Accessibility and secondary structure determination of a target nucleic acid molecule

Abstract

The disclosure concerns a method (400) for determining the accessibility of a locus of a target nucleic acid molecule (304), comprising the steps of a) immobilizing (410) target nucleic acid molecules (304) on a substrate (302), b1) contacting (412) the target nucleic acid molecules (304) with first adapters (306) under conditions allowing binding of the first adapters (306) to the locus of the target nucleic acid molecules (304), wherein each of the first adapters (306) comprises a first target binding site (308), which is capable of binding to the locus of the target nucleic acid molecules (304), and a first docking site (310), which is capable of binding to a corresponding first docking site (314) of a first imager (312), b2) contacting (414) the first adapters (306) with first imagers (312) under conditions allowing binding of the first adapters (306) to the first imagers (312), wherein each of the first imagers (312) comprises the corresponding first docking site (314) and a detectable label (316), c) performing time lapse imaging (416) on the immobilized target nucleic acid molecules (304) under conditions allowing detecting the detectable label (316), and d) performing image processing (418) including localizing a detected label to the first target binding site (308) to establish the accessibility of the locus of the target nucleic acid molecule (304). The disclosure further concerns a use of the method (400) for improving a mathematical model for predicting a secondary structure of a target nucleic acid molecule (304), for developing a drug and evaluating a drug candidate based on antisense oligonucleotides, for pre-tox screening of an antisense oligonucleotide (ASO) drug or an ASO drug candidate, and for optimization of an antisense oligonucleotide (ASO) drug candidate. The disclosure still further concerns a kit for determining a secondary structure of a target nucleic acid molecule (304).

Description

[0001] Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0002] 1 / 68

[0003] ACCESSIBILITY AND SECONDARY STRUCTURE DETERMINATION OF A TARGET NUCLEIC ACID MOLECULE

[0004] The present disclosure pertains to the field of structural and biophysical elucidation of biological molecules. In particular, the present disclosure discloses a method and kit for probing the structural dynamic and drugability of a target nucleic acid molecule. Beyond that, the method of present disclosure can be used to validate mathematical models to determine secondary and higher order structures of a target nucleic acid molecule, and the method can be used to find and optimize chemical matter which modulates this.

[0005] RNA, DNA, and nucleic acids (NAs) in general are subject of gene therapy, vaccination, as well as scaffolds to engineer antisense oligonucleotides (ASO). Because they regulate and orchestrate the structure and function of cells, NAs are a class of drug targets on its own. The interaction analysis of random chemical matter with a drug target is core competence in drug discovery as well as in the definition of diagnostic tool compounds. Whereas medicines which target proteins are most established, only a few therapies which target human RNAs could be developed to medicines. Risdiplam, a small molecule splice modifier to treat Spinal Muscle Athrophy, is the first U.S. Food and Drug Administration (FDA) approved small drug molecule which modulates the expression of a human protein on the level of RNA. Structured proteins populate a small conformational space, many offer a discrete / pre-formed binding pocket or interface which is structured selectively. A distinctly structured and rigid particle has physical properties which are correspondingly unique and constant in space and time. A high copy number of such particles enables the observation of even small spatial physical modulations with compounds, in any case this is required if one intends to monitor target engagements with macroscopic methods in binding assays (FRET assay, SPR, etc.). Structural knowledge of uniformly structured particles can be gathered with X-ray crystallography (X-ray), cryogenic electron microscopy (cryoEM) or nuclear magnetic resonance spectroscopy (NMR). However, all these methods, also require a coherently structured ensemble of particles to extract atomistic-structural information shared among many individual molecules. The physical property contrast of organic matter in liquids is too low, therefore it requires a physical amplification by averaging many similarly structured particles (a coherently ordered single crystal, structural reconstruction of imaging data). It is the combination of structural knowledge and physical interaction which guides an informed drug discovery approach; the transformation of hit classes into lead structures, into clinical candidates and to finally generate potent medicines. Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0006] 2 / 68

[0007] NAs are a space of macromolecules which behave structurally, physically and functionally differently. From a material science point of view they remind of liquid crystalline phases. To some extent, they are locally ordered, overall they are dynamic and even better structured regions are only held together by the multitude of weak-transient but highly cooperative interactions: base pairings. It makes these molecules non-uniformly shaped and structured in space and time. The macromolecular backbone of single stranded NAs is fluidic, only four bases make their chemical constitution likewise monotone, at least if compared to the pool of 21 amino acids of mostly chiral molecules. The likelihood to observe inter- and intramolecular hybridizations of bases scales with the target length, and a multitude of inter- or supramolecular hybridization likelihoods coexists. Consequently, a complex multitude of states and secondary structural motifs are competitively populated and lead to a poorly defined and dynamic structural landscape of NAs. NAs are not particles which are uniformly structured in space and time, accordingly, the physical properties between similar constituted NAs vary over time. It is a challenge to establish a classical biophysical binding assay as it is established for structured proteins. Because different regions or segments along the NA sequence have different likelihoods to expose a target site, the interaction constant is a local property which depends on the wider dynamical nature of the NA. As the dissociation constant (kD) of a drug target complex is the statistical likelihood to find a binder on the target site, the kD depends on the timely presence or accessibility of the targeted binding site (meaning, the concentration of the target site is only weakly related with the amount of NAs used in the assay). The “pocket” or binding site can thereby be a transient and induced structure, and differently structured binders are likely to target entirely different regions along different regions of the sequence. This structural dynamic nature of NAs makes it difficult to screen, compare and structurally optimize compounds interacting with NAs with methods established in biophysics. The target is a variable and its structure a function over time, the physical properties of the target vary, and macroscopic observations are an average of a multitude of many differently structured molecules and states.

[0008] The here disclosed invention overcomes these challenges. 1) The structural dynamic of the target is region-selectively probed in a reversible manner and over a larger time observation window. Different binding sites along the entire sequence are systematically probed for their statistical likelihood to be engaged by a binder. 2) Single molecule recognition events are counted to determine the probability of a molecule to be found on the target and enable to measure the biophysical interaction constant. Alternatively, the Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0009] 3 / 68 probability of a molecule to be found on the target can be assessed for example by mean fluorescence brightness or the rate of change of fluorescence signal. 3) As the binding site is known, structural knowledge is gathered. 4) Any modulation by the near physical environment or binder, which are expected to modulate the structural dynamic of the target, can be identified.

[0010] Different approaches to modulate the biological activity of RNAs, including mRNA, miRNA, and ncRNA by developing binders specifically attaching to those RNAs and thereby altering their structure and function have been developed. These binders can be small molecules, peptides, antisense oligonucleotides, proteins, or CRISPR guide RNAs (Disney et al, Nature Reviews Drug Discovery 2022).

[0011] The development of these binders is greatly facilitated by the knowledge of secondary structure for the RNA, and thus the accessibility of the various domains of the RNA. Furthermore, e.g., in the field of genomic medicine, DNA needs to be loaded into cells. Here, heavily folded ssDNAs are expected to clump and to resist the cellular loading machinery. Therefore, insight into the secondary structure can help optimize the loading efficiency.

[0012] In research as well as medical diagnostics, fluorescence in situ hybridization (FISH) has become a relevant tool. It relies on hybridizing fluorescently labeled oligonucleotides (FISH probes) to RNA (or in variants, DNA) present in fixed cells or tissue. Knowledge of the accessibility of different loci along the target RNA is helpful in designing optimal FISH probes for both permanent and transient binding.

[0013] The mechanical structure and functionality of RNA molecules is determined by its secondary and tertiary structure: domains can be flexibly fluid, can be under tension, or they are paired intramolecularly with other domains of the RNA. There are cooperative effects when these domains are formed. These cooperative effects make it difficult to predict the secondary structure computationally. Additionally, the strongly charged RNA backbone makes the structure sensitive to the presence and strength of counter ions and makes it difficult to compute. Especially long RNAs tend to form transient states and phases where liquid domains transiently pair with other domains of the same and other RNA molecules. The strength of the secondary structure is a measure of how likely it is that an RNA participates in liquid-liquid phase separations and regulatory functions: The Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0014] 4 / 68 intramolecular folding competes with intermolecular folding. These effects are difficult to compute and remain complex.

[0015] Many proteins are known to interact and / or be regulated by RNAs. RNAs are thought to spatially orchestrate the interaction of transient complexes. Taken together, it can be said that the rigidity of a given RNA is a parameter which is heavily influenced by factors which are complex and thus difficult to capture in a model. Nonetheless, many approaches exist that are helpful for the analysis of short RNAs, e. g., NLIPACK, UNAFold, or Mfold. Capturing the aspects mentioned above is a challenge for experimental assays as well.

[0016] Multiple experimental methods for DNA and RNA secondary structure and accessibility determination are known. Said methods include, among others, X-ray crystallography, nuclear magnetic resonance (NMR) spectroscopy, cryo-electron microscopy (cryo-EM), small-angle X-ray scattering (SAXS) (Zhang et al., Nature Methods 2022, doi: 10.1038 / s41592-022-01623-y) and gel electrophoresis-based probing methods (PARIS) (Lu et al., Cell 2016, doi: 10.1016 / j.cell.2016.04.028), as well as microarray- 1 ike probing methods (Fratczak et al, Biochemistry 2011 , doi: 10.1021 / bi200463p; Southern et al, Oligonucleotides as therapeutic agents, Ciba Foundation Symposium 209, 1997; PCT / GB02 / 01245). It is noteworthy that PCT / GB02 / 01245 describes ensemble visualization of immobilized RNAs, probed optionally with fluorescent oligonucleotides, in single frame images. In addition, sequencing-based methods are known, like Selective 2'- hydroxyl acylation analyzed by primer extension sequencing (SHAPE-seq) and many more. The latter are differentiated into enzymatic and chemical probing techniques, and techniques which are nucleotide specific or generalized (backbone). The methods have been reviewed in Stroebel et al., Nat. Rev. Gen 2018, doi: 10.1038 / s41576-018-0034-x. Among the chemical probing methods, PARIS and SHAPE crosslink or label single RNA nucleotides in situ and analyze RNA after extraction. Afterwards, the data reflect RNA structures populated in their natural context.

[0017] Other methods are performed in vitro. The methods are ensemble methods, and thus the results are aggregated data over many molecules; the resulting data are based on mean values and do not show distributions of the conformational landscape. Mechanical properties, binding accessibility or oligonucleotide hybridization accessibility are not probed except in the microarray- 1 ike methods, where RNA is captured via oligonucleotides and thus the mechanics of the captured region have an influence. Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0018] 5 / 68

[0019] For instance, a technique termed SiM-KARTS is known from Rinaldi et al., Nat. Comm. 2016, doi: 10.1038 / ncomms9976. This technique is used for investigating isolated aspects of RNA secondary structure by observing binding kinetics of a transiently binding, fluorophore labeled short RNA oligonucleotide.

[0020] The methods of the prior art have the drawback that they yield indirect and macroscopically averaged measurement values, which are often biochemically amplified, leading to significant noise. The combination of these problems leads to a situation where the secondary structure dynamics, energy landscape, and conformational space of nucleic acid molecules, particularly RNA, cannot be resolved. Another drawback is that the global secondary structure of the target molecule is not assessed. Other drawbacks include complex procedures, which also results in high costs.

[0021] The present disclosure seeks to overcome the drawbacks mentioned above. An objective of the present disclosure is to provide methods for determining increasing experimental accuracy while allowing high throughput, preferably at low cost. Another objective resides in the provision of a method that can be used for the provision of more reliable mathematical models affording, e.g., secondary structure prediction of nucleic acid molecules. Still another objective resides in the provision of a method allowing the assessment of a secondary structure of a target nucleic acid molecule under physiological conditions. Another objective of the present disclosure is to provide methods for determining relative accessibility of NAs on target molecules with high throughput and at low cost. Still other objectives reside in the provision of improved methods for drug discovery.

[0022] It is therefore desirable to provide a solution by means of which accuracy of secondary structure or accessibility determination may be increased and / or speed up, preferably, while enabling a high throughput, preferably at low cost. Another objective resides in the provision of a method that can be used for the provision of more reliable mathematical models affording, e.g., secondary structure prediction of nucleic acid molecules.

[0023] This is solved by the objects of the independent claims. Advantageous embodiments are indicated in the dependent claims and in the description.

[0024] In a first aspect, the invention concerns a method for determining the accessibility of a locus of a target nucleic acid molecule, comprising the steps of a) immobilizing target Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0025] 6 / 68 nucleic acid molecules on a substrate, b1) contacting the target nucleic acid molecules with first adapters under conditions allowing binding of the first adapters to the locus of the target nucleic acid molecules, wherein each of the first adapters comprises a first target binding site, which is capable of binding to the locus of the target nucleic acid molecules, and a first docking site, which is capable of binding to a corresponding first docking site of a first imager, b2) contacting the first adapters with first imagers under conditions allowing binding of the first adapters to the first imagers, wherein each of the first imagers comprises the corresponding first docking site and a detectable label, c) performing time lapse imaging on the immobilized target nucleic acid molecules under conditions allowing detecting the detectable label, and d) performing image processing including localizing a detected label to the first target binding site to establish the accessibility of the locus of the target nucleic acid molecule.

[0026] In a further aspect, the invention concerns a use of the method for improving a mathematical model for predicting a secondary structure of a target nucleic acid molecule, wherein step d) further comprises assessing an accessibility score for the first target binding site of the localized detected labels and comparing the accessibility score with the secondary structure of the target nucleic acid molecule.

[0027] In a further aspect, the invention concerns a use of the method for developing a drug and evaluating a drug candidate based on antisense oligonucleotides, wherein a high accessibility of a locus of a target molecule corresponds to a favorable target region of an antisense oligonucleotide, for example for downregulating the translation of a gene.

[0028] In a further aspect, the invention concerns a use of the method for pre-tox screening of an antisense oligonucleotide (ASO) drug or an ASO drug candidate, wherein off-target binding of an ASO of the ASO drug or the ASO drug candidate to an RNA molecule that is not a clinical target is indicative of a risk of adverse side effects of the ASO.

[0029] In a further aspect, the invention concerns a use of the method for optimization of an antisense oligonucleotide (ASO) drug candidate, wherein multiple variants of the ASO drug candidate, preferably wherein the ASO drug candidate comprises modified bases, are evaluated to determine the variant with the best interaction with the target molecule.

[0030] In a further aspect, the invention concerns a kit for determining a secondary structure of a target nucleic acid molecule, comprising a substrate, immobilizing reagents adapted to Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0031] 7 / 68 immobilize target nucleic acid molecules on the substrate, first adapters, first imagers, wherein the first adapters are adapted to bind to the target nucleic acid molecule and to the first imagers, the first adapters each comprising a first target binding site, which is capable of binding to the target nucleic acid molecules, and a first docking site, which is capable of binding to a corresponding first docking site of a first imager, each of the first imagers comprising the corresponding first docking site and a detectable label.

[0032] An accessibility of a binder, particularly single stranded nucleic acids (NAs), to a specific site or locus of target nucleic acid molecule is mainly determined by secondary structure. By the self-folding base pairing interactions of the target nucleic acid molecule, a binding site of a binder may become less accessible to the binder. Therefore, full accessibility means that the affinity of the binder to the target nucleic acid molecule is the same as for its isolated binding site, while reduced accessibility means that the affinity of the binder to the target nucleic acid molecule is reduced when (hypothetically) compared to the isolated binding site.

[0033] A target nucleic acid molecule is composed of nucleotides, which have a 5-carbon sugar, a phosphate group, and a nitrogenous base as the monomer components. The two main classes of nucleic acids are deoxyribonucleic acid (DNA) and ribonucleic acid (RNA). DNA is usually composed of the nucleotides cytosine [C], guanine [G], adenine [A] and thymine [T], RNA is usually composed of the nucleotides cytosine [C], guanine [G], adenine [A] and uracil [II]). However, DNA and RNA molecules may also comprise other nucleotides. Target nucleic acid molecules typically consist of a number of the same molecules.

[0034] A secondary structure of a target nucleic acid molecule considers the base pairing interactions within a single target nucleic acid in solution or between two target nucleic acids in solution. The secondary structures of DNAs and RNAs are different. DNA usually exists as fully base paired double helices, while biological RNA is normally single stranded and often forms complex and intricate base-pairing interactions due to its increased ability to form hydrogen bonds stemming from the additional hydroxyl group in the ribose sugar. However, single stranded DNA is also known to exist, e. g. in some viruses, such as phages. Alternatively, single strands of DNA can be obtained by denaturing a DNA double strand. The size of the target nucleic acid molecule is not particularly limited. For instance, the target nucleic acid molecule may have a length of up to 100 nucleotides, such as up to 250 nucleotides, up to 500 nucleotides, up to 750 nucleotides, up to 103nucleotides, up to 104nucleotides, or up to 5 x 104nucleotides. The solution may be selected such that the Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0035] 8 / 68 secondary structure of the target nucleic acid molecule is essentially preserved. The solution may comprise buffers of physiological pH and / or high salt contents to stabilize the secondary structure. Particularly optional is a solution comprising or consisting of cell lysate. Exemplary buffers comprise 2.5 mM to 15 mM Tris-HCI and 10 mM to 750 mM salt, such as NaCI, MgCh, Na2HPO4 / Na^PCL, at a pH ranging from 6.8 to 8.2, optionally containing 0.1 mM to 1.5 mM EDTA. Alternatively, or in addition, a PBS buffer supplemented with 10 mM to 750 mM salt, such as NaCI, MgCh, Na2HPC>4 / Na^PCL, at a physiological pH, such as a pH ranging from 6.6 to 7.4, may be used.

[0036] A substrate is any surface on which target nucleic acid molecules may be immobilized. In addition, the substrate may offer suitable detection modalities, e. g. coverslips made of glass transparent to fluorescent light to enable detection based on fluorescent signals. Exemplary substrates include coverslips, particularly microscopy coverslips, e. g. made of glass, and flow chambers. Immobilization may be performed in different ways and includes binding between different molecules, such as covalent binding, e. g. between biotin and streptavidin, and hybridization, e. g. between complementary nucleic acid strands, including complementary DNA-DNA, DNA-RNA, or RNA-RNA strands. The skilled person is aware of possibilities for immobilizing target nucleic acid molecules on a surface, e.g. using different residues, and chooses the type of immobilization and / or substrate according to the particular needs, based on, among other elements the detectable label and / or detection equipment.

[0037] Target nucleic acid molecules may be also immobilized via DNA origami structures. A DNA origami structure comprises a target nucleic acid molecule, which is bound with at least one staple strand, in particular at least one DNA staple strand to the substrate, for example via biotin-streptavidin interaction, as known in the art. The conformation of the DNA origami is fixed, and the target nucleic acid molecule may be exactly positioned on it, so that the target nucleic acid molecule has at least one region with essentially the same secondary structure. The staple strand can have length of about 10 to about 500 nucleic acids, such as about 15 to about 450 nucleic acids, about 20 to about 400 nucleic acids, about 25 to about 350 more nucleic acids, about 30 to about 300 nucleic acids, about 50 to about 250 nucleic acids, or about 100 to about 200 nucleic acids.

[0038] Conditions allowing binding, particularly hybridization, and unbinding, particularly dehybridization, of two nucleic acid strands are well known in the state of the art. Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0039] 9 / 68

[0040] The adapters may be DNA or RNA molecules that can optionally be modified by different base and backbone modifications such as 2’-O-MOE or locked nucleic acids. With a length of, e. g. 5-35 nucleotides (nt). Adapters bind to a specific portion, the target binding site, of the target nucleic acid molecule as well as to a further specific portion, the docking site of the imager. Said specific portions may be directly connected or via a spacer. DNA adapters are optional since they may mitigate the hard-to-control imager binding kinetics inherent to RNA-RNA or RNA-DNA hybridizations. Adapters can be designed to transiently bind the target molecule or permanently bind the target molecule. They can also be designed to transiently or permanently bind the imager. Permanent binding to the target molecule and transient binding to the imager is optional. The adapters can serve as reporters for secondary structure determination, and (simultaneously) as prototypes for binders.

[0041] The target nucleic acid molecule may comprise one or more easily accessible regions, including loops or poly-A tails and one or more badly accessible regions. In general, a high number, such as at least 125 percent or more compared to an average signal, such as 130 percent or more, 135 percent or more, 140 percent or more, or 150 percent or more compared to the average signal, and optionally a long duration, of the generated signal, preferably a fluorescence signal, indicates that the target nucleic acid molecule being probed is easily accessible for binding, preferably hybridization. The high number of the generated signal and / or the average signal, preferably of all signals provided by the method, can include the arithmetic mean of the values. Accordingly, the high number of the generated signal can include the arithmetic mean of the high number that is at least 125 percent or more of the average signal. Preferably, the average signal is the arithmetic mean of all signals provided by the assay. Conversely, a lower number, such as at least 75 percent or less compared to an average signal, and optionally a short duration, of the generated signal, preferably a fluorescence signal, indicates that the target nucleic acid molecule being probed is badly accessible for binding, preferably hybridization.

[0042] Preferably, a short duration of the signal, preferably of the fluorescence signal, is 75 percent or less, such as 50 percent or less, compared to the average signal duration of all signals provided by the method. Preferably, a long duration of the signal, preferably of the fluorescence signal, is 125 percent or more, 130 percent or more, 135 percent or more, 140 percent or more, or 150 percent or more, compared to the average signal duration of all signals provided by the method. Preferably the average signal duration is the arithmetic mean of all signals provided by the method. The expression “all signals provided by the Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0043] 10 / 68 method” can include all measurements that have been performed for determining the accessibility of a locus of a target nucleic acid molecule.

[0044] The imagers have a corresponding docking site, which binds, preferably hybridizes, to the docking site of the adapter. The corresponding docking site may be a nucleic acid strand, such as a DNA or RNA strand, preferably with a length of 5-50 nt, such as 8-40 nt, 10-30 nt, 12-20 nt, or 14-18 nt. The imager further comprises a detectable label that is bound directly to the corresponding docking site or via a spacer.

[0045] Suitable detectable labels are known in the art and include, among others, a fluorescent label and radioactive label. The type of detectable label may be selected according to the specific requirements. Fluorescent labels are easy to handle but often bulky, which can interfere with or prevent the binding of complementary nucleic acids. This may be at least partially avoided by employing a spacer between a nucleic acid designated for binding, particularly hybridization, and the detectable label. Radioactive labels, particularly18F, are usually less bulky than other detectable labels but are difficult to handle. Fluorescent labels are particularly preferred.

[0046] Spacer molecules or spacers may be used to provide a spatial distance between the target binding site and the docking site of the adapters and / or between the corresponding docking site and the detectable label of the imager. Suitable spacer molecules are well known in art.

[0047] Time lapse imaging concerns a detection method allowing the time-resolved image acquisition of detectable labels, e. g. fluorescent labels. Image acquisition may include the use of microscopy, such as time-resolved microscopy image acquisition, wherein, e. g. on-, and optionally off-, times of detectable labels are determined and usually localized to a specific position on the substrate. An exemplary and optional time lapse imaging method includes fluorescence microscopy or fluorescence super-resolution microscopy. Super-resolution microscopy relates to a class of microscopy methods which circumvent the diffraction limit of light and achieve spatial resolutions below 250 nm. State-of-the-art methods may achieve a resolution close to molecular resolution (< 5 nm). There are several implementations of super-resolution microscopy of which the most prominent ones are STED microscopy, structured illumination and single-molecule localization microscopy methods, including dSTORM, PALM or DNA-PAINT. Super-resolution microscopy is outlined in detail, among others, in Hell, S. W. et al. The 2015 super-resolution Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0048] 11 / 68 microscopy roadmap. J Phys D Appl Phys 48, doi:https: / / doi.org / 10.1088 / 0022- 3727 / 48 / 44 / 443001 (2015); and Lelek, M. et al. Single-molecule localization microscopy. Nat Rev Methods Primers 1 , doi:10.1038 / s43586-021-00038-x (2021).

[0049] The step of image processing including localizing a detected label to the first target binding site to establish the secondary structure of the target nucleic acid molecule may comprise simultaneous, preferably high throughput, analysis of a plurality of, preferably the same, target nucleic acid molecules that are all immobilized on the substrate. In this regard, it is optional that a plurality of target nucleic acid molecules is immobilized to the substrate, such as 102or more, preferably 103or more, more preferably 104or more, such as 5 x 104to 106target nucleic acid molecules. The presence of a detected label, i. e. a generated signal, preferably fluorescence signal, indicates that a binding event takes place, i.e. the target nucleic acid molecule being probed binds a known adapter sequence, which in turn binds to an imager. In some embodiments, the number of binding events over the substrate, and optionally also the duration of the generated signal, preferably fluorescence signal, provides a measure of the accessibility or secondary structure. Particularly, a high number, and optionally a long duration, of the generated signal, preferably a fluorescence signal, indicates that the target nucleic acid molecule being probed is easily accessible for binding, preferably hybridization. This is particularly the case for a single stranded nucleic acid region, prevailing, e. g., for RNA loop regions of sufficient length of, e. g. 5 or more nucleotides, such as 6 nt or more, 7 nt or more, 8 nt or more, 9 nt or more, 10 nt or more nucleotides. The use of second adapters with a modified binding region in comparison to the first adapter can be used for refining the information about the binding region. For instance, the second adapters may have a different length than the first adapters and / or may be shifted, e. g., for one nucleic acid, in comparison to the first adapters. Preferably, image processing comprises the typical DNA- Paint workflow (Schnitzbauer et al., Nat Protoc 2017, doi: 10.1038 / nprot.2017.024) that yields on- and off-times for the docking sites on all target molecules identified, as well as their average values and distributions across the dataset, and the number of binding events per molecule. Optionally, for data-analysis more sophisticated methods such as the Hidden Markov Model (HMM) can be applied, known from, e. g., DOI: 10.1038 / ncomms9976, doi10.1038 / nbt.3246. Preferably, the overall binding signal, for example fluorescence intensity, provides a measure for accessibility or secondary structure. Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0050] 12 / 68

[0051] Colocalization may refer to cooccurrence, preferably spatial overlap, of signals of two detected labels, and optionally also correlation, in which signals of two detected labels not only overlap with one another but codistribute in proportion to one another. Colocalization can also refer to the overlay / superimposition of signals, e.g. fluorescent signals.

[0052] The step of image processing including localizing a detected label to the first target binding site to establish the secondary structure of the target nucleic acid molecule may further comprise, among others, a target molecule scan (designating an experiment comprising a plurality of binding site probing rounds, which may afford the comparison of the plurality of binding sites of one target molecule) aggregating all measurements from various binding site probing (i.e. target molecule locus accessibility) rounds, to generate a reference signal that can replace the separate reference docking site measurement mentioned above. The target molecule is localized and registered to exist if at least one specific signal from a binding site probing round was recorded, wherein a binding site probing round refers to one measurement round of acquiring a time-lapse dataset for one binding site. Alternatively, a colocalization score may be calculated. For instance, a colocalization score may be determined for each target molecule by dividing the total on- time by the total expected on-time. The expected total on-time can be elucidated by a calibration measurement (where short nucleotide sequences, e. g. 5 to 12 nt, such as 6 to 11 nt, 7 to 10 nt, or 8 to 9 nt, of the same sequence without secondary structure are probed). Alternatively, the colocalization score may be determined by multiplying ideal on- times by the measurement time and corrected for the imager concentration. A mean accessibility score averaged over the target molecules registered, as well as the distribution can be calculated. Still alternatively, the threshold method may be used to determine the colocalization score. For each target molecule, the localizations of imager strands to the binding site of the adapter strands are counted. Target molecules with counts above a defined threshold are counted as colocalized. Still alternatively, the nearest neighbor method may be used for determining the colocalization score. Localizations belonging to single molecules will be clustered for both a reference and target channel. The center-of-mass is then calculated for each detected cluster in both channels. The distance from each reference to its nearest target is measured (nearest- neighbor distance). The distances are plotted in a histogram resulting in two distinct distributions. One population is a characteristic short distance when references and target signal colocalize. The second population results from random distribution that depends on the overall density of molecules on the surface. Finally, the ratio of reference molecules that colocalize with a target molecule and the total number of reference molecules results Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0053] 13 / 68 in the colocalization score. Based on the colocalization scores of all secondary strands, an average accessibility for each nucleotide within the target molecule may be calculated.

[0054] The kit may comprise one or more of buffers, preferably physiological buffers, allowing hybridization and / or dehybridization of the adapters and / or imagers, and / or a manual. The manual may provide complete instructions, e. g. exemplified for the K4 Aptamer HHRi RNA.

[0055] The disclosure offers the advantage that imager binding kinetics are more homogeneous and controllable due to indirect binding of the imagers, via adapters, to the target nucleic acid. Furthermore, particularly using DNA adapters has the advantage, that complexity and costs for performing the method are significantly reduced compared to using direct binding of the imager to the target nucleic acid without adapter. Particularly, the use of DNA adapters requires only one dye-modified oligonucleotide as imager, and a set of unmodified oligonucleotides acting as adapters. Specifically, dye-modified oligonucleotides are more complicated in preparation and thus more expensive, e. g. .approx, tenfold more expensive, than unmodified oligonucleotides. Using adapters, particularly DNA adapters, thus allows an easy and economic “brute-force” approach since whole sets of secondary strands with different lengths of the binding site can be produced and measured against the target molecule.

[0056] A further advantage is that the present method may be performed under physiological conditions. Thereby, the secondary structure determined may be most like the secondary structure found under natural conditions. Still a further advantage is the possibility to reuse the substrate with the target nucleic acid molecules immobilized in it, e. g. for different adapters and / or imagers, preferably different adapters, and unvarying imagers. This may be performed by washing the substrate with a buffer, optionally at an elevated temperature, affording dehybridization of the adapters and / or imagers.

[0057] The disclosure also offers the advantage that binding of adapters with altered sequence, preferably due to shifting the binding region for one or more nucleotides, and / or due to length modification, can be repeated for multiple regions, to scan over the whole target nucleic acid molecule. As it is usually a single molecule technique, wherein a plurality of same target nucleic acid molecules is immobilized in a single substrate, the heterogeneity in conformations can be evaluated - both over time and over molecules. The length of the region probed can be varied, adding richness to the data acquired and changing the Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0058] 14 / 68 perturbation of the cooperative secondary structure of the RNA - and both enthalpic and entropic binding aspects may be included. The perturbation is, like that of the binder, to be developed, thus being a part of the investigation, rather than a nuisance. An RNA molecule may be restructured upon binding due to cooperative effects, which may be considered in the results.

[0059] Optionally, the method is for further determining a secondary structure of the target nucleic acid molecule, wherein the locus is a first locus, the method further comprising repeating steps b1) to d) with second adapters and second imagers, wherein each of the second adapters comprises a second target binding site, which is capable of binding to a second locus of the target nucleic acid molecules, and a second docking site, which is capable of binding to a corresponding second docking site of a second imager, wherein each of the second imagers comprises the corresponding second docking site and a detectable label, wherein the second target binding site overlaps the first target binding site or is different therefrom, wherein optionally the second docking site and the second imager are identical to the first docking site and the first imager, wherein step d) includes localizing a detected label to the second target binding site. It is optional that before repeating steps b1) to d) with second adapters and second imagers, the substrate, having target nucleic acid molecules immobilized thereon, is washed one or more times with a buffer, preferably a physiological buffer, optionally at an elevated temperature affording dehybridization of the adapters and / or imagers. The buffer, preferably a physiological buffer, may be the same as used for hybridization. Alternatively, other method steps subsequent to step d), such as any of steps e), f), and g) may be repeated as well. In comparison to the first target binding site, the second target binding site has a different length and / or is shifted by at least one nucleotide. The second imager preferably corresponds to the first imager, thus facilitating and reducing costs of the method. This offers the advantage that the secondary structure of the target nucleic acid molecule may be better resolved. However, the second imager may have a different detectable label, preferably another fluorescence label, which is different from the first imager. In addition or alternatively, the first docking site and corresponding first docking site may be different from the second docking site and corresponding second docking site. Optionally, step d) further includes colocalizing the localized detected labels. This allows averaging, by any mathematical method including, but not limited to, determining the arithmetic mean, median, or geometric mean, of the signals provided by the detectable labels. This adds to accuracy, in particular in case of employing a high-throughput approach, wherein a plurality of target nucleic acid molecules is immobilized in the substrate. Optionally, before Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0060] 15 / 68 repeating steps b1) to d) the substrate is washed with a buffer, optionally at an elevated temperature. This may allow dehybridization of previously hybridized adapters and / or imagers.

[0061] Optionally, repeating steps b1) to d) is performed in a different reaction chamber and a different imaging chamber. Hence, repeating steps b1) to d) is performed as separate experiments in separate reaction and imaging chambers. For instance, repeating steps b1) to d) is performed on different wells on a micro titer plate, preferably a SBS (Society for Biomolecular Screening) well plate. Alternatively, repeating steps b1) to d) is performed in the same reaction chamber and the same imaging chamber, optionally by introducing an intermediate washing step for removing the first adapters. The optional washing step may take place after step b2) and e.g. before step c).

[0062] Optionally, step d) further includes co-localizing the localized detected labels

[0063] Optionally, in step a) the target nucleic acid molecules are immobilized to the substrate via a biotin residue or via hybridization to an oligonucleotide pre-deposited on the substrate, optionally wherein the oligonucleotide pre-deposited on the substrate allows immobilizing to an easily accessible region of the target nucleic acid molecule, such as a loop or a poly- A tail. Biotin-streptavidin (neutravidin) linkage can be used for synthesized biotinylated DNA or RNA, or via a biotinylated poly-T oligonucleotide hybridizing to the poly-A tail of an mRNA, or via hybridization of a biotinylated oligonucleotide with a sequence specific to the DNA or RNA probed. Other types of “click-chemistry”, designating a fast and reliable immobilization often performed as a single-pot reaction, may be used instead of biotinstreptavidin. The Immobilization may be to a biotin-BSA surface, to a biotin-PEG surface, or it may be via a DNA origami structure to the surface. The RNA may also be immobilized in a flow chamber or cell (by cell fixation) which leads to the measurement being performed in situ.

[0064] Optionally, the target nucleic acid molecules are provided from an in vitro transcription reaction or from cell lysate.

[0065] Optionally, the target nucleic acid molecules are immobilized to the substrate in situ. In situ substrate immobilization preferably means that the target nucleic acid molecule is anchored to a cell or a tissue, which in turn is anchored to the substrate. Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0066] 16 / 68

[0067] Optionally, the substrate is selected from the group consisting of a flow chamber, a coverslip, a well of a multi-well plate, preferably a well of a SBS multi-well plate, and a glass slide. These substrates are readily available for a multitude of different detection modalities, including, e. g., substrates made from different materials.

[0068] Optionally, a density of the target nucleic acid molecules on the substrate is 1-105molecules I pm2, preferably 1-10 molecules I pm2.

[0069] Optionally, the density of the target nucleic acid molecules on the substrate is 1-10 molecules I pm2, wherein step c) comprises performing a DNA-PAINT like experiment and step d) comprises DNA-PAINT analysis. DNA-points accumulation in nanoscale tomography (DNA-PAINT) is a fluorescence (super-resolution) microscopy technique in which short fluorescently labeled oligonucleotides (so-called imagers) transiently bind a target and thereby produce a short-lived diffraction-limited spot in a time-resolved microscopy image acquisition. Finding centers and aggregating over time produces results in a (super-resolution) image (Jungmann et al, Nano letters 2010, doi: 10.1021 / nl103427w). In addition to (super-resolution) image generation, the blinking kinetics can be used to extract further information (Jungmann et al., Nat Methods 2016, doi: 10.1038 / nmeth.3804). The method can be used to target RNA as well, where multiple adapters are hybridized to RNA in situ and targeted by imagers (RNA-PAINT). RNA- PAINT is a single-molecule RNA-domain probing technique based on DNA-PAINT. RNA molecules of interest or RNA target molecules are immobilized on a substrate, e. g. a microscopy coverslip, and the accessibility to a region of the RNA target molecules is probed analog to DNA-PAINT. The kinetic parameters deliver insights into the accessibility and secondary structure of the region: Minimal binding occurs when the region probed is present in a paired conformation. If it is single stranded but under tension there is more binding, and if it is flexibly fluid, binding events will be most prominent. These different grades of binding may be reflected through the number of detected labels and also the duration of appearance (so-called “on-time” of detected label) of the detected label. High numbers and high on-times show a single stranded, flexibly fluid RNA region. Binding (of the imager, direct or indirect via an adapter) to a specific region of the target molecule can be repeated for multiple regions, e.g. with the aim of scanning the whole RNA molecule. As it is a single molecule technique, the heterogeneity in conformations can be evaluated - both over time and over molecules. The length of the region probed can be varied, adding data acquired and changing the perturbation of the cooperative secondary structure of the RNA - and, preferably, both enthalpic and entropic binding Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0070] 17 / 68 aspects are included. The perturbation is like that of the binder to be developed, thus being a relevant part of the investigation rather than a nuisance. An RNA molecule may be restructured upon binding due to cooperative effects, which may be considered in the results.

[0071] Optionally, the density of the target nucleic acid molecules on the substrate is 10-105molecules I pm2, wherein step c) comprises imaging during equilibration of the binding of the first adapters to the locus of the target molecule, and step d) comprises analyzing the time-lapse signal of the time lapse imaging to retrieve an on-rate of the first adapters as a measure for the accessibility.

[0072] Optionally, step d) further comprises assessing an accessibility score for the first target binding site of the localized detected labels. The accessibility score may be a numeric value representing accessibility of a specific nucleotide (of the target nucleic acid molecule) to hybridization, or in other words, the energy to be used to enable hybridization. For instance, the accessibility score may be expressed by values ranging from 0.0 to 1.0, with, e. g., 0.0 indicating that the specific nucleotide does not hybridize at all (e.g. infinite energy is required) and 1.0 indicating that the specific nucleotide offer best possible hybridization conditions (e.g. zero energy is required). Alternatively, the accessibility score may be the raw average fluorescence intensity, or the mean off time. Based on the accessibility score the secondary structure of the target molecule can be determined. In this way, an accessibility score for each nucleotide of a region of the target nucleic acid molecule or the entire target nucleic acid molecule may be provided. The accessibility score may be determined experimentally by the present method and / or calculated, preferably using a software like NLIPACK (e.g. the NLIPACK Web App, California Institute of Technology, USA). The accessibility score may thus provide a comprehensible measure for characterizing the target nucleic acid molecule’s secondary structure. In addition, an experimentally determined accessibility score can be used to refine a corresponding calculated accessibility score, or vice versa.

[0073] Optionally, step d) further comprises comparing the assessed accessibility score for the first target binding site with a computed accessibility score. As mentioned above the computed accessibility score may be provided by NUPACK software.

[0074] Optionally, the detectable label is a fluorescent detectable label, and time lapse imaging is microcopy time lapse imaging. The suitable fluorescent detectable label is not particularly Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0075] 18 / 68 limited and may comprise, among others, one or more from Cy2, Cy3, Cy3B, Cy3.5, Cy5, Cy5.5, and Cy7. Preferably, Cy3B is used as the fluorescent detectable label. The use of a fluorescent detectable label and microscopy time lapse imaging facilitates handling and detection of the detectable label.

[0076] Optionally, the method further comprises, before step c), contacting the target nucleic acid molecules with reference imagers under conditions allowing binding of the reference imagers to the target nucleic acid molecules at a reference binding site, optionally wherein the reference binding site is an easily accessible region of the target nucleic acid molecules, such as a loop, or a poly-A tail of the target nucleic acid molecules, or a reference docking site being introduced into each of the target nucleic acid molecules, or a reference docking site comprised in an immobilizing agent of step a). The reference binding site can either be an additional part of the sequence added specifically as reference with an optimized sequence or alternatively as described here the sequence of the target molecule itself can be used. The reference binding site may be an artificial DNA or RNA sequence introduced into the target nucleic acid molecule, preferably at the 5’ end or 3’ end thereof, such as a poly A or poly T tail. The reference imager may be used to provide an unambiguous detection event, which may indicate permanent binding (of the reference imager) to the target nucleic acid molecule. Alternatively, the reference imager may transiently bind to the target nucleic acid molecule. Thereby, an internal standard may be provided. The use of a reference docking site, for, e. g., colocalizing a probe and reference signal, leads to deeper information on probed regions with low accessibility. For instance, a reference imager with a reference docking site on a region on the target nucleic acid molecule known to be well accessible can be imaged in a separate imaging round or simultaneously to the probing, e. g., the first or a second, imager, with a different color. Furthermore, reference imagers targeting an easily accessible region of the target molecule, or an adapter ensured to be present on the target molecule, are used to establish a ground truth of the position of target molecules. This way, e.g. the range of spatial inaccessibility and non-presence of an adapter can be determined.

[0077] Optionally, the target nucleic acid molecules, the first and second target binding sites, the first and second docking sites, and the corresponding first and second docking sites are independently selected from DNA, RNA, and nucleic acid analogues, such as a locked nucleic acid (LNA). Thereby, the properties and binding properties of the adapters and / or imagers can be readily adapted. Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0078] 19 / 68

[0079] Optionally, the target nucleic acid molecule is an RNA molecule and wherein the first and second target binding sites, the first and second docking sites, and the corresponding first and second docking sites are selected from DNA and LNA. This facilitates binding of the target nucleic acid molecule to the adapter, and binding of the adapter to the imager. Detection of the detectable label is more reliable. In addition, adapters and / or imagers can be easily prepared, making their preparation inexpensive.

[0080] Optionally, each of the first and second imagers has a length of 3 to 52 nucleotides, preferably 5-20 nucleotides and a length of the first docking site is 5-50 nucleotides. Preferably, each of the first and second imagers has a length of 6 to 11 nucleotides, such as 7 to 10 nucleotides or 8 to 9 nucleotides. Such relatively short imagers offer good detection performance and are inexpensive to produce. Preferably, a length of the first docking site 10-30 nucleotides or 20-25 nucleotides. This offers reliable results.

[0081] Optionally, the method further includes, preferably after step a) and before step d), adding a compound suspected to affect the secondary structure of the target nucleic acid molecule, step e) image processing including localizing a detected label to the first target binding site to probe an altered secondary structure of the target nucleic acid molecule, and g) comparing the secondary structure and the altered secondary structure. The compound may be added directly before step b1), b2), or c). Alternatively, the compound may be added during step b1), b2), or c). Thereby, the influence of the compound on the secondary structure and thus a potentially relevant activity, e. g. a modulating or inhibiting activity, of the compound may be easily assessed. This provides a new approach in the drug discovery process. In addition, the method allows high throughput screening, i. e. screening of a large number of compounds suspected to affect, such as to alter, e.g. by altering, inhibiting, or stabilizing, the secondary structure of the target nucleic acid molecule in a relatively short time and with fewer experimental resources.

[0082] Optionally, the compound suspected to affect, e.g. by altering, inhibiting, or stabilizing, the secondary structure of the target nucleic acid molecule is selected from an oligonucleotide, a drug, a ligand, a small molecule, and a protein. A small molecule or macromolecule preferably has low molecular weight of < 1000 g / mol, such as 400 to 900 g / mol, 500 to 800 g / mol, or 600 to 700 g / mol. For instance, drug discovery processes require novel chemical entities, which specifically interact with the RNA at the best possible affinity. Hits identified by high throughput (HT) screenings recorded in different series. Each is dissected and iteratively refined by improving atomistic recognition Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0083] 20 / 68 elements in the molecules in medicinal chemistry. This process requires localizing regions along the RNA that are essential for binding, and that contribute cooperatively. Ultimately, a molecule needs to be established that selectively binds to a certain RNA motif; the physical interaction strength is a measure which requires to be tuned to achieve affinity and selectivity. Due to the transient and complex nature of RNAs, in-vitro and in-vivo, the analytical culture which can guide such a process, is poor and / or lacking. Already known methods as they are established for structured target proteins tend to fail or deliver inconclusive results.

[0084] Optionally, steps b1) and b2) are performed in a single step, preferably, wherein the first adapters and first imagers (312) are contained in the same solution, and / or each of the adapters has a length of 5 to 55 nucleotides, preferably 5-35 nucleotides, preferably wherein the adapters are DNA adapters. This can simplify the implementation of the present method and provides reliable results.

[0085] Optionally, each of the adapters, has a length of 5 to 35 nucleotides, preferably the adapters are DNA adapters. The adapter may also have a length of 6 to 34 nt, such as 7 to 33 nt, 8 to 32 nt, 9 to 31 nt, 10 to 30 nt, 11 to 29 nt, 12 to 28 nt, 13 to 27 nt, 14 to 26 nt, 15 to 25 nt, 16 to 24 nt, 17 to 23 nt, 18 to 22 nt, 19 to 21 nt, or 20 nt. The length of the target binding site may be 5 to 30 nt, such as 6 to 29 nt, 7 to 28 nt, 8 to 27 nt, 9 to 26 nt, 10 to 25 nt, 11 to 24 nt, 12 to 23 nt, 13 to 22 nt, 14 to 21 nt, 15 to 20 nt, 16 to 19 nt, or 17 to 18 nt. Thereby, transient or stable binding of the adapters to the target nucleic acid molecules can be adjusted smoothly.

[0086] Optionally, the first adapters are adapted to permanently bind the target nucleic acid molecule and to transiently bind the first imagers, or wherein the first adapters are adapted to permanently bind the target nucleic acid molecule and to permanently bind the first imagers. Transient binding and permanent binding are preferably understood as relative terms and it is preferred that permanent binding is at least 10 times longer, such as 20 times to 1000 times longer, than transient binding.

[0087] Optionally, the kit further comprises the target nucleic acid molecules, a set of first adapters adapted to bind to the target nucleic acid molecules, and a set of first imagers adapted to bind to the target nucleic acid molecules. The target nucleic acid molecules may be supplied already immobilized in a substrate. It is preferred that the target nucleic acid molecules comprise control target nucleic acid molecules, such as K4 Aptamer HHRi Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0088] 21 / 68

[0089] RNA or FMN-Riboswitch RNA. A set of first adapters may be adapted to bind to at least one of loops 1 to 4 of the K4 Aptamer HHRi RNA or a region adjacent thereto. For instance, the set of first adapters may comprise the nucleic acids of any of tables 2 to 4 hereinafter. Thereby, a standard is provided. The K4 Aptamer HHRi is also known as FMN-Riboswitch and can be present either in its natural form as K4 Aptamer HHRi RNA or in a corresponding DNA version.

[0090] In yet another aspect, the disclosure concerns a method for determining a secondary structure of a target nucleic acid molecule, comprising the steps of a) immobilizing target nucleic acid molecules on a substrate, b) contacting the target nucleic acid molecules with first imagers under conditions allowing binding of the first imagers to the target nucleic acid molecules, wherein each of the first imagers comprises a first target binding site, which is capable of binding to the target nucleic acid molecules, and a detectable label, c) performing time lapse imaging on the immobilized target nucleic acid molecules under conditions allowing detecting the detectable label, and d) performing image processing including localizing a detected label to the first target binding site to establish the secondary structure of the target nucleic acid molecule, e) optionally repeating steps b) to d) with second imagers each having a second binding site, which is capable of binding to the target nucleic acid molecules, wherein the second target binding site overlaps with the first target binding site or is different from it, wherein step f) includes localizing a detected label to the second target binding site and, optionally colocalizing the localized detected labels. Direct binding of the imager to the target nucleic molecule has the advantage that the secondary structure probing has a better resolution along the target molecule. Alternatively or in addition, the two methods for determining a secondary structure of a target nucleic acid molecule may be combined. For instance, at least one adapter and imager pair, i. e. a first adapter and a first imager, are used, wherein further rounds (for the same or a different target nucleic acid region to be probed) are performed with imagers alone, i. e. with imagers already including the target binding site. This allows the procedure to be conducted more accurately.

[0091] Optionally, the target nucleic acids, adapters and imagers are in solution. The solution may be designed in a way that essentially preserves the secondary structure of the target nucleic acid molecule. The solution may comprise buffers of physiological pH and / or high salt contents for stabilizing the secondary structure. Particularly optional is a solution comprising or consisting of cell lysate. The solution may also have a physiological temperature of, e. g., about 37°C, such as from 36°C to 40°C, 36,5°C to 39°C, or 37°C to Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0092] 22 / 68

[0093] 38°C. Preferably the entire assay is conducted under physiological conditions, particularly a physiological temperature. More preferably, parts of the method are performed in vivo.

[0094] Optionally, the substrate having target nucleic acid molecules immobilized thereon in step a) is reused several times, which allows the subsequent method steps to be repeated. This may be conducted by washing the substrate having target nucleic acid molecules immobilized thereon with a buffer, preferably a physiological buffer, optionally at an elevated temperature, which allows dehybridization of the adapters and / or imagers. The buffer, preferably a physiological buffer, may be the same as that used for hybridization. The skilled person is aware of suitable dehybridization conditions and buffers.

[0095] Optionally, the method may be inverted. Instead of immobilizing the target nucleic acid molecules and determining, e.g. visualizing, the first adapters binding thereto, the first adapters can be immobilized, and binding of the target nucleic acid molecules can be determined, e.g. visualized. A Visualization of the target nucleic acid molecules needs labeling, which can be performed either by in vitro transcription of the target molecule with labeled nucleotides, or by potentially multiple reference binding sites.

[0096] Optionally, target nucleic acid molecules can be mRNA molecules immobilized by pulldown, e.g. from cell lysate on a substrate, such as a microscopy coverslip. Pulldown as used herein can include the extraction of the mRNA directly out of a biological sample, such as a cell or a tissue, and subsequent surface immobilization of the molecules of interest, such as a particular subset or portion of the mRNA contained the biological sample. Thereby it can be assured, or at least the probability can be increased, that the mRNA molecules are folded natively, and the measured secondary structure or accessibility of the mRNA molecules reflects the structure under physiological conditions. This contrasts with in vitro transcribed mRNA molecules, which tend to fold to the native secondary structure with a lower probability. Alternatively the analysis of mRNA or the respective target molecule can be performed directly in cells or tissue.

[0097] Unless expressly stated otherwise in the description, an average is the arithmetic mean of a plurality of values.

[0098] Unless expressly stated otherwise in the description, any measurement method or analysis is performed at standard conditions or environmental conditions, including measurement at room temperature, such as about 20°C to 25°C, about 21 °C to 24°C, or Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0099] 23 / 68 about 22°C to 23°C, optionally, a pressure of about 96,0 kPa to 106,4 kPa, such as about 98,0 kPa to 105,0 kPa, about 100,0 kPa to 102,0 kPa, or about 101,3 kPa.

[0100] Further details and advantages will now be explained in more detail with reference to the following examples and optional embodiments with reference to the figures.

[0101] Fig. 1 shows the concept of DNA-PAINT,

[0102] Fig. 2 shows the TIRF illumination mode,

[0103] Fig. 3 shows DNA origami structures,

[0104] Fig. 4 shows an experimental design to analyze hybridization kinetics in single molecules using DNA origami structures,

[0105] Fig. 5 shows an analysis of DNA-DNA interactions,

[0106] Fig. 6 shows an analysis of RNA-DNA interactions,

[0107] Fig. 7 shows an analysis of RNA-RNA interactions,

[0108] Fig. 8 shows an analysis of DNA binding kinetics on hairpin constructs,

[0109] Fig. 9 shows images of selected DNA origami structures of a transient stem hairpin,

[0110] Fig. 10 shows hybridization of short DNA imager strands against loop and stem regions, Fig. 11 shows DNA binding kinetics of 8 and 9 nt DNA imagers to different loop sizes, Fig. 12 shows a secondary adapter design,

[0111] Fig. 13 shows the DNA version of the K4 Aptamer HHRi with biotin modification at the 5’- end and a colocalization sequence,

[0112] Fig. 14 shows a DNA-PAINT image of a colocalization sequence and adapter signal, Fig. 15 shows the binding sites of 20 nt secondary adapters within the DNA construct of the K4 Aptamer HHRi,

[0113] Fig. 16 shows the degree of colocalization for each adapter strand for 20 nt, 15 nt and 10 nt adapter lengths,

[0114] Fig. 17 shows a schematic representation of accessibility score calculations,

[0115] Fig. 18: shows a comparison of experimental data to a theoretical accessibility score, Fig. 19 shows the binding sites of 15 nt secondary adapters within the DNA construct, Fig. 20 shows the binding sites of 10 nt secondary adapters within the DNA construct, Fig. 21 schematically shows a construct of a target nucleic acid molecule being probed, Fig. 22 schematically shows a method for determining a secondary structure of a target nucleic acid molecule,

[0116] Fig. 23 shows experimental accessibility compared to theoretical / calculated accessibility of adapters, Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0117] 24 / 68

[0118] Fig 24 shows the accessibility or secondary structure of the K4 Aptamer HHRi at different temperatures,

[0119] Fig. 25 shows the influence of cellular environment on accessibility or secondary structure.

[0120] Fig. 26 shows the influence of Tetracycline on accessibility or secondary structure, and Fig. 27 shows the analysis of mRNA of green fluorescent protein (EGFP) and Luciferase (FLuc).

[0121] 1 Miscellaneous

[0122] 1.1 Introduction

[0123] 1.1.1 Construct of a target nucleic acid molecule being probed

[0124] First referring to Fig. 21, a construct 300 of a target nucleic acid molecule 304 being probed is schematically shown. A target nucleic acid molecule 304, in the present case an RNA molecule, is immobilized via a reference binding site 320 and biotin 162 residue on a substrate 302, such as a glass slide, which is treated with streptavidin for the binding of biotin 162. The reference binding site 320 enables binding of a reference imager 318 including a detectable label 316. In a loop region of the target nucleic acid molecule 304 a first adapter 306 binds via a first target binding site 308, e. g., a DNA strand of 20 nucleotides, to the target nucleic acid molecule 304. The first adapter 306 also comprises a first docking site 310, e. g., a DNA strand of 10 nucleotides that is directly fused to the first target binding site 308 and capable of binding to a corresponding first docking site 314, e. g. a DNA strand of 10 nucleotides, of a first imager 312. The first imager 312 comprises a detectable label 316 directly attached to the corresponding first docking site 314. Interaction of the target nucleic acid molecule 304 with a compound 322 suspected to alter the secondary structure of the target nucleic acid molecule 304 may affect, e. g. inhibit, the binding of the first adapter 306 thereby suggesting that the compound 322 may function as an inhibitor.

[0125] With further reference to Fig. 22, a method for determining a secondary structure of a target nucleic acid molecule 304 is schematically shown. The method 400 comprises the steps of a) immobilizing 410 target nucleic acid molecules 304 on a substrate 302, b1) contacting 412 the target nucleic acid molecules 304 with first adapters 306 under conditions allowing binding of the first adapters 306 to the target nucleic acid molecules 304, wherein each of the first adapters 306 comprises a first target binding site 308, which is capable of binding to the target nucleic acid molecules 304, and a first docking site 310, which is capable of binding to a corresponding first docking site 314 of a first imager 312, Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0126] 25 / 68 b2) contacting 414 the first adapters 306 with first imagers 312 under conditions allowing binding of the first adapters 306 to the first imagers 312, wherein each of the first imagers 306 comprises the corresponding first docking site 310 and a detectable label 316, c) performing time lapse imaging 416 on the immobilized target nucleic acid molecules 304 under conditions allowing detecting the detectable label 316, d) performing image processing 418 including localizing a detected label to the first target binding site 308 to establish the secondary structure of the target nucleic acid molecule 304, and the optional step d) assessing 426 an accessibility score for the first target binding site of the localized detected labels and comparing 426 the accessibility score with the secondary structure of the target nucleic acid molecule 304.

[0127] The sequence of steps 412, 414, 416, 418, and the optional steps 422, 424, and 426, may be repeated in step 420 with second adapters and second imagers, wherein each of the second adapters comprises a second target binding site, which is capable of binding to the target nucleic acid molecules 304, and a second docking site, which is capable of binding to a corresponding second docking site of a second imager, wherein each of the second imagers comprises the corresponding second docking site and a detectable label, wherein the second target binding site overlaps with the first target binding site or is different from it, wherein step d) includes localizing a detected label to the second target binding site.

[0128] Optionally, the method 400 may encompass, before step c), contacting 422 the target nucleic acid molecules 304 with reference imagers 318 under conditions allowing binding of the reference imagers 318 to the target nucleic acid molecules 304 at a reference binding site 320. Thereby, a colocalization sequence may be introduced into the target nucleic acid molecules 304, e. g., upstream or downstream the target nucleic acid molecule 304, preferably upstream, allowing the provision of an internal standard affording comparison of binding events with the internal standard and / or normalizing binding events with the internal standard.

[0129] Optionally, preferably after step a) and before step d), adding 424 a compound 322 suspected to affect, e.g. to alter or stabilize, the secondary structure of the target nucleic acid molecule 304, followed by step e) image processing and comparing 426, wherein image processing includes localizing a detected label to the first target binding site 318 to establish an altered secondary structure of the target nucleic acid molecule 304, and Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0130] 26 / 68 comparing includes comparing the secondary structure and the altered secondary structure.

[0131] 1.1.2 Basic concept of DNA-PAINT

[0132] Point Accumulation for Imaging in Nanoscale Topography (PAINT) is a super-resolution single molecule localization microscopy approach to create blinking by introducing diffusing probes in solution that specifically and transiently interact with the molecular target. Fig. 1 shows the DNA-PAINT concept 10. The basic approach using, e. g., Nile red or peptide-based probes, is limited by the availability of specific probes providing suitable interaction kinetics (Sharonov, A. & Hochstrasser, R. M. Wide-field subdiffraction imaging by accumulated binding of diffusing probes. Proc Natl Acad Sci II S A 103, 18911-18916, doi:10.1073 / pnas.0609643104 (2006); and Kiuchi, T., Higuchi, M., Takamura, A., Maruoka, M. & Watanabe, N. Multitarget super-resolution microscopy with high-density labeling by exchangeable probes. Nat Methods 12, 743-746, doi:10.1038 / nmeth.3466 (2015)). DNA-PAINT is an adaptation of this technique to leverage the application to any target that can be labeled with a short DNA strand (9-20 nt) (cf. e. g. Jungmann, R. et al. Single-molecule kinetics and super-resolution microscopy by fluorescence imaging of transient binding on DNA origami. Nano Lett 10, 4756-4761 , doi:10.1021 / nl103427w (2010); Jungmann, R. et al. Multiplexed 3D cellular super-resolution imaging with DNA- PAINT and Exchange-PAINT. Nat Methods 11 , 313-318, doi:10.1038 / nmeth.2835 (2014)). In DNA-PAINT short fluorophore-coupled oligonucleotides, which are used as imager strands, transiently hybridize to their complementary target sequences (or docking strand) that is attached to the molecule of interest (Fig. 1). The programmability of DNA hybridization offers control over binding 24 and unbinding 26 kinetics of the imager strand 12 which can be adjusted to the desired resolution and imaging speed. With the imager strands 12 in solution the docking strand is replenished with an unlimited supply of imager strands 12 eliminating photobleaching as a potential issue). Since imager strands 12 are fluorescent in solution it is required to use a microscope modality that offers optical sectioning to minimize background fluorescence. Typically, a Total Internal Reflection Fluorescence (TIRF) microscope setup is used for DNA-PAINT imaging which illuminates only a few hundred nanometers deep into the sample, as can be derived from Fig. 2 showing the TIRF illumination mode 28. As with other single molecule localization microscopy techniques the achievable spatial resolution is primarily limited by the number of photons detected from one binding event according to: Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0133] 27 / 68

[0134] With oi0Cbeing the localization precision, the standard deviation of the PSF o0and the number of photons N 2.

[0135] With further reference to Fig. 1 , short fluorophore-coupled oligonucleotides, which are used as imager strands 12, are freely diffusing in solution and capable of binding their complementary sequence (docking site), here incorporated in a DNA origami structure 14. The binding and unbinding of imager strands 12 create on / off fluorescent switching patterns 16, 18 as seen in the graph showing intensity 20 vs. time 22 (bottom). The duration that an imager strand 12 is bound to the docking site is denoted as bright time (T&) and the time between two binding events is referred to as dark time (TD) (Jungmann, R. et al. Single-molecule kinetics and super-resolution microscopy by fluorescence imaging of transient binding on DNA origami. Nano Lett 10, 4756-4761 , doi:10.1021 / nl103427w (2010)).

[0136] With further reference Fig. 2, the TIRF illumination mode 28 is shown in detail: The excitation laser path 30 passes through the side of a high numerical aperture, oil immersion objective lens 32 and is refracted such that it exceeds the critical angle at the interface between glass slide 36 and specimen 40. Consequently, the laser beam is reflected at the glass slide 36 / specimen 40 interface and an evanescent wave 38 excites the volume close to the surface of the glass slide 36 (100 - 300 nm). TIRF illumination dramatically improves the signal to noise ratio since only imager strands 12 that diffuse close to the glass slide are excited.

[0137] 1.1.3 DNA hybridization kinetics in DNA-PAINT

[0138] The binding and unbinding of the imager strands 12 generate typical intensity traces as shown in Fig. 1. The time span the imager is bound to the docking strand is referred to as on-time or bright time (T&) while the time between binding events is called off-time or dark time (TD). The bright time is independent of the imager concentration and depends on the stability of the imager and docking strand duplex. Main factors that affect the bright time are oligonucleotide length, GC-content, temperature, and the buffer’s salt content (e. g. Na+or Mg2+concentrations). By measuring the bright time, the off-rate (koff) of the imager strand 12 can be calculated as follows:

[0139] The dark time (TD) depends on the imager strand 12 concentration (c;) and an on-rate (kon) which is characteristic for each imager-docking strand pair. The on-rate can be Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0140] 28 / 68 determined by experimentally measuring the dark time with known imager concentration and is then calculated using the following equation:kon = ^ ciTD [3]

[0141] Docking-imager strand 12 pairs can be designed such that they provide distinguishable kinetic traces. These have been successfully employed for multicolor imaging (Wade, O. K. et al. 124-Color Super-resolution Imaging by Engineering DNA-PAINT Blinking Kinetics. Nano Lett 19, 2641-2646, doi:10.1021 / acs.nanolett.9b00508 (2019)). Furthermore, molecules can be counted by measuring the imager strand 12 binding frequency and comparing them to a calibration (qPAINT)(Jungmann, R. et al. Quantitative super-resolution imaging with qPAINT. Nat Methods 13, 439-442, doi:10.1038 / nmeth.3804 (2016)).

[0142] 1.2 DNA origami

[0143] The DNA origami concept is disclosed in Rothemund, P. W. Folding DNA to create nanoscale shapes and patterns. Nature 440, 297-302, doi:10.1038 / nature04586 (2006). Historically, DNA-PAINT was first demonstrated on so-called DNA origami nanostructures (Jungmann, R. et al. Single-molecule kinetics and super-resolution microscopy by fluorescence imaging of transient binding on DNA origami. Nano Lett 10, 4756-4761, doi:10.1021 / nl103427w (2010)). DNA origami represents a method to fold DNA into well- defined structures, such as rectangles, as may be derived from Fig. 3 showing the creation of DNA origami structures. Their main advantage which is used in this project is the addressability of DNA origami structures, acting as a molecular breadboard on the nanoscale. We use a rectangular DNA origami structure with dimensions of 70x100 nm. At a grid size of 5 nm, single-stranded DNA oligonucleotides can be positioned with high precision. We use these DNA oligonucleotides as DNA-PAINT binding sites or alternatively for capturing and immobilization of constructs to be investigated.

[0144] DNA origami structures 14 consist of a DNA scaffold 52 with a length of approx. 8000 nt. Typically around 200 short oligonucleotides, so called staples 54 or staple strands, are hybridized to the scaffold 52 and link regions in such a way that a defined nanostructure is created. After annealing 56, the staples 54 fold the scaffold 52 into a defined shape as shown in box 58. The figure shows the folding of a rectangular shaped DNA origami structure 14. Depending on the set of staple strands arbitrary structures and shapes can be designed and folded (Dey, S. et al. DNA origami. Nature Reviews Methods Primers 1, 13, doi:10.1038 / s43586-020-00009-8 (2021)). Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0145] 29 / 68

[0146] 1.3 Numerical methods for secondary structure predictions

[0147] For RNA secondary structural predictions from sequences, the potential Watson-Crick base pairs within the RNA sequence are taken into consideration. Many different algorithms exist to determine the secondary structure (Zhang, J., Fei, Y., Sun, L. & Zhang, Q. C. Advances and opportunities in RNA structure experimental determination and computational modeling. Nat Methods 19, 1193-1207, doi:10.1038 / s41592-022-01623-y (2022)). Free energy-based algorithms are a subset of algorithms that arrive at a predicted secondary structure by using experimentally derived parameters to minimize the free energy of the desired structures. As such, NLIPACK (Wayment-Steele, H. K. et al. RNA secondary structure packages evaluated and improved by high-throughput experiments. Nat Methods 19, 1234-1242, doi:10.1038 / s41592-022-01605-0 (2022)) is an example for such an algorithm. Free energy-based algorithms have a high accuracy for sequence lengths under 200 nt but have the drawback that errors in the free energy calculation accumulate which leads to predictions for longer lengths becoming more inaccurate. Additionally, they cannot account for other factors influencing RNA folding in vivo such as co-transcriptional folding effects or interactions with RNA binding proteins. Furthermore, the possibility to predict dynamic structural behavior is very limited.

[0148] 1.4 Analyzing super-resolution data in Picasso

[0149] The analysis of the acquired single molecule localization data is performed using a dedicated software package called “Picasso”. Picasso is an open-source software project developed in the lab of Prof. Dr. Ralf Jungmann at the Max Planck Institute for Biochemistry in Martinsried / Munich and is optimized towards DNA-PAINT data analysis (Schnitzbauer, J., Strauss, M. T., Schlichthaerle, T., Schueder, F. & Jungmann, R. Superresolution microscopy with DNA-PAINT. Nat Protoc 12, 1198-1228, doi:10.1038 / nprot.2017.024 (2017)). Picasso consists of multiple modules: “Picasso Localize” is used to identify and fit single-molecule spots. The resulting localizations can be rendered and postprocessed in “Picasso Render”. Here, sample drift can be corrected and localizations belonging to the same source can be clustered. Furthermore, time traces and binding kinetics can be determined, and multiple datasets can be aligned for further analysis. Recently a new Picasso module “Picasso-server” has been developed to enable automated processing of acquired microscopy data (Strauss, M. T. Picasso-server: a community-based, open-source processing framework for super-resolution data. Commun Biol 5, 930, doi:10.1038 / s42003-022-03909-5 (2022)). Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0150] 30 / 68

[0151] 2 Secondary structure prediction of RNA

[0152] 2.1 Initial concept - Direct transient binding of imager strands 12

[0153] 2.1.1 Introduction

[0154] In order to probe the secondary structure of RNA we use short fluorophore-coupled imager strands 12 to probe the whole RNA sequence. Depending on the secondary structure of the targeted regions we expect distinguishable binding kinetics. Highly structured regions are expected to yield no or very short binding events while unstructured regions are revealed by longer and more frequent imager binding events. As a read-out for the binding kinetics, we determine the bright time rb) and dark time (TD) for each imager strand 12 that is designed to bind the RNA construct.

[0155] 2.1.2 Single binding site measurements

[0156] 2.1.2.1 Introduction

[0157] Before directly analyzing the construct’s secondary structure, it is necessary to understand the kinetics of unstructured DNA / RNA regions. To accomplish that, we are evaluating different combinations of DNA-DNA, DNA-RNA, and RNA-RNA interactions for the same docking / imager sequence pair. The sequence are short single stranded sequences without any secondary structure, so they will serve as a reference for downstream measurements on more structured regions. Furthermore, understanding these interactions will provide the crucial parameters for the imager strand 12 design that will be later applied on the RNA construct.

[0158] 2.1.2.2 Experimental design

[0159] We are using a DNA origami structure 14 that is immobilized on a glass coverslip via BSA-biotin and neutravidin. The DNA origami density on the surface is adjusted so that thousands of structures can be recorded per field of view but still clearly separated. The DNA origami structure 14 is equipped with a capture strand extension to specifically immobilize DNA strands that carry a complementary DNA sequence as may be derived from Fig. 4 showing an experimental design 60 to analyze hybridization kinetics in single molecules using DNA origami. The intermediary strand consists of the complementary sequences that bind to the DNA origami and the docking strand sequence for kinetic analysis (e. g., P3 9nt). The free docking strand is then targeted using a complementary fluorophore coupled DNA imager strand 12. Furthermore, since the initial RNA construct comprises a poly-A tail, we evaluated the immobilization of poly-A on the DNA origami structure 14 that is extended with a poly-T capture strand. Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0160] 31 / 68

[0161] 2.1.2.3 Poly-A intermediary strand immobilization on DNA origami structures 14

[0162] To evaluate the planned immobilization strategy for the RNA constructs, a DNA origami structure 14 was designed with a single Poly-T tail in the center. An intermediary strand, composed of a Poly-A sequences and the DNA-PAINT docking strand P3 9nt, was hybridized to the DNA origami’s poly-T. The DNA origami was additionally modified with a 4-corner pattern that is used for their specific detection. We successfully detected the P3 docking site overlaying with the 4-corner DNA origami pattern proofing that the intermediary strand was successfully immobilized on the DNA origami via poly-A as shown in Fig. 4c. Approximately 80 % of the DNA origami structures 14 were hybridized to a poly-A intermediary strand.

[0163] In particular, in Fig. 4a) DNA origami design is shown. A hexagon in the center represents the position of the capture strand 62. Other hexagons represent DNA origami detections strands 64. Fig. 4b) shows that the DNA origami is modified with a ssDNA Poly-T capture strand 66, which is used to hybridize 70 a Poly-A DNA intermediary strand 74 displaying a P3 9nt binding site 72 acting as docking site complementary to the DNA imager 68. Fig. 4c) shows an overlay signal of the 4-corner DNA origami detection strands (bright small spots) and P3 docking strand signal (darker spots centered by the bright small spots) with a scale of 100 nm (scale bar).

[0164] Used oligonucleotide:

[0165] P3 8 nt DNA imager: AATGAAGA - ATTO655

[0166] P3 9 nt DNA intermediary strand:

[0167] TCTTCATTATTAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA

[0168] DNA origami extension: CCAATAGCTCATCGTAGGAATCATGGCATCAATTTTTTTTTTTTTTTTTTTTT

[0169] 2.1.2.4 DNA-DNA hybridization kinetics

[0170] First, we measured kinetics of DNA-DNA hybridization and compared two different imager strands 12, length 8 nt and 9 nt, as may be derived from Fig. 5 showing the analysis of DNA-DNA interactions 80. Both strands were evaluated on the same docking strand that was displayed on the DNA origami structure 14. The mean bright time of the 9 nt imager was 360 ms and therefore longer than that of the 8 nt imager (240 ms).

[0171] In particular, Fig. 5a shows a DNA intermediary strand carrying a P3 9 nt docking strand, which is (reversibly) hybridized 70 to a DNA-capture strand on a DNA origami structure Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0172] 32 / 68

[0173] 14. Fig. 5b shows hybridization kinetics of 8 nt (grey spots 86) and 9 nt (bright spots 88) DNA imager strands 12 to a 9 nt docking strand. Each spot 86, 88 represents the mean bright time (rb) and dark time (TD) of a single binding site shown in a graph depicting ( -Bright) ( ) ^S. (^Darfc) (s).

[0174] Used oligonucleotides:

[0175] 8 nt DNA imager: AATGAAGA - ATTO655

[0176] 9 nt DNA imager: TAATGAAGA - ATTO655

[0177] 9 nt DNA intermediary strand: TCTTCATTATT TTTGCGGTACACAGCTACCA DNA origami extension: TGGTAGCTGTGTACCGCAAA

[0178] 2.1.2.5 DNA-RNA hybridization kinetics

[0179] Next, we measured the hybridization kinetics of short DNA imager strand 12 to an RNA docking strand and vice versa. Using an RNA docking strand and DNA imager strand 12 we could not detect any interactions. However, when using the same sequence with a DNA docking strand and an RNA imager strand 12 longer bright times were observed compared to the DNA / DNA interactions as may be derived from Fig. 6 showing an analysis of RNA-DNA interactions.

[0180] This observation could be explained by results from previous studies that investigated RNA / DNA interaction depending on the pyrimidine (C, T) and purine (A, G) content (Hershel H. Lackey, Z. C., Joel M. Harris, Eric M. Peterson, and Jennifer M. Heemstra. Single-Molecule Kinetics Show DNA Pyrimidine Content Strongly Affects RNA: DNA and TNA:DNA Heteroduplex Dissociation Rates. ACS Synthetic Biology 9, 249-253, doi:10.1021 / acssynbio.9b00471 (2020)). These studies showed that in heteroduplex interaction a higher number of pyrimidines in the RNA strand yielded a dissociation rate significantly higher than purine-rich ones. The 8 nt imager strand 12 we used in this experiment is 87.5% composed of purines which would explain the lack of interaction when used as DNA, and longer binding events when used as RNA.

[0181] In particular, Fig. 6a) shows a DNA intermediary strand carrying a P3 9 nt docking strand that is hybridized to a DNA capture strand on a DNA origami structure 14. The docking strand is targeted using an RNA imager strand 12. Fig. 6b shows hybridization kinetics of 8 nt (grey spots 86) and 9 nt (bright spots 88) RNA imager strands 96, 98 to a P3 9nt binding site 72 acting as a docking strand. Each spot represents the mean bright time (tB) Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0182] 33 / 68 and dark time (TD) of a single binding site shown in a graph depicting (TBright>) (s) 82 vs. (TDark) (s) 84.

[0183] This experiment showed that the pyrimidine / purine content has a striking influence on the hybridization kinetics of short sequences. Thus, when using DNA imager strands 12 this factor must be considered together with the length and GC content of the sequences. Potentially this can be circumvented by using RNA imager strands 12 that are investigated in the following section.

[0184] Used oligonucleotides:

[0185] 8 nt DNA imager: AATGAAGA - ATTO655

[0186] 9 nt DNA imager: TAATGAAGA - ATTO655

[0187] 8 nt RNA imager: AAUGAAGA - ATTO655

[0188] 9 nt RNA imager: UAAUGAAGA - ATTO655

[0189] 9 nt DNA intermediary strand: TCTTCATTATT TTTGCGGTACACAGCTACCA

[0190] 9 nt RNA intermediary strand: UCUUCAUUAUU UUUGCGGUACACAGCUACCA DNA origami extension: TGGTAGCTGTGTACCGCAAA

[0191] 2.1.2.6 RNA-RNA hybridization kinetics

[0192] With both docking strand and imager strand 12 as RNA the detected interactions are an order of magnitude longer compared to DNA / DNA and DNA / RNA. Furthermore, the distribution of bright times is very broad, ranging from 100 ms to 10 s as may be derived from Fig. 7 showing an analysis of DNA binding kinetics 100. The measured dark time is in the same order of magnitude as the DNA / DNA interactions. If RNA imager strands 12 were used later for secondary structure analysis the sequences needed to be designed shorter (e. g., 6 or 7 nt) or with lower GC content to reduce the bright time to the range of 0.1-1s.

[0193] In particular, Fig. 7a shows an RNA intermediary strand 94 carrying a P3 9 nt docking strand is hybridized to a DNA capture strand 64 on a DNA origami structure 14. The docking strand is targeted using an RNA imager strand 92. Fig. 7b shows hybridization kinetics of 8 nt (grey spots) and 9 nt (bright spots) RNA imager strands 12 to a 9 nt RNA docking strand. Each point represents the mean bright time (rb) and dark time (TD) of a single binding site shown in a graph depicting (TBright) (s) 82 vs. (TDarfe) (s) 84. Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0194] 34 / 68

[0195] Used oligonucleotides:

[0196] 8 nt RNA imager: AAUGAAGA - ATTO655

[0197] 9 nt RNA imager: UAAUGAAGA - ATTO655

[0198] 9 nt RNA intermediary strand: UCUUCAUUAUU UUUGCGGUACACAGCUACCA- DNA origami extension: TGGTAGCTGTGTACCGCAAA

[0199] 2.1.3 Hairpin construct measurements

[0200] In the previous experiments we analyzed hybridization kinetics of short linear regions. However, RNAs are primarily composed of double-stranded regions and single-stranded loop regions. Therefore, we next measured the effect of double-stranded regions (stem) and curved regions (loops) on the imager strand 12 binding kinetics. The method can be compared to competitive binding assays as known from drug discovery. We monitor kinetically inter- vs. intramolecular base pairing via single molecule sensitive fluorescent readout. We started using a simplified setup and designed two hairpin constructs with different stem region stabilities, one transient (Tmelt = 22°C) and one stable (Tmelt = 38°C) and measured the hybridization kinetics of imager strands 12 that are complementary to the loop and stem region, respectively (Fig. 8). The hairpins were designed to directly fold into the DNA origami structure 14. As described above, DNA origami structures 14 consist of a long scaffold strand with a length of approx. 8000 nt and around 200 short staple strands, folding the scaffold strand into the desired shape. For the hairpin construct measurements presented in this section, we added sequence extensions at the 5’ and 3’ end of the hairpin construct to function as staple strands that directly fold into the DNA origami structures 14. The respective unmodified staple strands at these positions are excluded. We would expect that imager strands 12 against the stem region show binding to the transient stem hairpin but no binding to the stable stem version. For the imager strand 12 complementary to the loop region, we expect signal for both hairpins.

[0201] With reference to Fig. 8, an analysis of DNA binding kinetics 100 on hairpin constructs is shown. Fig. 8a shows a hairpin forming oligonucleotide 102, 104 that is directly incorporated into the DNA origami structure 14 during the folding process. The loop region 104 and stem region 102 are targeted separately with two complementary DNA imager strands 68 respectively. Fig. 8b shows the DNA origami design. Four spots show incorporated hairpin structures 106 and several further spots show DNA origami detection sequences 108. Fig. 8c shows the structure of transient and stable stem hairpin Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0202] 35 / 68 constructs 110, 112 with free energies of -8.5 and -17.1 kcal / mol respectively (calculated with NUPACK).

[0203] With further reference to Fig. 9 images of selected DNA origami structures 14 of the transient stem hairpin construct 110 of Fig. 8 having a loop region 122 are depicted. The loop signals 126 and stem signals 128 are expected to colocalize with the DNA origami detection sequences. However, most detection sequences show no or very little colocalizing signals as may be further derived from frame signals 124, overlay 132 (and negative control 130). Reasons for that could be poor incorporation of the hairpin into the DNA origami structures 14 or a lack of interaction of the imager strands 12 with the hairpin structure. To average out the incorporated hairpins per DNA origami structure 14 we measured all detected binding events per structure and compared them to a negative control imager strand 12, that has no complementary part within the hairpin sequence (Fig. 10). For the transient stem construct we saw an increased number of binding events for both the stem and loop region compared to the negative control. The loop region showed approximately twice the number of binding events compared to the transient stem region. In contrast, when measuring the stable stem region, the number of binding events per DNA origami structure 14 is almost the same as in the negative control. Moreover, there is no difference between the stem and loop region.

[0204] In particular, Fig. 9 shows images of a selected DNA origami structure 14 with an immobilized hairpin construct comprising a transient stem and a 11 nt loop region. The design of the DNA origami structure 14 is depicted in Fig. 8b. The negative control 130 was performed using an imager sequence not complementary to the hairpin construct. Scale bar: 100 nm.

[0205] With further reference to Fig. 10, hybridization of short DNA imager strands against loop and stem regions 140 is shown. Two different constructs (with transient and stable stems 110, 112) were analyzed using imagers that are complementary to their loop 126 and stem region 128, respectively. An imager strand 12 without complementarity was applied as a negative control to calibrate the background signal. For quantification, the total number of binding events 142 were counted per DNA origami structure 14 as shown in the graphs depicting the binding events 142 vs. count 144 for the negative control TGTGTGT 146, stem GAGAGAG 148, and loop AGGAGGA 150, respectively. The statistics are listed in table 5 below. Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0206] 36 / 68

[0207] In conclusion, we observed a signal in the loop and stem region for the transient stem hairpin design. Since twice the number of binding events were detected in the loop region compared to the stem, we could successfully distinguish the structured and unstructured region of that construct. However, no significant signal was detected in the stable stem hairpin construct. This result indicates that imager strand 12 binding kinetics are not only affected by double or single stranded binding regions but also by the stability of the nearby stem regions.

[0208] Used oligonucleotides:

[0209] 7 nt DNA imager: TGTGTGT - Cy3B

[0210] 7 nt DNA imager: GAGAGAG - Cy3B

[0211] 7 nt DNA imager: AGGAGGA - Cy3B

[0212] Transient stem construct: AA GAGAGAG AA TCCTCCT AA CTCTCTC AA

[0213] Stable stem construct: AA GG GAGAGAG GG AA TCCTCCT AA CC CTCTCTC CC AA

[0214] 2.1.4 Experimental design with biotin-DNA-hairpin structures

[0215] Using the DNA hairpin construct displayed on DNA origami structures 14 as performed in the previous experiments had some disadvantages: The incorporation of DNA strands is not 100 %, thus it is unknown how many constructs are present per DNA origami structure 14. Therefore, the analysis required to be performed on whole DNA origami structures 14 rather than single hairpin constructs. Additionally, each different hairpin construct requires a folding of a new DNA origami, making the process less work efficient.

[0216] Therefore, we developed a different experimental setup that includes biotinylated DNA strands as shown in Fig. 11a. These constructs include an additional DNA-PAINT docking strand at the 5’-end (colocalization sequence) that is used to detect each construct. Consequently, each DNA hairpin displaying signal at the DNA-PAINT detection strand also contains a hairpin structure and it is possible to analyze the kinetics on single molecules. Moreover, the biotinylated DNA strands can be directly attached to the surface via BSA-biotin and neutravidin, without the need of prior DNA origami folding.

[0217] 2.1.4.1 Loop size dependence on imager binding kinetics

[0218] In the previous experiment we discovered that a stable stem region prevents hybridization of short DNA imager strands 12 at the stem as well as at the loop region. In the following experiment, we aim to find out which loop size is required to successfully detect imager binding events and how binding frequencies, and bright times change when further Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0219] 37 / 68 extending the loop size. Therefore, we designed the stem to be stable, so the hairpin is not opening during the measurement. We hypothesize that binding events happen more frequently and longer upon increasing loop lengths since the tension of the loop is reduced.

[0220] Fig. 11b and Table 6 show the measured kinetics for loop sizes ranging from 11 nt to 23 nt. We did not detect any signal for 11 nt and 12 nt loop sizes. From a loop size of 13 nt on and higher, we began to detect short binding events when using 8 and 9 nt imager strands 12. Interestingly, for 8 nt imager strands 12 the bright time did not significantly change from 13 to 17 nt (40 - 50 ms) until a loop size of 23 nt when it increased to 104 ms. We observed a stronger effect on the dark time, which was reduced gradually as the loop size increased, meaning a higher binding frequency. For 9 nt imager strands 12, we detected a slightly different behavior. The bright time increased from 41 to 87 ms for 13 - 17 nt loop sizes and 173 ms for 23 nt. The dark time of the 9 nt imager shows a stronger reduction compared to the 8 nt imager.

[0221] In particular, Fig. 11 shows DNA binding kinetics of 8 and 9nt DNA imagers to different loop sizes 160. Fig. 11a shows a DNA hairpin construct having biotin 162 as detectable label. The construct is composed of a stable stem region and loop region. The oligonucleotide is modified with a biotin 162 at the 5’-end followed by a colocalization sequence 164 to specifically detect the constructs. The minimum size of the loop region is 11 nt, and it is extended by two T-spacers 106, each ranging from 1 x T to 6 x T. Fig. 11b shows the binding kinetics of 8 and 9 nt imagers 86, 88 complementary to the loop region. Data pairs show bright times 82 (left data) and dark times (TD) 84 (right data) calculated for each hairpin oligonucleotide. Middle lines represent the median, boxes the quartiles and whiskers the outliers of the data. Median values are listed in Table 6.

[0222] Used oligonucleotides:

[0223] P3 8 nt DNA imager: AATGAAGA - Cy3B

[0224] P3 9 nt DNA imager: TAATGAAGA - Cy3B

[0225] OxT Loop sequence: biotin - CTCTCTCTCTC

[0226] TGCGGTACACAGC TCTTCATTACT GCTGTGTACCGCA

[0227] 1xT Loop sequence: biotin - CTCTCTCTCTC TGCGGTACACAGC

[0228] T TCTTCATTACT GCTGTGTACCGCA

[0229] 2xT Loop sequence: biotin - CTCTCTCTCTC TGCGGTACACAGC T TCTTCATTACT T GCTGTGTACCGCA Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0230] 38 / 68

[0231] 3xT Loop sequence: biotin - CTCTCTCTCTC TGCGGTACACAGC TT TCTTCATTACT T GCTGTGTACCGCA

[0232] 4xT Loop sequence: biotin - CTCTCTCTCTC TGCGGTACACAGC TT TCTTCATTACT TT GCTGTGTACCGCA

[0233] 5xT Loop sequence: biotin - CTCTCTCTCTC TGCGGTACACAGC TTT TCTTCATTACT TT GCTGTGTACCGCA

[0234] 6xT Loop sequence: biotin - CTCTCTCTCTC TGCGGTACACAGC TTT TCTTCATTACT TTT GCTGTGTACCGCA

[0235] 12xT Loop sequence: biotin - CTCTCTCTCTC TGCGGTACACAGC TTTTTT TCTTCATTACT TTTTTT GCTGTGTACCGCA

[0236] 2.1.4.2 Effect of imager strand 12 length on loop binding kinetics When using short imager strands 12 (8 and 9 nt) it is possible to detect loops that have a relatively large size of at least 13 nt if the surrounding region has a stable secondary structure. Shorter loops are not detectable with reasonable camera exposure times in the range of 30 ms. Therefore, we evaluated if loops < 13 nt can be detected using longer imager strands 12. We extended the imager strands 12 to 10 nt and 11 nt length and evaluated both on a loop size of 11 nt. We were able to detect significant binding events. However, these binding events are some orders of magnitudes longer (tens of seconds) compared to usual DNA-PAINT events (0.1 - 1 s). When bright times exceed the time scale of tens of seconds it is practically challenging to measure them using DNA-PAINT because of photobleaching before dissociation and the requirement of very long measurements to obtain sufficient statistics. A plausible explanation for these very long bright times would be that 10 and 11 nt imager strands 12 might cause the loop to completely open, so the hybridization kinetics are comparable to linear single stranded DNA. In conclusion, loop sizes of < 11 nt are generally difficult to detect using the transient binding approach.

[0237] Used oligonucleotides:

[0238] 7 nt DNA imager complementary to loop region: AGGAGGA - Cy3B 7 nt DNA imager complementary to stem region: GAGAGAG - Cy3B Transient stem hairpin oligonucleotide: AAGAGAGAGAAAGGAGGAAACTCTCTCAA Stable stem hairpin oligonucleotide: AAGGGAGAGAGGGAATCCTCCTAACCCTCTCTCCCAA Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0239] 39 / 68

[0240] Summary and conclusion

[0241] We first quantified binding kinetics of unstructured DNA / DNA, RNA / DNA and RNA / RNA interactions. When using a DNA imager against an RNA docking strand, we could not detect interactions due to the high purine content of the imager strand sequence. This effect will make a DNA imager sequence design against the RNA constructs more challenging. Alternatively, RNA imagers can be used, which, in the testing, provided very long bright times albeit broadly distributed. The stability of the RNA imagers needs to be optimized by shortening the imager length or GC content. Disadvantages of using RNA imagers are the higher price for synthetic fluorophore conjugated oligonucleotides as well as the long lead times. To lower the price and production time, amine modified RNA oligonucleotides could be purchased and conjugated in-house to Cy3B fluorophores.

[0242] Secondly, we investigated the binding of short imager strands 12 to loop and stem regions both composed of DNA. We were able to distinguish stem and loop regions for a hairpin structure that were designed with a semi-stable (transient) stem. However, in terms of a stable stem interaction, neither stem nor loop could be detected. Furthermore, we analyzed the binding kinetics of the same imager strands 12 to variable loop lengths ranging from 11 to 23 nt. While 11 and 12 nt loop were not detected at all, we observed short binding events starting from 13 nt loops. The bright times were relatively short for all loops besides 23 nt which showed significantly longer binding events and binding frequencies.

[0243] The application of directly conjugated imager strands comes with several disadvantages: Firstly, this approach requires fluorophore-conjugated oligonucleotides which are expensive when ordered directly or work-intensive when conjugated in-house. Furthermore, performing intelligent design of imager sequences is required for each target construct. This is due to the narrow range of unstructured regions that yield appropriate binding kinetics with bright times in the range of 50 ms to 5 s. These issues strongly limit this approach for screening nucleic acid secondary structures in a brute-force and high- throughput manner. In the following chapter we show an adapted approach enabling costefficient, high-throughput screening.

[0244] 2.2 Secondary adapter strands

[0245] 2.2.1 Introduction

[0246] The disadvantages of using direct imager strands 12 would limit the analysis of whole constructs. To overcome these limitations, we propose a new concept that includes Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0247] 40 / 68 secondary adapter strands as may be derived from the secondary adapter design 170 shown in Fig. 12). Secondary adapter strands are unmodified single-stranded oligonucleotides (DNA or RNA) made of two components. One part is designed complementary to a section of the target constructs and the second part is identical for all adapter strands and represents a binding site for DNA-PAINT imaging.

[0248] In particular, Fig. 12 shows that one part is complementary to a region of the target construct. It can be designed with different lengths (e. g. a 10 nt, 15 nt or 20 nt sequence binding to target 172), which ultimately determines the resolution of the obtained secondary structure. The second part is an optimized imaging sequence 174 which is bound, e. g., by a fluorophore-coupled imager strand in solution.

[0249] We established an adapter strand which has two different regions. One end is capable of binding optimized imager strands 12 on the millisecond scale. The other end consists of sequences which are complementary to segments of the construct. Therefore, the adapter strand is designed to undergo two independent reactions: At high rate and time resolution the adapter can be localized, and regions which are selective to the construct varied in terms of length and sequence. Even very slow folding and unfolding events become amendable with this. The basic assumption made when designing this concept is that the binding probability of the adapter strands depends on the RNA structure. One factor is the accessibility which can be quantified as the probability that a certain region of an RNA strand is single stranded 16. The second parameter affecting the binding probability is the stability of the forming adapter / target duplex which depends on the length and GC content of the adapter strand. A third parameter, based on the results in section 2.1.3, is the tension of a single stranded fragment that depends on the loop size. When decreasing the adapter length, the binding probability becomes more selective towards accessibility and tension of the RNA structure.

[0250] The DNA-PAINT binding site at the adapter strands is identical for all adapter strands and the sequences with the highest on-rate can be used, leading to a strong signal while keeping the background and unspecific binding low (Strauss, S. & Jungmann, R. Up to 100-fold speed-up and multiplexing in optimized DNA-PAINT. Nat Methods 17, 789-791, doi:10.1038 / s41592-020-0869-x (2020). Furthermore, the same dye-modified imager strand cannot only be used within the same set of adapter strands but also for different sets of adapter strands, keeping the costs low and allowing a brute-force and high- throughput approach. Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0251] 41 / 68

[0252] 2.2.2 Proof of concept of secondary adapters

[0253] 2.2.2.1 Experimental design and workflow

[0254] For an initial proof of concept using secondary adapter strands, we focused on analyzing a DNA construct with a length of 109 nt as may be derived from Fig. 13 showing a DNA version of the K4 Aptamer HHRi as DNA construct. DNA constructs were prepared using a folding protocol designed for K4 Aptamer HHRi (compare section 3). The DNA constructs were immobilized via biotin on a BSA-biotin neutravidin surface. The density of the DNA constructs on the surface was adjusted to obtain a clear spatial separation of constructs.

[0255] With reference to Fig. 3 the DNA version of the K4 Aptamer HHRi 200 has a biotin 162 modification at the 5’-end and a colocalization sequence 164. Grey highlighted regions depict four accessible regions or loops 1-4202, 204, 206, 208 for which we expect higher probability of secondary adapter binding. Only unstructured regions with a length > 5 nt are considered to be accessible. The DNA sequence of the K4 Aptamer HHRi 200 is shown in Table 1.

[0256] After immobilization of the DNA constructs on the BSA-biotin neutravidin surface, one secondary adapter strand is added. Incubating only one secondary adapter strand per measurement prevents secondary adapter strands affecting each other. Depending on the length of the secondary adapter strands it cannot be excluded that the secondary adapter strands are bound permanently to the construct of interest and affect the secondary structure of the RNA when analyzing subsequent adapter strands. Additionally, it would then limit the minimal step size to the length of the secondary adapter strands. The actual measurement consists of two imaging rounds. The first imaging round is a DNA-PAINT measurement of the docking strand close to the biotin 162 anchor. These localizations identify the positions of the DNA constructs. The second imaging round is a DNA-PAINT measurement of the binding site at the adapter strand, identifying DNA-constructs with a secondary bound adapter. After both imaging rounds the localizations of both rounds are overlayed / colocalized and analyzed as shown in Fig. 14.

[0257] Fig. 14 shows a DNA-PAINT image of a colocalization sequence and adapter signal. Single constructs are identified by the colocalization sequence signal. Then the percentage of constructs that have a colocalizing adapter signal (or colocalization percentage 212) is calculated. In this region the colocalization is at 60%. The whole field Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0258] 42 / 68 of view (512x512 pixels of 130 nm) contains on average ten thousand K4 Aptamer HHRi 200 structures. Scale bar: 200 nm.

[0259] We used three sets of secondary adapter designs, each covering the entire DNA construct. The only difference between the sets of secondary adapters is the length of the sequence complementary to the DNA construct. In this experiment we used lengths of 20 nt, 15 nt and 10 nt. The secondary adapter strands are complementary to the DNA construct as described in Fig. 13. The complete list of secondary adapter strand sequences can be found in Table 2, Table 3, and Table 4. The chosen step-size for each dataset was 5 nt. Fig. 15, Fig. 19, and Fig. 20 visualize the binding sites of each adapter sequence of the 20 nt, 15 nt and 10 nt sets within the DNA construct respectively. The readout parameter used for the downstream analysis is the percentage of DNA constructs yielding a signal from hybridized secondary adapter strands (Fig. 13). The overall hypothesis is, that for regions within the constructs with a large fraction of unstructured, single-stranded parts, the percentage of colocalizing spots is higher compared to regions which are more structured and double-stranded.

[0260] Figs. 15, 19, and 20 shows binding sites of 20 nt (Fig. 15), 15 nt (Fig. 19), and 10 nt (Fig. 20) secondary adapters within the DNA construct. Dark grey highlighted areas depict the target sequences of each 20 nt, 15 nt, and 10 nt secondary adapter that was used.

[0261] 2.2.2.2 Secondary adapter strands - Qualitative results

[0262] Fig. 16 shows the degree of colocalization for each adapter strand 230 for 20 nt, 15 nt and 10 nt adapter lengths 236, 238, 240 (adapter number 232 vs. % of colocalizing spots 234). For longer adapter lengths the colocalizations are generally higher. Peaks in all three graphs correspond to the largest three unstructured regions. For the 10 nt adapters 240 only the two largest peaks are visible with overall lower colocalization. The scattered line in the 10 nt adapter 240 plot represents a repeated measurement 244 of a subset of sequences to demonstrate reproducibility. Continuous lines show adapters 236, 238, 240 subjected to full adapter screening 242.

[0263] The experimental results are shown in Fig. 16, separated for each set of secondary adapter strands 236, 238, 240. The percentage of colocalization is plotted against the position of the secondary adapter strand within the construct. The secondary adapter strands within one set are numbered, starting with number 1 at the 3’-end of the DNA construct. For secondary adapters with a length of 20 nt, we observe that the positions Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0264] 43 / 68 yielding a high percentage of colocalization correspond to major accessible regions within the construct. The construct has four accessible (> 5 nt) loop regions as illustrated in Fig. 13, with sizes, 6 nt, 7 nt, 11 nt and 8 nt, respectively. The three peaks where the 20 nt adapter strands show high colocalization can be assigned as follows: 1) adapter strands that simultaneously bind loop 1 and 2; 2) binding the largest accessible loop 3; and 3) binding to loop 4 with 8 nt size.

[0265] When decreasing the adapter length to 15 nt we obtained less signal from the first and third peak compared to 20 nt adapters. With the shortest adapters (10 nt), only the middle peak is visible while the overall colocalization is highly reduced. This indicates that 15 nt and 10 nt are more selective towards longer unstructured regions while 20 nt adapters require only short unstructured regions for binding as they are sufficiently long to bind two or more short accessible regions simultaneously. The overall lower colocalization with decreasing adapter strand lengths can be explained by the lower stability of the forming duplex.

[0266] To validate the reproducibility of the measurements, we repeated parts of the measurement with 10 nt secondary adapter strands as shown in Fig. 16, confirming that the method is reproducible.

[0267] 2.2.2.3 Accessibility score

[0268] The experimental data shown in the previous section have so far been discussed on a qualitative level, however a quantitative analysis is required for validation of our method. Based on numerical calculations performed with NLIPACK we developed an accessibility score. The accessibility score is designed to indicate the degree of unstructured, single stranded regions for a given section of the sequence within the construct. NLIPACK (as described in section 1.1.3) calculates for each nucleotide within the construct the probability of being unbound as may be derived from Fig. 17.

[0269] Fig. 17 shows a schematic representation of accessibility score calculations 250. NLIPACK simulations result in each base of a DNA structure 252 being assigned a probability to be unbound 254. The probabilities for each base within one binding segment for an adapter are averaged to get an accessibility score 256 that is representative of how accessible a region in the construct is. More unstructured regions have higher accessibility scores. Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0270] 44 / 68

[0271] Based on the probability scores for each nucleotide, we calculated the average of the probabilities to be unbound for the hybridization sequence for each secondary adapter strand. That average value is termed accessibility score and has a value close to 1 for unstructured regions such as single-stranded loop regions and a value close to 0 for highly structured regions such as double-stranded stem regions. Depending on the ratio between structured and unstructured regions within a given sequence, the accessibility score can yield all values between 0 and 1 .

[0272] 2.2.2.4 Comparison of experimental data with accessibility score Reference is made to Fig. 18 showing a comparison of experimental data 262 to theoretical accessibility score 256.

[0273] In Fig. 18, we compare the experimental data 262 with our theoretical accessibility score 256. The calculated accessibility score 256 results in three major unstructured regions, and therefore overlaps well with the experimental data for 20 nt and 15 nt adapter strands 236, 248, 240 but correlates less with the experimental data of 10 nt adapters. For shorter sequence the accessibility score 256 is less reliable because of the overall low duplex stability which is not considered. These results demonstrate the first successful proof of concept for the use of secondary adapter strands to investigate the secondary structure of nucleic acids.

[0274] 2.2.3 Conclusion

[0275] We developed the concept of the secondary adapter strands and demonstrated that we can analyze the secondary structure of a DNA construct with a length of 109 nt. The secondary adapter strands thereby solve the disadvantages associated with the use of direct dye-conjugated imager strands 12 (compare section 2.1). This new approach promises low-cost, brute-force analysis of the nucleic acid secondary structures.

[0276] 3 Materials and methods

[0277] 3.1 Secondary structure predictions and accessibility score for DNA construct

[0278] All RNA secondary structure predictions for calculations of the accessibility scores were performed using NLIPACK with a 600 mM NaCI concentration and a temperature of 21 °C. Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0279] 45 / 68

[0280] 3.2 RNA folding

[0281] The refolding protocol used was previously established for K4 Aptamer HHRi RNA (Serganov, A., Huang, L. & Patel, D. J. Coenzyme recognition and gene regulation by a flavin mononucleotide riboswitch. Nature 458, 233-237, doi:10.1038 / nature07642 (2009). Materials:

[0282] RNase free 1M Tris-HCL pH 8.0 solution (Invitrogen, AM9855G), RNase free 1M Tris-HCL pH 7.0 solution (Invitrogen, AM9850G), UltraPure™ DNase / RNase-Free Distilled Water (Invitrogen, 10977015), RNase free 2M KCI (Invitrogen, AM9640G), RNase free 1M MgCh (Invitrogen, AM9530G).

[0283] Procedure

[0284] Thaw desalted RNA sample on ice. Prepare 1 M Tris pH 7.4 by mixing 8 ml 1 M Tris-pH 7.0 with 2 ml Tris-pH 8.0 according to vendor recommendation (Thermo Fisher). Prepare RNA sample solution with 3 pM RNA, 50 mM Tris-HCI pH 7.4 and 100 mM KCI with 2 / 3 of the final volume and keep on ice. Prepare refold solution with 50 mM Tris-HCL pH 7.4, 100 mM KCI and 6 mM MgCI2 (1 / 3 of the final volume) and keep on ice. Place the RNA sample for 2 min at 95°C followed by snap cooling for 3min on ice. Combine 2 / 3 RNA sample volume with 1 / 3 RNA refold solution and incubate for 10 min at 37°C. Keep RNA at RT until further use and refold RNA samples ideally just before the experiment.

[0285] 3.3 Sample preparation for single molecule experiments Materials: ibidi glass bottom slide VI 0.5 (ibidi, cat. no. 80607),

[0286] 5 M NaCI (Invitrogen, AM9760G), UltraPure™ DNase / RNase-Free Distilled Water (Invitrogen, 10977015), PBS pH 7.4 (10x) (gibco, 70011-036), 2 M KCI (Invitrogen, AM9640G), 1 M MgCI2 (Invitrogen, AM9530G),

[0287] 90 nm Standard Gold Nanoparticles (Cytodiagnostics, G90-100), BSA (Sigma-Aldrich, cat. no. A4503-10g), NeutrAvidin™ Biotin Binding (Thermo Scientific, 31000), 0.5 M EDTA pH 8.0 (Invitrogen, AM9260G), 1 M Tris-HCI pH 8.0 (Invitrogen, AM9856). Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0288] 46 / 68

[0289] Buffer solutions:

[0290] Buffer A: 10 mM Tris-HCI and 100 mM NaCI at pH 8.0,

[0291] Buffer B: 5 mM Tris-HCI, 10 mM MgCI2 and 1 mM EDTA at pH 8.0,

[0292] Buffer C: 1 * PBS at pH 7.2 supplemented with additional 500 mM NaCI,

[0293] BSA-Biotin solution: 0.5 mg / ml BSA-biotin in Buffer A,

[0294] Neutravidin solution: 0.5 mg / ml neutravidin in Buffer A,

[0295] Gold-nanoparticle solution: 1:10 dilution of 90 nm gold nanoparticles in Buffer C,

[0296] DNA construct solution: 50 nM folded biotinylated DNA construct in Buffer B,

[0297] DNA Origami solution: 1:100 dilution of folded DNA origamis in Buffer B,

[0298] 100x Trolox: 100 mg of Trolox, 430 pl of methanol and 345 pl of NaOH (1 M) in 3.2 ml of H2O,

[0299] 40x PCA: 154 mg of PCA in 10 ml of water, adjusted to pH 9.0 with NaOH,

[0300] 100x PCD: 9.3 mg of PCD and 13.3 ml of buffer (50% glycerol stock in 50 mM KCI, 1 mM EDTA and 100 mM Tris-HCI, pH 8.0),

[0301] Imaging solution for DNA origami experiments: 500 pM-2 nM of DNA-PAINT imaging strands with either Cy3B or ATTO655 on the 5’ end, 1xPCA, IxTrolox, 1xPCD in Buffer C, Imaging solution 1 : 5 nM AGGAGGA-Cy3B (R1-Cy3B), 1xPCA, Ix Trolox, 1xPCD in Buffer C,

[0302] Imaging solution 2: 50 nM secondary adapter strand, 10 nM GAGAGAG-Cy3B (R3 - Cy3B), 1xPCA, Ix Trolox, 1xPCD in Buffer C.

[0303] Sample preparation for biotin and DNA origami experiments in section 2.1 :

[0304] Pipette BSA-biotin solution into flow cell,

[0305] Incubate for 3 min,

[0306] Wash flow cell with 200 pl Buffer A,

[0307] Add Neutravidin solution into flow cell,

[0308] Incubate for 3 min,

[0309] Wash flow cell with 200 pl Buffer A,

[0310] Wash flow cell with 200 pl Buffer B,

[0311] Add DNA origami solutions or biotin-DNA (25-50 pM) to the flow cells,

[0312] Incubate DNA origami for 3 min or 2 min for biotin-DNA,

[0313] Wash flow cell with 200 pl Buffer B,

[0314] Wash flow cell with 200 pl Buffer C,

[0315] Add Gold-nanoparticle solution to flow cell,

[0316] Incubate for 5 min,

[0317] Wash flow cell with 200 pl Buffer C, Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0318] 47 / 68

[0319] Add respective imaging solution for desired experiments.

[0320] Sample preparation for secondary adapter experiments in section 2.1: Pipette BSA-biotin solution into flow cell, Incubate for 3 min,

[0321] Wash flow cell with 200 pl Buffer A,

[0322] Add Neutravidin solution into flow cell, Incubate for 3 min,

[0323] Wash flow cell with 200 pl Buffer A, Wash flow cell with 200 pl Buffer B, Add Construct solutions to the flow cells, Incubate for 2 min,

[0324] Wash flow cell with 200 pl Buffer B, Wash flow cell with 200 pl Buffer C, Add Gold-nanoparticle solution to flow cell, Incubate for 5 min,

[0325] Wash flow cell with 200 pl Buffer C,

[0326] Add imaging solution 1 to flow cell,

[0327] Imaging round 1 (20000 frames, 100 ms exposure time), Wash flow cell with 2 ml Buffer C, Add imaging solution 2 to flow cell,

[0328] Imaging round 2 (40000 frames, 30 ms exposure time).

[0329] 3.4 Data processing pipeline for DNA construct data

[0330] After data acquisition Picasso Localize is used to detect single molecule spots and calculate their centroids (localizations). The detection threshold of the spot finding algorithm is adjusted for each measurement to only include specific binding events. A least square fitting algorithm is used to determine the centroids of the detected spots / binding events. In Picasso Render data from both channels are drift corrected using a redundant cross-correlation algorithm with a 500-frame segmentation and aligned against each other using the same gold beads in both channels. Localizations that are in in the same location in consecutive frames are linked in both channels in a one-pixel distance around each localization with a maximum allowed dark frame gap of two. Spots showing repetitive DNA-PAINT binding events throughout the measurement are selected in the channel containing the data from the colocalization sequence. Their coordinates are Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0331] 48 / 68 used in the other channel to determine the DNA-PAINT binding parameters of the signal from the secondary adapter strands in the same locations as the constructs.

[0332] The number of binding events (n) for each selected site in the secondary adapter channel is filtered (5<n<1000) and used to determine the percentage of picked localizations that have a relevant signal in both colocalizing sequence and secondary adapter channels.

[0333] 3.5 DNA construct sequences

[0334] Table 1 : Sequence of the DNA construct. Table 2: 20 nt adapter strand sequences

[0335] Table 3: 15nt adapter strands Tautz & Schuhmacher Law GAT1101P11WO 31. Oktober2024

[0336] 49 / 68

[0337] Table 4: 10 nt adapter strands Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0338] 50 / 68

[0339] Table 5: Statistics for Fig. 10.

[0340] Transient stem

[0341] Stable stem

[0342] Table 6: Median values of data shown in Fig. 11.

[0343] 4 High throughput ensemble readout approach of K4 Aptamer HHRi

[0344] The accessibility of different regions of the K4 Aptamer HHRi was tested under different temperature and buffer conditions (compare Fig. 23).

[0345] As an alternative to using DNA-PAINT as the basic acquisition mode, adapters and imagers can be modified such that they bind either meta stably or permanently. The signal from each construct can no longer be separated and thus provides a readout signal that is Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0346] 51 / 68 averaged over all constructs, leading to an ensemble averaged fluorescence intensity, which can be averaged over time. If the acquisition is started upon addition of adapters and / or imagers, the binding constants of the adapters can be elucidated.

[0347] Accessibility scans were performed on the K4 Aptamer HHRi, with kinetics measurements of the 18 adapters of Table 2, 19 adapters of Table 3 and 20 adapters of Table 4 for adapter lengths of 20, 15 and 10 nt, respectively. The corresponding figure 18 shows a good agreement with the theoretically calculated accessibility, with some significant differences. Using kinetics as a measure for accessibility has the advantage that a higher throughput can be generated than for the DNA-PAINT-like embodiment of the method: More target molecules are probed per field of view, at the expense of the single-molecule information. This is advantageous for some applications. Importantly, the kinetics of the interaction between short single stranded probes and larger constructs cannot be observed using SPR or other methods in the state of the art, showcasing the usefulness of this method.

[0348] In particular, experimental accessibility 502 and theoretical accessibility 504, respectively, against the adapter number 500 (adapters 20nt_4xR3_1 to 20nt_4xR3_18) are shown, three adapters 510, 512, 514 (20nt_4xR3_9, 20nt_4xR3_10, and 20nt_4xR3_12) are highlighted and further tested for evaluating the fluorescence intensity 522 over time 520 shown in seconds, with adapter 512 having Kon= 0.12 • 106M'1s'1, adapter 510 having Kon= 0.11 • 106M'1s'1, and adapter 514 having Kon« 0.05 • 106M'1s'1and an equilibrium not yet reached after 800 seconds. Error margins for the Konof adapters 512 and 514 are shown in Fig. 23.

[0349] The experimental accessibility 502 is calculated as the intensity in the target channel divided by the intensity in the reference channel, datapoints from one measurement are normalized.

[0350] 4.1 The sample of 4 was tested under different conditions.

[0351] In the following experiments, the accessibility score was retrieved by dividing the signal brightness of the target channel by that of the reference channel. The resulting values were normalized to 1 . Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0352] 52 / 68 a) To compare the accessibility or secondary structure of the K4 Aptamer HHRi at different temperatures, the same experiment was performed at room temperature (22 °C), and at elevated temperature of 37°C (compare Fig. 24). Temperature was controlled using an ibidi silver-line Plate Heater. The plot shows the experimental accessibility 502 over the adapter number 500 at a first temperature 530 of 21°C and a second temperature 532 of 37°C. For the higher temperature the loss of the side loops around adapters 4 and 16 is evident. b) Cell Lysate

[0353] To study the effect of the cellular environment on accessibility or secondary structure, the experiment was compared between different buffer compositions (compare Fig. 25). The standard Buffer C 540 was compared to standard Buffer C complemented with 10 mg / ml HeLa Cell Lysate 542. Similar to the temperature elevation, the addition of cell lysate renders the side loops (around adapters 4 and 16) inaccessible. c) Tetracycline

[0354] Another point was the accessibility dependence on potentially the presence of molecules that interact with the target molecule (compare Fig. 26). To that end, the experiment was performed under the addition of various concentrations of Tetracycline using five adapters 550, 552, 554, 556, 558 of table 3 (adapter 550 (15nt_4xR3_4), adapter 552 (15nt_4xR3_7), adapter 554 (15nt_4xR3_9), adapter 556 (15nt_4xR3_10), and adapter 558 (15nt_4xR3_14)). A slight but insignificant decrease of experimental accessibility with increasing tetracycline concentration 560 (in pM) can be observed.

[0355] 4.2 Protocols:

[0356] Fold construct: Follow the RNA refolding protocol of section 3.2. For 100 pl end volume use 61.6 pl sample solution and 5 pl RNA construct (1 pM). After the snap cooling use 33.4 pl Refolding solution. The end concentration of the RNA construct is then 50 nM.

[0357] PolyT -biotin solution: 5 nM PolyT-Biotin in Buffer B

[0358] Construct solution: Dilute the folded construct to 5 nM in Buffer B

[0359] Imager solution: 2.5 nM Adapter, 2.5 nM Poly T imager (TTTTTTTTTTT-Atto655), 5 nM Adapter imager (AGAGAGAGAGAGAGA-Cy3B, named 15nt R3-Cy3B). A separate imager solution needs to be prepared for each adapter probed. Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0360] 53 / 68

[0361] Procedure:

[0362] Add 20mul BSA-Biotin to each well of a 96-well plate with #1.5 thickness glass bottom, compatible with high-resolution microscopy. Prepare as many wells as adapters are probed (e.g. 18, 19, and 20 adapters for the accessibility scans using adapters of Table 2, 3, and 4, respectively).

[0363] Centrifuge the plate for a very short time, so that the liquid is not sticking to the side of the wells

[0364] Incubate for 3-5min

[0365] 2x Wash each well with 75pl Buffer A

[0366] Add 40pl Neutravidin solution Incubate for 3-5min

[0367] 1x Wash each well with 75pl Buffer A 2x Wash each well with 75pl Buffer B Add 75pl Poly T biotin solution Incubate for 10m in

[0368] 2x Wash each well with 75pl Buffer B

[0369] Add 75pl Construct solution Incubate for 10m in

[0370] 1x Wash each well with 75pl Buffer B

[0371] 2x Wash each well with 75pl Buffer C, and

[0372] Wash with 75ul imager solution corresponding to the adapter to be probed in the respective well once before adding with 75pljmager solution again to each well.

[0373] Load the 96 well plate onto a TIRF microscope and capture images of each well using a laser and filterset for detecting Cy3B and not Atto655 as well as for detecting Atto655 and not Cy3B.

[0374] RNA refolding protocol:

[0375] Refolding protocol established based on previously described procedures for K4 Aptamer HHRi RNA (compare Serganov, A., Huang, L. and Patel, D.J. (2009) Coenzyme recognition and gene regulation by a flavin mononucleotide riboswitch. Nature, 458, 233- 237; and Vicens, Q., Mondragon, E., Reyes, F.E., Coish, P., Aristoff, P., Berman, J., Kaur, H., Kells, K.W., Wickens, P., Wilson, J. et al. (2018) Structure-Activity Relationship of Flavin Analogues That Target the Flavin Mononucleotide Riboswitch. ACS Chem Biol, 13, 2908-2919). Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0376] 54 / 68

[0377] Material:

[0378] RNase free 1M Tris-HCL pH 8.0 solution (Invitrogen, AM9855G) RNase free 1M Tris-HCL pH 7.0 solution (Invitrogen, AM9850G) UltraPure™ DNase / RNase-Free Distilled Water (Invitrogen, 10977015) RNase free 2M KCI (Invitrogen, AM9640G) RNase free 1M MgCI2 (Invitrogen, AM9530G) Thermo-shaker (e.g. Qlnstruments, BioShake iQ), and PCR clean 1 ,5mL tubes.

[0379] Procedure:

[0380] Thaw desalted RNA sample on ice. Prepare 1M Tris pH 7.4 by mixing 8mL 1M Tris-pH 7.0 with 2mL Tris-pH 8.0 according to vendor recommendation (Thermo Fischer). Prepare RNA sample solution with 3uM RNA, 50mM Tris-HCI pH 7.4 and 100mM KCI with 2 / 3rd of the final volume and keep on ice. Prepare refold solution with 50mM Tris-HCL pH 7.4, 100mM KCI and 6mM MgCh (1 / 3rd of the final volume) and keep on ice. Place the RNA sample for 2min at 95°C followed by snap cooling for 3min on ice. Combine 2 / 3rdRNA sample volume with 1 / 3rd RNA refold solution and incubate for 10min at 37°C. Keep RNA at RT until further use and refold RNA samples ideally just before the subsequent assay.

[0381] Analysis:

[0382] Intensity approach:

[0383] For each measurement the sum of the intensity values for each frame are calculated

[0384] The sum of the intensity values of the frames corresponding to 15nt-R3-Cy3B fluorescence intensity (the adapter strand signal) is divided by the sum of the intensity values of the frames corresponding to Poly-A-Atto655 fluorescence intensity (the reference signal) for each position in each well. The resulting value may be used as an accessibility metric.

[0385] For multiple positions in the same well the average and the standard deviation are calculated and plotted. Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0386] 55 / 68

[0387] For each measurement the signal is normalized across all adapters and measurement conditions to be compared. This removes the dependence on differential sensitivity or excitation power for the Cy3B and Att655 related images and results in data points between 0 and 1.

[0388] 5. EGFP and FLuc

[0389] As shown in Fig. 27, the mRNA of green fluorescent protein (EGFP) and Luciferase (FLuc) were analyzed with 36 adapters and 83 adapters (compare Tables 7 and 8), respectively, spanning the whole range of > 1kB of the mRNAs.

[0390] While for EGFP, experimental and theoretical results show some alignment and especially the well-accessible regions (adapters 9 (20nt_4xR3_9) and 25 (20nt_4xR3_25), named 560 and 562 in Fig. 27) can be recovered, the theoretical predictions for FLuc break down and show that in silico predictions are not reliable enough, and experimental validation, for example using the method presented herein, is necessary( compare adapters 38 (20nt_4xR3_38) and 72 (20nt_4xR3_72) (named 564 and 566), respectively).

[0391] Table 7: 20 nt adapter strand sequences against EGFP Tautz & Schuhmacher Law GAT1101P11WO 31. Oktober2024

[0392] 56 / 68

[0393] Table 8: 20 nt adapter strand sequences against FLuc Tautz & Schuhmacher Law GAT1101P11WO 31. Oktober2024

[0394] 57 / 68 Tautz & Schuhmacher Law GAT1101P11WO 31. Oktober2024

[0395] 58 / 68 Tautz & Schuhmacher Law GAT1101P11WO 31. Oktober2024

[0396] 59 / 68

[0397] Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0398] 60 / 68

[0399] Reference signs

[0400] 10 DNA Paint concept

[0401] 12 imager strands

[0402] 14 DNA Origami (nano)structure

[0403] 16 on fluorescent switching pattern

[0404] 18 off fluorescent switching pattern

[0405] 20 time

[0406] 22 intensity

[0407] 24 binding

[0408] 26 unbinding

[0409] 28 TIRF illumination mode

[0410] 30 excitation laser path

[0411] 32 objective lens

[0412] 34 immersion liquid

[0413] 36 glass slide

[0414] 38 evanescent wave

[0415] 40 specimen

[0416] 50 creation of DNA origami (nano)structure

[0417] 52 scaffold DNA

[0418] 54 staples

[0419] 56 annealing

[0420] 58 box

[0421] 60 experimental design

[0422] 62 DNA capture strand

[0423] 64 DNA origami detection strands

[0424] 66 poly T DNA capture strand

[0425] 68 DNA imager

[0426] 70 reversible hybridization

[0427] 72 P3 9nt binding site

[0428] 74 Poly A DNA intermediary strand

[0429] 80 Analysis of DNA-DNA and DNA-RNA interactions

[0430] 82 ^Bright (S)

[0431] 84 ^Dark (S)

[0432] 86 8 nt DNA imager

[0433] 88 9 nt DNA imager Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0434] 61 / 68

[0435] 90 analysis of RNA-RNA interactions

[0436] 92 RNA imager

[0437] 94 RNA intermediary strand

[0438] 96 8 nt RNA imager strands

[0439] 98 9 nt RNA imager strands

[0440] 100 analysis of DNA binding kinetics

[0441] 102 variable stem

[0442] 104 binding site

[0443] 106 Incorporated hairpin structures

[0444] 108 DNA origami detection sequences

[0445] 110 transient stem hairpin construct

[0446] 112 stable stem hairpin construct

[0447] 120 images of a selected DNA origami structure

[0448] 122 loop region

[0449] 124 frame signals

[0450] 126 loop signals

[0451] 128 stem signals

[0452] 130 negative control

[0453] 132 overlay

[0454] 140 hybridization of short DNA imager strands against loop and stem regions

[0455] 142 number of binding events

[0456] 144 count

[0457] 146 negative control (TGTGTGT)

[0458] 148 stem (GAGAGAG)

[0459] 150 loop (AGGAGGA)

[0460] 160 DNA binding kinetics of 8 and 9nt DNA imagers to different loop sizes

[0461] 162 biotin

[0462] 164 colocalization sequence

[0463] 166 T spacers

[0464] 170 secondary adapter design

[0465] 172 10 nt, 15 nt or 20 nt sequence binding to target

[0466] 174 (optimized) imaging sequence

[0467] 200 DNA version of the K4 Aptamer HHRi

[0468] 202, 204, 206, 208 loops 1 to 4 Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0469] 62 / 68

[0470] 212 colocalization percentage

[0471] 220 binding sites of 20 nt secondary adapters within the DNA construct

[0472] 230 degree of colocalization for each adapter strand

[0473] 232 adapter number

[0474] 234 % of colocalizing spots

[0475] 236 20 nt adapters

[0476] 238 15 nt adapters

[0477] 240 10 nt adapters

[0478] 242 full adapter screening

[0479] 244 repeat measurements

[0480] 250 schematic representation of accessibility score calculations

[0481] 252 DNA structure

[0482] 254 probability to be unbound

[0483] 256 accessibility score

[0484] 260 comparison of experimental data to theoretical accessibility score

[0485] 262 experimental data

[0486] 270 binding sites of 15 nt secondary adapters within the DNA construct

[0487] 280 binding sites of 10 nt secondary adapters within the DNA construct

[0488] 300 construct

[0489] 302 substrate

[0490] 304 target nucleic acid molecule

[0491] 306 first adapter

[0492] 308 first target binding site

[0493] 310 first docking site

[0494] 312 first imager

[0495] 314 corresponding first docking site

[0496] 316 detectable label

[0497] 318 reference imager

[0498] 320 reference binding site

[0499] 322 compound

[0500] 400 method

[0501] 410 immobilizing Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 2024

[0502] 63 / 68

[0503] 412, 414, 422 contacting

[0504] 416 performing time lapse imaging

[0505] 418 performing image processing

[0506] 420 repeating 424 adding

[0507] 426 comparing

[0508] 500 adapter number

[0509] 502 experimental accessibility

[0510] 504 theoretical accessibility 510, 512, 514 adapters to K4 Aptamer HHRi

[0511] 530 first temperature

[0512] 532 second temperature

[0513] 540 Buffer C

[0514] 542 Buffer 0 / HeLa Cell Lysate 550, 552, 554, 556, 55 adapters to tetracycline

[0515] 560, 562, 564, 566 adapters to EGFP and FLuc

Claims

Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 202464 / 68Claims1. A method (400) for determining the accessibility of a locus of a target nucleic acid molecule (304), comprising the steps of a) immobilizing (410) target nucleic acid molecules (304) on a substrate (302), b1) contacting (412) the target nucleic acid molecules (304) with first adapters (306) under conditions allowing binding of the first adapters (306) to the locus of the target nucleic acid molecules (304), wherein each of the first adapters (306) comprises a first target binding site (308), which is capable of binding to the locus of the target nucleic acid molecules (304), and a first docking site (310), which is capable of binding to a corresponding first docking site (314) of a first imager (312), b2) contacting (414) the first adapters (306) with first imagers (312) under conditions allowing binding of the first adapters (306) to the first imagers (312), wherein each of the first imagers (312) comprises the corresponding first docking site (314) and a detectable label (316), c) performing time lapse imaging (416) on the immobilized target nucleic acid molecules (304) under conditions allowing detecting the detectable label (316), and d) performing image processing (418) including localizing a detected label to the first target binding site (308) to establish the accessibility of the locus of the target nucleic acid molecule (304).

2. The method (400) of claim 1 , wherein the method is for further determining a secondary structure of the target nucleic acid molecule (304), wherein the locus is a first locus, the method further comprising repeating steps b1) to d) with second adapters and second imagers, wherein each of the second adapters comprises a second target binding site, which is capable of binding to a second locus of the target nucleic acid molecules (304), and a second docking site, which is capable of binding to a corresponding second docking site of a second imager, wherein each of the second imagers comprises the corresponding second docking site and a detectable label (316), wherein the second target binding site overlaps the first target binding site (308) or is different therefrom, wherein optionally the second docking site and the second imager are identical to the first docking site and the first imager, wherein step d) includes localizing a detected label to the second target binding site.

3. The method (400) of any of claims 1-2, wherein in step a) the target nucleic acid molecules (304) are immobilized to the substrate (302) via a biotin residue (162) or viaTautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 202465 / 68 hybridization to an oligonucleotide pre-deposited on the substrate (302), optionally wherein the oligonucleotide pre-deposited on the substrate (302) allows immobilizing to an easily accessible region of the target nucleic acid molecule (304), such as a loop or a poly-A tail.

4. The method (400) of any of claims 1-3, wherein the density of the target nucleic acid molecules (304) on the substrate (302) is 1-10 molecules I pm2, wherein step c) comprises performing a DNA-PAINT like experiment and step d) comprises DNA-PAINT analysis.

5. The method (400) of any of claims 1-4, wherein the density of the target nucleic acid molecules (304) on the substrate (302) is 10-10000 molecules I pm2, wherein step c) comprises imaging during equilibration of the binding of the first adapters (306) to the locus of the target molecule (304), and step d) comprises analyzing the time-lapse signal of the time lapse imaging (416) to retrieve an on-rate of the first adapters (306) as a measure for the accessibility.

6. The method (400) of any of claims 1-5, wherein step d) further comprises assessing an accessibility score for the first target binding site (308) of the localized detected labels.

7. The method (400) of any of claims 1-6, further comprising, before step c), contacting the target nucleic acid molecules (304) with reference imagers (318) under conditions allowing binding of the reference imagers (318) to the target nucleic acid molecules (304) at a reference binding site (320), optionally wherein the reference binding site (320) is an easily accessible region of the target nucleic acid molecules (304), such as a loop, or a poly-A tail of the target nucleic acid molecules (304), or a reference docking site being introduced into each of the target nucleic acid molecules (304), or a reference docking site comprised in an immobilizing agent of step a).

8. The method (400) of any of claims 1-7, further including, preferably after step a) and before step d), adding a compound (322) suspected to affect the secondary structure of the target nucleic acid molecule (304), step e) image processing including localizing a detected label to the first target binding site (308) to probe an altered secondary structure of the target nucleic acid molecule (304), and g) comparing the secondary structure and the altered secondary structure.Tautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 202466 / 689. The method (400) of any of claims 1-8, wherein steps b1) and b2) are performed in a single step, preferably, wherein the first adapters (306) and first imagers (312) are contained in the same solution, and / or each of the adapters (306) has a length of 5 to 55 nucleotides, preferably 5-35 nucleotides, preferably wherein the adapters (306) are DNA adapters.

10. The method (400) of any of claims 1-9, wherein the first adapters (306) are adapted to permanently bind the target nucleic acid molecule (304) and to transiently bind the first imagers (312), or wherein the first adapters (306) are adapted to permanently bind the target nucleic acid molecule (304) and to permanently bind the first imagers (312).

11. A use of the method (400) of any of claim 1-10 for improving a mathematical model for predicting a secondary structure of a target nucleic acid molecule (304), wherein step d) further comprises assessing an accessibility score for the first target binding site (308) of the localized detected labels and comparing the accessibility score with the secondary structure of the target nucleic acid molecule (304).

12. A use of the method (400) of any of claims 1-10 for developing a drug and evaluating a drug candidate based on antisense oligonucleotides, wherein a high accessibility of a locus of a target molecule corresponds to a favorable target region of an antisense oligonucleotide, for example for downregulating the translation of a gene.

13. A use of the method (400) of any of claim 1-10 for pre-tox screening of an antisense oligonucleotide (ASO) drug or an ASO drug candidate, wherein off-target binding of an ASO of the ASO drug or the ASO drug candidate to an RNA molecules that is not a clinical target is indicative of a risk of adverse side effects of the ASO.

14. A use of the method (400) of any of claim 1-10 for optimization of an antisense oligonucleotide (ASO) drug candidate, wherein multiple variants of the ASO drug candidate, preferably wherein the ASO drug candidate comprises modified bases, are evaluated to determine the variant with the best interaction with the target molecule.

15. A kit for determining a secondary structure of a target nucleic acid molecule (304), comprisingTautz & Schuhmacher Law GAT1101 P11WO 31. Oktober 202467 / 68 a substrate (302), immobilizing reagents adapted to immobilize target nucleic acid molecules (304) on the substrate (302), first adapters (306), first imagers (312), wherein the first adapters (306) are adapted to bind to the target nucleic acid molecule (304) and to the first imagers (312), the first adapters (306) each comprising a first target binding site (308), which is capable of binding to the target nucleic acid molecules (304), and a first docking site (310), which is capable of binding to a corresponding first docking site (314) of a first imager (312), each of the first imagers (312) comprising the corresponding first docking site (314) and a detectable label (316).

Images