22 relative to subsequent base pairs are measured. The heat scale for each measurement is centered on the mean of the control MD trajectory measured for 5-nanoseconds of four unmodified RNA double helixes having the same base sequence as the siRNA and sensor double helixes of the HIV and AML constructs. The heat scale extends to three standard deviations measured on the same control trajectory. The inventors observed significant deviations from normal RNA values. In these plots, construct 1 is the HIV construct and construct 2 is the AML construct. [Figure 8d] Figure 8d shows the base pair parameters obtained from 5-nanosecond molecular dynamics simulations of the sensor and siRNA double helix. Slide parameters for base pairs 1-22 relative to subsequent base pairs are measured. The heat scale for each measurement is centered on the mean of control MD trajectories measured for 5-nanoseconds of four unmodified RNA double helixes having the same base sequence as the siRNA and sensor double helixes of the HIV and AML constructs. The heat scale extends to three standard deviations measured on the same control trajectory. The inventors observed significant deviations from normal RNA values. In these plots, construct 1 is the HIV construct and construct 2 is the AML construct. [Figure 8e]Figure 8e shows the base pair parameters obtained from 5-nanosecond molecular dynamics simulations of the sensor and siRNA double helix. The shift parameters of base pairs 1-22 relative to subsequent base pairs are measured. The heat scale for each measurement is centered on the mean of the control MD trajectory, measured for 5-nanoseconds on four unmodified RNA double helixes with the same base sequences as the siRNA and sensor double helixes of the HIV and AML constructs. The heat scale extends to three standard deviations measured on the same control trajectory. The inventors observed significant deviations from normal RNA values. In these plots, construct 1 is the HIV construct and construct 2 is the AML construct. [Figure 8f] Figure 8f shows the base pair parameters obtained from 5-nanosecond molecular dynamics simulations of the sensor and siRNA double helix. The propeller parameters are defined for base pairs 1-23. The heatscale for each measurement is centered on the mean of control MD trajectories measured for 5-nanoseconds on four unmodified RNA double helixes with identical base sequences to the siRNA and sensor double helixes of the HIV and AML constructs. The heatscale extends to three standard deviations measured on the same control trajectory. The inventors observed significant deviations from normal RNA values. In these plots, construct 1 is the HIV construct and construct 2 is the AML construct. [Figure 8g] Figure 8g shows the base pair parameters obtained from 5-nanosecond molecular dynamics simulations of the sensor and siRNA double helix. Shear parameters are defined for base pairs 1-23. The heatscale for each measurement is centered on the mean of control MD trajectories measured for 5-nanoseconds on four unmodified RNA double helixes with the same base sequences as the siRNA and sensor double helixes of the HIV and AML constructs. The heatscale extends to three standard deviations measured on the same control trajectory. In these plots, construct 1 is the HIV construct, and construct 2 is the AML construct. [Figure 8h]Figure 8h shows the base pair parameters obtained from 5-nanosecond molecular dynamics simulations of the sensor and siRNA double helix. Stagger parameters are defined for base pairs 1-23. The heat scale for each measurement is centered on the mean of control MD trajectories measured for 5-nanoseconds on four unmodified RNA double helixes with the same base sequences as the siRNA and sensor double helixes of the HIV and AML constructs. The heat scale spans three standard deviations measured on the same control trajectory. In these plots, construct 1 is the HIV construct and construct 2 is the AML construct. [Figure 8i] Figure 8i shows the base pair parameters obtained from 5-nanosecond molecular dynamics simulations of the sensor and siRNA double helix. Stretch parameters are defined for base pairs 1–23. The heatscale for each measurement is centered on the mean of control MD trajectories measured for 5-nanoseconds on four unmodified RNA double helixes with identical base sequences to the siRNA and sensor double helixes of the HIV and AML constructs. The heatscale spans three standard deviations measured on the same control trajectory. In these plots, construct 1 is the HIV construct and construct 2 is the AML construct. [Figure 8j] Figure 8j shows the base pair parameters obtained from 5-nanosecond molecular dynamics simulations of the sensor and siRNA double helix. The tilt parameters of base pairs 1-22 relative to subsequent base pairs are measured. The heat scale for each measurement is centered on the mean of the control MD trajectory measured for 5-nanoseconds for four unmodified RNA double helixes with the same base sequences as the siRNA and sensor double helixes of the HIV and AML constructs. The heat scale extends to three standard deviations measured on the same control trajectory. In these plots, construct 1 is the HIV construct and construct 2 is the AML construct. [Figure 8k]Figure 8k shows the base pair parameters obtained from 5-nanosecond molecular dynamics simulations of the sensor and siRNA double helix. Buckle parameters are defined for base pairs 1–23. The heat scale for each measurement is centered on the mean of control MD trajectories measured for 5-nanoseconds on four unmodified RNA double helixes with the same base sequences as the siRNA and sensor double helixes of the HIV and AML constructs. The heat scale spans three standard deviations measured on the same control trajectory. In these plots, construct 1 is the HIV construct and construct 2 is the AML construct. [Figure 8l] Figure 8l shows the base pair parameters obtained from 5-nanosecond molecular dynamics simulations of the sensor and siRNA double helix. Opening parameters are defined for base pairs 1-23. The heatscale for each measurement is centered on the mean of control MD trajectories measured for 5-nanoseconds using four unmodified RNA double helixes with identical base sequences to the siRNA and sensor double helixes of the HIV and AML constructs. The heatscale spans three standard deviations measured using the same control trajectory. In these plots, construct 1 is the HIV construct, and construct 2 is the AML construct. [Figure 9a-b] Figures 9a-9b (SEQ ID NOs. 70-72) show the RNAi activity, sensor strand design, and sequence of MCL-1-targeted Cond-siRNAs that sense CBFβ-MYH11. Figure 9a shows the RNAi activity of construct IV variants in cells expressing unrelated transcripts, fusion activators, MYH11, or CBFβ transcripts. Figure 9b shows the sensor strand binding sites on each transcript. [Figure 9c] Figure 9c (SEQ ID NOs. 70-72) shows the RNAi activity, sensor strand design, and sequence of the MCL-1-targeted Cond-siRNA that senses CBFβ-MYH11. Figure 9c shows the sequence map of IV.3. Yellow highlighting indicates regions with optimized chemically modified motifs (regions A, B, C, and D). [Figure 10]Figures 10a-10b show the relative RNAi activity of different input RNAs (construct I.1 (Figure 10a) and construct III.1 (Figure 10b)). Target expression data in Figures 5c and 5f were re-normalized against the expression levels of cells transfected with Cond-siRNA but expressing completely mismatched input strands. The results show that target expression levels in cells expressing matched input strands were lower than those in cells expressing mismatched input strands. Figure 10a: For construct I.1, cells expressing 5' mismatched and perfectly matched input strands showed decreased target expression. Figure 10b: For III.1, cells expressing fused input strands showed decreased target expression. [Figure 11] Figure 11 shows Northern blots of previous generation RNAi trigger designs transfected into HCT116 cells at a concentration of 1 nM for 24 hours. The secondary structure of the trigger is illustrated. Green dots represent 2'-O-methyl RNA bases. Blue dots represent DNA. White dots represent RNA. Black arrows represent Dicer products. The results show that double-stranded RNA with adjacent 2'-O-methyl modified double helix reduced the Dicer product. [Figure 12a] Figure 12a (SEQ ID NOs. 73-75) shows an example of the Cond-siRNA disclosed herein. Figure 12a shows the structure of the construct, where the sensor domain and siRNA domain are shaded in gray, and the sensor double helix is ​​enclosed in a box. [Figure 12b] Figure 12b (SEQ ID NOs. 73-75) shows an example of the Cond-siRNA disclosed herein. Figure 12b shows core chain regions I, II, and III and exonuclease blocking regions I and II in the core chain. [Figure 12c] Figure 12c (SEQ ID NOs. 73-75) shows an example of the Cond-siRNA disclosed herein. Figure 12c shows sensor chain regions I, II, and III. [Modes for carrying out the invention]

[0010] Detailed explanation An approach for developing programmable, conditionally activated siRNAs (Cond-siRNAs) is disclosed herein. These simple riboswitches can maintain their integrity in mammalian cytosol for several days and can detect intracellular RNA transcripts from genes on a specific input strand via foothold-mediated strand substitution. Upon input strand detection, the Cond-siRNAs act as a potent RNAi trigger, silencing identified target genes with sequences completely independent of the input strand. 9 It can release [a certain substance]. As demonstrated in the examples, the switching activity of dozens of Cond-siRNA variants was tested in human adherent cells to identify the necessary and sufficient chemical modification motifs that enable good device performance across a variety of input:output combinations. Some optimized Cond-siRNAs achieved over 90% silencing of target genes (protein expression compared to baseline) in cells expressing sequence-matched RNA transcripts and strongly suppressed background RNAi activity (<25% knockdown) in cells expressing mismatched input strands. Thus, a method for substantially improving the performance of strand substitution switches in living mammalian cells is provided herein. Cond-siRNA technology provides a practical and versatile platform for gene-expression-activated RNAi smart drugs.

[0011] The Cond-siRNA construct disclosed herein comprises two domains, a sensor domain and an RNAi domain, linked together by two fragments of a core strand. This structure is obtained by complementary binding of the sensor strand to the 3' and 5' fragments of the core strand to form a sensor double helix, and by complementary binding of the guide strand to the central fragment of the core strand to form an RNAi double helix. As shown in Figure 1, the entire core strand sequence does not bind complementaryly to the sensor strand or the guide strand. Rather, the first fragment in the core strand between the 3' fragment that binds to the sensor strand and the central fragment that binds to the guide strand remains single-stranded to form a first "bridge" connecting the first end of the sensor domain and the first end of the RNAi domain. Similarly, the second fragment in the core strand between the 5' fragment that binds to the sensor strand and the central fragment that binds to the guide strand remains single-stranded to form a second "bridge" connecting the second end of the sensor domain and the second end of the RNAi domain.

[0012] In this specification, the terms “complementary binding” or “complementarily binding” mean that two single strands form base pairs with each other to form a double helix. However, a certain percentage of mismatch between the two single strands is acceptable as long as a stable double helix is ​​formed. For example, in some embodiments, the sensor double helix or RNAi double helix may have mismatches of about 1%, about 2%, about 3%, about 4%, about 5%, about 6%, about 7%, about 8%, about 9%, about 10%, about 15%, about 20%, about 25%, about 30%, about 35%, about 40%, about 45%, or about 50%.

[0013] A. Input chain The input strand for Cond-siRNA is a "trigger" that switches on Cond-siRNA and is typically an intracellular RNA transcript that is expressed at relatively high levels in targeted cells (e.g., cancer cells) and relatively low levels in untargeted cells (e.g., normal cells). Based on the design of the disclosed Cond-siRNA construct, RNAi is switched on only in targeted cells; in untargeted cells, RNAi remains in the off state. In targeted cells, the input strand is expressed at levels at least 2-fold, at least 5-fold, at least 10-fold, at least 20-fold, at least 30-fold, at least 40-fold, at least 50-fold, at least 60-fold, at least 70-fold, at least 80-fold, at least 90-fold, or at least 100-fold higher than in untargeted cells. Alternatively, in targeted cells, the input chain is expressed at levels of at least 50, at least 100, at least 200, at least 300, at least 400, at least 500, at least 600, at least 700, at least 800, at least 900, or at least 1000 transcripts; in untargeted cells, the input chain is expressed at levels of less than 50, less than 40, less than 30, less than 20, or less than 10 transcripts. Preferably, untargeted cells do not have detectable expression of the input chain.

[0014] The input strand includes mRNA, miRNA, or non-coding RNA, such as a long non-coding RNA, RNA fragment, or viral RNA transcript. For therapeutic use, the input RNA is typically an RNA transcript expressed in cells that are causing disease progression and are therefore targeted for RNAi therapy. Untargeted cells are typically those in which silencing of the RNAi target may cause side effects. To treat diseases or conditions in which the input RNA is overexpressed in targeted cells, Cond-RNAi is designed so that the sensor strand has a sequence complementary to the input RNA sequence. Administration of Cond-RNAi induces the binding of the input RNA to the sensor strand, which in turn induces the dissociation of the RNAi double helix from the sensor double helix in targeted cells, thereby activating the RNAi that targets the disease or condition. In untargeted cells, Cond-RNAi remains inactive.

[0015] B. Structure design As disclosed herein, a Cond-siRNA construct for specific base pairing of the input strand and target is designed using the repeating protocol illustrated below: 1. Obtain a guide strand sequence for the RNAi domain from previously validated siRNA, literature, or siRNA design tools. In some embodiments, the guide strand has a size of 21 nt. 2. Dicer substrates are created from selected guide strands by adding four G / C-rich bases to the 5' end of the guide strand. In some embodiments, the Dicer substrate has a size of 23 bp. Using Nupack (RNA strand, parameters from Mathews et al. 1999, some dangle treatment), it is confirmed that RNAi double helix formation occurs with a probability of over 95% in 1 nM guide (antisense) and sense strands. 3. From the sequence of the input biomarker, a list of all possible sensor segments that are antisense to the input strand is generated. In some embodiments, the sensor segments are 31-33 nt in size. For the CBFβ(CBFB)-MYH11 fusion sequence, only sensor segments that approximately satisfy the parameters shown in Figure 3b were considered. 4. Sensor sequences were ranked for uniqueness in the transcriptome of target animals using NCBI BLAST. For human cancer cell lines, the BLASTn algorithm was used to sequence-check against human transcripts and genome collections. Where possible, sensor segments with sequence complementarity of more than 17 bases to known or predicted RNA transcripts and complete overhang complementarity were excluded. 5. Starting from the most unique sensor segment, the core strand sequence is constructed according to the desired structural parameters for Cond-siRNA. The core strand has the sequence morphology 5'-B-C3-P-C3-A-3', where A and B are complementary to the 5' and 3' ends of the putative double-stranded domain of the sensor strand, P is complementary to the putative guide strand, and C3 is the C3 linker. 6. Using Nupack, rank the thermodynamic stability of the double helix formed between the sensor strand segments and their corresponding 5' and 3' core strand overhangs. Use RNA strands, parameters from Mathews et al. 1999, with some degree of dangling treatment. Ideally, more than 95% of the strands should form base pairs at a strand concentration of 1 nM. Also verify that the core strand does not have a large degree of internal secondary structure. 7. Select the best construct (arrangement of guide chain, core chain, and sensor chain) generated in steps 1-6. 8. Add the chemical modifications disclosed herein. 9. Optimize the placement of LNA modifications using an oligonucleotide design tool, such as "LNA Oligo™ Prediction" (www.qiagen.com / us / service-and-support / learning-hub / technologies-and-research-topics / lna / custom-lna-design-and-applications / lna-design-tools-calculators / lna-oligo-tm-prediction / ) or "LNA Oligo Optimizer" (www.qiagen.com / us / service-and-support / learning-hub / technologies-and-research-topics / lna / custom-lna-design-and-applications / lna-design-tools-calculators / lna-oligo-optimizer / ). Attach the LNA modifications to the sensor strand with approximately one LNA every 3-4 base pairs. Use the "LNA Oligo Optimizer" tool to ensure that the LNA pattern used does not result in secondary structures or self-base pairing interactions with a score above 60. Minimize the self-complementarity score and self-base pairing score as much as possible. Use the "LNA Tm Prediction" tool to determine the Tm of the modified LNA oligo when it forms base pairs with RNA. Use the "LNA Oligo Optimizer" tool to select a configuration that maximizes Tm while avoiding self-base pairing interactions and secondary structures with a score above 60.

[0016] C. Core Chain Design Characteristics As shown in Figure 12, in some embodiments, the core chain has one or more of the following characteristics.

[0017] The 5' and 3' ends of the core chain have one or more of the following characteristics: a. The terminal base on 5' is 2'-F, 2'-O-methyl, or another modified base that is resistant to nuclease cleavage. b. The terminal base on 3' is 2'-F, 2'-O-methyl, or another modified base that is resistant to nuclease cleavage. The three terminal bases on c.5' have the pattern MRM, where M is a modified base (2'-O-methyl, 2'-F) and R is an RNA base. The three terminal bases on d.3' have the pattern MRM, where M is a modified base (2'-O-methyl, 2'-F) and R is an RNA base. The three terminal bases on e.3' and 5' do not have consecutive PS skeleton modifications. f. The portion of the core strand that forms base pairs with the sensor strand has an alternating chemical modification pattern (MR)n. g. In the above, M is a chemically modified base that does not reduce the double-stranded Tm when compared with an equivalent RNA base. h. Any combination of the above, wherein the 5' and 3' ends of the core chain have at least one of the features a-g. i. The 3' and 5' regions I and III of the core strand that form base pairs with the sensor strand are as follows: (a) The formula is entirely made up of pattern (M)n, where M is a 2'-O-methyl or 2'-F modified base, or (b) At least 50% of the bases in this region are 2'-O-methyl or 2'-F, and Phosphorothioates are not present in 30%, 50%, 80%, or 100% or less of the skeletal linkage.

[0018] The exonuclease-blocking region I at the 5' end of the core strand (where three bases form base pairs with the 5' end of the guide strand) has one or more of the following characteristics: aM * + * The M pattern is one in which M is a 2'-modified base (e.g., 2'-O-methyl or 2'-F), * The pattern is where PS is a skeletal bond, and + is an LNA base or another 2'-4' bridged base. b. As defined above, M * + * +Pattern, c.+ * + * +Pattern, dR * + * The M pattern is one in which R is an RNA base. eR * + * +Pattern, f.+ * M * M pattern, and It is a pattern of ga~f, and in the formula, * However, the pattern can be either a PS bond or an unmodified (phosphodiester) bond.

[0019] The exonuclease blocking region II of the core chain has one or more of the following characteristics: aM * The R pattern is such that M is a 2'-modified base (e.g., 2'-O-methyl or 2'-F), * However, the pattern is a PS bond, where R is an LNA base or another 2'-4' bridged base. b. As defined above, M * M pattern, c.+ * M pattern, d.+ * An R pattern, where R is an RNA base, e.+ * +Pattern, fM * + Pattern, and It is a pattern of ga~f, and in the formula, * However, the pattern can be either a PS bond or an unmodified (phosphodiester) bond.

[0020] In this specification, "LNA" means locked nucleic acids as widely used in the art. See, for example, www.glenresearch.com / products / dna-rna-nucleosides-analogs-and-supports / backbone-modification / locked-analog-phosphoramidites.html and en.wikipedia.org / wiki / Locked_nucleic_acid. In addition to LNA modification, 2'-O-methyl modification, and 2'-F modification, other base modification techniques are known in the art. See, for example, www.glenresearch.com / browse / nucleoside-analog-phosphoramidites. In some embodiments, glycol nucleic acids may be used. 61,62 .

[0021] D. Sensor Chain Design Features As shown in Figure 12, in some embodiments, the sensor chain has one or more of the following features: The sensor chain has a 5' overhang, a 3' overhang, or both.

[0022] The bases of the sensor strand have one or more of the following characteristics: (1) at least 25% of the bases are neither RNA nor DNA, (2) at least 50% of the bases are neither RNA nor DNA, (3) at least 75% of the bases are neither RNA nor DNA, (4) 100% of the bases are neither RNA nor DNA, (5) at least one of the bases is LNA or an LNA analog, (6) at least one of the three terminal bases is LNA or an LNA analog, (7) 10% to 50% of the bases are LNA or an LNA analog, (8) 25% to 100% of the bases are LNA or an LNA analog, (9) one or more non-LNA bases are: (a) 2'-o-methyl, (b) 2'-fluoro, (c) 2'-MOE, (d) glycol nucleic acid 61,62 , and other variants shown in (e) www.glenresearch.com / browse / nucleoside-analog-phosphoramidites.

[0023] Phosphothioate bonds can be present in various locations, for example, between the 5' or 3' terminal base and the second base from the terminal, between the terminal and the third base from the terminal, between the terminal and the fifth base from the terminal, between the terminal and the eighth base from the terminal, on 50% of the skeletal bonds, between all bases, and / or at the junctions between regions I and II and between regions II and III.

[0024] End modifications at the 5' end, the 3' end, or both include one or more of the following: (1) triethylene glycol spacers or hexaethylene glycol spacers, (2) C3 spacers, (3) inverted dT, (4) amine linkers, (5) other linkers or end modifications known in the art, e.g., listed at eu.idtdna.com / pages / products / custom-dna-rna / oligo-modifications and www.glenresearch.com / browse / labels-and-modifiers, and (6) modifications may be used to link the 3' and 5' ends to other chemical parts (e.g., antibodies, gold or other metal nanoparticles, polymer nanoparticles, dendrimer nanoparticles, small molecules, single-chain or branched fatty acids, peptides, proteins, aptamers, and other nucleic acid chains and nucleic acid nanostructures).

[0025] As shown in Figure 3, region C of the sensor chain has one or more of the following characteristics: (1) less than 50% of the skeletal bonds are phosphorothioate (PS) bonds; (2) more than 50% of the bases are chemically modified to resist nuclease degradation or to increase the melting temperature (Tm) of the double helix; (3) 100% of the bases are chemically modified to resist nuclease degradation and to increase Tm; (4) approximately 10% to 50% of the bases are LNA or other chemically modified bases with 2'-4' bridges that substantially increase Tm. See, for example, www.glenresearch.com / products / dna-rna-nucleosides-analogs-and-supports / backbone-modification / locked-analog-phosphoramidites.html and en.wikipedia.org / wiki / Locked_nucleic_acid.

[0026] In some embodiments, non-LNA modifications include 2'-O-methyl and 2'-F, as well as other modifications disclosed in www.glenresearch.com / browse / nucleoside-analog-phosphoramidites. In some embodiments, glycol nucleic acids may be used. 61,62 .

[0027] In one aspect, the programmable, conditionally activated RNAi (e.g., Cond-siRNA) disclosed herein has transfection concentrations of less than 0.1 nM in mammalian cells, compared to transfection concentrations of more than 10 nM for other RNAi molecules known in the art. Cond-siRNA remains active for extended periods, for example, at least 12 hours, at least 24 hours, at least 36 hours, at least 48 hours, at least 60 hours, at least 72 hours, at least 84 hours, or at least 96 hours. In some embodiments, Cond-siRNA remains active for 30 days or less, 60 days or less, or 90 days or less.

[0028] In this specification, the term “programmable” means that the Cond-RNAi constructs disclosed herein are designed to allow for changes in the sequence of the input strand without substantially altering the secondary and tertiary structures of the constructs. Further design principles are disclosed in U.S. Patent No. 9,115,355, which are incorporated herein by reference.

[0029] In some embodiments, the dissociation of the RNAi duplex from the input strand-sensor duplex is amplified by mismatches or fluctuation pairs between the core strand and the sensor strand. For an RNAi trigger to have potent RNAi activity, it needs to completely dissociate from the discard duplex formed by the input RNA and the sensor strand. Previous schemes for conditional RNAi often featured an activated RNAi trigger that remained bound to the input signal (e.g., mRNA) via Watson-Crick base pairing. 14,15,17,18,37 During the development of conditional RNAi triggers, it was found that ligation of a Dicer substrate to an adjacent chemically protected double-stranded RNA domain significantly reduced RNAi activity (Figure 2). A possible cause for this is the presence of Dicer inhibitory proteins, such as PACT. 38 This could be binding to the extended double helix. In any case, the dissociation of the RNAi double helix from the input strand-sensor double helix was important in enabling the simultaneous optimization of RNAi suppression in the off state and RNAi efficacy in the on state.

[0030] In some embodiments, the Cond-siRNA disclosed herein has a single-construct design comprising a sensor double helix ligated to an RNAi double helix. Existing designs for conditional RNAi triggers and other DNA circuits have a single-construct design for translating the input strand sequence to an independent output sequence. 14,15,39 or multi-structure 17One of the schemes has been characterized by the fact that single-construction translators should theoretically be more efficient in signal detection and transmission. However, the disadvantage is that the RNAi trigger must be hidden within the construct, creating an opportunity for false RNAi activation due to construct degradation.

[0031] In some embodiments, the Cond-siRNAs disclosed herein are chemically modified. For strand-substituted sensor strands, the double-strand domain of the sensor strand has an LNA modification, a 2'-O-methyl modification, or both. In some embodiments, the double-strand domain of the sensor strand does not have a phosphorothioate (PS) modification. In some embodiments, either or both ends of the core strand or the protective strand are modified with either PS or 2'-O-methyl. In some embodiments, thermodynamically stabilizing modifications generally improve the suppression of background activation, while thermodynamically destabilizing modifications, such as the PS backbone, may increase erroneous activity when used extensively in the double-strand region. In some embodiments, chemical modifications, including LNA modifications, 2'-O-methyl modifications, PS modifications, or combinations thereof, are located in the foothold domain. These modifications in the foothold domain improve the affinity for base pairing and nuclease resistance of the single-strand overhang.

[0032] In this specification, the binding partner to the sensor chain may also be referred to as the “protective chain.” The protective chain is homologous to the input chain. The protective chain can be the core chain. The sensor chain is complementary to the input chain. Binding of the input chain to the sensor chain replaces the protective chain.

[0033] In some embodiments, the chemical modifications include: (i) the sensor chain is modified with LNA and 2'-O-methyl, where the single-stranded foothold region has a PS skeleton modification, but the base-paired double-stranded region does not; (ii) either or both ends of the protective chain have a 2'-O-methyl modification, or both of (i) and (ii).

[0034] In some embodiments, the 3' and 5' terminal regions of the core strand are modified with PS or 2'-OMe such that the Cond-siRNA construct is highly stable when base-paired with the sensor strand, but vulnerable to degradation when base-pairing is not. Degradation of the single-stranded overhang can be halted at the ends of the double helix by an exonuclease blocking domain.

[0035] Complementing targeted drug delivery with self-regulating "smart drugs" that can activate or deactivate RNAi activity in response to cell-specific biochemical signals offers a means to overcome current limitations. For example, RNAi smart drugs that can detect and respond to viral RNA transcripts have been able to potentially eliminate persistent viral infections by killing virus-infected cells through viral RNA-activated silencing of host genes essential for survival. This pharmacodynamics-centric approach is well-suited to the application of nucleic acid switches based on foothold-mediated strand substitution, because these switches can sense and respond to specific base sequences in the input strand's DNA or RNA. 3 .

[0036] In one aspect, the Cond-siRNA disclosed herein is a self-assembled molecular mechanical transducer consisting of a sensor strand, a guide strand, and a core strand, as shown in Figures 1a, 1c, and 1d. These three strands form base pairs, forming a double crossover construct consisting of a 23 bp sensor double helix linked to a 23 base pair (bp) RNAi double helix. 20 It forms a structure in which the sensor strand inhibits the enzymatic processing of the RNAi double helix, thereby keeping RNAi activity switched off in the mammalian cytosol (Figure 1a).

[0037] In some embodiments, Cond-siRNA constructs are further modified by chemical modification to improve their switching performance. Specifically, the Cond-siRNAs disclosed herein have reduced unwanted RNAi activity when the construct is in its off state and improved switching from the off state to the on state in the presence of intracellular RNA transcripts with the correct sequence.

[0038] Inhibition of the enzymatic processing of RNAi double helix occurs in several synergistic ways when combined. For example, the space occupied by the sensor double helix largely overlaps with the space required for the RNAi double helix to bind to Dicer's dsRBD and to the endonuclease domain (Figure 1a, right). 21 In another example, the PAZ domain of Dicer has a stabilizing interaction with the 5' terminal phosphate and 3' double base overhang on the end of the canonical dsRNA trigger. 22 The assembled Cond-siRNA prevents these interactions because the 5' and 3' ends of the core strand extend through the ends of the RNAi double helix to the middle of the sensor double helix (Figure 1c, 1d). The RNAi double helix is ​​23 bp long. This length is too long for Dicer to cleave the RNAi double helix for RNAi loading via the alternative pathway. 23 Furthermore, the RNAi pathway proteins (TRBP and Argonaut (1-4)) required for alternative loading 21 Interactions with ) may also be hindered by the abnormal tertiary structure of the construct. Furthermore, to prevent premature degradation of the sensor double helix or unintended separation of the sensor strand from the RNAi double helix, the entire sensor strand and other critical sites on the construct are thermodynamically stabilized and protected from nuclease activity through extensive chemical modifications. 24 (Figures 1c, 1d).

[0039] To release and activate the RNAi double helix, intracellular RNA transcripts induce the separation of the sensor strand from the core strand via foothold-mediated strand displacement (Figure 1b). In some embodiments, the intracellular RNA transcript is derived from an internal source of the target, e.g., a metabolite. The displacement can begin at the 3' or 5' end of the sensor strand. Sequence specificity of input strand detection is strongly dependent on the strand displacement kinetics for foothold stability. 25 This strengthens the foothold region and the double-strand region due to the need to replace the core chain with a completely complementary one.

[0040] The region where the core strand formed a double helix with the sensor strand becomes 3' and 5' overhangs extending from the RNAi double helix after the sensor strand is separated. These overhangs are cytozolenuclease 26 It is degraded by (V in Figure 1b and Figure 2), which then terminates at the end of the RNAi double helix with a chemically modified nuclease-blocking domain (highlighted in yellow in Figure 3a, construct I.1). This leaves an active RNAi trigger for processing by Dicer and RISC loading (VI in Figure 1b, siRNA).

[0041] The chain composition used in the intracellular degradation of chemically modified single-stranded overhangs was as follows: Array (5'->3') Guide chain A: mCmG CGUCUGAGGGAUCUCUAGU UACCUU (Sequence ID 76) Guide chain B: mCmG+CGUCUGAGGGAUCUCUAGU+TACCUU (Sequence ID 77) 3' Passenger segment: cccucagacg mc * mg * 9s idT (Sequence ID 78) 5' Passenger Segment 0 (control): c3 mG * mG * mU AACUmAGAmGAmU (Sequence ID 79) 1: C G A C G A G C U C A U C A c3mG * mG * mU AACUmAGAmGAmU (Sequence ID 80) 2: 18s * C * G * A * C * G * A * A * G * C * U * C * A * U * C * c3mG * mG * mU AACUmAGAmGAmU (Sequence No. 81) 3: 18s * C * G * A * C * G * A * A * G * C * U * C * A U C c3mG * mG * mU AACUmAGAmGAmU (Sequence No. 82) 4: 18s[[ID=七十六]] * [[ID=七十七]]C[[ID=七十八]] * [[ID=七十九]]G[[ID=八十]] * [[ID=八十一]]A[[ID=八十二]] * [[ID=八十三]]C G A G C U C A U C c3mG[[ID=八十四]] * [[ID=八十五]]mG[[ID=八十六]] * [[ID=八十七]]mU AACUmAGAmGAmU[[ID=八十八]] (Sequence No. 83) [[ID=八十九]] [[ID=九十]]Northern probe: ATCTCTAGTTACC[[ID=九十一]] (Sequence No. 84) [[ID=九十二]] [[ID=九十三]]L: Ambion decade marker[[ID=九十四]] [[ID=九十五]] [[ID=九十六]]

[0042] [[ID=九十七]] [[ID=九十八]]Abbreviations:[[ID=九十九]] [[ID=一百]]9s: Triethylene glycol spacer[[ID=一百零一]] [[ID=一百零二]]18s: Hexaethylene glycol spacer[[ID=一百零三]] It should be noted that there may be some inaccuracies in the translation of the Japanese part "ノーザンプローブ:ATCTCTAGTTACC" as the context is rather technical and specific terms might need more in-depth domain knowledge for a more precise rendering. The above translation is a best effort based on the given text.C3: C3 Spacer idT: inverted dT * : Phosphothioate bond.

[0043] As shown in Figure 2, the sample with guide strand A had sufficient loading and exposure to visualize all bands. Lane 0 is the control strand (15 nucleotides) without an overhang. Passenger 1 has a single detectable band at approximately 15 nt, indicating a decrease in the amount of full-length passenger strand, which indicates rapid processive degradation of the overhang. Passengers 2 and 3 have multiple bands throughout the 15–27 nt size, which indicates slow, non-processive loss of nucleotides, consistent with the presence of PS binding throughout the overhang. Passenger 4 has two bands near 15 nt and shows more full-length product compared to lane 0, which indicates slow initial degradation due to end protection before rapid processive degradation of the overhang when 5' end protection is lost.

[0044] Several Cond-siRNAs are used in antisense oligonucleotide (ASO) and other oligonucleotide therapeutics. 24 It was designed using a chemically modified motif similar to that used to protect the conserved HIV(tat / rev) mRNA gene sequence. Construct I.1 (Figure 3a) was designed using the conserved HIV(tat / rev) mRNA gene sequence. 27 The system was programmed to detect the following. The activator sequences for constructs I and II are shown in Table 1 below. Regions intended to align with the sensor chain are bold and italicized; regions intended to align with the footholds are underlined; segments complementary to the sensor chain are capitalized; segments mismatched with the sensor chain are lowercase.

[0045] [Table 1]

[0046] The construct III.1 (Figure 3a) was programmed to detect the acute myeloid leukemia (AML)-associated fusion oncogene sequence (CBFβ-MYH11) 28,29 The activator sequences for constructs III and IV are shown in Table 2 below. Regions intended to align with the sensor strand are in bold and italic; regions intended to align with the primer are underlined; segments complementary to the sensor strand are in uppercase; segments mismatched to the sensor strand are in lowercase.

[0047]

Table 2

[0048] To enable a direct readout of RNAi activity, constructs I.1 and construct III.1 targeted Renilla luciferase mRNA carrying a biologically irrelevant target sequence (derived from the U5 region of HIV) in its 3’UTR.

[0049] To predict realistic molecular conformations, all-atom molecular dynamics (MD) simulations of I.1 and III.1 30 using a hybrid Amber force field that uses a combination of previously published parameters to describe the chemical modifications in our design 31-33The experiments were carried out in an explicit solvent using [a specific method / tool]. MD-optimized models of both constructs (Figures 1c, 1d) show structural distortions in the sensor and RNAi double helices compared to ideal A-form RNA helical parameters (Figure 4). This distortion is due to (1) extensive chemical modifications and (2) structural distortions inherent in the double crossover motif (Figures 1c, 1d, 4, and Tables 3a and 3b). Tables 3a and 3b below show the mean base pair parameters of the sensor and siRNA double helices over a 5-nanosecond molecular dynamics trajectory. The mean and standard deviation values ​​for each base pair parameter for each double helice were calculated from the data shown in Figure 4. For comparison, the mean and standard deviation were also calculated for unconnected RNA double helices with the same sequence composition as the siRNA and sensor double helices in the HIV and AML constructs (both constructs had the same sequence in the siRNA double helices).

[0050] [Table 3a]

[0051] [Table 3b]

[0052] Despite the distortion, all intended base pairs were maintained throughout the simulation, and the relative alignment of the sensor and RNAi double helixes in both models was consistent with steric protection from Dicer cleavage.

[0053] For empirical testing of the constructs, component strands were purchased from a commercial supplier, and Cond-siRNA was assembled using thermal annealing in 1×PBS. Non-denaturing polyacrylamide gel (PAGE) analysis showed that both constructs, assembled with high purity, were capable of sequence-specific detection of the input RNA and could release RNAi double helixes by isothermal chain substitution at 37°C in 1×PBS buffer (Figures 5a, 5d).

[0054] The RNAi activity of Cond-siRNA in both the off (guide, core, and sensor strands) and on (guide and core strands only) states was measured using a dual luciferase assay. Varying amounts of the construct were co-transfected into HCT116 cells with immobilized amounts of DNA vectors encoding (a) a dual luciferase reporter and (b) short input RNA sequences complementary (matched) or non-complementary (mismatched) to the sensor strand (Tables 1 and 2).

[0055] In cells that did not express the input strand or expressed a completely mismatched input strand, the off-state Cond-siRNA significantly reduced RNAi activity, while the on-state construct proved to be a potent RNAi trigger (supported by Figure 5b and 5e, as shown in Figure 6, which is the result of a Northern blot assay). To activate RNAi, the off-state construct was transfected into cells expressing either the matched or mismatched input strand (Figures 5c and 5f). Input RNA expression levels were assessed by Northern blotting (Figure 4). While the variability in background (off-state) RNAi activity showed reduced apparent statistical significance when the data were standardized against target expression in cells without transfected siRNA (Figures 5c and 5f), another analysis directly comparing the matched and mismatched input strands suggested a significant increase in RNAi activity with the matched input strand (Figure 10).

[0056] The different sensor strand designs of I.1 and III.1 appeared to result in some functional differences in RNAi activation. Construct I.1 had 5' and 3' sensor strand overhangs of 5 nucleotides. RNAi activation was unaffected by input strand mismatches to the 5' overhang but completely eliminated by mismatches to the 3' overhang (Figures 5c, 5f, and 10). Therefore, only the 3' sensor overhang functioned as a foothold for strand substitution. In MD simulations, strand substitution from the 5' foothold of construct I.1 initially proceeded inward, while substitution from the 3' foothold proceeded outward (Figures 1c, 1d). Therefore, steric hindrance may play a role in preventing strand substitution from the 5' foothold. Other factors, such as accessibility to the binding site on the input RNA or differences in the stability of base pairing between the 5' and 3' footholds, may also contribute to preventing strand substitution from the 5' foothold.

[0057] Despite having a similarly strong on-state RNAi trigger (Figures 5b, 5e), construct III.1 showed higher RNAi activation (Figure 5f) than construct I.1 (Figure 5c). A plausible explanation is that the higher thermodynamic stability of the 8-base 3' foothold in III.1 leads to a faster strand displacement kinetics.

[0058] Further optimization of chemical modifications in key regions of the construct was performed (Figure 3a, highlighted region). The off and on state RNAi activity of construct II variants was measured with different patterns of chemical modifications at the 3' and 5' ends of the core strand (Figure 3b, region A, adjacent to the central nick in the sensor double helix), as well as at the exonuclease blocking domain of the RNAi double helix (Figure 3a, region B) (Figures 7, 8, constructs II.1-5).

[0059] For region B, the results indicated that thermodynamic stability at the ends of the RNAi double helix may be important for suppressing RNAi activity in the off state. Substitution of the 2'-OMe base with a locking nucleic acid (LNA) in the center of the nuclease blockade module (region C in Figure 9c) reduced background RNAi activity without impairing RNAi efficacy (Figure 5g, II.1 vs. II.2).

[0060] For region A, the 3' and 5' ends of the core strand need to be chemically modified and protected to terminate the RNAi activity in the off state. When the PS backbone linkage was replaced with a phosphodiester linkage, background RNAi activity increased significantly (Figure 5g, II.2 has three consecutive terminal PSs, II.3 has two alternating PSs, and II.4 has a single terminal PS), and the ability to regulate RNAi activity was completely lost with the unmodified ends (II.0).

[0061] The problem with using PS linkages in region A is that each PS is a stereocenter with two possible enantiomer conformations. Unfortunately, the more nuclease-resistant Sp conformation is destabilizing for RNA base pairing. 34 To improve the thermodynamic stability of the sensor double helix and reduce racemic heterogeneity, PS backbone modifications were replaced with 2'-OMe base modifications (Figure 5h, 8c: constructs II.5-7). The use of three consecutive (II.5), two alternating (II.6), or one terminal (II.7) 2'-OMe bases at the end of each core strand all significantly reduced background RNAi activity. Furthermore, all three motifs increased on-state RNAi potency beyond the PS-modified analogs (Figure 5h, II.5-7 vs. II.2, on-state).

[0062] To determine whether the 2'-OMe modification enables RNAi activation by intracellular RNA transcripts, RNAi activation for II.1, II.2, II.6, and II.7 was compared using a two-step transfection protocol (Figure 5i). II.6 and II.7 showed significantly reduced background RNAi activity, rejected activation by mismatched RNA transcripts, and exhibited increased RNAi knockdown of the cymbidium luciferase target in cells expressing a sequence-matched input strand to the sensor strand (Figure 5i, approximately 50% reduction in target expression at a 2nM Cond-siRNA concentration).

[0063] To generalize the optimization of region A, a version of construct III with a 2'-OMe protected core strand end was constructed (Figure 8d, III.2), and its activation by intracellular RNA transcripts was tested (Figure 5j). Compared to the original PS-protected variant (III.1), III.2 showed significantly reduced background RNAi activity in cells expressing the mismatched input strand and better RNAi activation in cells expressing the correct input strand (Figure 5j, III.1 vs. III.2). In particular, removal of the terminal 2'-OMe modification from only one of the core strand ends (III.3, the 3' overhang of the core strand) almost completely disrupted the regulation of RNAi activity (Figure 5j, III.3).

[0064] Optimizing the core strand also improves the sensor strand. The initial constructs (I.1, II.1, III.1) used sensor strands completely modified with PS, LNA, and 2'-OMe modifications in all regions. PS modifications in the double-stranded region of the sensor strand may increase false activation by reducing the stability of base pairing. Other researchers have also used 2'-OMe modifications alone, without LNA, to stabilize the strand substitution switch for operation in mammalian cells. 8,19A series of construct III variants disclosed herein were tested, in which a 2'-OMe-protected core strand (region A) was combined with sensor strands having one of the following in the double-strand region: LNA+2'-OMe+PS (Figure 5k, 8e, III.2, III.4), LNA+2'-OMe (III.5), 2'-OMe only (III.6), or no modification (III.7). The results showed that the RNA version (III.7) and the 2'-OMe-only version (III.6) of the sensor strand failed to suppress unintended RNAi activity (Figure 5k). The optimal variant was III.5, which had LNA and 2'-OMe bases in the double-strand region but no PS modification. At a 2nM Cond-siRNA concentration, construct III.5 achieved over 90% knockdown of the *Cypripedium macranthos* target in cells expressing the correct input RNA, with background RNAi activity at only 20% in cells expressing the mismatched input RNA (approximately a 10× reduction in target expression between matched and mismatched RNAs). Statistically significant RNAi activation appeared to decrease to as low as 80 pM of Cond-siRNA.

[0065] Using insights gained from optimizing the CBFβ-MYH11 sensor chain in construct III, construct IV was developed. Construct IV is an endogenous apoptosis inhibitor that is crucial for the survival of AML cells (not HCT116 cells used in the dual luciferase experiment) and MCL-1 35This is a Cond-siRNA that senses CBFβ-MYH11 and has an RNAi domain that targets it. Three versions of construct IV were tested for RNAi activation in response to CBFβ-MYH11, CBFβ, or MYH11 sequences in transcripts. To ensure that the sensor strand continues to activate the fusion sequence, the sensor strand (the same as that used in construct III) was designed to be complementary to the MYH11 sequence in its foothold region and complementary to the CBFβ sequence in its double-strand region (Figure 9b). Three versions of construct IV were tested (Figure 8f). IV.1 had a modification pattern similar to prototype construct III.1; construct IV.2 had optimized regions A and B and the sensor strand motif. IV.3 had a pattern previously found to have little effect on RNAi efficacy. 36 Further 2'-O-methyl modifications were introduced in the guide strand. Surprisingly, IV.2 and IV.3 showed significantly improved suppression of background RNAi activity and improved RNAi switching compared to IV.1 (Figure 9c). IV.2, in fusion sequence-expressing cells, exhibited background RNAi at 2nM Cond-siRNA, resulting in approximately 60% reduction in the *Cordyceps* readout target, but approximately 90% reduction in *Cordyceps* itself. IV.3 had almost no detectable background RNAi, but also exhibited lower activated RNAi activity (approximately 60% reduction in *Cordyceps* at 2nM).

[0066] In some embodiments, the chemical modifications of the Cond-siRNA constructs disclosed herein include one or more of the following: A. Use of a sensor chain in the highlighted region C in Figure 3a, “Region for screening chemical modifications,” having one or more of the following features: a. Less than 50% of the skeletal bonds are phosphorothioate (PS) bonds. b. More than 50% of the base has been chemically modified to resist nuclease degradation or to increase the melting temperature (Tm) of the double hemisphere. c. 100% of the bases are chemically modified to resist nuclease degradation and increase Tm. d. Approximately 10% to 50% of the bases are locating nucleic acids (LNAs) or other chemically modified bases with 2'-4' crosslinks that substantially increase Tm. B. The 5' and 3' ends of the core chain have one or more of the following characteristics: a. The terminal base on the 5' side is 2'-F, 2'-O-methyl, or another modified base that is resistant to nuclease cleavage. b. The terminal base on 3' is 2'-F, 2'-O-methyl, or another modified base that is resistant to nuclease cleavage. The three terminal bases on c.5' have the pattern MRM, where M is a modified base (2'-O-methyl, 2'-F) and R is an RNA base. The three terminal bases on d.3' have the pattern MRM, where M is a modified base (2'-O-methyl, 2'-F) and R is an RNA base. The three terminal bases on e.3' and 5' do not have consecutive PS skeleton modifications. f. The portion of the core chain that forms base pairs with the sensor chain has an alternating chemical modification pattern (MR)n. The above characteristics indicate that gM is a chemically modified base that does not reduce the double-stranded Tm when compared to an equivalent RNA base. h. Any combination of the 5' and 3' ends of the core chain having at least one of the features a-g. i. The 3' and 5' regions of the core strand that form base pairs with the sensor strand are as follows: (a) The pattern (M) is entirely made up of n, where M is a 2'-O-methyl or 2'-F modified base, or (b) At least 50% of the bases in this region are 2'-O-methyl or 2'-F, and 30%, 50%, 80%, or less than 100% of the skeletal bonds are not phosphorothioates. C. The core strand has three bases that form a base pair with the 3' end of the guide strand, each possessing one or more of the following characteristics: aM* + * The M pattern is one in which M is a 2'-modified base (e.g., 2'-O-methyl or 2'-F), * The pattern is where the PS bond is, and the + is an LNA base or another 2'-4' bridged base. b. As defined above, M * + * +Pattern c.+ * + * +Pattern dR * + * The M pattern is a pattern in which R is an RNA base. eR * + * +Pattern f.+ * M * M pattern It is a pattern of ga~f, and in the formula, * However, the pattern can be either a PS bond or an unmodified (phosphodiester) bond. D. Guide chain having one or more of the following characteristics: a. 30% to 95% of the bases are chemically modified bases (2'-O-methyl, 2'-F, LNA, 2'-4' bridged bases). b. The two terminal bases on 5' have been chemically modified. The two terminal bases on c.5' have at least one LNA. The two terminal bases on the d.5' side are PS-linked. Both terminal bases on e.3' have been chemically modified. f. Approximately 5% to 50% of the skeletal connections are PS (polysynthetic phosphate). The base adjacent to the g.Dicer cleavage site has not been chemically modified.

[0067] In summary, this specification discloses a programmable, conditionally activated siRNA (Cond-siRNA). This device outputs an active RNA interference trigger to a target gene only when detecting RNA transcripts from different input genes. The improved performance of the strand substitution sensor strand and the realization of conditionally activated RNAi are partly attributable to five key design principles.

[0068] Firstly, for an RNAi trigger to have potent RNAi activity, it needs to be completely dissociated from the discard double helix formed by the input RNA and the sensor strand. Previous schemes for conditional RNAi often featured an activated RNAi trigger to which the input signal (e.g., mRNA) remained bound via Watson-Crick base pairing. 15,17,18,37,60 During the development of previous generations of conditional RNAi triggers, it was found that ligation of a Dicer substrate to an adjacent 2'-O-methyl-modified double-stranded RNA domain significantly reduced RNAi activity (Figure 11). A possible cause for this is the presence of Dicer inhibitory proteins, such as PACT. 38 This could be due to binding to the extended double helix. However, the dissociation of the RNAi double helix from the input strand-sensor double helix was necessary to achieve simultaneous optimization of RNAi suppression in the off state and RNAi efficacy in the on state.

[0069] The array used in the construct shown in Figure 11 is as follows:

[0070] [Table 4]

[0071] Secondly, past designs for conditional RNAi triggers and other DNA circuits have been single constructs for translating input sequences to independent output sequences. 15,16 or multi-structure 17One of the schemes has been characterized. Single-construct translators should theoretically be more efficient in signal detection and transmission. However, the disadvantage is that the RNAi trigger must be hidden within the construct, creating an opportunity for false RNAi activation due to construct degradation. The success of the disclosed Cond-siRNA designs demonstrates that the advantages of single-construct translators can be utilized while effectively controlling the risk of false activation.

[0072] Thirdly, extensive chemical modifications are important for the proper function of the strand-substituted sensor strand and, consequently, the Cond-siRNA. Examples of optimization experiments have shown that: 1) the double-strand domain of the sensor strand must have both LNA and 2'-O-methyl modifications, but not PS modifications; 2) the ends of the protective strand must be modified with either PS or 2'-O-methyl; 3) thermodynamically stabilizing modifications generally improve the suppression of background activation, while thermodynamically destabilizing modifications, such as the PS backbone, can actually increase erroneous activity if used extensively in the double-strand region; 4) Although chemical modifications in the foothold domain were not tested, combinations of LNA, 2'-O-methyl, and PS modifications are beneficial because they improve the affinity for base pairing and nuclease resistance of the single-strand overhang.

[0073] Fourth, endogenous RNA degradation machines can be effective tools for construct switching. As shown in the examples, the chemically modified 3' and 5' terminal regions of the core strand were highly stable when base-paired with the sensor strand, but vulnerable to degradation when not base-paired. This differential sensitivity to nuclease activity can be achieved with PS-modified ends, but it appeared to be more effective using the 2'-OMe modification. As demonstrated herein, degradation of single-stranded overhangs can be stopped at the ends of the double helix by the exonuclease blocking domain. Trimming of the disclosed RNAi double helix using this scheme produced a strong RNAi trigger from the DX secondary structure. Similar schemes may be useful in the dynamic recognition of other nucleic acid nanostructures.

[0074] Finally, the separation of the sensor domain and RNAi domain into two distinct double helixes was crucial for intracellular stability, programmability, and ease of technological development. Firstly, there was no overlap in base pairing between the two domains, and therefore no competing secondary structure conformations. This ensured a large margin of thermodynamic stability and simplified the programming of new input and output sequences. Secondly, the dimensions of the double helix and the linking chemistry at the crossover point were configured to minimize strain in the tertiary structure (Figure 1c). This further enhanced thermodynamic stability. Thirdly, since the double helixes did not overlap, chemical modifications could be applied to the sensor domain without compromising the compatibility of the RNAi domain.

[0075] The following examples are provided to better illustrate the claimed invention and should not be construed as limiting the scope of the invention. To the extent that specific materials are mentioned, this is for illustrative purposes only and is not intended to limit the invention. It will be apparent to those skilled in the art that various equivalents, variations, and modifications may be made without departing from the scope of the invention, and it will be understood that such equivalents are included in this specification. Furthermore, all references cited in this disclosure are incorporated herein in their entirety as if they were fully presented herein. [Examples]

[0076] Example 1: Structure Design A Cond-siRNA was designed using a repeating protocol to form specific base pairs for the input strand and target. A suitable 21nt guide strand sequence as the RNAi domain was obtained from previously identified siRNAs, literature, or siRNA design tools. A 23bp Dicer substrate was created by adding four G / C-rich bases to the 5' end of the guide strand. The RNAi double helix was formed with a probability of over 95% in 1nM guide (antisense) and sense strands using Nupack (RNA strand, Mathews) 59 This was verified using the parameters disclosed by [the company] (and some degree of dangle processing).

[0077] A list of all possible 31-33 nt sensor segments (antisense relative to the input strand) was generated from the input biomarker sequence. For the CBFβ-MYH11 fusion sequence, only sensor segments that approximately met the parameters shown in Figure 9b were considered. Sensor sequences were ranked in terms of uniqueness in the target animal transcriptome using NCBI BLAST. The BLASTn algorithm was used to verify the sequences of human cancer cell lines against human transcript + genome collections. Where possible, sensor segments with sequence complementarity of more than 17 bases and complete overhang complementarity to known or predicted RNA transcripts were excluded.

[0078] Starting with the most unique sensor segment, core chain sequences were selected to conform to the desired structural parameters for Cond-siRNA. For example, the core chain sequence is 5'-B-C3-P-C3-A-3' (A and B are complementary to the 5' and 3' ends of the putative double-strand domain of the sensor chain, P is complementary to the putative guide chain, and C3 is the C3 linker). The thermodynamic stability of the double helix formed between the sensor chain segment and the corresponding 5' and 3' overhangs of the core chain was ranked using Nupack. Mathews et al. 59 RNA strands with the parameters disclosed were used after some degree of dangling. Ideally, at a strand concentration of 1 nM, more than 95% of the strands should form base pairs. The core strand was also confirmed to not have a large internal secondary structure.

[0079] The best designed guide, core, and sensor strands were selected and chemically modified. The locations of LNA modifications were optimized using Exiqon's oligonucleotide design tool (www.exiqon.com / oligo-tools). In the sensor strand, approximately one LNA modification was made every 3-4 bases. Using the LNA Oligo Optimizer tool, it was confirmed that this LNA pattern did not result in secondary structures or autobase-pairing interactions with scores exceeding 60. The autocomplementarity and autobase-pairing scores were optimized as much as possible.

[0080] Example 2: Chain synthesis The chains containing LNA bases were synthesized by Exiqon Inc (now a division of Qiagen). The chains without LNA were synthesized by GE Life Sciences Dharmacon (now a division of the Horizon Discovery Group). All chains were ordered to be purified by PAGE or HPLC, as recommended by the manufacturers.

[0081] Example 3: Assembly of Cond-siRNA Cond-siRNA was assembled by thermal annealing in 1× phosphate-buffered saline (PBS). The construct may or may not be purified. The quality of the construct can be evaluated by non-denaturing gel electrophoresis at 4°C using 1× TBE and 10%-15% PAGE.

[0082] Without purification, sensor, core, and guide strands in a molar ratio of 1.1:1.00:1.1 were mixed at a pH of approximately 7.0 at a concentration of 50 nM or 100 nM in 1 × PBS. Using a slightly excess of sensor and guide strands prevented the production of double helix formations of guide and core strands with constitutive RNAi activity. A PCR thermocycler was used under the following conditions: Heat the lid to 105°C. • Maintain at 85°C for 30 seconds to denature the chain. Cool to 50°C at a rate of 0.1°C / second. Maintain at 50°C for 45 minutes. Cool to 37°C at a rate of 0.02°C / second. • Rapidly cool to 4°C and maintain that temperature.

[0083] For purification, the sensor strand, core strand, and guide strand were mixed under the annealing conditions described above and assembled at a set concentration of 1 μM in 1 × PBS. The construct was then loaded onto a Bio-Rad mini protean 10% native PAGE gel using TBE buffer and electrophoresed at 125 V and 4 °C for approximately 45 minutes. Bands corresponding to Cond-siRNA were visualized and excised under UV lamp illumination.

[0084] The excised bands were extracted by electrolytic elution using a Harvard Apparatus Electroprep system according to the manufacturer's instructions. The gel fragments were placed in 0.5 mL chambers sealed with 100 K MWCO filter membranes and 2 K MWCO filter membranes. The constructs were eluted through the 100 K MWCO membrane and captured in an adjacent 0.5 mL chamber formed by the 100 K MWCO membrane and a second 2 K MWCO membrane. Elution was performed in 0.1 M Na2HOP4 buffer (approximately pH 7.0) at 4°C for approximately 45 minutes. The power was set to maintain a constant current of 15 mA with a voltage cutoff of 65 V.

[0085] The concentration of the purified construct was quantified using a Bio-Rad ChemiDoc XRS+ Imager and calculated by comparing it to a known concentration of Cond-siRNA standard using SYBR Gold staining on non-denaturing PAGE.

[0086] The construct is in its best condition immediately after assembly or purification. The construct can also be divided and stored indefinitely at -80°C. However, freeze-thaw cycles impair the quality of the construct and can lead to disassembly. Disassembled constructs can be reassembled by repeating thermal annealing immediately before assay.

[0087] The assembly yield was high, and purification did not necessarily improve the performance of the constructs, so unpurified constructs were used in various tests.

[0088] Example 4: Chain substitution assay The construct was prepared at a set concentration of 50 nM and combined 1:1 with 50 nM oligonucleotide activator (or PBS as control) in PBS buffer at 37°C to obtain a mixture containing 25 nM input signal and construct. The construct-input strand combination was then incubated in a PCR thermocycler at 37°C for 4 hours. The samples were collected and immediately frozen to -80°C in 1× native PAGE loading dye. At the end of the experiment, all samples were rapidly thawed and analyzed by non-denaturing PAGE.

[0089] Example 5: Generation of a dual luciferase reporter and activator plasmid All clones were constructed by annealing DNA oligos for specific inserts using standard molecular biology protocols, and then ligating them to designated sites in the parent vector. The accuracy of all constructs was verified by DNA sequencing.

[0090] Example 6: PsiCHECK Dual Luciferase Reporter The DNA oligos shown below were annealed and ligated to the XhoI and NotI sites of the psiCHECK 2 (Promega) dual luciferase reporter. Nucleotides in bold are sense target sequences. Lowercase nucleotides indicate the 5' overhang of the restriction site.

[0091] [ka]

[0092] Example 7: Input RNA transcript for signal activation experiment The activator sequence was expressed as part of a chimeric tRNA transcript. The first portion is shown below as the mature sequence, modified with a CCA at the 3' end. 5 tRNA Lys3It consists of the following. CCA prevents endonuclease-like cleavage by the pre-tRNA processing enzyme tRNAse Z. The tRNA Pol III promoter is an internal promoter and is included within the coding sequence of tRNAmpDNA.

[0093] tRNA Lys3 To clone the parent plasmid containing the first 69 nucleotides, the NruI restriction site 48 The process terminated there. Digestion of the parent plasmid by NruI generates a blunt end immediately after nucleotide tRNA 69. The annealed duplicate oligo encodes the remaining modified tRNA nucleotides and the subsequent specific activation sequences. Each activation sequence is a 12-base tetraloop (GGCGCAAGCC) (Sequence ID 107) It terminates with a T6 run that codes for the Pol III termination sequence. The U4+ RNA transcript sequences are listed below.

[0094] For constructs I and II, tRNA Lys3 Leader array, 5'->3':

[0095] [ka]

[0096] The sequences in bold are the binding sites for Northern blot probes. The activator sequences are listed in Tables 1 and 2. The Northern blot probes are listed in Table 5.

[0097] [Table 5]

[0098] Example 8: Tissue Culture All analyses were performed using HCT 116 colorectal cancer cells. Without the addition of antibiotics, cells were maintained at 37°C in a humidified 5% CO2 incubator using McCoy's 5A basic medium (Irvine Scientific, USA) supplemented with 10% fetal bovine serum (FBS), 1.5 mM L-glutamine (Irvine Scientific, USA), and 10 mM pirubate (Irvine Scientific, USA).

[0099] Example 9: Northern blot analysis Activator expression was analyzed using a 6-well plate with 250 μL of OptiMEM and 250 μL of 1:50 diluted Lipofectamine 2000, each containing 2 μg of plasmid DNA. Liposomes were formed according to the manufacturer's instructions and added to cells with 2 mL of fresh complete medium. The medium was changed after 18 hours, then at least once daily, and 6 hours before RNA recovery. Analysis of off-state and pre-activated (on-state) cond-siRNA was performed similarly; the indicated amount of RNAi complex was added to 2 μg of pBluescript plasmid (carrier) in 250 μL of OptiMEM.

[0100] All RNA was recovered using 1000 μL of RNA Stat-60 (Tel-Test, Inc.) and processed according to the manufacturer's instructions, with the addition of a second organic extraction using a 1:1 phenol:chloroform extract before precipitation; the RNA pellet was washed twice with 70% ethanol, then excess ethanol was evaporated and the RNA was resuspended in RNAse-free TE, pH 6.8.

[0101] For Northern nucleotide analysis, 15 μg of all RNA was electrophoresed on 8% (for activators) or 12% (for cond-siRNA) urea-PAGE gel (15 cm) along with a 32P-labeled Ambion decade marker. The gel was electroblotted onto Hybond XL (GE Healthcare Life Sciences), pre-hybridized, and hybridized at 37°C using Sigma Perfect Hyb Plus with 5–10 picomoles of P32-5' end-labeled oligo probes. The blot was washed at 37°C with 4–5 changes of 2×SSC / 1%SDS. Old oligo probes were removed from the membrane using serial hybridization according to the manufacturer's instructions and confirmed by re-exposure before re-hybridization unless otherwise specified. U6 snRNA hybridization was used as a control. The cloning procedure, oligos, and all probe sequences are listed in the table.

[0102] Example 10: Dual luciferase assay A dual luciferase assay was performed using the Promega Dual-Luciferase Reporter Assay System according to the manufacturer's instructions. The RNAi target sequence was cloned into the 3'UTR of the sea urchin luciferase gene on the psiCHECK-2(Promega) vector, and firefly luciferase was used as a control.

[0103] Cells were incubated and transfected in 48-well cluster plates. Cells were seeded one day before transfection and transfected at 50% confluence. Each experiment was repeated at least three times to obtain biological replicas. Figures 5b–5h show the single-step transfection protocol, and Figures 5i–5k and 9a show the two-step transfection protocol.

[0104] Single-step cotransfection protocol : For each experiment, a master mix of the psiCHECK (Promega Corporation) reporter plasmid in Opti-MEM (Thermo Fisher Scientific) was prepared. This master mix was divided, and either the pBluescript (Agilent) control or activator plasmid was added. The fresh mixture was then divided again for the addition of Cond-siRNA complexes at varying concentrations. Finally, a 1:50 dilution of Lipofectamine 2000 (L2K) was added to the plasmid + Cond-siRNA mixture in a 1:1 volume ratio to achieve the manufacturer's recommended 1:100 dilution of L2K, and incubated at room temperature according to the manufacturer's recommendations.

[0105] Sufficient mixtures (3.3x volume required) were prepared under each experimental condition (combination of activator and cond-siRNA at specific concentrations), and transfected in three separate wells as technical replicas.

[0106] Therefore, 40 μL of transfection mixture consisting of 16 μL (psiCHECK and activator plasmid in OptiMEM), 4 μL (50×Cond-siRNA in 1×PBS buffer), and 20 μL (1:50 dilution L2K) was added to each well of a 48-well plate. PBS was calcium- and magnesium-free phosphate-buffered saline treated with DEPC (diethyl pyrocarbonate) to remove all RNAse activity.

[0107] Immediately before transfection, the medium in each well was replaced with 160 μL of fresh medium, and then 40 μL of transfection mixture containing 40 ng of psiCHECK-2 double luciferase reporter plasmid, 120 ng of pBluescript or activator expression plasmid, and a Cond-siRNA complex at the indicated concentrations was added to a final volume of 200 μL / well.

[0108] Two-step transfection protocol : Using this protocol, we obtained the data shown in Figures 5i-5k and Figure 9a. Transfection with target and activator plasmids for 1 hour to 8 hours.

[0109] For two-step transfection, a master mix of the psiCHECK (Promega Corporation) reporter plasmid in Opti-MEM (Thermo Fisher Scientific) was prepared. This master mix was divided, and then either the pBluescript (Agilent) control or activator plasmid was added. Lipofectamine 2000 (L2K) at a 1:50 dilution was added to the plasmid mixture at a 1:1 volume ratio to achieve the manufacturer's recommended 1:100 dilution of L2K, and the mixture was incubated at room temperature according to the manufacturer's recommendations to form a lipoplex.

[0110] Under each experimental condition, a sufficient mixture was prepared and transfected in three separate wells as technical replicas. Thus, 40 μL of transfection mixture, consisting of 20 μL (psiCHECK and activator plasmid in OptiMEM) and 20 μL (1:50 dilution L2K), was added to each well in a 48-well plate.

[0111] Immediately before transfection, the medium in each well was replaced with 160 μL of fresh medium, and then 40 μL of transfection mixture containing 40 ng of psiCHECK-2 double luciferase reporter plasmid and 120 ng of pBluescript or activator expression plasmid was added to a final volume of 200 μL / well.

[0112] The transfection mixture was removed and the cells were gently washed with culture medium for approximately 2 hours after 6 hours. 160 µl of fresh medium was added to each well, and cell incubation was continued until the second transfection.

[0113] Transfection with Cond-siRNA complex 2, time - 0 Eight hours after transfection 1, Cond-siRNA was transfected at the specified variable concentrations in experiments using RNAiMAX reagent (ThermoFisher). For each experimental condition, sufficient amounts of Cond-siRNA at each concentration were prepared to perform three experiments with each target / activator combination in PBS. Each Cond-siRNA dilution was mixed with an equal volume of RNAiMAX in 1:50 OptiMEM and incubated at room temperature to form a lipoplex according to the manufacturer's instructions. Specifically, 40 μL of transfection mixture consisting of 20 μL of 10× final concentration Cond-siRNA (8 μL of PBS + 12 μL of OptiMEM) and 20 μL of 1:50 diluted RNAiMAX was added to each well of a 48-well plate.

[0114] maintain In the single-step protocol, time 0 is defined as the point at which the cotransfection mixture is added to the cells, while in the two-step protocol, time 0 is defined as the point at which the Cond-siRNA complex is added (transfection number 2). The culture medium was changed 18 hours after transfection, then at least once daily thereafter, and 6 hours before lysate preparation.

[0115] Lysate preparation For each experiment, the 48-well plate was removed at the specified time. The culture medium was carefully aspirated from each well. The wells were then washed once with 1×PBS, aspirated, and dried. 100 μL of 1× Promega Passive Lysis Buffer was added to each well. The plate was then covered with aluminum foil and frozen at -80°C or placed on a shaker for gentle stirring (approximately 70 rpm) at room temperature for about 30 minutes. If frozen, the cells were thawed on the shaker by gentle stirring for at least 30 minutes prior to the double luciferase assay. Visual inspection of the wells confirmed that the cells were sufficiently lysed before the assay.

[0116] Assay Cell lysates were assayed using the Dual-Luciferase Reporter Assay Kit (Promega) according to the manufacturer's instructions. Sea lice luciferase values ​​were standardized against firefly luciferase in each technical replica (each well). Values ​​for a single biological replica were obtained by averaging three experiments. All graphs show the results of experiments with at least three independent biological replicas.

[0117] Example 11: Molecular Dynamics Simulation The atomic model of Cond-siRNA is used in Nucleic Acid Builder. 49 And built using custom scripts, Accelrys (now BIOVIA, a division of Dassault Systems) Cerius 45 The package was edited and appropriate chemical modifications were made.

[0118] Hybrid force field (FF) RNA 50 , 2'-O-methyl 51 LNA 52 and phosphorothioate 53 The parameters were created by combining previously reported Amber force field parameters for the modifications. Previous reports did not provide a parameter set for LNA thymidine. The FF parameters for the LNA sugar ring were derived from the LNA force field, and the base parameters were derived from the Amber03 force field. The charge was calculated using the RESP ESP charge Derive (RED) server (q4md-forcefieldtools.org / REDServer / ). FF parameters for non-DNA components, e.g., C3 linker, terminal amine modification, and terminal PEG linker, were derived from GAFF FF. 54 Obtained from: All structures were placed in a periodic box with a 15 Å space on each side, and then solvated with TIP3 water. 55 First, Mg to neutralize half of the charge. 2+ Add ions, then neutralize the remaining half with Na + Na was added. Finally, Na+ and Cl - Ions were added to a concentration of 150 mM.

[0119] Molecular dynamics simulations were performed on an NVIDIA K80 GPU using the LAMMPS13 GPU-compatible release (December 21, 2016). The structure was first minimized by steepest descent, then by a conjugate gradient algorithm over 500 steps, and then equilibrated by MD simulation at 310K using an NVT ensemble over a 10ps process with a 1fs time step. Next, a 10ps NPT MD was performed on the resulting structure at 310K, 350atm to relax the periodic box and ensure a net positive pressure. A nose-Hoover thermostat with a 100fs time constant was used for the NVT simulation. A nose-Hoover barostat with a 1ps time constant was used for the NPT simulation. Next, a 20ns MD was performed on the equilibrated structure at 310°K (NVT ensemble, 1fs time step).

[0120] To obtain the structure shown in Figure 9, the structure with the lowest potential energy was extracted from the MD trajectory, and the minimization of the conjugate gradient energy was applied over 500 steps. UCSF Chimera package 57 The construct was visualized using X3DNA. 58 The helical parameters were calculated from the simulation trajectory using [a specific method / tool]. The invention as described in the original claims of the patent application is listed below. [1] A programmable, conditionally activatable siRNA (Cond-siRNA) construct comprising a sensor strand, a core strand and a guide strand, (i) The sensor chain and the core chain form a sensor double chain, (ii) The guide strand and the core strand form an RNAi double helix, (iii) The sensor double helix is ​​linked to the RNAi double helix to form a single structure. Cond-siRNA construct. [2] The Cond-siRNA construct according to [1], wherein the sensor chain has an overhang that is not complementary to the core chain, and the overhang of the sensor chain is capable of binding complementaryly to the input chain to form a foothold, thereby causing the sensor chain to be replaced from the core chain. [3] One or more of the following chemical modifications A through D: A. Use of a sensor chain in which the highlighted region A has one or more of the following features a-d: a. Less than 50% of the skeletal bonds are phosphorothioate (PS) bonds; b. More than 50% of the base has been chemically modified to resist nuclease degradation or to increase the double-strand melting temperature (Tm); c. 100% of the bases are chemically modified to resist nuclease degradation and to increase Tm; d. Approximately 10% to 50% of the bases are Loc nucleic acids (LNAs) or other chemically modified bases having 2'-4' crosslinks that substantially increase the Tm; B. The 5' and 3' ends of the core chain have one or more of the following characteristics a to i: a. The terminal base on the 5' side is a 2'-F base, a 2'-O-methyl base, or another modified base that is resistant to nuclease cleavage; b. The terminal base on 3' is a 2'-F base, a 2'-O-methyl base, or another modified base that is resistant to nuclease cleavage; The three terminal bases on c.5' are pattern MRM, where M is a modified base (2'-O-methyl base, 2'-F base) and R is an RNA base; The three terminal bases on d.3' are pattern MRM, where M is a modified base (2'-O-methyl base, 2'-F base) and R is an RNA base; The three terminal bases on e.3' and 5' do not have consecutive PS skeleton modifications; f. The portion of the core chain that forms base pairs with the sensor chain has an alternating chemical modification pattern (MR)n; g. The above characteristics, wherein M is a chemically modified base that does not reduce the double-stranded Tm when compared to an equivalent RNA base; h. The 5' and 3' ends of the core chain are any combination having at least one of the features a to g; i. The 3' and 5' regions of the core strand that form base pairs with the sensor strand, (a) The pattern is (M)n, where M is a 2'-O-methyl base or a 2'-F base, or (b) At least 50% of the bases in this region are 2'-O-methyl bases or 2'-F bases, and 30%, 50%, 80%, or 100% or less of the skeletal bonds are not phosphorothioates; C. A core strand in which three bases that form a base pair with the 3' end of the guide strand have one or more of the following characteristics a to g: aM * + * It is an M pattern, where M is a 2'-modified base (e.g., a 2'-O-methyl base or a 2'-F base). * The pattern is where is a PS bond, and + is an LNA base or another 2'-4' bridged base; bM * + * +Pattern, M, * And + is defined as a, pattern; c.+ * + * +Pattern; dR * + * M is a pattern where R is an RNA base; eR * + * +Pattern; f.+ * M * M pattern; It is a ga~f pattern, * The pattern may be either a PS bond or an unmodified (phosphodiester) bond; D. A guide chain having one or more of the following characteristics a-g: a. 30% to 95% of the bases are chemically modified bases (2'-O-methyl bases, 2'-F bases, LNA, 2'-4' bridged bases); b. The two terminal bases on 5' have been chemically modified; The two terminal bases on c.5' have at least one LNA. The two terminal bases on the d.5' side are joined by a PS bond; Both terminal bases on e.3' have been chemically altered; f. Approximately 5% to 50% of the skeletal connections are PS; g. The base adjacent to the Dicer cleavage site has not been chemically altered. A Cond-siRNA construct as described in [1] or [2], including the following: [4] A Cond-siRNA construct according to any one of [1] to [3], wherein the double-stranded domain of the sensor chain is LNA modified, 2'-O-methyl modified, or both. [5] A Cond-siRNA construct according to any one of [1] to [4], wherein the double-stranded domain of the sensor chain is not modified with phosphorothioate (PS). [6] A Cond-siRNA construct according to any one of [1] to [5], wherein one or both ends of the core chain are either a PS modification or a 2'-O-methyl base. [7] A Cond-siRNA construct according to any one of [2] to [6], wherein the chemical modification, including LNA modification, 2'-O-methyl modification, PS modification, or a combination thereof, is located in the foothold domain. [8] A Cond-siRNA construct according to any one of [1] to [7], wherein the sensor double helix is ​​23 bp and the RNAi double helix is ​​23 bp. [9] A method for activating synthetic RNAi, comprising administering a Cond-siRNA construct described in any of [1] to [8] to a target, wherein the input strand binds to the sensor strand, causing the sensor strand to be replaced by the Cond-siRNA, thereby activating the RNAi in the Cond-siRNA construct.

[10] The method according to [9], wherein the input strand is an intracellular RNA transcript.

[11] A method for treating a disease or condition, comprising administering a Cond-siRNA construct according to any of [1] to [8] to a subject in need thereof, wherein the input strand binds to the sensor strand, causing the sensor strand to be replaced from the Cond-siRNA, thereby activating RNAi in the Cond-siRNA construct, the RNAi targeting the disease or condition.

[12] The method according to

[11] , wherein the input strand is an intracellular RNA transcript.

[0121] References 1. Seeman, N. C. DNA in a material world. Nature 421, 427-431 (2003). 2. Guo, P. The emerging field of RNA nanotechnology. Nat Nano 5, 833-842 (2010).​​​4. Benenson, Y. Biomolecular computing systems: principles, progress and potential. Nat Rev Genet 13, 455-468, doi: http: / / www.nature.com / nrg / journal / v13 / n7 / suppinfo / nrg3197_S1.html (2012). 5. Yurke, B., Turberfield, A. J., Mills, A. P., Simmel, F. C. & Neumann, J. L. A DNA-fuelled molecular machine made of DNA. Nature 406, 605-608, doi:http: / / www.nature.com / nature / journal / v406 / n6796 / suppinfo / 406605a0_S1.html (2000). 6. Srinivas, N. et al. On the biophysics and kinetics of toehold-mediated DNA strand displacement. Nucleic Acids Research, doi:10.1093 / nar / gkt801 (2013). 7. Green, A. A. et al. Complex cellular logic computation using ribocomputing devices. Nature 548, 117-121, doi:10.1038 / nature23271 http: / / www.nature.com / nature / journal / v548 / n7665 / abs / nature23271.html#supplementaryinformation (2017). 8. Groves, B. et al. Computing in mammalian cells with nucleic acid strand exchange. Nat Nano 11, 287-294, doi:10.1038 / nnano.2015.278 http: / / www.nature.com / nnano / journal / v11 / n3 / abs / nnano.2015.278.html#supplementaryinformation (2016). 9. Setten, R. L., Rossi, J. J. & Han, S. P. The current state and future directions of RNAi based therapeutics Nature Reviews Drug Discovery, In press (2019). 10. Bobbin, M. L. & Rossi, J. J. RNA Interference (RNAi)-Based Therapeutics: Delivering on the Promise? Annual Review of Pharmacology and Toxicology 56, 103-122, doi:10.1146 / annurev-pharmtox-010715-103633 (2016). 11. Adams, D. et al. Patisiran, an RNAi Therapeutic, for Hereditary Transthyretin Amyloidosis. New England Journal of Medicine 379, 11-21, doi:10.1056 / NEJMoa1716153 (2018). 12. Benenson, Y., Gil, B., Ben-Dor, U., Adar, R. & Shapiro, E. An autonomous molecular computer for logical control of gene expression. Nature 429, 423-429, doi:http: / / www.nature.com / nature / journal / v429 / n6990 / suppinfo / nature02551_S1.html (2004). 13. Kumar, D., Kim, S. H. & Yokobayashi, Y. Combinatorially Inducible RNA Interference Triggered by Chemically Modified Oligonucleotides. Journal of the American Chemical Society 133, 2783-2788, doi:10.1021 / ja1107436 (2011). 14. Han, S. P., Barish, R. D. & Goddard, W. A. (Google Patents, 2015). 15. Han, S.-P., Goddard III, W. A., Scherer, L. & Rossi, J. J. Signal activatable constructs and related components compositions methods and systems. USA patent US9725715B2 (2015). 16. Hochrein, L. M., Ge, T. J., Schwarzkopf, M. & Pierce, N. A. Signal Transduction in Human Cell Lysate via Dynamic RNA Nanotechnology. ACS Synthetic Biology, doi:10.1021 / acssynbio.8b00424 (2018). 17. Hochrein, L. M., Schwarzkopf, M., Shahgholi, M., Yin, P. & Pierce, N. A. Conditional Dicer Substrate Formation via Shape and Sequence Transduction with Small Conditional RNAs. Journal of the American Chemical Society 135, 17322-17330, doi:10.1021 / ja404676x (2013). 18. Bindewald, E. et al. Multistrand Structure Prediction of Nucleic Acid Assemblies and Design of RNA Switches. Nano Letters 16, 1726-1735, doi:10.1021 / acs.nanolett.5b04651 (2016). 19. Chatterjee, G., Chen, Y.-J. & Seelig, G. Nucleic Acid Strand Displacement with Synthetic mRNA Inputs in Living Mammalian Cells. ACS Synthetic Biology, doi:10.1021 / acssynbio.8b00288 (2018). 20. Li, X., Yang, X., Qi, J. & Seeman, N. C. Antiparallel DNA Double Crossover Molecules As Components for Nanoconstruction. Journal of the American Chemical Society 118, 6131-6140, doi:10.1021 / ja960162o (1996). 21. Ha, M. & Kim, V. N. Regulation of microRNA biogenesis. Nature Reviews Molecular Cell Biology 15, 509, doi:10.1038 / nrm3838 (2014). 22. MacRae, I. J. et al. Structural Basis for Double-Stranded RNA Processing by Dicer. Science (New York, N.Y.) 311, 195-198, doi:10.1126 / science.1121638 (2006). 23. Kim, D.-H. et al. Synthetic dsRNA Dicer substrates enhance RNAi potency and efficacy. Nat Biotech 23, 222-226, doi:http: / / www.nature.com / nbt / journal / v23 / n2 / suppinfo / nbt1051_S1.html (2005). 24. Khvorova, A. & Watts, J. K. The chemical evolution of oligonucleotide therapies of clinical utility. Nat Biotech 35, 238-248, doi:10.1038 / nbt.3765 (2017). 25. Srinivas, N. et al. On the biophysics and kinetics of toehold-mediated DNA strand displacement. Nucleic Acids Research 41, 10641-10658, doi:10.1093 / nar / gkt801 (2013). 26. ORBAN, T. I. & IZAURRALDE, E. Decay of mRNAs targeted by RISC requires XRN1, the Ski complex, and the exosome. RNA (New York, N.Y.) 11, 459-469, doi:10.1261 / rna.7231505 (2005). 27. Hope, T. J. & Trono, D. Structure, Expression, and Regulation of the HIV Genome, <http: / / hivinsite.ucsf.edu / InSite?page=kb-00&doc=kb-02-01-02> (2000). 28. Kundu, M. & Liu, P. P. Function of the inv(16) fusion gene CBFB-MYH11. Current Opinion in Hematology 8, 201-205 (2001). 29. Look, A. T. Oncogenic Transcription Factors in the Human Acute Leukemias. Science (New York, N.Y.) 278, 1059-1064, doi:10.1126 / science.278.5340.1059 (1997). 30. Plimpton, S. Fast Parallel Algorithms for Short-Range Molecular Dynamics. Journal of Computational Physics 117, 1-19, doi:http: / / dx.doi.org / 10.1006 / jcph.1995.1039 (1995). 31. Condon, D. E. et al. Optimization of an AMBER Force Field for the Artificial Nucleic Acid, LNA, and Benchmarking with NMR of L(CAAU). The Journal of Physical Chemistry B 118, 1216-1228, doi:10.1021 / jp408909t (2014). 32. Lind, K. E., Sherlin, L. D., Mohan, V., Griffey, R. H. & Ferguson, D. M. in Molecular Modeling of Nucleic Acids Vol. 682 ACS Symposium Series Ch. 3, 41-54 (American Chemical Society, 1997). 33. Aduri, R. et al. AMBER Force Field Parameters for the Naturally Occurring Modified Nucleosides in RNA. Journal of Chemical Theory and Computation 3, 1464-1475, doi:10.1021 / ct600329w (2007). 34. Iwamoto, N. et al. Control of phosphorothioate stereochemistry substantially increases the efficacy of antisense oligonucleotides. Nat Biotech advance online publication, doi:10.1038 / nbt.3948 http: / / www.nature.com / nbt / journal / vaop / ncurrent / abs / nbt.3948.html#supplementaryinformation (2017). 35. Glaser, S. P. et al. Anti-apoptotic Mcl-1 is essential for the development and sustained growth of acute myeloid leukemia. Genes & Development 26, 120-125, doi:10.1101 / gad.182980.111 (2012). 36. Collingwood, M. A. et al. Chemical Modification Patterns Compatible with High Potency Dicer-Substrate Small Interfering RNAs. Oligonucleotides 18, 187-200, doi:10.1089 / oli.2008.0123 (2008). 37. Yin, P. & Pierce, N. A. Triggered RNAi. USA patent US8318921B2 (2008). 38. Lee, H. Y., Zhou, K., Smith, A. M., Noland, C. L. & Doudna, J. A. Differential roles of human Dicer-binding proteins TRBP and PACT in small RNA processing. Nucleic Acids Research, doi:10.1093 / nar / gkt361 (2013). 39. Han, S. P., Goddard, W. A., SCHERER, L. & Rossi, J. J. (Google Patents, 2015). 40. Silverman, S. K. Control of macromolecular structure and function using covalently attached double-stranded DNA constraints. Molecular BioSystems 3, 24-29, doi:10.1039 / B614116A (2007). 41. Engelen, W., Janssen, B. M. G. & Merkx, M. DNA-based control of protein activity. Chemical Communications 52, 3598-3610, doi:10.1039 / C5CC09853J (2016). 42. Mukherjee, P., Leman, L. J., Griffin, J. H. & Ghadiri, M. R. Design of a DNA-Programmed Plasminogen Activator. Journal of the American Chemical Society 140, 15516-15524, doi:10.1021 / jacs.8b10166 (2018). 43. Colasanti, A. V., Lu, X.‐J. & Olson, W. K. Analyzing and Building Nucleic Acid Structures with 3DNA. e4401, doi:doi:10.3791 / 4401 (2013). 44. Zadeh, J. N. et al. NUPACK: Analysis and design of nucleic acid systems. Journal of Computational Chemistry 32, 170‐173, doi:10.1002 / jcc.21596 (2011). 45. Mathews, D. H., Sabina, J., Zuker, M. & Turner, D. H. Expanded sequence dependence of thermodynamic parameters improves prediction of RNA secondary structure11 Edited by I. Tinoco. Journal of Molecular Biology 288, 911‐940, doi:https: / / doi.org / 10.1006 / jmbi.1999.2700 (1999). 46. Camacho, C. et al. BLAST+: architecture and applications. BMC Bioinformatics 10, 421‐421, doi:10.1186 / 1471‐2105‐10‐421 (2009). 47. Tolstrup, N. et al. OligoDesign: optimal design of LNA (locked nucleic acid) oligonucleotide capture probes for gene expression profiling. Nucleic Acids Research 31, 3758‐3762 (2003). 48. Scherer, L. J., Frank, R. & Rossi, J. J. Optimization and characterization of tRNA‐shRNA expression constructs. Nucleic acids research 35, 2620‐2628, doi:10.1093 / nar / gkm103 (2007). 49. Macke, T. J. & Case, D. A. in Molecular Modeling of Nucleic Acids Vol. 682 ACS Symposium Series Ch. 24, 379‐393 (American Chemical Society, 1997). 50. Duan, Y. et al. A point‐charge force field for molecular mechanics simulations of proteins based on condensed‐phase quantum mechanical calculations. Journal of Computational Chemistry 24, 1999‐2012, doi:10.1002 / jcc.10349 (2003). 51. Aduri, R. et al. AMBER Force Field Parameters for the Naturally Occurring Modified Nucleosides in RNA. Journal of Chemical Theory and Computation 3, 1464‐1475, doi:10.1021 / ct600329w (2007). 52. Condon, D. E. et al. Optimization of an AMBER Force Field for the Artificial Nucleic Acid, LNA, and Benchmarking with NMR of L(CAAU). The Journal of Physical Chemistry B 118, 1216‐1228, doi:10.1021 / jp408909t (2014). 53. Lind, K. E., Sherlin, L. D., Mohan, V., Griffey, R. H. & Ferguson, D. M. in Molecular Modeling of Nucleic Acids Vol. 682 ACS Symposium Series Ch. 3, 41‐54 (American Chemical Society, 1997). 54. Wang, J., Wolf, R. M., Caldwell, J. W., Kollman, P. A. & Case, D. A. Development and testing of a general amber force field. Journal of Computational Chemistry 25, 1157‐1174, doi:10.1002 / jcc.20035 (2004). 55. Mark, P. & Nilsson, L. Structure and Dynamics of the TIP3P, SPC, and SPC / E Water Models at 298K. The Journal of Physical Chemistry A 105, 9954‐9960, doi:10.1021 / jp003020w (2001). 56. Plimpton, S. Fast Parallel Algorithms for Short‐Range Molecular Dynamics. Journal of Computational Physics 117, 1‐19, doi:http: / / dx.doi.org / 10.1006 / jcph.1995.1039 (1995). 57. Pettersen, E. F. et al. UCSF Chimera-A visualization system for exploratory research and analysis. Journal of Computational Chemistry 25, 1605‐1612, doi:10.1002 / jcc.20084 (2004). 58. Colasanti, A. V., Lu, X.‐J. & Olson, W. K. Analyzing and Building Nucleic Acid Structures with 3DNA. e4401, doi: doi:10.3791 / 4401 (2013). 59. Mathews et al., Expanded Sequence Dependence of Thermodynamic Parameters Improves Prediction of RNA Secondary Structure, J. Mol. Biol. (1999) 288: 911-940. 60. U.S. Patent No. 9,029,524. 61. Meggers, et al., Acc. Chem. Res. (2010) 43(8): 1092-1102. 62. Schlegel et al., J. Am. Chem. Soc. (2017) 139(25): 8537-8546.

Claims

1. A conditionally activatable small interfering RNA (Cond-siRNA) construct for use in reducing the expression of target genes within cells, comprising a sensor strand, a core strand, and a guide strand, (i) The sensor chain and the core chain form a sensor double chain, (ii) The guide strand and the core strand form an RNAi double helix, (iii) The sensor double helix is ​​linked to the RNAi double helix to form a single structure; The input strand binds to the sensor strand, causing the substitution of the sensor strand from the Cond-siRNA, thereby activating the RNAi in the Cond-siRNA construct; and, (a) Less than 50% of the skeletal bonds within the sensor double-strand domain of the sensor chain are phosphorothioate (PS) bonds; (b) The terminal base at the 5' end of the core chain is a 2'-F base or a 2'-O-methyl base, and / or the terminal base at the 3' end of the core chain is a 2'-F base or a 2'-O-methyl base; and / or (c) The three bases of the core strand that have formed a base pair with the 3' end of the guide strand are M ＊ + ＊ M pattern (wherein M is a 2'-F base or a 2'-O-methyl base) ＊ (where + is a PS bond, and + is an LNA base or another 2'-4' bridged base); The aforementioned target gene is associated with a disease or disorder; Cond-siRNA construct.

2. The Cond-siRNA construct according to claim 1, wherein the sensor strand has an overhang that is not complementary to the core strand, and the overhang of the sensor strand binds complementaryly to the input strand to form a foothold, thereby causing the sensor strand to be replaced from the core strand.

3. The above Cond-siRNA underwent the following chemical modifications A to D: A. The double-stranded domain of the sensor chain has one or more of the following features a to c: a. More than 50% of the base has been chemically modified to resist nuclease degradation or to increase the double-strand melting temperature (Tm); b. 100% of the bases are chemically modified to resist nuclease degradation and to increase Tm; c. Approximately 10% to 50% of the bases are Loc nucleic acids (LNAs) or other chemically modified bases having 2'-4' crosslinks that substantially increase the Tm; B. The 5' and 3' ends of the core chain have one or more of the following characteristics a to g: a. The three terminal bases of the 5' are pattern MRM, where M is a modified base (2'-O-methyl base, 2'-F base) and R is an RNA base; b. The three terminal bases of the 3' are pattern MRM, where M is a modified base (2'-O-methyl base, 2'-F base) and R is an RNA base; c. The three terminal bases at 3' and 5' do not have consecutive PS skeleton modifications; d. The portion of the core chain that forms base pairs with the sensor chain has an alternating chemical modification pattern (MR) n; e. The above characteristics, wherein M is a chemically modified base that does not decrease the double-stranded Tm when compared with an equivalent RNA base; f. The 5' and 3' ends of the core chain are any combination having at least one of the features a to e; g. The 3' and 5' regions of the core strand that form base pairs with the sensor strand are (a) The pattern (M) is n, where M is a 2'-O-methyl base or a 2'-F base, or (b) At least 50% of the bases in this region are 2'-O-methyl bases or 2'-F bases, and 30%, 50%, 80%, or 100% or less of the skeletal bonds are not phosphorothioates; C. A core strand in which three bases that form a base pair with the 3' end of the guide strand have one or more of the following characteristics a to f: a. M ＊ + ＊ +Pattern; b. + ＊ + ＊ +Pattern; c.R ＊ + ＊ an M pattern, where R is an RNA base; d. R ＊ + ＊ +Pattern; e. + ＊ M ＊ M pattern; f. Patterns a to e, ＊ The pattern may be either a PS bond or an unmodified (phosphodiester) bond; D. A guide chain having one or more of the following characteristics a to g: a. 30% to 95% of the bases are chemically modified bases (2'-O-methyl bases, 2'-F bases, LNAs, 2'-4' bridged bases); b. The two terminal bases of the 5' position are chemically modified; c. The two terminal bases of the 5' position have at least one LNA; d. The two terminal bases of the 5' are PS-linked; e. Both terminal bases of the 3' position are chemically modified; f. Approximately 5% to 50% of the skeletal connections are PS; g. The base adjacent to the Dicer cleavage site is not chemically modified. A Cond-siRNA construct according to claim 1, having one or more of the following.

4. The Cond-siRNA construct according to claim 1, wherein the double-stranded domain of the sensor strand has LNA modification, 2'-O-methyl modification, or both.

5. The Cond-siRNA construct according to claim 1, wherein the double-stranded domain of the sensor strand does not have phosphorothioate (PS) modification.

6. The Cond-siRNA construct according to claim 2, wherein the Cond-siRNA has a chemical modification of the foothold domain of the sensor strand, including LNA modification, 2'-O-methyl modification, PS modification, or a combination thereof.

7. The Cond-siRNA construct according to any one of claims 1 to 6, wherein the sensor double helix is ​​23 bp.

8. The Cond-siRNA construct according to claim 1, wherein the RNAi double helix is ​​23 bp.

9. The Cond-siRNA construct according to any one of claims 1 to 6, wherein the input strand is mRNA, miRNA, or non-coding RNA.

10. The Cond-siRNA construct according to any one of claims 1 to 6, wherein the use comprises contacting the cells with an effective amount of Cond-siRNA, the concentration of the effective amount of Cond-siRNA being 0.016 nM to 2 nM.

11. The Cond-siRNA construct according to any one of claims 1 to 6, wherein the knockdown activity of the Cond-siRNA in the "on" state is 90% or more.

12. The Cond-siRNA construct according to any one of claims 1 to 6, wherein the use comprises contacting the cells with an effective amount of Cond-siRNA, the effective amount being 2 nM.

13. The Cond-siRNA construct according to any one of claims 1 to 6, wherein the knockdown activity of the Cond-siRNA in the "off" state is less than 25%.

14. The Cond-siRNA construct according to claim 13, wherein the use comprises contacting the cells with an effective amount of Cond-siRNA, the effective amount being 2 nM.

15. The Cond-siRNA construct according to any one of claims 1 to 6, wherein the input strand is an intracellular RNA transcript.

16. The Cond-siRNA construct according to any one of claims 1 to 6, wherein the guide strand can hybridize with at least a portion of the mRNA of the target gene.

17. The Cond-siRNA construct according to any one of claims 1 to 6, wherein the activation of the RNAi comprises, after the replacement of the sensor strand, digesting the 5' and 3' overhangs of the core strand in the RNAi double helix with one or more endogenous nucleases to generate a double helix that can be processed by Dicer.

18. The Cond-siRNA construct according to any one of claims 1 to 6, wherein the cells are mammalian cells.

Images