Programmable conditional SIRNA and its uses

Abstract

Disclosed herein are programmable, conditionally activated small interfering RNA constructs (Cond-siRNAs) and methods of preparing and using them as therapeutic agents. Cond-siRNAs include sensor strands, core strands, and guide strands that cross to form sensor duplexes and RNAi duplexes that are connected to each other to form a single structure. When the input strands bind to the sensor strands, the Cond-siRNAs are activated and release RNAi targeting the desired gene.

Description

[0001] Priority declaration

[0002] This application claims priority to U.S. Provisional Application No. 62 / 717,686, filed on August 10, 2018, and U.S. Provisional Application No. 62 / 811,183, filed on February 27, 2019, the contents of which are incorporated herein by reference in their entirety, including the drawings.

[0003] Government Interest Statement

[0004] This invention was made in part with government support from the National Science Foundation (NSF) under Grant Nos. NSF EFRI-ODISSEI 1332411, and CMMI-SNM 1120890. The government has certain rights in this invention.

[0005] background

[0006] Nucleic Acid Nanotechnology 1-3 and biomolecular computing 4 The long-term goal is to develop conditionally activated oligonucleotide therapeutics that can detect and respond to cellular expression of specific genes. 3，4 Based on toehold-mediated strand displacement 5，6 Nucleic acid switches have performed logical operations and have been used in bacteria 7 and mammalian cells 3，8 RNA transcripts have been detected in , but conditional activation of oligonucleotide drugs by RNA transcripts in mammalian cells has not been convincingly demonstrated. Significant challenges include poor suppression of background drug activity, weak ON-state drug efficacy, overlapping input and output sequences, high design complexity, short device life (<24 hours), and high device concentrations required (>10 nM).

[0007] In the past decade, synthetic RNAi triggers such as small interfering RNA (siRNA) have emerged 10 ) has become a ubiquitous tool in biological research, and extensive basic and clinical development efforts recently led to FDA approval of the first RNAi drug, ONPATTRO 11 Despite the rapid development of the drug development pipeline and the availability of excipients and delivery technologies for targeting ligands, 9 Although RNAi has been widely developed, the difficulty in delivering RNAi agents to specific populations of disease-related cells continues to limit the potential of RNAi therapeutics. 3，8，16-19 , but for chain permutation switches 12-15Repeated attempts to develop programmable, conditionally activated RNAi agents have not convincingly demonstrated the expected effects. Therefore, novel conditionally activated siRNAs (Cond-siRNAs) are provided herein to overcome the problems in the art. Summary of the invention

[0008] On the one hand, the present disclosure relates to a programmable, conditionally activated small interfering RNA construct (Cond-siRNA) in an OFF state, the construct comprising a sensor strand, a core strand, and a guide strand, wherein the sensor strand and the core strand are complementary bound to form a sensor duplex, the guide strand and the core strand are complementary bound to form an RNAi duplex, and the sensor duplex and the RNAi duplex are connected to each other to form a single structure. In some embodiments, the sensor duplex and the RNAi duplex are connected to each other via the core strand. In some embodiments, the sensor strand is complementary bound to a fragment on the 5' and 3' of the core strand to form the sensor duplex, and the guide strand is complementary bound to a fragment in the middle of the core strand to form the RNAi duplex, and the middle fragment does not contain any 5' or 3' sequence of the core strand, so that the sensor duplex and the RNAi duplex are connected to each other via two different fragments of the core strand. In some embodiments, the sensor domain of the Cond-siRNA construct comprises a sensor duplex formed by complementary binding of the sensor strand to the 3' and 5' fragments of the core strand, and a sensor overhang that is not paired with the core strand. The sensor overhang is located at the 3' end or the 5' end of the sensor strand. In some embodiments, the sensor strand, the core strand, and the guide strand form the single construct via self-assembly when in contact with each other. In some embodiments, the sensor duplex comprises 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 comprises 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, or 30 base pairs.

[0009] In some embodiments, the Cond-siRNA is chemically modified to further improve the stability of the OFF state and / or the separation efficiency when turned on when in contact with the input strand. For example, the bases of the sensor strand in the duplex region are modified by LNA, 2'-O-methyl, or both, but not by phosphorothioate (PS); one or both ends of the core strand are modified by PS, 2'-O-methyl, or both; or the single-stranded overhang of the sensor strand is modified by LNA, 2'-O-methyl, PS, or any combination thereof. In some embodiments, the sensor domain and the RNAi domain are modified by different chemical modifications.

[0010] The OFF state Cond-siRNA is activated or turned on by bringing it into contact with an input strand, wherein one end of the input strand forms a toehold with the overhanging end of the sensor strand, so as to induce displacement of the sensor strand from the core strand through complementary binding of the input strand to the sensor strand, thereby forming a waste duplex, whereby the RNAi duplex is completely separated from the construct.

[0011] BRIEF DESCRIPTION OF THE DRAWINGS

[0012] This application contains at least one drawing executed in color. Copies of this application with one or more color drawing(s) will be provided by the Patent and Trademark Office upon request and payment of the necessary fee.

[0013] 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 RNAi duplexes with the X-ray crystal structure of Giardia Dicer shows the huge steric conflict between the sensor duplex and Dicer. Figure 1b shows RNAi activation via strand displacement. When the complementary input RNA meets the Cond-siRNA (I), the input forms a foothold with a 3' or 5' single-stranded overhang on the sensor strand (II), resulting in strand displacement (III). The displacement causes the sensor strand to separate from the RNAi duplex (IV). Cellular nucleases remove the core strand overhang on the RNAi duplex (V), leaving an active RNAi trigger for Dicer processing (VI). Figures 1c and 1d show molecular dynamics (MD) optimization models of constructs I.1 and III.1. The green box arrow indicates the direction of strand displacement from the 3' to 5' end of the sensor strand.

[0014] Figure 2(SEQ ID NOS: 77 and 113) show the intracellular degradation of chemically modified single-stranded overhangs. Northern blots show the gradual degradation of 5' overhangs protected by thiophosphate (PS) on the test constructs. The test constructs (Dicer substrates with segmented passenger strands and multiple 5' overhangs) were transfected into HCT116 cells for 24 hours. Total RNA was extracted and analyzed via Northern blots. Two similar sets of samples were analyzed, showing similar results at different loading concentrations.

[0015] Figures 3a-3f show sequence maps and optimized region maps of the prototype construct (Figure 3a) (SEQ ID NOS: 1-12), Construct II variants (Figures 3b, 3c) (SEQ ID NOS: 13-36), Construct III variants (Figures 3d, 3e) (SEQ ID NOS: 37-60), and Construct IV variants (Figure 3f) (SEQ ID NOS: 61-69).

[0016] Figures 4a-4b show Northern blots of RNA input in HCT 116 cells. Northern blot assay probes 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 matching their common leader sequence. Lanes: (L) Ambion decade marker; (0) negative control of simulated transfected RNA; (1) fully matched input RNA; (2) 5' mismatched input; (3) fully mismatched input; (4) duplex mismatched activator (unused); (5) 3' mismatched activator. The expected size of the input RNA is 145-150nt. Figure 4b shows CBFB-MYH11 RNA transcripts. Lanes: (0) mock transfection; (1) tat / rev perfect match input (for comparison); (2) CBFB-MYH11 fusion; (3) MYH11 parental; (4) MYH11 parental. Successive panels show the same sample probed with mutant tRNALys3 probe, MYH11 probe, and CBFB probe. Expression levels of input RNA were comparable in all cohorts.

[0017] Figures 5a-5k show assembly, strand displacement, RNAi activity, and RNAi activation of constructs I, II, and III variants. Figures 5a, 5d show native PAGE of construct I and II assembly, and isothermal strand displacement at 37°C in 1X PBS buffer. Constructs I and II were disassembled by their respective inputs and were not affected by mismatched inputs. Control lanes are: I = input, C = construct, P = RNAi duplex, W = discarded duplex. Figures 5b, 5e show RNAi activity of constructs in the presence of mismatched input or empty input, and ON state constructs in the presence of mismatched input. Figures 5c, 5f show RNAi activity of OFF state constructs in cells expressing inputs g and h as shown, and OFF and ON state RNAi activity of construct II variants in cells expressing mismatched inputs. Figures 5i, 5j show RNAi activity of OFF state constructs II and III variants with different core strand modifications in cells expressing mismatched or matched inputs. Figure 5k shows the RNAi activity of construct III variants with different sensor chain modifications in cells with matched or mismatched input.

[0018] Figure 6 Dicer processing of preactivated Cond-siRNA constructs is shown. Northern blot assay probing Cond-siRNA guide strands obtained from HCT 116 cells 48 hours later. 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 reported herein; (3) and (4) OFF and ON states of the prototype HIV construct; (5) and (6) OFF and ON states of the AML construct. Arrows mark the location of Dicer-cleaved guide strands. Dicer cleavage products (~21nt guide strand fragments) were detected in RNA material extracted from cells transfected with the ON state Cond-siRNA, but not in cells transfected with the OFF state Cond-siRNA.

[0019] Figure 7 shows the use of x3DNA 43 The base pair parameters for the 23 base pairs in each helix were measured every 5 ps and plotted as a heat map. The base pairs are numbered from top (1) to bottom (23) in the direction shown below. The definition of base pair parameters is explained on the website x3dna.org / articles / seeing-is-understanding-as-well-as-believing.

[0020] Figures 8a-8l show the base pair parameters of the molecular dynamics simulation of the sensor and siRNA duplexes in 5 nanoseconds. Shear (8g), buckle (8k), stretch (8i), propeller (8f), stagger (8h), and opening (81) parameters are defined for base pairs 1-23. Shift (8e), tilt (8j), slide (8d), roll (8c), rise (8b), and twist (8a) parameters are measured for base pairs 1-22 and subsequent base pairs. The heat scale of each measurement is centered on the average value measured on the 5ns control MD trajectory of four normal unmodified RNA duplexes, whose base sequences are the same as the siRNA and sensor duplexes of HIV and AML constructs. The heat scale measured from the same control trajectory spans three standard deviations. We observed significant deviations from normal RNA values ​​in twist, rise, roll, slide, shift, and torsion measurements (8a-8f). In these panels, construct 1 is the HIV construct and construct 2 is the AML construct.

[0021] Figures 9a-9c (SEQ ID NOS: 70-72) show RNAi activity, sensor design, and sequences of CBFβ-MYH11 sensing, MCL-1 targeting Cond-siRNA. Figure 9a shows RNAi activity of construct IV variants in cells expressing unrelated, fusion activator, MYH11, or CBFB transcripts. Figure 9b shows the sensor strand binding position on each transcript. Figure 9c shows a sequence map of IV.3. Yellow highlights show regions with optimized chemical modification motifs (regions A, B, C, and D).

[0022] Fig.10 a-10b shows the expression of construct I.1 ( Fig.10 a) and III.1( Fig.10 Figure 5c and 5f show the relative RNAi activity of different RNA inputs. The target expression data of Figures 5c and 5f were renormalized for the expression levels of cells transfected with Cond-siRNA but expressing completely mismatched inputs. The results show that the target expression level in cells expressing matched inputs is lower than the target expression level in cells expressing mismatched inputs. Fig.10 a: For construct I.1, target expression was reduced in cells expressing 5' mismatched and perfectly matched inputs. Fig.10b: For III.1, cells expressing the fusion input have reduced target expression.

[0023] Fig.11 Northern blots of various previous RNAi trigger designs transfected into HCT1 16 cells at 1 nM concentration for 24 hours are shown. The secondary structure of the trigger is shown. Green bubbles represent 2'-O-methyl RNA bases. Blue dots represent DNA. White dots represent RNA. Black arrows represent Dicer products. The results show that the Dicer products of duplex RNA with adjacent 2'-O-methyl modified duplexes are reduced.

[0024] Figures 12A-12C (SEQ ID NOS: 73-75) show examples of Cond-siRNAs disclosed herein. Fig. 12A The structure of the construct is shown, with the sensor domain and siRNA domain shown in grey shading and the sensor duplex shown in the box. Fig. 12B Core strand regions I, II, and III are shown, as well as exonuclease blocking regions I and II in the core strand. Fig. 12C Sensor chain regions I, II, and III are shown. DETAILED DESCRIPTION OF THE INVENTION

[0026] Disclosed herein are methods for developing programmable, conditionally activated small interfering RNAs (Cond-siRNAs). These simple riboswitches can remain intact in the mammalian cytosol for days and detect cellular RNA transcripts of specific input genes via toehold-mediated strand displacement. Upon detection of the input, the Cond-siRNA can release a potent RNAi trigger 9 , silencing a specific target gene with a sequence completely independent of the input. As shown in the working examples, the switching activity of many Cond-siRNA variants was tested in human adherent cells to determine the necessary and sufficient chemical modification motifs so that good device performance can be achieved on different input: output combinations. Some optimized Cond-siRNAs achieved more than 90% target gene silencing (protein expression relative to baseline) in cells expressing sequence-matched RNA transcripts and strongly suppressed background RNAi activity (<25% knockout) in cells expressing mismatched inputs. Therefore, this article provides a method for substantially improving the performance of chain displacement switches in living mammalian cells. Cond-siRNA technology provides a practical and versatile platform for RNAi smart drugs activated by gene expression.

[0027] The Cond-siRNA construct disclosed herein comprises two domains, namely the sensor domain and the RNAi domain, which are connected to each other by two fragments of the core chain. This structure is obtained by the complementary binding of the sensor chain to the 3' fragment and the 5' fragment of the core chain to form a sensor duplex, and the complementary binding of the guide chain to the middle fragment of the core chain to form an RNAi duplex. As shown in Figure 1, it is not that the entire sequence of the core chain is complementary to the sensor chain or the guide chain. Instead, the first fragment between the 3' fragment bound to the sensor chain and the middle fragment bound to the guide chain in the core chain remains single-stranded to form a first "bridge", and the first "bridge" connects the first end of the sensor domain to the first end of the RNAi domain. Similarly, the second fragment between the 5' fragment bound to the sensor chain and the middle fragment bound to the guide chain in the core chain remains single-stranded to form a second "bridge", and the second "bridge" connects the second end of the sensor domain to the second end of the RNAi domain.

[0028] As used herein, the phrase "complementary binding" or "complementarily binding" refers to two single strands that base pair with each other to form a double-stranded duplex. However, as long as a stable double-stranded duplex can be formed, a certain percentage of mismatch between the two single strands is allowed. For example, in some embodiments, the sensor duplex or RNAi duplex has a mismatch 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%.

[0029] A. Input Chain

[0030] The input strand of Cond-siRNA is the "trigger" that turns on Cond-siRNA, and is typically a cellular RNA transcript that appears at a relatively high expression level in a group of targeted cells (e.g., cancer cells) and at a relatively low expression level in a group of non-targeted cells (e.g., normal cells). Based on the design of the disclosed Cond-siRNA construct, RNAi is turned on only in targeted cells; while in non-targeted cells, RNAi remains in the OFF state. In targeted cells, the expression level of the input strand is at least 2 times, at least 5 times, at least 10 times, at least 20 times, at least 30 times, at least 40 times, at least 50 times, at least 60 times, at least 70 times, at least 80 times, at least 90 times, or at least 100 times higher than in non-targeted cells. Alternatively, in targeted cells, the input strand is expressed at a level 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, at least 1000 transcripts; in non-targeted cells, the input strand is expressed at a level of less than 50, less than 40, less than 30, less than 20, or less than 10 transcripts. Preferably, non-targeted cells have no detectable expression of the input strand.

[0031] The input strand includes viral mRNA, miRNA, or non-coding RNA, such as long non-coding RNA, RNA fragment, or RNA transcript. For therapeutic use, the input RNA is typically an RNA transcript expressed in a group of cells causing disease progression, and is therefore targeted for RNAi therapy. Non-targeted cells are typically a group of cells, in which the silencing of the RNAi target can cause side effects that are not conducive to treatment. In order to treat a disease or condition 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. After applying Cond-RNAi, the combination of the sensor strand and the input RNA induces the separation of RNAi duplexes from the sensor duplexes in the targeted cells, thereby activating the RNAi of the targeted disease or condition. In non-targeted cells, Cond-RNAi remains inactive.

[0032] B. Construct Design

[0033] As disclosed herein, Cond-siRNA constructs are designed to specific pairings of input and target using an iterative approach, exemplified below:

[0034] 1. Obtain the guide strand sequence for the RNAi domain from previously validated siRNA, literature data, or siRNA design tools. In some embodiments, the guide strand is 21 nt in length.

[0035] 2. Generate Dicer substrate from selected guide strand by adding four G / C rich bases at 5' of guide strand. In some embodiments, the size of Dicer substrate is 23 bp. Use Nupack (RNA strand, Mathews et al. 1999 parameters, and some dangle processing) to check that RNAi duplex is formed with >95% probability at 1 nM concentration of guide strand (antisense strand) and sense strand.

[0036] 3. Based on the sequence of the input biomarker, generate a list of all possible sensor segments that are antisense to the input strand. In some embodiments, the size of the sensor segment is 31-33 nt. For the CBFβ-MYH11 fusion sequence, only sensor segments that approximately meet the parameters shown in Figure 3b are considered.

[0037] 4. Use NCBI BLAST to rank the uniqueness of sensor sequences in the transcriptome of the target animal. For human cancer cell lines, sequences were checked against human transcript and genome collections using the BLASTn algorithm. Where possible, sensor segments with sequence complementarity exceeding 17 bases and full overhang complementarity to known or predicted RNA transcripts were eliminated.

[0038] 5. Starting from the most unique sensor segment, the core strand sequence was constructed according to the desired structural parameters of the Cond-siRNA. The core strand had a sequence of the form 5'-B-C3-P-C3-A-3', where A and B were complementary to the 5' and 3' ends of the putative duplex domain of the sensor strand, P was complementary to the putative guide strand, and C3 was the C3 linker.

[0039] 6. Use Nupack to rank the thermodynamic stability of duplexes formed between sensor segments and their corresponding 5' and 3' core strand overhangs. Use the RNA strand parameters of Mathews et al., 1999, with some overhang handling. Ideally, >95% of the strands should be base paired at 1 nM strand concentration. Also verify that the core strand does not have extensive internal secondary structure.

[0040] 7. Select the best structure (guide, core, and sensor sequences) generated in steps 1-6.

[0041] 8. Addition of chemical modifications disclosed herein.

[0042] 9. Use oligonucleotide design tools such as “LNA Oligonucleotide Tm Prediction”

[0043] ( 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 the "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 / ) to optimize the position of the LNA modification. Add LNA modifications to the sensor strand at approximately 1 LNA every 3 to 4 bases. Use the "LNA Oligo Optimizer" tool to check that the LNA pattern used does not result in secondary structure or self-pairing interactions with scores above 60. Minimize self-complementarity and self-pairing scores as much as possible. Use the "LNA Tm Prediction" tool to check the Tm of the LNA-modified oligonucleotide when paired with RNA. Use the "LNA Oligo Optimizer" tool to select positions that maximize the Tm while avoiding self-pairing interactions and secondary structures with scores above 60.

[0044] C. Design features of the core chain

[0045] As shown in FIG. 12 , in some embodiments, the core strand has one or more of the following characteristics: The 5' and 3' ends of the core strand have one or more of the following characteristics:

[0046] a. The 5' terminal base is 2'-F, 2'-O-methyl, or other modified bases that resist nuclease cleavage.

[0047] b. The 3' terminal base is 2'-F, 2'-O-methyl, or other modified bases that resist nuclease cleavage.

[0048] c. The three terminal bases at 5' have the pattern MRM, where M is a modified base (2'-O-methyl, 2'-F) and R is an RNA base.

[0049] d. The three terminal bases at 3' have the pattern MRM, where M is a modified base (2'-O-methyl, 2'-F) and R is an RNA base.

[0050] e. The three terminal bases at 3' and 5' have no consecutive PS backbone modifications.

[0051] f. The portion of the core strand that base pairs with the sensor strand has an alternating chemical modification pattern (MR) n.

[0052] g. above, wherein M is a chemically modified base that does not reduce the duplex Tm compared to an equivalent RNA base.

[0053] h. Any combination of the above, wherein both the 5' and 3' ends of the core chain have at least one feature of ag.

[0054] i. The 3' and 5' regions I and III of the core strand that base pair with the sensor strand are:

[0055] (a) consists entirely of the pattern (M)n, where M is a 2'-O-methyl or 2'-F modified base, or

[0056] (b) At least 50% of the bases in the region are 2'-O-methyl or 2'F

[0057] Furthermore, up to 30%, 50%, 80%, or 100% of the backbone linkages are not phosphorothioates. The exonuclease blocking region I, which is attached to the 5' end of the core strand (wherein three bases are base paired with the 5' end of the guide strand), has one or more of the following characteristics:

[0058] aM*+*M pattern, where M is a 2' modified base (e.g., 2'-O-methyl or 2'-F), * is a PS backbone linker, and + is an LNA base or other 2'-4' bridging base,

[0059] bM*+*+ mode, as defined above,

[0060] c.+*+*+ mode,

[0061] dR*+*M pattern, where R is an RNA base,

[0062] eR*+*+ mode,

[0063] f.+*M*M pattern, and

[0064] Patterns in ga-f, where * can be either a PS backbone linkage or an unmodified (phosphodiester) backbone linkage.

[0065] The exonuclease blocking region II of the core strand has one or more of the following characteristics:

[0066] aM*R pattern, where M is a 2' modified base (e.g., 2'-O-methyl or 2'-F), * is a PS backbone linker, and R is an LNA base or other 2'-4' bridged base,

[0067] bM*M mode, as defined above,

[0068] c.+*M mode,

[0069] d. +*R pattern, where R is the RNA base,

[0070] e.+*+ mode,

[0071] fM*+ mode, and

[0072] Patterns in ga-f, where * can be either a PS backbone linkage or an unmodified (phosphodiester) backbone linkage.

[0073] As used herein, "LNA" refers to locked nucleic acids which are widely used in the art. See, e.g.,

[0074] www.glenresearch . com / products / dna-rna-nucleosides-analogs-and- supports / hackhone-modification / locked-analog-phosphoramidites.html

[0075] 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 also known in the art. See, for example, www.glenresearch.com / browse / nucleoside-analog-phosphoramidites In some embodiments, ethylene glycol nucleic acid can be used 61，62 .

[0076] D. Design features of the sensor chain

[0077] 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.

[0078] The bases of the sensor chain have one or more of the following characteristics: (1) at least 25% of the bases are not RNA or DNA, (2) at least 50% of the bases are not RNA or DNA, (3) at least 75% of the bases are not RNA or DNA, (4) 100% of the bases are not RNA or DNA, (5) at least one base is LNA or an LNA analog, (6) at least one 3-terminal base 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) the bases that are not LNA are one or more of the following: (a) 2'-O-methyl, (b) 2'-fluoro, (c) 2'MOE, (d) ethylene glycol nucleic acid 61，62 , and (e) in www.glenresearch.com / browse / nucleoside-analog-phosphoramidites Other variations shown in .

[0079] Phosphorothioate backbone linkages can be present at multiple positions, such as between the terminal base and the penultimate base, between the last 3 bases at the 5' or 3' terminus, between the last 5 bases at the 5' or 3' terminus, between the last 8 bases at the 5' or 3' terminus, at 50% of the backbone linkages, between all bases, and / or at the junctions between regions I and II and between regions II and III.

[0080] The terminal modifications at the 5' end, 3' end, or both include one or more of the following: (1) triethylene glycol or hexaethylene glycol spacer; (2) C3 spacer; (3) inverted dT; (4) amine linker; (5) other linkers or terminal modifications known in the art, such as

[0081] eu.idtdna.com / pages / products / custom-dna-rna / oligo-modifications, and those listed at www.glenresearch.com / browse / labels-and-modiffiers, and (6) modifications can be used to link the 3' and 5' ends to other chemical moieties such as 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.

[0082] As shown in FIG3 , region C of the sensor chain has one or more of the following features: (1) 50% or less of the backbone positions are phosphorothioate (PS) linkages, (2) 50% or more of the bases are chemically modified to resist nuclease degradation or to increase the melting point (Tm) of the duplex, (3) 100% of the bases are chemically modified to resist nuclease degradation and increase Tm, (4) about 10% to 50% of the bases are LNA or other chemically modified bases with 2'-4' bridging modifications that substantially increase Tm. See, e.g., www.glenresearch.com / products / dna-rna-nucleosides-analogs-and-supports / backbone-modification / locked-analog-phosphoramidites.html ,

[0083] and en.wikipedia.org / wiki / Locked_nucleic_acid.

[0084] In some embodiments, non-LNA modifications include 2'-O-methyl and 2'-F, as well as www.glenresearch.com / browse / nucleoside-analog-phosphoramidites In some embodiments, ethylene glycol nucleic acid can be used 61，62 .

[0085] On the one hand, compared with other RNAi molecules known in the art greater than 10nM, the transfection concentration of programmable conditionally activated RNAi (such as Cond-siRNA) disclosed herein in mammalian cells is less than 0.1nM. Cond-siRNA has activity for a long period of time, such as 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 has an activity of up to 30 days, up to 60 days, or up to 90 days.

[0086] As used herein, the term "programmable" means that the Cond-RNAi constructs disclosed herein are designed to allow the sequence of the input strand to be changed without substantially changing the secondary and tertiary structures of the constructs. Other design principles are disclosed in US 9,115,355, the contents of which are incorporated herein by reference in their entirety.

[0087] In some embodiments, the separation of the RNAi duplex from the input-sensor duplex is increased by mismatch or wobble pairing between the core strand and the sensor strand. The RNAi trigger needs to be completely separated from the discarded duplex formed by the input RNA and the sensor strand to have effective RNAi activity. Previous approaches to conditional RNAi typically feature an activated RNAi trigger that remains connected 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 Dicer substrates to adjacent chemically protected duplex RNA domains significantly reduced RNAi activity ( Figure 2 ). A possible reason for this is the binding of Dicer inhibitory proteins (such as PACT38) to the extended duplex. Regardless, separation of the RNAi duplex from the input-sensor duplex is critical to allow simultaneous optimization of OFF-state RNAi suppression and ON-state RNAi efficacy.

[0088] In some embodiments, the Cond-siRNA disclosed herein has a single construct design comprising a sensor duplex linked to an RNAi duplex. Existing designs of conditional RNAi triggers and other DNA circuits are based on a single construct. 14，15，39 or multiple constructs 17 One of the schemes is characterized by translating the input sequence into an independent output sequence. In theory, a single construct translator should be more efficient in signal detection and transduction. However, the disadvantage is that the RNAi trigger must be hidden in the construct, which creates opportunities for spurious RNAi activation caused by construct degradation.

[0089] In some embodiments, the Cond-siRNA disclosed herein is chemically modified. For strand displacement sensors, the duplex domain of the sensor chain has LNA modification, 2'-O-methyl modification, or both. In some embodiments, the duplex domain of the sensor chain does not have phosphorothioate (PS) modification. In some embodiments, one or both ends of the core chain or the protection chain are modified with one of PS or 2'-O-methyl. In some embodiments, thermodynamically stable modifications generally improve the inhibition of background activation, while thermodynamically unstable modifications (such as PS backbones) can increase false activity when widely used in duplex regions. In some embodiments, there are chemical modifications in the foothold domain, including LNA modification, 2'-O-methyl modification, PS modification, or a combination thereof. These modifications in the foothold domain improve the base pairing affinity and nuclease resistance of the single-stranded overhang.

[0090] As used herein, a binding partner of a sensor chain may also be referred to as a "protection chain." A protection chain is a chain homologous to an input chain. A protection chain may be a core chain. A sensor chain is a chain complementary to an input chain. An input that binds to a sensor chain displaces the protection chain.

[0091] In some embodiments, the chemical modification comprises: (i) modification of the sensor strand with LNA and 2'-O-methyl, wherein the single-stranded toehold region has a PS backbone modification and the base-paired duplex region does not have a PS backbone modification, (ii) one or both ends of the guard strand have a 2'-O-methyl modification, or both (i) and (ii).

[0092] In some embodiments, the 3' and 5' terminal regions of the core strand are modified with PS or 2'-OMe, making the Cond-siRNA construct highly stable when base-paired with the sensor strand, but easily degraded when not paired. The degradation of the single-stranded overhang can be terminated at the end of the duplex by the nuclease blocking domain.

[0093] Complementing targeted drug delivery with self-regulating "smart drugs" that can activate or inactivate RNAi activity in response to cell-specific biochemical signals provides a means to overcome current limitations. For example, RNAi smart drugs that can detect and respond to viral RNA transcripts can kill virus-infected cells via viral RNA-activated silencing of survival-essential host genes, potentially eliminating persistent viral infections. This pharmacodynamic-centric approach is well suited for the application of toehold-mediated strand displacement-based nucleic acid switches, as these switches can sense and respond to specific base sequences in DNA or RNA inputs. 3 .

[0094] In one aspect, the Cond-siRNA disclosed herein is a self-assembled molecular mechanical transducer composed of a sensor strand, a guide strand, and a core strand, as shown in Figures 1a, 1c, and 1d. The base pairs of these three strands together form a double crossover construct. 20 The double crossover construct consists of a 23-base pair (bp) sensor duplex coupled to a 23-bp RNAi duplex. In the assembled configuration, the sensor inhibits enzymatic processing of the RNAi duplex, thereby keeping RNAi activity OFF in the mammalian cytosol (Figure 1a).

[0095] In some embodiments, the Cond-siRNA construct is further modified by chemical modification to improve its switching performance. Specifically, when the construct is in its OFF state, the unwanted RNAi activity of the Cond-siRNA disclosed herein is reduced, and in the presence of a cellular RNA transcript with the correct sequence, the switching from the OFF to the ON state is improved.

[0096] Inhibition of enzymatic processing by RNAi duplexes occurs in multiple ways that can have synergistic effects in combination. For example, during binding to RNAi duplexes, the position of the sensor duplexes greatly overlaps with the space required for the dsRBD and endonuclease domains of Dicer (Figure 1a, right). 21 In another example, the PAZ domain of Dicer interacts with the 5' terminal phosphate and 3' dibasic overhang of the canonical dsRNA trigger end to stabilize it. 22 . Assembled Cond-siRNAs prevent these interactions because the 5' and 3' ends of the core strand extend beyond the ends of the RNAi duplex into the middle of the sensor duplex (Figure 1c, 1d). The RNAi duplex is 23-bp long. This makes it too long to bypass Dicer cleavage for RNAi loading via alternative pathways 23 In addition, the abnormal tertiary structure of this construct may also hinder alternative loading pathways (TRBP and Argonaute(1-4) 21 ) required interactions with RNAi pathway proteins. Additionally, to prevent premature degradation of the sensor duplex or accidental separation of the sensor from the RNAi duplex, the entire sensor strand and other key sites on the construct were thermodynamically stabilized and protected from nuclease activity through extensive chemical modifications. 24 (Figure 1c, 1d).

[0097] To release and activate the RNAi duplex, the cellular RNA transcript induces the sensor strand to separate from the core strand via toehold-mediated strand displacement ( FIG. 1 b ). In some embodiments, the cellular RNA transcript is derived from an internal source of the subject, such as a metabolite. The displacement can be initiated from a toehold formed at the 3' or 5' end of the sensor strand. The sequence specificity of the input detection is ensured by the strong dependence of the strand displacement kinetics on the toehold stability in the toehold region. 25 , while in the duplex region the sequence specificity of input detection is ensured by the need to displace the fully complementary core strand.

[0098] After separation from the sensor strand, the displaced regions of the core strand become 3' and 5' overhangs extending from the RNAi duplex. The overhangs are degraded by cytosolic nucleases 26 (Fig. 1b V and Figure 2 ), which is terminated at the end of the RNAi duplex by a chemically modified nuclease blocking domain (Fig. 3a, yellow highlight in construct I.1). This leaves an active RNAi trigger for Dicer processing and RISC loading (Fig. 1b, VI, siRNA).

[0099] The strand compositions used for intracellular degradation of chemically modified single-stranded overhangs are as follows:

[0100] Sequence (5'->3')

[0101] Guide A: mCmG CGUCUGAGGGAUCUCUAGU UACCUU (SEQ ID NO: 76)

[0102] Guide B: mCmG+CGUCUGAGGGAUCUCUAGU+TACCUU (SEQ ID NO: 77)

[0103] 3' passenger segment: cccucagacg mc*mg*9s idT (SEQ ID NO: 78)

[0104] 5' Passenger section

[0105] 0 (Control): (SEQ ID NO: 79)c3 mG*mG*mU AACUmAGAmGAmU

[0106] 1: CG ACGAGCUCAUCA c3mG*mG*mU AACUmAGAmGAmU (SEQ ID NO: 80)

[0107] 2: 18s*C*G*A*C*G*A*A*G*C*U*C*A*U*C* c3mG*mG*mU AACUmAGAmGAmU (SEQ IDNO: 81)

[0108] 3: 18s*C*G*A*C*G*A*A*G*C*U*C*AUC c3mG*mG*mU AACUmAGAmGAmU (SEQ ID NO: 82)

[0109] 4: 18s*C*G*A*C GAGCUCAUC c3mG*mG*mU AACUmAGAmGAmU (SEQ ID NO: 83)

[0110] Nothern probe: ATCTCTAGTTACC (SEQ ID NO: 84)

[0111] L: Ambion's ten-year landmark

[0112] abbreviation:

[0113] 9s: triethylene glycol spacer

[0114] 18s: Hexaethylene glycol spacer

[0115] C3: C3 spacer

[0116] idT: inverse dT

[0117] *: Phosphorothioate backbone linkage.

[0118] like Figure 2 As shown, the sample with guide strand A had sufficient loading and exposure to visualize all bands. Lane 0 shows the position of the control strand without an overhang (15 nucleotides). Passenger 1 has a reduced amount of full-length passenger strand, with one detectable band at ~15nt, indicating rapid progressive degradation of the overhang. Passengers 2 and 3 have multiple bands and streaks in the size range of 15 to 27nt, indicating a slow, non-progressive loss of nucleotides, consistent with the presence of a PS backbone connection across the overhang. Passenger 4 shows a higher amount of full-length product compared to 0, with two bands visible near 15nt, indicating that end protection results in a slower initial degradation rate, followed by rapid progressive loss of the overhang once 5' end protection is lost.

[0119] Several Cond-siRNAs were designed with chemical modification motifs similar to those used to protect antisense oligonucleotides (ASOs) and other oligonucleotide therapeutics. 24 Construct I.1 (Figure 3a) was programmed to detect conserved HIV (tat / rev) mRNA gene sequences27 The activator sequences for constructs I and II are shown below in Table 1. Regions that should align with the sensor strand are bold and italicized; regions that should align with the toehold are underlined; segments that are complementary to the sensor strand are capitalized; segments that do not match the sensor strand are lowercase.

[0120]

[0121] Construct III.1 (Figure 3a) was programmed to detect an acute myeloid leukemia (AML)-associated fusion oncogene sequence (CBFB-MYH11) 28，29 The activator sequences for constructs III and IV are shown below in Table 2. The regions that should align with the sensor strand are bold and italicized; the regions that should align with the toehold are underlined; the segments that are complementary to the sensor strand are capitalized; the segments that do not match the sensor strand are lowercase.

[0122]

[0123]

[0124] To allow a direct readout of RNAi activity, constructs I.1 and III.1 both targeted Renilla luciferase mRNA with a biologically irrelevant target sequence in its 3'UTR (from the U5 region of HIV).

[0125] To predict the actual molecular conformations, atomistic molecular dynamics (MD) simulations of I.1 and III.1 30 The hybrid amber force field was used in explicit solvent using previously published parameters. 31-33 The combination of describes the chemical modifications used in our design. The MD optimized models of the two constructs (Figures 1c, 1d) show that the sensor and RNAi duplexes have structural deformations compared to the ideal A-form RNA helix parameters (Figure 4). This deformation is attributed to (1) the widely used chemical modifications and (2) the structural strain inherent in the double cross motif (Figures 1c, 1d, 4, and Tables 3a and 3b). The following Tables 3a and 3b show the average base pair parameters of the sensor and siRNA duplexes in the 5 nanosecond molecular dynamics trajectory. The mean and standard deviation values ​​of each base pair parameter of each marked duplex were calculated based on the data shown in Figure 4. For comparison, the mean and standard deviation of unconnected RNA duplexes with the same sequence composition as the siRNA and sensor duplexes in the HIV and AML constructs (the two constructs have the same sequence in the siRNA duplex) were also calculated.

[0126] Table 3a

[0127]

[0128]

[0129] Table 3b

[0130]

[0131] Despite the deformation, all expected base pairs were maintained throughout the simulations, and the relative dispositions of the sensor and RNAi duplexes in both models were consistent with steric protection from Dicer cleavage.

[0132] To empirically test the constructs, component strands were purchased from commercial sources and Cond-siRNA was assembled using thermal annealing in 1X PBS. Non-denaturing polyacrylamide gel (PAGE) analysis showed that both highly pure assembled constructs were able to perform sequence-specific detection of input RNA and release RNAi duplexes via isothermal strand displacement at 37°C in 1X PBS buffer (Figures 5a, 5d).

[0133] The RNAi activity of Cond-siRNA in the OFF (guide strand, core strand, and sensor strand) and ON (guide strand and core strand only) states was measured using a dual luciferase assay. Different amounts of the constructs were co-transfected with a fixed amount of a DNA vector to HCT116 cells, which encoded (a) a dual luciferase reporter and (b) a short RNA input sequence that was complementary (matched) or non-complementary (mismatched) to the sensor strand (Tables 1 and 2).

[0134] In cells expressing null or completely mismatched inputs, the OFF-state Cond-siRNA had significantly reduced RNAi activity, whereas the ON-state constructs proved to be effective RNAi triggers (Figure 5b, 5e). Figure 6 For RNAi activation, the OFF state construct was transfected into cells expressing matched or mismatched input (Figures 5c, 5f). The expression level of the input RNA was assessed using Northern blot (Figure 4). Although fluctuations in background (OFF state) RNAi activity levels reduced the apparent statistical significance when the data were normalized to target expression in cells without transfected siRNA (Figures 5c and 5f), another analysis directly comparing matched and mismatched inputs showed that RNAi activity was significantly increased under matched input (Figure 5c). Fig.10 ).

[0135] The different sensor designs of I.1 and III.1 appear to result in some functional differences in RNAi activation. Construct I.1 has 5 nucleotide 5' and 3' sensor strand overhangs. RNAi activation is not affected by mismatches with the 5' overhang input, but is completely abolished by mismatches with the 3' overhang (Figures 5c, 5f, and 10). Therefore, only the 3' sensor overhang serves as a toehold for strand displacement. In MD simulations, strand displacement from the 5' toehold of construct I.1 will first proceed toward the interior of the construct, while displacement from the 3' toehold will proceed toward the exterior (Figures 1c, 1d). Therefore, steric hindrance may play a role in preventing strand displacement from the 5' toehold. Other factors, such as binding site accessibility on the RNA input or differences in base pairing stability between the 5' and 3' toeholds, may also contribute to preventing strand displacement from the 5' toehold.

[0136] Although construct III.1 has a similarly potent ON-state RNAi trigger (Fig. 5b, 5e), it shows stronger RNAi activation (Fig. 5f) than construct I.1 (Fig. 5c). A plausible explanation is that the higher thermodynamic stability of the 8-base 3' toehold of III.1 leads to faster strand displacement kinetics.

[0137] Further optimization of chemical modifications was performed in key regions of the construct (Figure 3a, highlighted areas). The OFF and ON state RNAi activities of construct II variants were measured with different patterns of chemical modifications at the 3' and 5' ends of the core strand (Figure 3b, region A, flanking the central nick of the sensor duplex) and in the exonuclease blocking domain of the RNAi duplex (Figure 3a, region B). Figure 7 , 8, construct II.1-5).

[0138] For region B, the results show that the end thermodynamic stability of the RNAi duplex may be important for inhibiting OFF-state RNAi activity. Replacing the 2'-OMe modification at the center of the nuclease blocking module with a locked nucleic acid (LNA) modification (Fig. 9c region C) reduced background RNAi activity without compromising RNAi efficacy (Fig. 5g, II.1 and II.2).

[0139] For region A, to turn off the OFF-state RNAi activity, the 3' and 5' ends of the core strand need to be protected by chemical modification to turn off the OFF-state RNAi activity. When the PS backbone bond was replaced by a phosphodiester bond, the background RNAi activity increased significantly (Figure 5g, II.2 has 3 consecutive terminal PS, II.3 has 2 alternating PS, and II.4 has a single terminal PS), while the unmodified ends completely lost the ability to control RNAi activity (II.0).

[0140] The problem with using PS bonds in region A is that each PS is a stereocenter with two possible enantiomeric conformations. Unfortunately, the more nuclease-resistant Sp conformation destabilizes RNA base pairing34. To improve the thermodynamic stability of the sensor duplex and reduce racemic heterogeneity, the PS backbone modification was replaced by 2'-OMe base modifications (Fig. 5h, 8c constructs II.5-7). Using three consecutive (II.5), two alternating (II.6), or one terminal (II.7) 2'-OMe bases at each core strand end significantly reduced background RNAi activity. In addition, all three motifs increased ON state RNAi efficacy over PS-modified analogs (Fig. 5h, II.5-7 vs. II.2, ON state).

[0141] To determine whether 2'-OMe modification allows cellular RNA transcripts to activate RNAi, RNAi activation of II.1, II.2, II.6, and II.7 was compared using a two-step transfection protocol (Figure 5i). Both II.6 and II.7 showed significantly reduced background RNAi activity, no activation of mismatched RNA transcripts, and increased RNAi knockdown of the Renilla luciferase target in cells expressing sequence-matched input of the sensor (Figure 5i, target expression was reduced by ∼50% at 2 nM Cond-siRNA concentration).

[0142] To test the general applicability of the optimization of region A, a version of construct III with 2'-OMe protected core strand ends was created (Figure 8d, III.2) and tested for activation by cellular RNA transcripts (Figure 5j). Compared to the original PS-protected variant (III.1), III.2 had significantly reduced background RNAi activity in cells expressing mismatched inputs and better RNAi activation in cells expressing the correct inputs (Figure 5j, III.1 vs. III.2). Notably, 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 eliminated control of RNAi activity (Figure 5j, III.3).

[0143] After the core strand was optimized, the sensor strand was also improved. The initial constructs (I.1, II.1, III.1) used a fully modified sensor strand with PS, LNA, and 2'-OMe modifications in all regions. PS modifications in the sensor duplex region can increase false activation by reducing base pairing stability. In addition, other researchers have used only 2'-OMe modifications without LNA to stabilize the strand displacement switch operating in mammalian cells. 8，19. A series of construct III variants disclosed herein were tested, which combined a 2'-OMe protected core strand (region A) with a sensor strand having LNA+2'-OMe+PS (Figures 5k, 8e, III.2, III.4), LNA+2'-OMe (III.5), 2'-OMe only (III.6), or no modification (III.7) in the duplex region of the sensor strand. The results showed that the RNA (III.7) and 2'-OMe only (III.6) versions of the sensor strand were unable to suppress unexpected RNAi activity (Figure 5k). The optimal variant was III.5, which had LNA and 2'-OMe modified bases in the duplex region, but no PS modification. At 2 nM Cond-siRNA concentration, construct III.5 had only 20% background RNAi activity in cells expressing mismatched RNA input, and achieved over 90% knockdown of the Renilla target in cells expressing the correct RNA input (~10X reduction in target expression between matched and mismatched). Statistically significant RNAi activation was still seen at Cond-siRNA as low as 80 pM.

[0144] Using insights gained from optimization of the CBFB-MYH11 sensor of construct III, construct IV was developed. Construct IV is a CBFB-MYH11 sensing Cond-siRNA with an RNAi domain targeting MCL-1, an endogenous inhibitor of apoptosis that is essential for the survival of AML cells 35 (but not HCT116 cells used for dual luciferase experiments). Three versions of construct IV were tested for RNAi activation in response to transcripts with CBFB-MYH11, CBFB, or MYH11 sequences. To ensure that the sensor was activated on the fusion sequence, the sensor strand (the same as used in construct III) was designed to be complementary to the MYH11 sequence in its toehold-forming region and to the CBFB sequence in its duplex region (Figure 9b). Three versions of construct IV were tested (Figure 8f). IV.1 had a modification pattern similar to that of the prototype construct III.1; construct IV.2 had optimized regions A, B, and sensor strand motifs. IV.3 Additional 2'-O-methyl modifications in the guide strand using a pattern previously found to have little effect on RNAi efficacy 36Unexpectedly, both IV.2 and IV.3 significantly improved suppression of background RNAi activity and improved RNAi switching compared to IV.1 (Fig. 9c). IV.2 had background RNAi, which resulted in a ~60% reduction in the Renilla readout target at 2 nM Cond-siRNA, but showed a ~90% reduction in Renilla in fusion sequence expressing cells. IV.3 had almost no detectable background RNAi, but also had lower activated RNAi activity (~60% reduction at 2 nM Renilla).

[0145] In some embodiments, chemical modifications of the Cond-siRNA constructs disclosed herein include one or more of the following:

[0146] A. The sensor chain used in the highlighted area C in Figure 3a "Areas for screening chemical modifications" has one or more of the following features:

[0147] a. 50% or less of the backbone positions are phosphorothioate (PS) linkages.

[0148] b. 50% or more of the bases are chemically modified to resist nuclease degradation or to increase the melting point (Tm) of the duplex

[0149] c. 100% of the bases are chemically modified to resist nuclease degradation and increase Tm

[0150] d. About 10%-50% of the bases are locked nucleic acids (LNA) or other chemically modified bases with 2'-4' bridging modifications that substantially increase Tm.

[0151] B. The 5' and 3' ends of the core strand have one or more of the following characteristics:

[0152] a. The 5' terminal base is 2'-F, 2'-O-methyl, or other modified bases that resist nuclease cleavage

[0153] b. The terminal base at the 3' end is 2'-F, 2'-O-methyl, or other modified bases that resist nuclease cleavage

[0154] The three terminal bases at c.5' have the pattern MRM, where M is the modified base (2'-O-methyl, 2'-F) and R is the RNA base

[0155] d. The three terminal bases at 3' have the pattern MRM, where M is the modified base (2'-O-methyl, 2'-F) and R is the RNA base

[0156] e. The three terminal bases at 3' and 5' do not have continuous PS backbone modifications

[0157] f. The core strand that base pairs with the sensor strand has an alternating chemical modification pattern (MR)

[0158] g. above, wherein M is a chemically modified base that does not reduce the duplex Tm compared to the equivalent RNA base

[0159] h. Any combination, wherein the 5' and 3' ends of the core chain have at least one feature of ag

[0160] i. The 3' and 5' regions of the core strand that base pair with the sensor strand are:

[0161] (a) consists entirely of the pattern (M)n, where M is a 2'-O-methyl or 2'-F modified base, or

[0162] (b) at least 50% of the bases in the region are 2'-O-methyl or 2'F, and up to 30%, 50%, 80%, or 100% of the backbone linkages are not phosphorothioates.

[0163] C. The core strand that has three bases paired with the 3' end of the guide strand has one or more of the following characteristics:

[0164] aM*+*M pattern, where M is a 2' modified base (e.g., 2'-O-methyl or 2'-F), * is a PS backbone linker, and + is an LNA base or other 2'-4' bridging base

[0165] bM*+*+ mode, as defined above

[0166] c.+*+*+ mode

[0167] dR*+*M pattern, where R is an RNA base

[0168] eR*+*+ mode

[0169] f.+*M*M mode

[0170] Patterns in ga-f, where * can be either a PS backbone linkage or an unmodified (phosphodiester) backbone linkage

[0171] D. The guide chain has one or more of the following characteristics:

[0172] a. 30% to 95% of the bases are chemically modified bases (2'-O-methyl, 2'-F, LNA,

[0173] 2'-4' bridge base)

[0174] b. The two terminal bases at the 5' end are chemically modified

[0175] c. The two terminal bases on the 5' end have at least one LNA

[0176] d. The two terminal bases on the 5' end are connected by the PS backbone

[0177] e. Both terminal bases at the 3' end are chemically modified

[0178] f. ~5% to 50% of the backbone linkages are PS

[0179] g. Wherein the bases flanking the indicated site of Dicer cleavage are not chemically modified.

[0180] In summary, disclosed herein are programmable, conditionally activated small interfering RNAs (Cond-siRNAs). The device outputs an active RNA interference trigger against a target gene only when RNA transcripts from different input genes are detected. The improved performance of strand displacement sensors and the realization of conditionally activated RNAi are partly attributed to five key design principles.

[0181] First, the RNAi trigger needs to be completely dissociated from the discarded duplex formed by the input RNA and the sensor strand to have effective RNAi activity. Existing protocols for conditional RNAi typically feature an activated RNAi trigger that remains attached to the input signal (e.g., mRNA) via Watson-Crick base pairing. 15，17，18，37，60 During the development of the previous generation of conditional RNAi triggers, it was found that ligation of Dicer substrates to adjacent 2'-O-methyl-modified duplex RNA domains significantly reduced RNAi activity ( Fig.11 ). The possible cause is Dicer inhibitory proteins (such as PACT) 38 Binding to the extended duplex. However, separation of the RNAi duplex from the input-sensor duplex is necessary to achieve simultaneous optimization of OFF-state RNAi repression and ON-state RNAi efficacy.

[0182] Used for Fig.11 The sequence of the construct is as follows:

[0183]

[0184]

[0185] Second, the design of conditional RNAi triggers and other DNA circuits in the past has been based on single constructs. 15，16 or multiple constructs 17One of the schemes is a feature that the input sequence is translated into an independent output sequence. Theoretically, a single construct translator should be more efficient in nature in signal detection and transduction. However, a disadvantage is that the RNAi trigger must be hidden in the construct, which creates opportunities for the false RNAi activation caused by construct degradation. The success of the disclosed Cond-siRNA design shows that the advantage of a single construct translator can be utilized while effectively controlling the risk of false activation.

[0186] Third, extensive chemical modifications are key to the proper functioning of strand displacement sensors and even Cond-siRNA. Working examples of optimization experiments show that: 1) the duplex domain of the sensor strand needs to have both LNA and 2'-O-methyl modifications, but not PS modifications; 2) the ends of the protective strands need to be modified with either PS or 2'-O-methyl; 3) thermodynamically stable modifications generally improve the suppression of background activation, while thermodynamically unstable modifications (such as PS backbones) can actually increase false activity when widely used in duplex regions; 4) although chemical modifications were not tested in the toehold domain, combinations of LNA, 2'-O-methyl, and PS modifications are beneficial because they improve base pairing affinity and nuclease resistance of single-stranded overhangs.

[0187] Fourth, endogenous RNA degradation mechanisms can be an effective tool for construct switching. As shown in the working examples, the chemically modified 3' and 5' terminal regions of the core chain are highly stable when base-paired with the sensor chain, but are easily degraded when not paired. This differential sensitivity to nuclease activity can be achieved by PS-modified ends, but appears to be more effective when modified with 2'-OMe. As demonstrated herein, the degradation of single-stranded overhangs can be terminated at the end of the duplex by the nuclease exonuclease blocking domain. Using this scheme to trim the disclosed RNAi duplexes yields an effective RNAi trigger from the DX secondary structure. Similar schemes may be useful in the dynamic reconfiguration of other nucleic acid nanostructures.

[0188] Finally, the separation of the sensor and RNAi domains in two different duplexes is important for intracellular stability, programmability, and ease of engineering development. First, there is no base pairing overlap between the two domains and no competing secondary structure conformations. This ensures a large degree of thermodynamic stability and simplifies the programming of new input and output sequences. Secondly, the size of the duplex and the bonding chemistry at the intersection are configured to minimize strain in the tertiary structure (Figure 1c). This further improves thermodynamic stability. Third, since the duplexes do not overlap, chemical modifications can be applied to the sensor domain without compromising the compatibility of the RNAi domain.

[0189] The following examples are provided to better illustrate the claimed invention and should not be construed as limiting the scope of the present invention. To the extent that specific materials are mentioned, they are for illustrative purposes only and are not intended to limit the present invention. It will be apparent to those skilled in the art that multiple equivalents, changes, and modifications may be made without departing from the scope of the present invention, and it should be understood that such equivalent embodiments will be included herein. In addition, all references cited in this disclosure are incorporated herein by reference as if fully set forth herein.

[0190] Example 1: Construct design

[0191] Cond-siRNAs are designed for specific pairing of input and target using an iterative approach. Appropriate 21-nt guide strand sequences for RNAi domains are obtained from previously validated siRNAs, literature sources, or siRNA design tools. A 23-bp Dicer substrate is generated from the selected guide strand by adding four G / C-rich bases at the 5' end of the guide strand. Nupack (RNA strand, Mathews 59 Published parameters, with some overhang processing) were used to confirm that RNAi duplexes were formed with >95% probability at 1 nM of both guide (antisense) and sense strands.

[0192] Based on the sequence of the input biomarker, a list of all possible 31 to 33 nt sensor segments (antisense to the input) was generated. For the CBFβ-MYH11 fusion sequence, only sensor segments that approximately met the parameters shown in Figure 9b were considered. Sensor sequences in the transcriptome of the target animal were ranked for uniqueness using NCBI BLAST. For human cancer cell lines, sequences were checked against the human transcript + genome collection using the BLASTn algorithm. Where possible, sensor segments with sequence complementarity exceeding 17 bases and complete overhang complementarity to known or predicted RNA transcripts were eliminated.

[0193] Starting with the most unique sensor segment, the core strand sequence is selected based on the desired structural parameters of the Cond-siRNA. For example, the core strand has a sequence of the form 5'-B-C3-P-C3-A-3', where A and B are complementary to the 5' and 3' ends of the putative duplex domain of the sensor strand, P is complementary to the putative guide strand, and C3 is the C3 linker. The thermodynamic stability of the duplexes formed between the sensor strand segments and their corresponding 5' and 3' core strand overhangs is ranked using Nupack. RNA strands with the parameters published by Mathews et al. 59 are used with some overhang processing. Ideally, >95% of the strands should be base paired at 1 nM strand concentration. The core strand is also checked to confirm that it does not have a large amount of internal secondary structure.

[0194] The best construct with the above designed guide chain, core chain, and sensor chain sequences was selected and chemically modified. Exiqon's oligonucleotide design tool ( www.exiqon.com / oligo-tools ) was used to optimize the positions of LNA modifications. Approximately 1 LNA modification per 3 to 4 bases was added to the sensor strand. The LNA Oligonucleotide Optimizer tool was used to check that the LNA pattern did not result in secondary structure or self-pairing interactions with a score above 60. Self-complementarity and self-pairing scores were optimized to the greatest extent possible.

[0195] Example 2: Chain synthesis

[0196] Strands with LNA bases were synthesized by Exiqon Inc (now a division of Qiagen). Strands without LNA were synthesized by GE Life Sciences Dharmacon (now a division of Horizon Discovery Group). All ordered strands were PAGE or HPLC purified according to the manufacturer's recommendations.

[0197] Example 3: Assembly of Cond-siRNA

[0198] Cond-siRNAs are assembled by heat annealing in 1X phosphate buffered saline (PBS). Constructs can be assembled with or without purification. Assembly quality can be assessed using native gel electrophoresis on 10% to 15% PAGE in 1X TBE at 4°C.

[0199] For unpurified assembly, mix the sensor strand, core strand, and guide strand in 1XPBS pH ~7.0 at a concentration of 50 nM or 100 nM, with a molar ratio of 1.1 to 1.00 to 1.1. Using a slight excess of sensor strand and guide strand helps prevent the generation of constitutively RNAi-active guide + core duplexes. Use a PCR thermal cycler with the following program:

[0200] Heat the lid to 105°C

[0201] Hold at 85°C for 30 seconds to denature the chains

[0202] Cool down to 50℃ at a rate of 0.1℃ / sec

[0203] Keep at 50℃ for 45 minutes

[0204] Cool down to 37°C at a rate of 0.02°C / sec

[0205] Cool to 4°C at maximum speed and hold.

[0206] For the assembly to be purified, the sensor strand, core strand, and guide strand were mixed and assembled in 1xPBS at a nominal concentration of 1uM using the above annealing protocol. The assembled construct was then loaded onto a Bio-Rad mini protean 10% native PAGE gel in TBE buffer and run at 125V at 4°C for ~45 minutes. The band corresponding to Cond-siRNA was visualized under UV illumination and excised.

[0207] According to the manufacturer's instructions, the excised bands were extracted by electroelution using the Harvard Apparatus Electroprep system. The gel slice was placed in a 0.5 mL chamber sealed by a 100K MWCO filter and a 2K MWCO filter. The construct was eluted by a 100K MWCO membrane and captured in an adjacent 0.5 mL chamber formed by a 100K MWCO membrane and a second 2K MWCO membrane. Elution was performed at 4 °C for 45 minutes in a 0.1M Na2HOP4 buffer (~ pH 7.0). The power supply was set to maintain a constant current of 15 mA, and the cut-off voltage was 65 V.

[0208] The concentration of the purified construct was calculated by comparison with a Cond-siRNA standard of known concentration on native PAGE stained with SYBR Gold and quantified using a Bio-Rad ChemiDoc XRS+ Imager.

[0209] The assembled construct is best used immediately after assembly or purification. The construct can also be stored in aliquots at -80°C indefinitely. However, freeze-thaw cycles compromise construct quality and cause construct disassembly. Disassembled constructs can be reassembled by repeating thermal annealing immediately prior to experimentation.

[0210] Unpurified constructs were used for many tests because assembly yields were already high and purification did not always improve the performance of the constructs.

[0211] Example 4: Strand displacement assay

[0212] Preassembled constructs were prepared at a nominal concentration of 50 nM and combined with 50 nM oligonucleotide activator (or PBS for control) at 37°C in PBS buffer at 1:1 to obtain a mixture with 25 nM input signal and construct. The construct-input combination was then incubated in a PCR thermal cycler at a constant 37°C for 4 hours. Samples were collected at designated time points and immediately frozen at -80°C in 1X native PAGE loading dye. After the experiment, all samples were quickly thawed and analyzed using non-denaturing PAGE.

[0213] Example 5: Generation of Dual Luciferase Reporter and Activator Plasmids

[0214] All clones were generated using standard molecular biology protocols by annealing DNA oligonucleotides specific for the insert followed by ligation into the specified site of the parental vector. The accuracy of all constructs was verified by DNA sequencing.

[0215] Example 6: PsiCHECK Dual Luciferase Reporter

[0216] The DNA oligonucleotides indicated below were annealed and ligated into the XhoI and NotI sites of the psiCHECK 2 (Promega) dual luciferase reporter. The nucleotides in bold constitute the sense target sequence. The lowercase nucleotides represent the restriction site 5' overhang.

[0217] HIV U5 region targets

[0218]

[0219] MCL-1 Targets

[0220]

[0221] Example 7: RNA input transcripts for signal activation experiments

[0222] The activator sequence is expressed as part of the chimeric tRNA transcript. The first part consists of a modified 5 tRNA Lys3 The complete mature sequence is shown below. CCA prevents endonucleolytic cleavage by the tRNA precursor (pre-tRNA) processing enzyme (tRNAse Z). The tRNA Pol III promoter is internal and contained within the coding sequence of the tRNA DNA.

[0223] To clone tRNA Lys3 The first 69 nucleotides of the parent plasmid are terminated at the NruI restriction site. 48 The parent plasmid was digested with NruI, followed by nucleotide tRNA 69 to generate a blunt end. The annealed overlapping oligonucleotides encoded the remaining modified tRNA nucleotides, followed by the designated activation sequence. Each activation sequence terminated with a 12-base tetraloop (GGCGCAAGCC) (SEQ ID NO: 107), followed by a T6 run encoding the Pol III termination sequence U4+ RNA transcript sequence, as shown below.

[0224] For constructs I and II, tRNA Lys3Leader sequence, 5'->3':

[0225] (SEQ ID NO: 108) - activator sequence. The sequence in bold is the binding site for the Northern blot probe. The activator sequences are listed in Tables 1 and

[0226] Table 2. Table 5 below lists Northern blot probes.

[0227]

[0228]

[0229] Example 8: Tissue Culture

[0230] All analyses utilized HCT 116 colorectal cancer cells. Cells were maintained using McCoy's 5A basal medium (Irvine Scientific, USA) supplemented with 10% fetal bovine serum (FBS), 1.5 mM L-glutamine (Irvine Scientific, USA), and 10 mM pyruvate (Irvine Scientific, USA) without antibiotics and stored in a 37°C, 5% CO2 humidified incubator.

[0231] Example 9: Northern blot analysis

[0232] Activator expression analysis was performed in 250 μL OptiMEM and 250 μL 1:50 diluted Lipofectamine2000 using 2 μg plasmid DNA in 6-well plates. Liposomes were allowed to form according to the manufacturer's instructions and added to cells with 2 mL of fresh complete medium. The medium was changed at 18 hours, and depending on the length of the transfection, it was changed at least once every subsequent day and changed 6 hours before RNA harvest. OFF and pre-activated (ON) c-siRNA were analyzed similarly; however, the RNAi complex of the specified amount was added to 2 μg pBluescript plasmid as a carrier in 250 μL OptiMEM.

[0233] Total RNA was harvested using 1000 μL RNA Stat-60 (Tel-Test, Inc) and processed according to the manufacturer's instructions and a second organic extraction using 1:1 phenol:chloroform extraction before precipitation; RNA pellets were washed twice with 70% ethanol, then the excess ethanol was evaporated and resuspended in RNAse-free TE (pH 6.8).

[0234] For Northern analysis, 15 μg of total RNA was run on 8% (for activators) or 12% (for c-siRNA) urea-PAGE gels (15 cm) with 32P-labeled Ambion decade markers. The gels were electroblotted onto Hybond XL (GE Healthcare Life Sciences), prehybridized and hybridized at 37°C using Sigma Perfect Hyb Plus, and hybridized with 5-10 pmoles of P32-5' end-labeled oligonucleotide probe. The blots were washed 4-5 times with 2X SSC / 1% SDS at 37°C. When performing serial hybridizations, unless otherwise stated, the old oligonucleotide probe was removed from the membrane according to the manufacturer's instructions and checked by re-exposure before re-hybridization. Hybridization of U6 snRNA was used for loading control. The cloning procedures, oligonucleotides, and all probe sequences are listed in the table.

[0235] Example 10: Dual luciferase assay

[0236] Dual luciferase assays were 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 Renilla luciferase gene on the psiCHECK-2 (Promega) vector, and firefly luciferase was used as a reference control.

[0237] Cells were incubated and transfected in 48-well cluster plates. Cells were seeded one day before transfection and transfected at 50% confluency. Each experiment was repeated at least three times to obtain biological replicates. Single-step transfection scheme for Figures 5b-5h, two-step transfection scheme for Figures 5i-5k and Figure 9a.

[0238] Single-step co-transfection protocol:

[0239] For each experiment, a premix of psiCHECK (Promega Corporation) reporter plasmid in Opti-MEM (Thermo Fisher Scientific) was generated. The premix was divided into equal portions and either a pBluescript (Agilent) control or an activator plasmid was added. The new mix was then divided again to add different concentrations of Cond-siRNA complexes. 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.

[0240] For each experimental condition (combination of activator and c-siRNA at a specific concentration), prepare enough mixture (3.3x the amount required) to transfect 3 separate wells as technical replicates.

[0241] Therefore, each well in a 48-well plate received 40 μL of transfection mixture, consisting of 16 μL (psiCHECK and activator plasmids in OptiMEM), 4 μL (50X Cond-siRNA in 1X PBS buffer), and 20 μL (1:50 dilution of L2K). PBS is phosphate-buffered saline without calcium or magnesium, treated with DEPC (diethylpyrocarbonate) to remove any RNAse activity.

[0242] Just before transfection, the medium in each well was replaced with 160 μL of fresh medium, and then 40 μL of transfection mixture was added to a final volume of 200 μL / well, containing: 40 nanograms (ng) of psiCHECK-2 dual-luciferase reporter plasmid, 120 ng of pBluescript or activator expression plasmid, and the indicated concentrations of Cond siRNA complexes.

[0243] Two-step transfection protocol

[0244] This protocol was used to generate the data for Figures 5i-5k and Figure 9a.

[0245] Target and activator plasmid transfection 1, time - 8 hours

[0246] For two-step transfection, a premix of psiCHECK (Promega) reporter plasmid in Opti-MEM (Thermo Fisher Scientific) was generated. The premix was divided into equal portions and then pBluescript (Agilent) control or an activator plasmid was added. A 1:50 dilution of Lipofectamine 2000 (L2K) was added to the plasmid mixture in a 1:1 volume ratio to achieve the manufacturer's recommended 1:100 dilution of L2K and incubated at room temperature to form lipoplexes according to the manufacturer's recommendations.

[0247] For each experimental condition, sufficient mixture was prepared to transfect 3 separate wells as technical replicates. Thus, each well in a 48-well plate received 40 μL of transfection mixture, consisting of 20 μL (psiCHECK and activator plasmids in Opti-MEM), and 20 μL (1:50 dilution of L2K).

[0248] Just before transfection, the medium in each well was replaced with 160 μL of fresh medium, and then 40 μL of transfection mix was added to a final volume of 200 μL / well, containing: 40 nanograms (ng) of psiCHECK-2 dual-luciferase reporter plasmid, and 120 ng of pBluescript or activator-expressing plasmid.

[0249] Remove the transfection mixture and wash gently with medium for about 2 hours after 6 hours. Add 160ul of fresh medium to each well and continue cell incubation until the second transfection.

[0250] Cond-siRNA complex transfection 2, time-0

[0251] 8 hours after transfection 1, RNAiMAX reagent (Thermo Fisher) was used to transfect Cond-siRNA at different concentrations specified in the experiment. For each experimental condition, enough Cond-siRNA of each concentration was prepared for three technical replicates of each target / activator combination in PBS. Each Cond-siRNA dilution was mixed with an equal volume of 1:50 RNAiMAX in OptiMEM and incubated at room temperature to form liposome complexes according to the manufacturer's instructions. Specifically, each well in a 48-well plate received 40 μL of transfection mixture, consisting of 20 μL Cond-siRNA (8 μL PBS+12 μL Opti-MEM) at a final concentration of 10X, and 20 μL 1:50 diluted RNAiMAX.

[0252] maintain

[0253] For the one-step protocol, time 0 is marked as the time when the co-transfection mixture was added to the cells, and for the two-step protocol, time 0 is the time when the Cond-siRNA complex (transfection #2) was added. The culture medium was changed 18 hours after transfection, at least once every day thereafter, and changed 6 hours before preparing the lysate.

[0254] Lysate preparation

[0255] At the designated time point for each experiment, remove the 48-well plate. Carefully aspirate the culture medium from each well. Then wash the wells once with 1XPBS and blot dry. Add 100 μL 1X Promega passive lysis buffer to each well. Then cover the plate with aluminum foil and freeze at -80°C or place on a shaker and gently stir (~70rpm) at room temperature for ~30 minutes. If frozen, thaw the cells by gently stirring on a shaker for at least 30 minutes before the dual luciferase assay. Before the assay, visually inspect the wells to ensure that the cells are fully lysed.

[0256] Test Stirrup

[0257] According to the manufacturer's instructions, dual luciferase reporter assay kit (Promega) was used to analyze cell lysates. Renilla luciferase values ​​were standardized to firefly luciferase in each technical repeat (per well). The technical three replicates in the experiment were averaged to obtain a single biological replicate value. All figures represent the results of at least three independent biological replicate experiments.

[0258] Example 11: Molecular dynamics simulation

[0259] Atomic model of Cond-siRNA using Nucleic Acid Builder 49 and custom scripts built using Accelrys (now BIOVIA, a division of Dassault Systèmes) Cerius 45 The package was edited to generate the appropriate chemical modifications.

[0260] By combining previously reported RNA 50 , 2'-O-methyl 51 , LNA 52 , and phosphorothioates 53 The modified amber force field parameters generated a heterogeneous force field (FF). No parameter set for LNA thymidine was previously reported. 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. Charges were calculated using the RESPESP Charge Derivation (RED) server (q4md-forcefieldtools.org / REDServer / ). FF parameters for non-DNA components such as the C3 linker, terminal amine modifications, and terminal PEG linkers were taken from the GAFF FF. 54 . Place all structures in a period box with a spacing of Then dissolve with TIP3 water 55 First add Mg 2+ ions to neutralize half of the charge, and then add Na + to neutralize the other half of the charge. Finally, add additional Na + and Cl - ions to a concentration of 150 mM.

[0261] Molecular dynamics simulations were run using LAMMPS13 GPU-compatible release (release 21 December 2016) on an NVIDIA K80 GPU. The structure was first minimized using steepest descent followed by 500 minimization steps using the conjugate gradient algorithm and then equilibrated by MD simulations at 310 K using the NVT ensemble over the course of 10 ps using a 1 fs time step. The resulting structure was then subjected to 10 ps of NPT MD at 310 K, 350 atm to relax the periodic box and ensure a positive net pressure. For the NVT simulations, a Nosé-Hoover thermostat was used with a time constant of 100 fs. For the NPT simulations, a Nosé-Hoover barostat was used with a time constant of 1 ps. The equilibrated structure was then subjected to 20 ns of MD at 310°K (NVT ensemble, 1 fs time step).

[0262] To obtain the structure shown in Figure 9, the structure with the lowest potential energy was extracted from the MD trajectory and conjugate gradient energy minimization was applied for 500 steps. 57 Visualization of constructs. Calculation of helical parameters from simulation trajectories using X3DNA 58 .

[0263] References

[0264] 1. Seeman, NCDNA in a material world. Nature 421, 427-431 (2003).

[0265] 2. Guo, P. The emerging field of RNA nanotechnology. Nat Nano 5, 833-842 (2010).

[0266] 3. Chen, Y.-J., Groves, B., Muscat, RA & Seelig, G. DNA nanotechnology from the test tube to the cell. Nat Nano 10, 748-760, doi: 10.1038 / nnano.2015.195 (2015).

[0267] 4.Benenson,Y.BiomoleGular computing systems:principles,progress andpotential.Nat Rev Genet 13,455-468,doi:http: / / www.nature.com / nrg / journal / v13 / n7 / suppinfo / nrg3197_S1.html(2012).

[0268] 5.Yurke,B.,Turberfield,A.J.,Mills,A.P.,Simmel,F.C.&Neumann,J.L.A DNA-fuelled moleGular machine made of DNA.Nature 406,605-608,doi:http: / / www.nature.com / nature / journal / v406 / n6796 / suppinfo / 406605a0_S1.html(2000).

[0269] 6.Srinivas,N.et al.On the biophysics and kinetics of toehold-mediatedDNA strand displacement.Nucleic Acids Research,doi:10.1093 / nar / gkt801(2013).

[0270] 7.Green,A.A.et al.Complex cellular 1ogic computation usingribocomputing devices.Nature 548,117-121,doi:10.1038 / nature23271 http: / / www.nature.com / nature / journal / v548 / n7665 / abs / nature23271.html# supplementaryinformation (2017).

[0271] 8.Groves,B.et al.Computing in mammalian cells with nucleic acidstrand exchange.Nat Nano11,287-294,doi:10.1038 / nnano.2015.278 http: / / www.nature.com / nnano / journal / v11 / n3 / abs / nnano.2015.278.html# supplementaryinformation (2016).

[0272] 9.Setten,R.L.,Rossi,J.J.&Han,S.P.The current state and futuredirections of RNAi based therapeutics Nature Reviews Drug Discovery,In press(2019).

[0273] 10.Bobbin,M.L.&Rossi,J.J.RNA Interference(RNAi)-Based Therapeutics:Delivering on the Promise?Annual Review of Pharmaco1ogy and Toxico1ogy56,103-122,doi:10.1146 / annurev-pharmtox-010715-103633(2016).

[0274] 11.Adams,D.et al.Patisiran,an RNAi Therapeutic,for HereditaryTransthyretin Amyloidosis.New England Journal of Medicine379,11-21,doi:10.1056 / NEJMoa1716153(2018).

[0275] 12.Benenson,Y.,Gil,B.,Ben-Dor,U.,Adar,R.&Shapiro,E.An autonomousmolecular computer forlogical control of gene expression.Nature 429,423-429,doi:http: / / www.nature.com / nature / journal / v429 / n6990 / suppinfo / nature02551_S1.ht ml(2004).

[0276] 13.Kumar,D.,Kim,S.H.&Yokobayashi,Y.Combinatorially Inducible RNAInterference Triggered by Chemically Modified Oligonucleotides.Journal of theAmerican Chemical Society 133,2783-2788,doi:10.1021 / ja1107436(2011).

[0277] 14.Han,S.P.,Barish,R.D.&Goddard,W.A.(Google Patents,2015).

[0278] 15.Han,S.-P.,Goddard III,W.A.,Scherer,L.&Rossi,J.J.Signal activatableconstructs and related components compositions methods and systems.USA patentUS9725715B2(2015).

[0279] 16.Hochrein,L.M.,Ge,T.J.,Schwarzkopf,M.&Pierce,N.A.SignalTransduction in Human Cell Lysate via Dynamic RNA Nanotechnology.ACSSynthetic Biology,doi:10.1021 / acssynbio.8b00424(2018).

[0280] 17.Hochrein,L.M.,Schwarzkopf,M.,Shahgholi,M.,Yin,P.&Pierce,N.A.Conditional Dicer Substrate Formation via Shape and Sequence Transductionwith Small Conditional RNAs.Journal of the American Chemical Society 135,17322-17330,doi:10.1021 / ja404676x(2013).

[0281] 18.Bindewald.E.et al.Multistrand Structure Prediction of Nucleic AcidAssemblies and Design of RNA Switches.Nano Letters16,1726-1735,doi:10.1021 / acs.nanolett.5b04651(2016).

[0282] 19.Chatterjee,G.,Chen,Y.-J.&Seelig,G.Nucleic Acid Strand Displacementwith Synthetic mRNA Inputs in Living Mammalian Cells.ACS Synthetic Biology,doi:10.1021 / acssynbio.8b00288(2018).

[0283] 20.Li,X.,Yang,X.,Qi,J.&Seeman,N.C.Antiparallel DNA Double CrossoverMolecules As Components for Nanoconstruction.Journal of the American ChemicalSociety 118,6131-6140,doi:10.1021 / ja960162o(1996).

[0284] 21.Ha,M.&Kim,V.N.Regulation of microRNA biogenesis.Nature ReviewsMolecular Cell Biology 15,509,doi:10.1038 / nrm3838(2014).

[0285] 22.MacRae,I.J.et al.Structural Basis for Double-Stranded RNAProcessing by Dicer.Science(NewYork,N.Y.)311,195-198,doi:10.1126 / science.1121638(2006).

[0286] 23.Kim,D.-H.et al.Synthetic dsRNA Dicer substrates enhance RNAipotency and efficacy.Nat Biotech 23,222-226,doi:http: / / www.nature.com / nbt / journal / v23 / n2 / suppinfo / nbt1051_S1.html(2005).

[0287] 24.Khvorova,A.&Watts,J.K.The chemical evolution of oligonucleotidetherapies of clinical utility.Nat Biotech 35,238-248,doi:10.1038 / nbt.3765(2017).

[0288] 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).

[0289] 26.ORBAN,T.I.&IZAURRALDE,E.Decay of mRNAs targeted by RISC requiresXRN1,the Ski complex,and the exosome.RNA(New York,N.Y.)11,459-469,doi:10.1261 / rna.7231505(2005).

[0290] 27.Hope,T.J.&Trono,D.Structure,Expression,and Regulation of the HIVGenome,<http: / / hivinsite.ucsf.edu / InSite?page=kb-00&doc=kb-02-01-02>(2000).

[0291] 28.Kundu,M.&Liu,P.P.Function of the inv(16)fusion gene CBFB-MYH11.Current Opinion in Hematology 8,201-205(2001).

[0292] 29.Look,A.T.Oncogenic Transcription Factors in the Human AcuteLeukemias.Science(New York,N.Y.)278,1059-1064,doi:10.1126 / science.278.5340.1059(1997).

[0293] 30.Plimpton,S.Fast Parallel Algorithms for Short-Range MolecularDynamics.Journal of Computational Physics 117,1-19,doi:http: / / dx.doi.org / 10.1006 / jcph.1995.1039(1995).

[0294] 31.Condon,D.E.et al.Optimization of an AMBER Force Field for theArtificial Nucleic Acid,LNA,and Benchmarking with NMR of L(CAAU).The Joumalof Physical Chemistry B 118,1216-1228,doi:10.1021 / jp408909t(2014).

[0295] 32.Lind,K.E.,Sherlin,L.D.,Mohan,V.,Griffey,R.H.&Ferguson,D.M.inMolecular Modeling of Nucleic Acids Vol.682ACS Symposium Series Ch.3,41-54(American Chemical Society,1997).

[0296] 33.Aduri,R.et al.AMBER Force Field Parameters for the NaturallyOccurring Modified Nucleosides in RNA.Journal of Chemical Theory andComputation 3,1464-1475,doi:10.1021 / ct600329w(2007).

[0297] 34.Iwamoto,N.et al.Control of phosphorothioate stereochemistrysubstantially increases the efficacy of antisense oligonucleotides.NatBiotech advance online publication,doi:10.1038 / nbt.3948 http: / / www.nature.com / nbt / journal / vaop / ncurrent / abs / nbt.3948.html#supplementaryinformation (2017).

[0298] 35.Glaser,S.P.et al.Anti-apoptotic Mcl-1is essential for thedevelopment and sustained growth of acute myeloid leukemia.Genes&Development26,120-125,doi:10.1101 / gad.182980.111(2012).

[0299] 36.Collingwood,M.A.et al.Chemical Modification Patterns Compatiblewith High Potency Dicer-Substrate Small Interfering RNAs.Oligonucleotides 18,187-200,doi:10.1089 / oli.2008.0123(2008).

[0300] 37.Yin,P.&Pierce,N.A.Triggered RNAi.USA patent US8318921B2(2008).

[0301] 38.Lee,H.Y.,Zhou,K.,Smith,A.M.,Noland,C.L.&Doudna,J.A.Differentialroles of human Dicer-binding proteins TRBP and PACT in small RNAprocessing.Nucleic Acids Research,doi:10.1093 / nar / gkt361(2013).

[0302] 39.Han,S.P.,Goddard,W.A.,SCHERER,L.&Rossi,J.J.(Google Patents,2015).

[0303] 40.Silverman,S.K.Control of macromolecular structure and functionusing covalently attached double-stranded DNA constraints.MolecularBioSystems 3,24-29,doi:10.1039 / B614116A(2007).

[0304] 41.Engelen,W.,Janssen,B.M.G.&Merkx,M.DNA-based control of proteinactivity.Chemical Communications 52,3598-3610,doi:10.1039 / C5CC09853J(2016).

[0305] 42.Mukherjee,P.,Leman,L.J.,Griffin,J.H.&Ghadiri,M.R.Design of a DNA-Programmed Plasminogen Activator.Journal of the American Chemical Society140,15516-15524,doi:10.1021 / jacs.8b10166(2018).

[0306] 43.Colasanti,A.V.,Lu,X.-J.&Olson,W.K.Analyzing and Building NucleicAcid Structures with 3DNA.e4401,doi:doi:10.3791 / 4401(2013).

[0307] 44.Zadeh,J.N.et al.NUPACK:Analysis and design of nuGleic acidsystems.Journal ofComputational Chemistry 32,170-173,doi:10.1002 / jcc.21596(2011).

[0308] 45.Mathews,D.H.,Sabina,J.,Zuker,M.&Turner,D.H.Expanded sequencedependence of thermodynamic parameters improves prediction of RNA secondarystructure 11Edited by I.Tinoco.Journal of Molecular Biology 288,911-940,doi:https: / / doi.org / 10.1006 / jmbi.1999.2700(1999).

[0309] 46.Camacho,C.et al.BLAST+:architecture and applications.BMCBioinformatics 10,421-421,doi:10.1186 / 1471-2105-10-421(2009).

[0310] 47.Tolstrup,N.et al.OligoDesign:optimal design of LNA(locked nucleicacid)oligonucleotide capture probes for gene expression profiling.NucleicAcids Research 31,3758-3762(2003).

[0311] 48.Scherer,L..J.,Frank,R.&Rossi,J.J.Optimization and characterizationof tRNA-shRNA expression constructs.Nucleic acids research 35,2620-2628,doi:10.1093 / nar / gkm103(2007).

[0312] 49.Macke,T.J.&Case,D.A.in Molecular Modeling of Nucleic AcidsVol.682ACS Symposium Series Ch.24,379-393(American Chemical Society,1997).

[0313] 50.Duan,Y.et al.A point-charge force field for molecular mechanicssimulations of proteins based on condensed-phase quantum mechanicalcalculations.Joumal of Computational Chemistry 24,1999-2012,doi:10.1002 / jcc.10349(2003).

[0314] 51.Aduri,R.et al.AMBER Force Field Parameters for the NaturallyOccurring Modified Nucleosides in RNA.Journal of Chemical Theory andComputation 3,1464-1475,doi:10.1021 / ct600329w(2007).

[0315] 52.Condon,D.E.et al.Optimization of an AMBER Force Field for theArtificial Nucleic Acid,LNA,and Bernchmarking with NMR of L(CAAU).The Journalof Physical Chemistry B 118,1216-1228,doi:10.1021 / jp408909t(2014).

[0316] 53.Lind,K.E.,Sherlin,L.D.,Mohan,V.,Griffey,R.H.&Ferguson,D.M.inMolecular Modeling of Nucleic Acids Vol.682ACS Symposium Series Ch.3,41-54(American Chemical Society,1997).

[0317] 54.Wang,J.,Wolf,R.M.,Caldwell,J.W.,Kollman,P.A.&Case,D.A.Developmentand testing of a general amber force field.Journal of Computational Chemistry25,1157-1174,doi:10.1002 / jcc.20035(2004).

[0318] 55.Mark,P.&NilsSon,L.Structure and Dynamics of the TIP3P,SPC,and SPC / E Water Models at298K.The Journal of Physical Chemistry A105,9954-9960,doi:10.1021 / jp003020w(2001).

[0319] 56.P1impton,S.Fast Parallel Algorithms for Short-Range Mo1ecularDynamics.Journal of Computational Physics117,1-19,doi:http: / / dx.doi.org / 10.1006 / jcph.1995.1039(1995).

[0320] 57.Pettersen,E.F.et al.UCSF Chimera-A visualization system forexploratory research and analysis.Journal of Computational Chemistry 25,1605-1612,doi:10.1002 / jcc.20084(2004).

[0321] 58.Colasanti,A.V.,Lu,X.-J.&Olson,W.K.Analyzing and Building NucleicAcid Structures with 3DNA.e4401,doi:doi:10.3791 / 4401(2013).

[0322] 59.Mathews et al.,Expanded Sequence Dependence of ThermodynamicParameters Improves Prediction of RNA Secondary Structure,J.Mol.Biol.(1999)288:911-940.

[0323] 60.U.S.Patent No.9,029,524.

[0324] 61.Meggers,et al.,Acc.Chem.Res.(2010)43(8):1092-1102.

[0325] 62.Schlegel et al.,J.Am.Chem.Soc.(2017)139(25):8537-8546.

Claims

1. A programmable conditionally activated small interfering RNA (Cond-siRNA) construct comprising a sensor strand, a core strand, and a guide strand, wherein: (i) the sensor strand and the core strand form a sensor duplex, wherein the terminal base at the 5' end of the core chain is 2'-F or 2'-O-methyl, and the terminal base at the 3' end of the core chain is 2'-F or 2'-O-methyl, (ii) the guide strand and the core strand form an RNAi duplex, and (iii) the sensor duplex is linked to the RNAi duplex to form a single structure, wherein (a) 50% or less of the backbone positions in the sensor chain duplex domain of the sensor chain are phosphorothioate (PS) linkages, or (b) The three bases of the core strand that are base paired with the 3' end of the guide strand have an M*+*M pattern, where M is a 2' modified base, * is a PS backbone connection, and + is an LNA base or other 2'-4' bridging base.

2. The Cond-siRNA construct according to claim 1, wherein The sensor chain has protruding ends that are not complementary to the core chain, and wherein the protruding ends on the sensor chain are capable of complementary binding to an input chain to form a toehold, thereby causing displacement of the sensor chain from the core chain.

3. The Cond-siRNA construct according to claim 1 or claim 2, comprising one or more of the following chemical modifications: A. The duplex domain of the sensor chain has one or more of the following features: a .50% or more of the bases are chemically modified to resist nuclease degradation or to increase the melting point (Tm) of the duplex; b. 100% of the bases are chemically modified to resist nuclease degradation and increase Tm; and c. 10%-50% of the bases are locked nucleic acids (LNA) or other chemically modified bases with 2'-4' bridge modifications that substantially increase Tm; B. The 5' and 3' ends of the core strand have one or more of the following characteristics: a. The three terminal bases on the 5' have the pattern MRM, where M is a base modified by 2'-O-methyl or 2'-F, and R is an RNA base; b. The three terminal bases on the 3' have the pattern MRM, where M is a base modified by 2'-O-methyl or 2'-F, and R is an RNA base; c. The three terminal bases at 3' and 5' have no continuous PS backbone modifications; d. The core strand portion that base pairs with the sensor strand has an alternating chemical modification pattern (MR) n; e. above, wherein M is a chemically modified base that does not reduce the duplex Tm compared to the equivalent RNA base; f. Any combination, wherein the 5' and 3' ends of the core chain have at least one feature of ae; and g. The 3' and 5' regions of the core strand that base pair with the sensor strand are: (a) consists entirely of the pattern (M)n, where M is a 2'-O-methyl or 2'-F modified base, or (b) at least 50% of the bases in the region are 2'-O-methyl or 2'F, and up to 30%, 50%, 80%, or 100% of the backbone linkages are not phosphorothioates; C. The core strand in which three bases are base paired with the 3' end of the guide strand has one or more of the following characteristics: aM*+*+ mode; b.+*+*+ mode; cR*+*M pattern, where R is an RNA base; dR*+*+ mode; e. *M*M pattern; and The pattern in fa-e, where * is a PS backbone linkage or an unmodified phosphodiester backbone linkage; and D. A wizard chain having one or more of the following characteristics: a. 30% to 95% of the bases are 2'-O-methyl or 2'-F modified bases, LNA, or 2'-4' bridged bases; b. wherein the two terminal bases at the 5' end are chemically modified; c. wherein the two terminal bases on the 5' side have at least one LNA; d. wherein the two terminal bases at the 5' end are connected by a PS backbone; e. wherein both terminal bases at the 3' end are chemically modified; f. 5% to 50% of the backbone linkages are PS; and g. The bases flanking the Dicer cleavage site are not chemically modified.

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

5. The Cond-siRNA construct according to claim 1 or 2, wherein The duplex domain of the sensor chain does not have phosphorothioate (PS) modifications.

6. The Cond-siRNA construct according to claim 1 or 2, wherein One or both ends of the core chain are modified with PS or 2'-O-methyl.

7. The Cond-siRNA construct according to claim 2, wherein The toehold domain comprises one or more chemical modifications selected from the group consisting of LNA modification, 2'-O-methyl modification, PS modification, and combinations thereof.

8. The Cond-siRNA construct according to claim 1 or 2, wherein The sensor duplex is 23 bp and the RNAi duplex is 23 bp.

9. A method for activating synthetic RNAi, comprising administering to a cell culture the Cond-siRNA construct of any one of claims 1 to 8, wherein: The input strand binds to the sensor strand, resulting in displacement of the sensor strand from the Cond-siRNA, thereby activating RNAi in the Cond-siRNA construct.

10. The method according to claim 9, wherein: The input strand is a cellular RNA transcript.

11. Use of the Cond-siRNA construct according to any one of claims 1 to 8 in the preparation of a drug for treating a disease or condition, wherein: The input strand binds to the sensor strand, resulting in displacement of the sensor strand from the Cond-siRNA, thereby activating RNAi in the Cond-siRNA construct, and wherein the RNAi targets the disease or the disorder.

12. The use according to claim 11, wherein The input strand is a cellular RNA transcript.

Images