WO2025186451A1 - Method for multivalent selection of nucleic-acid-based binders using multimerization scaffold with target-tailored geometry - Google Patents

Abstract

The present invention pertains to the field of biotechnology and bioengineering, specifically to the design and selection of multivalent binders for interrogating protein-protein interactions (PPIs). This invention focuses on novel methods for generating multivalent binders with user-defined spatial organization, using a combination of nucleic acid polymers and a target-tailored multimerization scaffold.

Description

METHOD FOR MULTIVALENT SELECTION OF NUCLEIC-ACID-BASED BINDERS USING MULTIMERIZATION SCAFFOLD WITH TARGET-TAILORED GEOMETRY

Field of the Invention

[0001] The present invention pertains to the field of biotechnology and bioengineering, specifically to the design and selection of multivalent binders for interrogating, interrupting or interfering with protein-protein interactions (PPIs). This invention focuses on novel methods for generating multivalent binders with user-defined spatial organization, using a combination of nucleic acid polymers and a target-tailored multimerization scaffold.

Background of the Invention

[0002] Multivalent binders represent an effective tool for interrogating protein-protein interactions (PPIs), including those of therapeutic and diagnostic significance. The arrangement of ligands into a target-specific spatial organization improves the strength and specificity of binding. However, current methods for developing multivalent binders rely on known natural ligands or monovalently selected ligands, along with prior structural information about their binding sites. The multimerization of aptamers using flexible linkers often requires extensive linker optimization and monovalent efficacy does not always translate to multivalent function.

[0003] Furthermore, multivalency allows for more efficient enrichment of binding polymers when the concentration of individual members in the random polymer library is very low (some sequences are represented as a single copy in the starting library), for example, in all multiplexed HTS systems. An example of a particularly successful multivalent directed evolution platform is Phage Display. However, Phage display does not provide control over the geometry of ligand presentation and does not allow the arrangement of different binding ligands into the multimeric assembly (all phages are homo-multivalent).

Summary of the Invention

[0004] To address the aforementioned limitations, the present invention relates to a novel method for selecting multivalent binders with user-defined spatial organization.

[0005] This method leverages a multivalent evolution of nucleic acid polymers within a target-tailored multimerization scaffold. The invention integrates the geometric constraints of multivalent ligand presentation into a regular SELEX (Systematic Evolution of Ligands by Exponential Enrichment) procedure, thereby selectively enriching binding units capable of engaging in multivalent interactions with the target.

[0006] In the present method, a multivalent combinatorial library of nucleic acid polymers is prepared by hybridizing a target-tailored multimerization scaffold with a pool of random nucleic acid polymers, which is then subjected to iterative cycles of affinitybased selections following standard SELEX methodology.

[0007] Multivalent assemblies of polymers capable of binding multivalently to the target exhibit slower dissociation rates and are selectively enriched within the selection library.

[0008] The method according to the invention enables the selection of binders with distinct kinetic properties and binding epitopes compared to those obtained through monovalent SELEX using the same starting library of nucleic acids. Moreover, the inventors surprisingly discovered that multivalent assemblies of binders selected using the present method demonstrate superior inhibitory properties and target selectivity compared to their monovalently selected counterparts, as evidenced by ACE2- competition biolayer interferometry (BLI) experiments.

[0009] Additionally, the inventors discovered the significance of a moderately rigid target-tailored multimerization scaffold by comparing it with a set of more flexible configurations.

[0010] Applications of this method span a wide array of fields within biotechnology. From targeted drug delivery and precision medicine to the development of highly specific diagnostic assays, the potential impact is vast and transformative with implications for the design of novel therapeutics, biomaterials, and biosensors.

Detailed Description of the Invention

[0011] The present invention firstly relates to a method for identifying nucleic acid-based binders having binding affinities for a user-defined target, where said target provides a target-tailored geometry-and-valency selection constraint present during a Systematic Evolution of Ligands by Exponential enrichment (SELEX) procedure, the method comprising:

- Preparation of a random or combinatorial library of nucleic acid polymers;

- Preparation of a target-specific multimerization scaffold where geometry, dimension and valency match the desired target binding-site(s);

- Hybridizing said random or combinatorial library with said scaffold to create a multimeric combinatorial library;

- Performing iterative rounds of affinity-based selections according to the SELEX procedure between said multimeric combinatorial library of nucleic acid polymers and the target to allow to enrich for binders capable of engaging in multivalent interactions with the target;

- Determination of the selected binder sequences by NGS (Next Generation Sequencing).

[0012] In some embodiments, the method according to the invention comprises the further steps of synthesis of the final candidate(s) and functional affinity verification(s).

[0013] The term "nucleic acid-based binders having binding affinities for a user-defined target" are also referred to as aptamers. Aptamers are short, single-stranded DNA or RNA molecules that can fold into unique three-dimensional structures enabling them to bind with high specificity and affinity to specific target molecules. These targets can be virtually any kind of molecule, including proteins, peptides, carbohydrates, small molecules, and even whole cells or viruses.

[0014] The method according to the invention comprises a Systematic Evolution of Ligands by Exponential enrichment (SELEX) procedure. SELEX is an experimental procedure of progressive enrichment of high-affinity aptamers by subjecting a large combinatorial library of ssRNA, ssDNA or modified nucleic acids to repeated cycles of affinity-mediated selection. This procedure was first proposed in 1990 independently by Ellington and Szostak and Terk and Gold. Since then, many variations of the protocol have been developed (cellSELEX, SNP-SELEX, minimal primer and primer-free SELEX, gSELEX, one-step aptamer selection using non-fouling porous hydrogel etc.). Typical SELEX library consists of a central randomized region (20-60 nt) flanked by primerbinding sites (~20 nt). Typical starting library size for SELEX experiments can range between 10¹³ to 10¹⁵ unique sequences (~20 pmol to ~2 nmol of DNA/RNA respectively). Selection of a high-affinity aptamer (micromolar range for small molecule target and nanomolar to picomolar range for a protein) with low off-target activity is considered to be a good selection result.

[0015] Below are the detailed steps involved in a typical SELEX procedure:

1. Preparation of a Nucleic Acid Library

The SELEX procedure begins with the creation of a vast and diverse library of singlestranded DNA or RNA molecules. This library contains a random sequence region flanked by constant sequences. The random region is where binding to the target can occur, and the constant regions facilitate amplification steps. The diversity of the library ensures a broad range of potential binders. 2. Binding Phase

The nucleic acid library is incubated with the target molecule under conditions that allow specific binding interactions. Non-specific binders are washed away, leaving only those oligonucleotides that have bound to the target.

3. Separation of Bound from Unbound Oligonucleotides

After the binding phase, a method is employed to separate the target-bound oligonucleotides from those that did not bind. The method of separation varies depending on the nature of the target but can include techniques like column chromatography, magnetic bead separation, or even gel electrophoresis.

4. Elution of Bound Oligonucleotides

Once the bound and unbound oligonucleotides are separated, the next step is to elute or release the bound oligonucleotides from the target. This is often achieved by changing the conditions (e.g., temperature, ionic strength) to disrupt the binding interaction, allowing the collection of bound sequences.

5. Amplification

The eluted, target-bound oligonucleotides are then subjected to amplification, typically via PCR (polymerase chain reaction) for DNA or reverse transcription PCR (RT-PCR) followed by PCR for RNA. This step generates a large quantity of DNA corresponding to the selected sequences, preparing the pool for the next round of selection.

6. Iteration

The amplified pool of sequences is then used for another round (also called “iterative cycle”) of binding, separation, elution, and amplification. With each round, conditions can be progressively made more stringent to increase the selection pressure for high-affinity binders. Typically, SELEX involves multiple rounds (often 5-15) to enrich the pool with sequences that have the highest affinity for the target.

7. Cloning and Sequencing

After several rounds of selection and amplification, the enriched pool of sequences is cloned into suitable vectors and transformed into bacteria for propagation. Individual clones are then sequenced to identify the oligonucleotide sequences that were enriched during the SELEX process.

8. Characterization

The identified sequences are synthesized and characterized for their binding properties to the target. This step often involves determining the affinity constants (e.g., K_d values) through various biochemical assays to confirm and quantify the interaction. [0016] Several variations of the basic SELEX process exist, tailored to specific applications or challenges. These include Cell-SELEX (for targeting whole cells), Photo- SELEX (using photo-crosslinkable nucleotides for covalent bond formation with the target), Counter-SELEX (to eliminate sequences binding to unwanted targets), and others designed to improve the efficiency or specificity of the selection.

[0017] The method according to the invention begins with the preparation of a random or combinatorial library of nucleic acid polymers. This library is comprised of a diverse array of nucleic acid sequences, which can be selected from the group consisting of DNA, RNA, base- or backbone-modified DNA, base- or backbone-modified RNA, peptide nucleic acids (PNAs), or a combination thereof.

[0018] In the context of the invention, the terms "random library" and "combinatorial library" refer to collections of diverse molecules used for screening and identifying compounds with specific desired properties, such as binding affinity to a target molecule. Although these terms are sometimes used interchangeably, they have distinct nuances in their definitions and applications.

[0019] A random library refers to a collection of molecules, typically nucleic acids (DNA or RNA), peptides, or other polymers, that have been synthesized or generated in such a manner as to encompass a wide, random assortment of sequences or structures. In the context of nucleic acids, a random library is created by incorporating a mixture of all four nucleotides (A, T/G, C, and G/T) at multiple positions within the molecules, resulting in a vast pool of sequences. This randomness allows for the potential binding of these molecules to various targets, making random libraries particularly useful for initial screening in SELEX (Systematic Evolution of Ligands by Exponential Enrichment) processes where the target might not be well-characterized, or a broad range of potential binders is desired.

[0020] A combinatorial library, while similar in diversity to a random library, is typically generated through systematic methods that combine sets of building blocks in every possible way to generate a vast array of distinct molecules. In nucleic acid libraries, this could involve the deliberate combination of subsets of nucleotide sequences or motifs known or predicted to have certain functional properties. In peptide libraries, it could involve the systematic variation of amino acids at specific positions within the peptide sequence.

[0021] In some embodiments, the library is constructed from functionalized nucleic acid polymers (FNAPs) or highly functionalized nucleic acid polymers (HFNAPs). FNAPs and HFNAPs are sequence-defined synthetic polymers that contain base-modified nucleotides. Various side-chain modifications are introduced into pyrimidine nucleobases (T and C) at the C5 position or purine nucleobases (A and G) at position 08. At these functionalization positions, side chains are oriented outward from the base pairing face and are unlikely to disrupt Watson-Crick base pairing. This allows the production of a combinatorial library of HFNAPs or FNAPs by DNA Ligase (T3 or T4)-catalyzed translation from functionalized tri- or tetranucleotide building blocks on the ssDNA template library strand.

[0022] A critical aspect of the invention is the preparation of a target-specific multimerization scaffold, which is designed to match the geometry, dimension, and valency of the desired target binding sites. The scaffold can be composed of various materials selected from the group consisting of DNA (cyclic single stranded DNA or linear single stranded DNA), RNA (cyclic single stranded RNA or linear single stranded RNA) or XNA including: LNA, PNA, TNA, FANA, GNA, CeNA and HNA, peptides, PEG dendrimers, synthetic polymers, or combinations thereof.

[0023] In some embodiments, the core of the target-specific multimerization scaffold is engineered to be rigid, ensuring the precise spatial arrangement and orientation of attached nucleic acid polymers. A scaffold is considered rigid when the dimensions used are well below the persistence length of the polymer. The use of a rigid scaffold ensures the spatial and geometric positions of the binders matches the target-tailored intended design after synthesis.

[0024] The scaffold can comprise flexible spacers as long as the global geometry stays within the spatial boundaries defined by the target.

[0025] As used herein, the term “core” is the part of the scaffold that matches the target geometry.

[0026] As used herein, the term “rigid” refers to the property of the scaffold to maintain its shape and structural integrity in solution and with or without applied forces or stresses. A rigid scaffold is one that does not deform or change its spatial arrangement easily when subjected to external pressures, movements, or binding events. The term “rigid” is known and understood in the field of nanostructure, see for instance WO201 7189870 or Ohtsuki, Shozo et al. “Folding of single-stranded circular DNA into rigid rectangular DNA accelerates its cellular uptake.” Nanoscale (2019).

[0001] The rigidity of the scaffold can be routinely assayed by modelization with molecular simulations and the plotting of the probability distribution function of the spatial freedom of the extremities (Do-Nyun Kim, Fabian Kilchherr, Hendrik Dietz, Mark Bathe, Quantitative prediction of 3D solution shape and flexibility of nucleic acid nanostructures, Nucleic Acids Research, Volume 40, Issue 7, 1 April 2012, Pages 2862-2868).

[0002] Because the scaffold is rigid, the mean and median spacing between two adjacent nucleic acid polymers cannot significantly change (preferably less than 10%, more preferably less than 5%, even more preferably less than 1%). The nucleic acid polymers present on the scaffold according to the invention have thus limited spatial freedom and mobility in order to restrict the mobility of the interacting target in space during the interaction.

[0027] In some embodiments, the target-specific multimerization scaffold’s geometry can be selected in the group consisting of points, lines, rays, line segments, triangles, which can be classified as equilateral, isosceles, scalene, right, acute, or obtuse, quadrilaterals that encompass squares, rectangles, parallelograms, rhombuses, trapezoids, and kites, other polygons including pentagons, hexagons, heptagons, octagons, nonagons, decagons, hendecagons, and dodecagons. Additionally, the patterns or shapes can be curved like ellipses, circles, semi-circles, arcs, sectors, segments, and annuluses, or more intricate patterns including ovals, hearts, infinity symbols («), spirals, sine and cosine waves, tessellations, lattices, grids, chevrons, checkerboards, herringbones, zigzags, honeycombs, and labyrinths, and combination therein.

[0028] In some preferred embodiments, the target-specific multimerization scaffold’s geometry is a triangle with vertices of between 5 to 15 nm, preferably 7 to 12 nm.

[0029] The method of the invention is applicable to a list of targets including multimeric protein complexes, homomultimeric or heteromultimeric protein complexes, viral envelope proteins, trimeric viral capsid protein complexes, cellular receptors, enzymes, transcription factors, ion channels, structural proteins, and biomolecules of therapeutic, diagnostic, and research significance, including but not limited to nucleic acids, peptides, carbohydrates, and lipid structures involved in critical biological functions and disease pathologies, enzymes, receptors, and other proteins of therapeutic and diagnostic significance. All of the above can be homomultimeric or heteromultimeric, e.g. subunits of the same identity or of different identity.

[0030] The possible targets for nucleic acid-based binders, span a wide range of biological molecules and structures. These targets can vary greatly in complexity, size, and function, reflecting the diversity of roles they play in biological systems. These targets can include Proteins, enzymes, receptors, transcription factors, structural proteins, transport proteins, DNA, RNA (including mRNA, rRNA, tRNA, and miRNA), multimeric protein complexes, protein-nucleic acid complexes, cell surface complexes, hormones, metabolites, drugs, viruses, bacteria, parasites, fungi, cancer cells, stem cells, and specific tissue types, and combination thereof.

[0031] Identifying the target-specific multimerization scaffold's geometry and valency involves a comprehensive approach that integrates structural biology insights with computational modeling and experimental validation. This process ensures that the scaffold precisely mimics the spatial arrangement and multivalency of the target molecule's binding sites, thereby maximizing the efficacy of the multivalent binders. Here's how this identification process typically unfolds:

[0032] Structural Analysis of the Target

Data Gathering: Collect detailed structural information on the target molecule, preferably from high-resolution techniques like X-ray crystallography, NMR spectroscopy, or cryo-electron microscopy. This data provides insights into the molecular architecture, binding sites, and the spatial relationship between multiple binding sites.

Identify Binding Sites and Valency: Analyze the structural data to identify potential binding sites and determine their valency, which refers to the number of distinct binding interactions a single molecule can engage in simultaneously. For targets with known ligands or substrates, the positions and orientations of these molecules can offer clues about valency and geometry.

[0033] Designing of the Scaffold:

[0034] - Measuring the dimensions of the target proteins from the model obtained from the structural data (cryoEM, or X-ray crystallography...) and design the multimerization scaffold to allow the orientation of the ligands in accordance with these dimensions.

[0035] The designing of the scaffold could also be performed by Computational Modeling and Prediction:

Molecular Modeling: Employ computational tools to model the target molecule and its interactions with potential binders. This step may involve molecular dynamics simulations, docking studies, and the use of artificial intelligence (Al) algorithms to predict how different scaffold geometries and valencies might interact with the target. Design Scaffold Geometry and Valency: Based on the computational predictions and structural analysis, design a scaffold that mirrors the target's binding site geometry and valency. This design process involves selecting a scaffold material (e.g., DNA, RNA, peptides) and configuring its shape and functional groups to align with the target sites.

[0036] Experimental Validation

Synthesis and Testing: Synthesize the designed scaffold and conduct experimental tests to evaluate its binding to the target. Techniques like biolayer interferometry, surface plasmon resonance, or fluorescence assays can quantify binding affinity and specificity, providing feedback on the scaffold's design.

Iterative Optimization: Use the results from experimental validation to refine the scaffold design. This may involve adjustments to the scaffold's geometry, flexibility, or valency to improve interaction with the target. Multiple iterations of modeling, synthesis, and testing may be necessary to achieve optimal binding characteristics.

[0037] Considerations for Scaffold Design

Flexibility vs. Rigidity: The scaffold's flexibility can impact its ability to maintain the desired geometry and engage in multivalent interactions. A balance may be needed between structural rigidity to preserve geometry and some flexibility to accommodate binding under physiological conditions. In preferred embodiments, the scaffold is rigid.

Valency Matching: The scaffold should present its binding sites in a manner that matches the valency of the target. This ensures that all parts of the scaffold can potentially engage with the target, maximizing binding strength and specificity. In some embodiments, the valency of the target-specific multimerization scaffold is defined to be between 2 and 10, preferably between 2 and 8, more preferably between 3 and 7.

Geometry Accuracy: The precise spatial arrangement of binding units on the scaffold is crucial for effective multivalent interactions. This involves not only the distance between binding sites but also their angular orientation relative to each other and the target.

[0038] The inventors have successfully applied the method according to the invention to generate potent multivalent binders (with Kd in nanomolar range) targeting trimeric viral capsid protein complexes, such as the SARS-CoV2 Spike trimer (see Examples and figures).

[0039] In some preferred embodiments, the target is a trimeric viral capsid protein, preferably the SARS-CoV2 Spike trimer or the HIV Env complex. [0040] These trimeric proteins play critical roles in the viral life cycle, particularly in the entry process into host cells, making them prime targets for interventions designed to block viral infection.

[0041] The Spike protein of SARS-CoV-2, the virus responsible for COVID- 19, is a trimeric glycoprotein that protrudes from the viral surface, giving the virus its characteristic crown-like appearance under a microscope. It is responsible for mediating the entry of the virus into host cells by binding to the ACE2 receptor on the cell surface. The Spike protein comprises two subunits: S1, which contains the receptor-binding domain (RBD) that directly interacts with ACE2, and S2, which facilitates membrane fusion.

[0042] Therapeutic and Diagnostic Relevance:

- Blocking Interaction: Binders that specifically target the RBD of the Spike protein can block its interaction with ACE2, thereby preventing viral entry and subsequent infection.

- Vaccine Development: Identifying unique epitopes within the Spike protein can aid in the design of vaccines that elicit a robust immune response.

- Diagnostics: High-affinity binders can serve as the basis for diagnostic tests to detect the presence of the virus in biological samples.

[0043] The Env complex of the Human Immunodeficiency Virus (HIV) consists of a trimer of heterodimeric spikes, each composed of three gp120 surface proteins and three gp41 transmembrane proteins. The gp120 component binds to the CD4 receptor on host T cells, a critical initial step for viral entry, followed by a conformational change that allows gp120 to interact with a coreceptor (CCR5 or CXCR4), ultimately leading to the fusion of the viral and cellular membranes mediated by gp41.

[0044] Therapeutic and Diagnostic Relevance:

- Inhibition of Entry: Binders targeting the gp120 or gp41 components of the Env complex can inhibit the virus's ability to bind to CD4 and coreceptors, blocking entry into host cells.

- Neutralizing Antibodies: Identifying binders that mimic or induce the production of broadly neutralizing antibodies can be pivotal for vaccine development, offering protection against a wide range of HIV strains.

- Diagnostics: Specific binders to the Env complex can be utilized in diagnostic assays to detect HIV infection, monitor viral load, or screen blood products.

[0045] Once the scaffold that matches the desired geometry and valency of the target binding sites is prepared, it can be hybridized with the random or combinatorial library to create a multimeric combinatorial library of nucleic acid polymers comprised of target- matched valency-and-geometry random multivalent binding candidates.

[0046] The prepared scaffold and the nucleic acid library (either random or combinatorial) are typically mixed in a solution that facilitates hybridization. The nucleic acid polymers in the library have sequences complementary to specific sites on the scaffold, allowing them to bind precisely to these sites. As the nucleic acid polymers from the library bind to the scaffold, they form a multimeric combinatorial library of nucleic acid polymers. This library is comprised of complexes where the spatial organization and valency of the nucleic acid polymers are dictated by the structure of the scaffold. These complexes are designed to match the target's geometry and valency.

[0047] The hybridization conditions, such as temperature, salt concentration, and pH, are optimized to favor the specific interaction between the scaffold and the nucleic acid polymers.

[0048] The outcome of the method according to the present invention is a diverse library of multivalent nucleic acid complexes, each with the potential to interact with the target in a multivalent manner. This multivalency can significantly increase the effective binding affinity (avidity) and specificity compared to monovalent interactions, due to the cooperative effects of multiple binding events occurring simultaneously.

[0049] This outcome is different from a standard monovalent SELEX procedure. The traditional monovalent SELEX yields binders that might not have a multivalent activity or activity when multimerized. Furthermore, the depth of sequence variety is limited in a monovalent SELEX procedure, producing binders that belong to the same family.

[0050] Following hybridization, this multimeric combinatorial library is then subjected to iterative rounds of affinity- based selections according to the SELEX procedure. This typically involves iterative rounds of binding, separation, and amplification, progressively enriching the library in favor of those complexes that exhibit the desired binding characteristics.

[0051] The number of rounds in a SELEX (Systematic Evolution of Ligands by Exponential enrichment) procedure can vary significantly depending on several factors, including the complexity of the target, the diversity of the initial nucleic acid library, the efficiency of the selection process, and the specific goals of the experiment. Typically, a SELEX process involves between 5 to 15 rounds of selection, but there are instances where fewer or greater rounds are necessary. [0052] For relatively simple targets or when high-affinity binders are easily enriched, fewer rounds (around 5-8) might be sufficient to isolate strong binders.

[0053] For more complex targets, or when the desired property is not just affinity but also specificity against a backdrop of closely related molecules, more rounds might be necessary. In such cases, 10-15 rounds, or sometimes even more, are not uncommon.

[0054] Each round of selection aims to enrich the pool of nucleic acids with those having higher affinity and specificity for the target. After a certain number of rounds, the process reaches a point of diminishing returns where additional rounds do not significantly increase the affinity or specificity of the selected ligands. At this stage, the process is usually stopped, and the most promising candidates are identified for further characterization.

[0055] It's also worth noting that the optimization of selection conditions throughout the SELEX process can affect the number of rounds needed. For example, increasing the stringency of selection conditions (e.g., reducing the concentration of target, altering salt concentrations, changing temperature) in later rounds can help to more quickly enrich high-affinity binders, potentially reducing the total number of rounds required.

[0056] In another aspect of the invention, the invention relates to the use of the Systematic Evolution of Ligands by Exponential enrichment (SELEX) procedure to identify nucleic acid-based binders having binding affinities for a user-defined target, wherein a random or combinatorial library of nucleic acid polymers is hybridized with a target-specific multimerization scaffold where geometry, dimension and valency match the desired target binding-site(s) before performing at least two iterative rounds of affinity-based selections according to the SELEX procedure.

[0057] In another aspect of the invention, the invention relates to nucleic acid-based binders capable of engaging in multivalent interactions with a target, wherein said nucleic acid-based binders are selected according to the method of the present invention.

[0058] In some embodiments, the nucleic acid-based binders capable of engaging in multivalent interactions with the target comprise at least about 10 nucleotides up to about 100 nucleotides. In some embodiments, the nucleic acid-based binders capable of engaging in multivalent interactions with the target have an equilibrium constant (Kd) of about 1 pM up to about 10.0 pM; about 1 pM up to about 1.0 pM; about 1 pM up to about 100 nM; about 1 pM up to about 10 nM; about 1 pM up to about 1 nM; about 1 pM up to about 100 pM. [0059] In another aspect of the invention, the invention relates to multivalent binding complex comprising a cyclic single-stranded DNA (cssDNA) multimerization scaffold and three binding units that sequence-specifically hybridize to the multimerization scaffold (see Fig. 7). Preferably this molecular modality is used in the method of the present invention.

[0060] The multimerization cssDNA scaffold comprises three binding unit anchoring sites spaced by three hinge regions. Binding unit anchoring sites can be 10-45 monomeric units long, with each unit being one of the four natural DNA nucleotides (A, T, G, C) or one of the modified bases (Super T (5-hydroxybutynl-2’-deoxyuridine), Super G (8-aza-7-deazaguanosine)). The hinge regions can be 1-40 monomeric units, or preferably 1-20 or more preferably 1-10 monomeric units long, where unit refers to a polymer monomeric repeat, preferably with each unit being one of the four natural DNA nucleotides (A, T, G, C) or one of the following modified monomers (abasic nucleotide (1’,2’-Dideoxyribose), iSp18 spacer (hexa-ethyleneglycol), iC3 spacer, iSp9 spacer(triethylene glycol)), iPC spacer (Photo-cleavable spacer)). The sequences of binding unit anchoring sites can be all identical, two identical and one orthogonal, or all three orthogonal.

[0061] The binding units comprise the scaffold-hybridization region, which is a DNA sequence complementary to one or several binding unit anchoring sites on the multimerization cssDNA scaffold, and a binding moiety connected to the scaffoldhybridization region via a flexible linker. The scaffold-hybridization region can be 8-80 monomeric units, or preferably 10-45 monomeric units long, with each unit being one of the four natural DNA nucleotides (A, T, G, C) or one of the modified bases (Super T, Super G). The flexible linker can be 1-20 monomeric units, or more preferably 1-10 monomeric units long, with each unit being one of the four natural DNA nucleotides (A, T, G, C) or a polymer, preferably one of the following modified monomers (abasic nucleotide, iSp18 spacer, iC3 spacer, iSp9 spacer, iPC spacer). The binding moiety can be DNA or RNA aptamer, FNAP or HFNAp binder, peptide, protein, or small molecule.

[0062] The programmability of this system allows for the creation of hetero-trivalent assemblies of binding units using the multimerization cssDNA scaffold with corresponding orthogonal binding unit anchoring sites. Additionally, the flexibility/rigidity of the multimerization cssDNA scaffold can be modulated using the aforementioned nonnatural monomers. [0063] The invention comprises broad applications across biotechnology and medicine.

By enabling the creation of multivalent binders that can simultaneously interact with multiple parts of a target molecule, this strategy significantly enhances the specificity and strength of these interactions when these identified binders are presented on a nanoparticle, material scaffold, carrier, drug etc. Here are several possible applications:

[0064] Therapeutic Agents

[0065] Viral Inhibition: For viruses like HIV or SARS-CoV-2, multivalent binders can be designed to block critical proteins involved in the virus's entry into host cells, such as the Spike protein in SARS-CoV-2 or the Env protein in HIV, preventing infection.

[0066] Cancer Therapy: Binders can target cancer-specific homo-multimeric and/or hetero-multimeric markers or receptors overexpressed on tumor cells (e.g., HER2 in breast cancer) to deliver therapeutic agents or to inhibit signaling pathways crucial for tumor growth and metastasis.

[0067] Autoimmune Diseases: By targeting specific immune cell receptors or cytokines involved in autoimmune responses, multivalent binders can modulate the immune system, offering a new approach to treating conditions like rheumatoid arthritis or multiple sclerosis.

[0068] Diagnostic Tools

[0069] Biomarker Detection: Multivalent binders can enhance the sensitivity and specificity of diagnostic assays for detecting biomarkers associated with diseases such as cancer, cardiovascular diseases, or viral infections, enabling early detection and monitoring of disease progression.

[0070] Pathogen Identification: The high specificity of multivalent binders makes them ideal for developing rapid, accurate diagnostic tests for identifying infectious pathogens, including bacteria, viruses, and fungi, crucial for timely treatment and containment of infectious diseases.

[0071] Drug Delivery Systems

[0072] Targeted Drug Delivery: Multivalent binders can be used to create targeted drug delivery systems that specifically direct therapeutic agents to diseased cells while minimizing exposure to healthy cells, reducing side effects and increasing therapeutic efficacy.

[0073] Controlled Release: The strong and specific binding to target molecules can also be exploited to design systems that release drugs in response to specific stimuli (e.g., changes in pH or the presence of certain enzymes) found in the disease microenvironment.

[0074] Research Tools

[0075] Protein-Protein Interactions: This method can facilitate the study of proteinprotein interactions by creating binders that can selectively stabilize or inhibit these interactions, helping to elucidate their roles in cellular processes and disease mechanisms.

[0076] Molecular Structure Studies: By binding to specific conformations of molecules, multivalent binders can aid in the study of molecular structures, dynamics, and conformational changes, providing insights into their functions and interactions.

[0077] Biosensors

[0078] Environmental Monitoring: Multivalent binders can be incorporated into biosensors designed to detect toxins, pollutants, or pathogens in the environment, offering a powerful tool for environmental monitoring and safety assessments.

[0079] Biological and Chemical Warfare Detection: The high specificity and sensitivity of these binders make them suitable for developing detection systems for biological and chemical warfare agents, enhancing defense capabilities and response strategies.

[0080] Regenerative Medicine

[0081] Stem Cell Modulation: By targeting specific surface markers or signaling pathways, multivalent binders can influence stem cell differentiation and proliferation, opening new avenues for tissue engineering and regenerative medicine.

[0082] Description of the figures:

[0083] Figure 1: DNA multimerization scaffold with target-tailored geometry for evolution of Functionalized Nucleic Acid Polymers (FNAPs). a. A scheme representing monovalent and multivalent selection strategies for the evolution of FNAPs. b. Multiple therapeutically relevant protein targets feature distinct and evolutionarily- conservative 3-fold symmetry, which can be utilized in the rational design of multimerization scaffolds. The SARS-CoV-2 S-protein (PDB ID: 5I08) and HIV Env complex (PDB ID: 7L8T) were used in this study. c. Two multimerization scaffold designs were considered: cssDNA and DNA dendrimer. dsDNA extensions that replace the FNAP binding units were added for AFM imaging. The scale bar for AFM images is 20 nm. d. Stochastic modeling of trivalent assemblies’ composition as a function of the frequency of binding sequences in the library. e. Given that binding sequences are sparse in the naive library, the initial stages of multivalent selection are dominated by monovalent binding. The affinity distribution is an intrinsic property of each random library and cannot be known beforehand. For the method to be universally applicable to different nucleic acid libraries, the multimerization scaffold is added from the very first selection round, as the number of pre-focusing monovalent selection rounds necessary varies for different nucleic acid libraries and depends on the target protein. f. Upon enrichment of binding sequences in the pool, multivalent selection becomes dominant. By definition, assemblies of binders with high binding cooperativity have slower off-rates and thus expand faster in the pool of sequences over the round of biopanning. Kmono - dissociation constant of the monovalent ligand unit; JLU - angular steric restriction factor between the ligand unit and nanoparticle’s core; m - number of binding pockets per target’s subunit; C_n - cooperativity factor.

[0084] Figure 2. Multivalent SELEX of trimeric FNAPs: design of template and trinucleotide libraries. a. A scheme of multivalent SELEX for trimeric FNAP assemblies. b. The structure of 8-mer and 12-mer template libraries with corresponding Sanger sequencing electrophoretic traces for direct and reverse sequencing runs. c. Structures of functionalized and non-functionalized trinucleotide building blocks for the production of FNAP library. d. Translation reaction of biotinylated 8-mer and 12-mer template libraries into corresponding FNAP libraries performed at different trinucleotide library concentrations. r8 and r12 (non-biotinylated 8-mer and 12-mer template libraries respectively) were loaded as a reference. e. 8% Native PAGE of the library prepared according to the method of the present invention (herein referred to as “MEDUSA” for Multivalent Evolved DNA-based SUpramolecular Assembly) prepared with an 8-mer FNAP library. f. 6% Native PAGE of the MEDUSA library prepared with a 12-mer FNAP library. g. AFM image of the MEDUSA library prepared with an 8-mer FNAP library.

[0085] Figure 3. Selection of SPIKE-protein-binding FNAPs via multivalent SELEX. a. Progress in SPIKE-binding selection for monovalently pre-focused and multivalent selection strategies. Bulk affinity of trivalent and monovalent FNAP libraries to trimeric SPIKE protein was assessed by quantifying the amount of FNAPs in the flow-through versus elution fraction using qPCR. b. Increase in bulk affinity of trimeric FNAP libraries to SPIKE-protein by comparing the MEDUSA assemblies prepared with FNAP library from selection round 1 versus selection round 7 for both selection strategies using Mass-Photometry. c. Progression of selection process indicated by the decrease in FNAP library complexity using NGS. d. MSA of the top 100 sequences from NGS data for two tested selection strategies with corresponding sequence abundances. e. Enrichment of three selected hits over the rounds of selection for two tested selection strategies using NGS. f. Sequences and side-chain structures of selected FNAPs. g. Variants of ml and m2 sequences for primer minimization studies. h. SPR sensorgrams characterizing the binding kinetics between surface-immobilized trimeric SPIKE protein and monomeric forms of ml and m2 binders. The concentrations of injected FNAPs were 18.75, 37.5, 75, 150, 300, and 600 nM.

[0086] Figure 4. Affinity characterization of MEDUSAs prepared with selected FNAPs. a. SPR sensorgrams that characterize the binding kinetics between trimeric SARS- CoV2 S-protein immobilized on the CM3 chip and mono-, bi- and trivalent supramolecular assemblies of selected FNAPs. As controls, assemblies prepared with non-modified (nm) and scrambled non-modified (snm) versions of correspondent binding units were used. The concentrations of injected assemblies were: 9.375, 18.75, 37.5, 75, 150 and 300 nM. b. SPR kinetic parameters for trivalent supramolecular assemblies prepared using side-chain-deficient variants of m2 and m11 sequences. Side chain modifications highlighted in gray play a crucial role in binding to the SARS-CoV2 S-protein. c. Competition ELISA assay indicates distinct binding specificity between FNAPs selected via multivalent and monovalent selection strategies. d. SPR binding kinetic parameters for hetero-trivalent FNAP assemblies prepared with combinations of ml, m2 and m11 FNAPs. e. Competition BLI sensorgrams depict trimeric SARS-CoV2 S-protein binding to dimeric ACE2-Fc, immobilized on the Protein A BLI probes. Assemblies of selected FNAPs were mixed with trimeric SARS-CoV2 S-protein at three increasing assembly concentrations. The decrease in mass transfer to the BLI probe indicates that the compound interferes with the ACE2-S-protein interaction. The reported RBD-binding aptamer (CoV-6C3) was used as a positive control. Assemblies of scrambled nonmodified variants of the selected sequences were used as negative controls. The gradient triangle indicates the increasing concentration of FNAP assembly. All measurements were performed in duplicates, and average signals were plotted with the standard deviation range highlighted.

[0087] Figure 5. Assessment of the effect of the scaffold configuration on the performance of the FNAP assemblies. a. cryoEM 2D class averages for MEDUSA prepared with m2 FNAP. b. A scheme depicting the base multimerization scaffold (2T), its linear (Lin) variant, and the variant with more flexible vertex regions achieved by substituting one of the “T” nucleotides with a hexaethylene glycol spacer (iSp18). Additionally, FNAPs with extended spacers, separating the binding region from the assembly's core, were synthesized for m2 and m11 FNAPs. Representative AFM images of assemblies prepared with 2T and Lin scaffolds and m2 FNAP are shown. c. Violin plots of end-to-end distances between FNAP-attachment points for different assemblies obtained from coarse-grained simulations. d. Violin plots of angle distribution between FNAP-attachment points and a . e. Competition BLI sensorgrams for FNAP assemblies of different scaffold and binding unit compositions. All measurements were performed in duplicates, and average signals were plotted with the standard deviation range highlighted. f. Heatmap of mean areas under the curve (AUC) calculated from the BLI data for each scaffold/FNAP combination.

[0088] Figure 6: OxDNA-generated models of assembly core prepared with 2T, iSp18 and Lin scaffold strands. The iSp18 linker was simulated by 4 abasic nucleotides.

[0089] Figure 7: Schematic depicting the multimerization cssDNA scaffold and binding unit assembly into a trivalent supramolecular structure. The hinge region of the multimerization scaffold can include 1 to 10 of N (N = A,T,G,C,) or 1-10 of X (X = abasic nucleotide, iSp18 spacer, iC3 spacer, iPC spacer, iSp9 spacer). The binding unit anchoring site can include 10-45 of N (N = A,T,G,C,) or 10-45 of X (X = Super G, Super T). Correspondingly, the scaffold-hybridization region can include 10-45 of N (N = A,T,G,C,) or 10-45 of X (X = Super G, Super T). The flexible linker that connects the binding moiety with the scaffold-hybridization region on the binding unit can be made of 1 to 10 of N (N = A,T,G,C,) or 1-10 of X (X = abasic nucleotide, iSp18 spacer, iC3 spacer, iPC spacer, iSp9 spacer). The binding moiety can be a DNA/RNA aptamer, FNAP/HFNAp binder, peptide, protein, or small molecule.

[0090] b. Binding unit anchoring sites on the multimerization cssDNA scaffold and the corresponding scaffold-hybridization regions on the binding units can be programmed to create hetero-trivalent assemblies of the desired composition. In the picture 1 , all anchoring sites have the same sequence, in the picture 2, two anchoring sites have the same sequence but the third have a different (orthogonal) sequence, in the picture 3 all three anchoring sites have different (orthogonal) sequences.

[0091] c. Modified nucleotides and spacers that can be used to modulate the properties of the hinge region and binding unit anchoring sites of the multimerization cssDNA scaffold, as well as the scaffold-hybridization region and flexible linker of the binding unit.

[0092] Figure 8: Assessment of the effect of the scaffold configuration on the performance of the FNAP assemblies, a, A schematic illustrating the sMEDUSA concept according to example 6, which features two conformational states in dynamic equilibrium, with the closed conformation being stabilized upon multivalent cis-interaction with the trimeric spike protein (shown in grey), b, Conformational state diagram for ml and m2 devices as a function of temperature and Mg²⁺ concentration. The condition used in the spike protein binding assays is indicated by an asterisk, c, Fluorescence intensity spectral scans for ml sMEDUSA. d, Fluorescence intensity spectral scans for m2 sMEDUSA. The respective spectral scans of sMEDUSAs in the presence of 250 mM MgCI2 are included as a reference for maximal FRET, e, 670/560 ratio change relative to sMEDUSA in buffer for ml and m2 sMEDUSAs in the presence of increasing concentrations of spike protein, as well as scrambled non-modified (smn) sMEDUSAs at 100 pg/mL spike concentration. Dashed lines represent 670/560 ratio change for ml and m2 sMEDUSAs in the presence of 100 pg/mL BSA. All FRET assays were performed in triplicates (n=3), with the average values plotted and standard deviation ranges highlighted.

[0093] Figure 9: FRET induction by cooperative multivalent binding in cis using guest sMEDUSA and host MEDUSAs of different scaffold configurations, a, A schematic of the experiment showing a set of “guest” sMEDUSAs with “binding units” featuring 80%, 50%, and 10% GC content. Correspondingly, a set of “host” MEDUSAs were prepared with three scaffold variants, each containing “binding units” complementary to those of the sMEDUSAs. b, Native PAGE analysis of “guest” sMEDUSAs and “host” MEDUSAs with different scaffold and “binding unit” compositions, c, Native PAGE analysis of host-guest MEDUSA pairs mixed at a 1:1 molar ratio, d, Representative fluorescence intensity spectra for sMEDUSAs with 80% GC content “binding units” upon binding to corresponding 2T and Lin “host” MEDUSAs. e, The change in the 670/560 fluorescence intensity ratio relative to sMEDUSA in buffer for each host-guest MEDUSA pair.

Examples:

[0094] Note: The term “MEDUSA” (for Multivalent Evolved DNA-based SUpramolecular Assembly) used in the example relates to the obtained construct using the hybrid molecular modality consisting of ad hoc multimerization scaffold with three binding units mirroring the trimeric configuration of class I viral fusion proteins) according to the method of the present invention. It can thus be used to refer to the target-specific multimerization scaffold, multimeric combinatorial library, or nucleic acid-based binders according to the invention.

Example 1 : DNA multimerization scaffold with target-tailored geometry for evolution of FNAPs

[0095] The inventors hypothesized that multimerization of the combinatorial nucleic acid library into trivalent supramolecular assemblies with target-tailored spatial organization could modify the selection regime, leading to the enrichment of sequences with different binding specificities compared to the monovalent selection process, See Fig. 1.

[0096] There are multiple DNA frameworks that allow achieving the desired geometry of ligand presentation. The inventors designed cssDNA and DNA dendrimer-based nanoscaffolds (See Fig. 1). Both scaffolds display comparable pairwise distances between ligand anchoring points (as determined from AFM data). However, cssDNA offers maximal simplicity and is dubbed the optimal option to keep the system as simple as possible and introduce the minimal amount of additional DNA parts into the selection process.

[0097] In conclusion, DNA nanotechnology offers a convenient platform for arranging ligands into defined structural organization and is fully compatible with the SELEX procedure.

Example 2: Multivalent SELEX of trimeric FNAPs: design of template and trinucleotide libraries [0098] See Fig.2, the inventors designed a multivalent SELEX procedure for trimeric

FNAP assemblies.

[0099] Hydrophobic side chains were used to increase the chemical diversity of the library (Fig. 2 c).

[0100] The inventors noticed from the initial studies on the base-functionalized nucleic acids SELEX that side-chain-functionalization density is not particularly important (David Liu 2018, structure-activity relationship studies), with most of the side chains not involved in target binding. Hence, the inventors designed a unique library architecture that sacrifices the density of functionalization in favor of enhanced sequence diversity so that the library would sample a greater space of secondary structures (Fig. 2b).

[0101] Furthermore, the inventors successfully validated the assembly of a trimeric combinatorial library of supramolecular assemblies of synthesized FNAPs using designed multimerization scaffolds with target specific geometry (see Fig. 2 e, f and g). Example 3: Selection of SPIKE-protein-binding FNAPs via multivalent SELEX.

[0102] The inventors compared monovalently pre-focused and multivalent selection strategies according to the present invention (see Fig. 3).

[0103] Three different hit sequences were selected based on their sequence dissimilarity score (MSA data): ml belongs to the family of sequences enriched via monovalent selections, while m2 and m11 are sequences uniquely enriched by multivalent selections (Figure 3d, f).

[0104] The inventors observed that multimerization of FNAP library into the library of trivalent FNAP assemblies drastically changed the types of sequences that were enriched from the same starting FNAP library comparing to monovalent procedure (see Fig. 3 d).

[0105] Moreover, a larger number of sequence-unrelated polymers were able to expand in the library during multivalent selection procedure (see Fig. 3 c, e ,d).

Example 4: Affinity characterization and functional activity of MEDUSAs prepared with selected FNAPs.

[0106] The inventors compared the binding affinity of the selected ml, m2, and m11 sequences (see Example 3) in monovalent, bivalent, and trivalent configurations using the cssDNA multimerization scaffold with three orthogonal FNAP-binding sites (see Fig. 4). [0107] Increasing the valency for the ml assembly did not result in any significant increase in binding strength, while multimerization of the m2 and m11 sequences led to approximately a tenfold increase in binding affinity (Fig. 4 a).

[0108] Trivalent FNAP assemblies prepared with ml, m2, and m11 sequences displayed different binding specificities. Sequences selected using multivalent SELEX possessed a different binding site compared to those selected using monovalent enrichment (Fig 4 c).

[0109] Trivalent assemblies of m2 and m11 FNAPs interfered with the SPIKE-protein- ACE2 interaction using BLI assay, while the ml assembly did not demonstrate any functional activity (Fig 4 e).

Example 5: Assessment of the effect of the scaffold configuration on the performance of the FNAP assemblies.

[0110] By assessing the effect of the scaffold configuration on the performance of the FNAP assemblies (see Fig. 5), the inventors observed that alteration of the original geometry and flexibility of ligand presentation diminishes the properties of m2 and m11 FNAP assemblies. Elongating the FNAP spacer from the original "TCC" to "TiSp18TCC" led to a significant decrease in the ability of such assemblies to compete with ACE2 binding for both m2 and m11, as well as for all tested scaffold assemblies. Additionally, a significant decrease in the performance of the FNAP assembly prepared with the m2 binding unit and Lin scaffold was observed compared to the original m2 assembly with the 2T scaffold.

[0111] Materials and Methods

[0112] Synthesis, purification and characterization of trinucleotide building blocks.

[0113] The synthesis of trinucleotide building blocks has been performed using the automated DNA/RNA synthesizer (Oligomaker ApS, Denmark) on a 200 nmol scale (CPG resin, 100 pcs, TAG Copenhagen AS, OM-11-96-01). The essential materials for custom DNA synthesis were dC(ac) amidite (TAG Copenhagen AS, C113080), dG(dmf) amidite (TAG Copenhagen AS, G115080), dA(bz) amidite (TAG Copenhagen AS, A111080), T amidite (TAG Copenhagen AS, T111080), NHS-Carboxy-dT amidite (Glen Research, 10-1535-XX). For incorporation of 5’ phosphate into trinucleotide sequences, Chemical Phosphorylation Reagent II (Glen Research, 10-1901-XX) has been used. Side-chain modifications were introduced into trinucleotides by coupling C5-carbxy-NHS- modified dT base at the 5’ end of trinucleotide followed by on-resin amidation with the library of primary amines. Two consecutive 6 min couplings were used for all non- standard phosphoramidites. For amidation, 300 equivalents of primary amine in 200 uL of DMSO:Acetonitrile = 10:90 were added onto the column after the coupling and detritylation of Chemical Phosphorylation Reagent II. All trinucleotides were synthesized DMT-off, cleaved and deprotected with 200 uL of 30% aqueous Ammonium Hydroxide overnight at RT. After the cleavage, the column was washed with additional 200 uL of 30% aqueous Ammonium Hydroxide solution. Ammonium Hydroxide was evaporated using vacuum concentrator (SpeedVac™ DNA130) at 65 oC for 1h. The remaining solution was centrifuged at 15,000 g for 10 min to remove the silica from CPG solid support that is typically carried over into the solution during ammonium hydroxide cleavage and deprotection. Trinucleotides were then purified by RP-HPLC. The gradient for functionalized trinucleotides: 0-20% of acetonitrile in 100 mM TEAA, pH 7.0 over 22 min, followed by 20-40% of acetonitrile in 100 mM TEAA, pH 7.0 over 8 min. For purification of non-functionalized trinucleotides, the following gradient has been used: 5- 20% of acetonitrile in 100 mM TEAA, pH 7.0 over 10 min, followed by 20-40% of acetonitrile in 100 mM TEAA, pH 7.0 over 2 min. The fractions of interest were collected and lyophilized. Purified trinucleotide building blocks were analyzed by MALDI-TOF mass spectrometry using 50 mg/mL 3-HPA, 10 mg/mL DAHC matrix. MS-spectra were acquired with a MALDI-TOF/TOF AutoFlex Speed (Bruker) mass spectrometer. The device was operated in the negative ion mode using reflective TOF.

[0114] Template library preparation and characterization

[0115] The template library for ligase-mediated translation was synthesized using standard phosphoramidite synthesis on a 200 nmol scale. Two mixtures of phosphoramidites were prepared in dry acetonitrile at a 200 mM total concentration with the following molar ratios of bases (User's Manual for PE Biosystems Models 392 and 394 DNA/RNA Synthesis):

[0116] Mixture 4 - DMT-dA(Bz):DMT-dT:DMT-dG(dmf):DMT-dC(Ac) = 1.5 : 1 : 1.15 : 1.25

[0117] Mixture 5 - DMT-dA(Bz):DMT-dT:DMT-dG(dmf):DMT-dC(Ac) = 0 : 1 : 1.15 : 1.25

[0118] The template library was synthesized DMT-off, cleaved/deprotected with a 40% aqueous solution of methylamine at 65°C for 1 hour. The library was eluted from controlled pore glass (CPG) resin with 300 pL of 30% acetonitrile, and the crude library was precipitated by adding 1/10 volume of 3 M NaAc and 3 volumes of absolute ethanol. The crude library was pelleted by incubating the solution at -20°C for 20 minutes and centrifugation at 15000 g, 4°C for 30 minutes. The pellet was then reconstituted in 300 pL of RF water, and the concentration was measured using absorbance at 260 nm. Two hundred micrograms of crude DNA were purified using 10% polyacrylamide gel electrophoresis (PAGE) with 8M urea.

[0119] Synthesis of tyramine-O-Pivalate

[0120] Tyramine (3 g, 21 mmol, 1 equiv) was dissolved in 40 mL of DCM:TFA = 1 :1 at 25°C. Pivaloyl chloride (2.69 mL, 21 mmol, 1 equiv) was added dropwise while stirring. The resulting brown mixture was stirred for 12 hours. The reaction mixture was concentrated under reduced pressure and purified by flash column chromatography (Silica, DCM:MeOH = 9:1 to 1 :1) to yield the product.

[0121] Circular ssDNA scaffold strand production and purification

[0122] Circular ssDNA was prepared by stepwise addition of linear 5’-phosphorylated DNA (l-DNA) and the splint strand into the T4 DNA Ligase mixture at low Mg(ll) concentrations. The initial reactions were set up in a 40 pL volume with 2 pM l-DNA, 30 pM splint strand, and 0.05x T4 Ligase buffer. After adding 20 U of T4 DNA Ligase (Fisher Scientific), the reaction was incubated at 22°C. New portions of l-DNA and splint strand were added to the reaction mixture at 30-minute intervals.

[0123] After the last addition, the reaction was left overnight at 22°C. Then, 2 pL of Exonuclease I (NEB) was added to 75 pL of the reaction mixture, and the reaction was incubated at 37°C for 1 hour. Exonuclease I was heat-inactivated at 80°C for 15 minutes. For scaled-up production, eight reactions were performed in parallel in PCR strips.

These reactions were later combined into a 1.5 mL tube, and the volume was reduced to approximately 200 pL using a vacuum concentrator (SpeedVac™ DNA130) at 65°C for 60 minutes. Next, DNA precipitation was carried out by adding 1/10 volume of 3 M NaAc and three volumes of absolute ethanol. The mixture was then incubated at -20°C for 30 minutes and centrifuged at 15,000 g, 4°C for 30 minutes. Subsequently, the pellet obtained from the precipitation was resuspended in a solution containing 30 pL of RF water, 30 pL of 8M urea, 60% glycerol, and 1x TBE loading buffer. This mixture was heated to 90°C for 5 minutes before loading onto the preparative PAGE gel.

[0124] DNA purification

[0125] To purify the circular ssDNA, FNAPs and template library, a 10% polyacrylamide gel electrophoresis (PAGE) with 8M urea was performed. Thirteen milliliters (13 mL) of the solution (10% acrylamide:bisacrylamide = 19:1 , 8M urea, 1x TBE (89 mM Tris-HCI, 89 mM boric acid, 2 mM EDTA, pH 8.0), 12 pL TEMED, and 80 pL of 10% APS) were used to cast one 1.5 mm gel (HasLab). Five microliters of the sample were loaded per well (1 - 15 pg of crude DNA), and the gel was run at 180 V for 1.5 hours (power supply - Labgene scientific, cell - Biometra). The bands were visualized by briefly exposing the gel to 260 nm UV light against the TLC plate. The band of interest was then excised, the gel slice was crushed, and the oligo was extracted using 2 mL of 200 mM NaCI, 40 mM TE buffer, pH 7.5. The first extraction step was carried out at 37°C overnight on a shaker. After the initial extraction, the gel slurry was filtered out using Freeze 'n Squeeze (BioRad, 7326165) spin filters, and the gel was subjected to a second extraction with 2 mL of RF water for 2 hours at 37°C on a shaker. The extracts were collected into a 15 mL tube. Two volumes of dry n-butanol were added to the extract (creating an organic layer on the top), vortexed, and removed three times. After the volume of the extract was reduced to 300 pL, purified oligos were precipitated using 3 M NaAc and EtOH. After centrifugation, the pellets were washed with cold 70% EtOH and air-dried. The pellets were then reconstituted in 75-100 pL of restriction enzyme-free water, and the concentration was measured using NanoDrop based on the absorbance at 260 nm.

[0126] Biotinylated template library preparation

[0127] For production of biotinylated ssDNA template library, eluted FNAP library from the previous selection round was pre-amplified in 8 of 25 pL Q5 DNA polymerase (NEB) PCR reactions using sub-saturation number of cycles according to qPCR. The amplicons were then purified using Monarch PCR & DNA Cleanup Kit (NEB, #T1030) following manufacturer’s protocol. The concentration of library was measured using Nanodrop and preparative PCR was performed on 64 of 25 pL Q5 DNA polymerase PCR reactions’ scale using sub-saturation number of cycles according to qPCR (typical matrix input was 0.3 pmol per reaction, which were amplified in 6 PCR cycles).

[0128] After thermocycle program is complete, reactions were pooled into 4 of 2 mL tubes (400 pL per tube) and amplicons were ethanol-precipitated as described previously. The pellets were resuspended in 10 pL of RF water and 10 pL of Novex TBE- Urea Sample Buffer (2X) (Invitrogen) per pellet yielding a total volume of 80 pL. The amplicons were denatured at 95 oC for 5 min and loaded onto 10% polyacrylamide gel with 8M Urea (10 pL per one well of 1.5 mm gel). The gel was run at 190 V for 1.5 h and the bands were visualized using UV-shadowing against a TLC plate. The lower molecular weight band that corresponds to the biotinylated template was excised and extracted as described before.

[0129] FNAP production via T3 DNA Ligase-mediated translation

[0130] The purified biotinylated library was translated into a functionalized nucleic acid polymer library using T3 DNA Ligase. For each 10 pL reaction, the following components were mixed in a PCR strip: 1 pL of 10x T4 RNA ligase reaction buffer (B0216SVIAL, NEB), 0.75 pL of 20 pM initiation (1.5 equiv) and phosphorylated termination primers (1 .5 equiv), 0.64 pL or 0.77 pL of 5 mM functionalized trinucleotide library mix (5 equiv or 6 equiv per each occurrence of the correspondent codon for 8-mer and 12-mer library, respectively), 1.8 pL or 2.13 pL of 5.4 mM non-functionalized trinucleotide library mix (5 equiv or 6 equiv per each occurrence of the correspondent codon for 8-mer and 12-mer library, respectively), 10 pmol of biotinylated template library. The mixture was subjected to the following thermocycler program: 95°C for 10 seconds, 65°C for 4 minutes, 65°C to 4°C at -0.1 °C/cycle (610 cycles).

[0131] After annealing, 1.2 pL of 10 mM ATP and 0.6 pL of T3 DNA Ligase were added, and reactions were incubated at 4°C for 12 hours and then for 2 hours at 16°C. Alkaline strand separation was performed using Dynabeads™ MyOne™ Streptavidin C1 magnetic beads (ThermoFisher). One microliter of 10 pg/pL stock bead suspension was used per 4 pmol of biotinylated template. The beads were washed three times with one volume of 1X Binding and Washing (B&W) Buffer (5 mM Tris-HCI, pH 7.5, 0.5 mM EDTA, 1 M NaCI) and suspended in one volume of 2X B&W Buffer (10 mM Tris-HCI, pH 7.5, 1 mM EDTA, 2 M NaCI). The translation reactions were then added to the bead suspension and incubated at room temperature for 1.5 hours in a rotary mixer. The supernatant was removed by magnetic separation, and the beads were washed three times with 1X B&W buffer (combined volume of translation mixture + original volume of 1% bead suspension) with a 5-minute incubation in the rotary mixer followed by 1-minute magnetic separation. After the last wash, the beads were resuspended in freshly prepared 20 mM NaOH (original volume of 1% bead suspension), and the FNAP library strand was eluted for 5 minutes in a rotary mixer. The elution was repeated a second time, the fractions were pooled, and 1/40 volumes of 1M HEPES (pH 7.3) were added to neutralize the base. The FNAP strands were purified using the Monarch PCR & DNA Cleanup Kit (NEB, #T1030) by adding 2 volumes of Cleanup Binding Buffer and 6 volumes of absolute ethanol. The column was then washed once with 250 pL of Wash buffer, and the FNAP strand was eluted with 10-20 pL of RF water and PAGE-purified as described previously.

[0132] Beads preparation

[0133] Dynabeads™ MyOne™ Streptavidin C1 magnetic beads were used for affinity selections. Prior to use, the beads were washed three times by adding one volume of 1x PBS followed by magnetic separation for 1 minute in a magnetic stand. Then, the beads were loaded with Twin-StrepTag-tagged stabilized SPIKE trimer (2P mutation) or BG505 SOSIP MD39 with N241 N289 at 20 pg of protein per 1 mg of resin at a 100 pg/mL protein concentration (two volumes to the original 1 % bead stock solution). The suspension was incubated on a rotary mixer at 4°C for 1.5 hours. The supernatant was then discarded, and the beads were washed ten times with two volumes of selection wash buffer (IxDPBS, 1 mM MgCI2, 0.1 mg/mL BSA, 0.005% Tween 20). This procedure yielded an immobilization level of 10 pg of SPIKE-trimer per 1 mg of beads and 8.5 pg of BG505 SOSIP per 1 mg of beads (40-50% of maximal load capacity). Finally, the beads were washed five times for 2 minutes with one volume of SW buffer (IxDPBS, 1 mM MgCI2, 0.1 mg/mL BSA, 0.005% Tween 20).

[0134] Affinity selections

[0135] A library of trivalent assemblies of FNAP was prepared by annealing the purified

FNAP library with the cssDNA scaffold strand in 3.3:1 molar ratio (330 nM FNAP and 100 nM cssDNA) in IxDPBS, 1 mM MgCI2.

[0136] The monovalent FNAP library was prepared at a final concentration of 330 nM in the same buffer and subjected to the same thermocycle program. After folding, both monovalent and multivalent libraries were subjected to negative selection against blank Dynabeads™ MyOne™ Streptavidin 01 magnetic beads by incubating the library with an equal amount of beads as used for the target protein immobilization for 1 hour at room temperature on the rotary mixer. Then, the libraries were incubated for 1 hour with target-coated beads. After incubation, the beads were washed three times with IxDPBS, 1 mM MgCI2, 0.1 mg/mL BSA, 0.005% Tween20 for 5 minutes per washing step on the rotary mixer. After the last wash, binders were eluted by heating the beads suspended in 50 pL of RF water for 10 minutes at 95 °C.

[0137] Production of selected functionalized nucleic acids

[0138] All selected hit sequences were synthesized on a 200 nmol or 40 nmol scale in 5 (for 8-mer sequences) or 7 (for 12-mer sequences) consecutive runs on the automated DNA/RNA synthesizer (Oligomaker ApS, Denmark). Each run ended with the addition of NHS-Carboxy-dT amidite (Glen Research, 10-1535-XX), followed by on-resin amidation with the corresponding primary amine.

[0139] Competition ELISA assay

[0140] Wells of the NuncTM MaxiSorpTM flat-bottom half-area 96-well plate (Thermo Fisher Scientific) were coated with 30 pL of 20 pg/mL SPIKE-protein trimer solutions in 1xPBS at 4°C overnight. The coating solution was discarded by decanting the plate, and the wells were washed three times with 100 pL of IxDPBS, 1 mM MgCI2, 0.1 mg/mL BSA, 0.005% Tween20. The wells were then blocked with 30 pL of 1% BSA in IxDPBS for 2 hours at 4°C. Afterward, the blocking solution was discarded by decanting the plate. To the blocked plate, 25 pL of 1000 nM unlabelled assembly were added followed by 25 pL of 100 nM Fluorescein-labelled assembly in IxDPBS, 1mM MgCI2. As a negative control, equivalent amount of salmon sperm DNA was added in place of unlabelled assembly. The plate was centrifuged at 2000 rpm for 2 min and incubated for 1h at RT. The solutions were removed by decanting the plate and the wells were washed three times by putting 100 pL of IxDPBS, 1mM MgCI2, 0.1 mg/mL BSA, 0.005% Tween20, letting in the well for 5 sec and removing by the multichannel. After the last wash, the plate was tapped firmly on the paper towel to remove the residual wash buffer. Then, 50 pL of RF MiQ were added, the plate was centrifuged at 2000 rpm for 2 min and placed into the pre-heated oven at 90°C for 5 min. The plate then was cooled on ice for 5 min, centrifuged and imaged with extended gain and 6.25 mm read hight (BioTek Cytation 5 Plate Imager).

[0141] Competition BLI assays

[0142] All experiments were performed at 25 oC on Gator BLI system. Running buffer was IxDPBS supplemented with 1 mM MgCI2. For competition assays, dimeric ACE2- Fc was diluted to 20 pg/mL and captured on the Protein A probes (GatorBio, PL168- 160001) using 600 sec loading time. After loading and 200 sec equilibration, loaded probes were dipped into the solution of pre-mixed SPIKE-protein at 50 nM with the MEDUSA or monomeric FNAP The highest tested concentrations for trivalent MEDUSA and monovalent FNAP were 500 nM and 1500 nM, respectively. Association phase was 300 sec followed by 600 sec of dissociation phase.

[0143] Surface Plasmon Resonance (SPR) assays

[0144] All SPR assays were performed at 25°C on the Biacore 8K+ system.

Approximately 1000 RU of trimeric SPIKE-protein was immobilized on the CM3 chip (Cytiva, BR-1005-36) using the EDC/NHS kit (Cytiva, BR100050). For this, a 10 pg/mL SPIKE-protein solution in 10 mM NaAc buffer pH 5.5 was used with a contact time of 290 seconds and a flow rate of 5 pL/min. Single-cycle kinetics runs were performed using 120 seconds for association and 300 seconds for dissociation phases. IxDPBS supplemented with 1 mM MgCI2 was used as the running buffer.

[0145] Mass-Photometry measurements [0146] To evaluate the binding capacity of the libraries of trivalent assemblies at different selection rounds, Refelyn Two MP mass photometer has been used. Firstly, a calibration curve was obtained using BSA solution. For this, 5 pL of 10 pg/mL of BSA in 1xPBS were introduced into 15 pL of 1xPBS previously deposited into one of the wells of a silicon gasket. For library measurements, 1.5 pL of 200 nM library of assemblies in 1xPBS, 1 mM MgCI2 were incubated with 3 uL of 62 ug/mL SPIKE protein in 1xPBS for 1 h. Then, 4.5 pL of the mixture were introduced into 15 pL of 1xPBS that were previously deposited into one of the wells of the silicon gasket. All single-molecule light scattering events were recorded for 180 sec.

[0147] High-throughput sequencing and data analysis

[0148] For sequencing, a sample of 30 fmol of FNAP library was taken from the elution fractions of selection rounds 3, 6, and 7 for both monovalent and multivalent selection strategies. Two consecutive PCR amplifications catalyzed by Q5 DNA polymerase were performed to install Illumina adaptors (perl) and index sequences (pcr2). The first PCR was performed in 2 reactions of 25 pL each using Q5 DNA polymerase. After 10 PCR cycles, the reactions were pooled, and the amplicons were purified using the Monarch PCR & DNA Cleanup Kit. Subsequently, the amplicons were purified using 8% native TBE PAGE. The second PCR was set up in a 50 pL reaction using Q5 DNA polymerase.

[0149] The indexed libraries were then purified using the Monarch PCR & DNA Cleanup Kit followed by 6% native TBE PAGE.

[0150] The concentrations were measured using the Qubit (Invitrogen™ Qubit™ Flex), and the indexed libraries were submitted for sequencing using the NovaSeq (multiplexed r3 E medusa, r6 E medusa, r7 E medusa, r3 E mono, r6 E focused medusa, and r7 E focused medusa; 20 million reads per library) and MiSeq (15 million reads for multiplexed r7 E medusa + r7 E focused medusa) systems.

[0151] Then, the homemade python script was used to calculate the frequencies of the sequences and cumulative Q-score for each sequence.

[0152] The top 100 sequences from the resulting file, along with their sequence counts, were used for multiple sequence alignment (MSA) using ClustalOmega, and the guide tree was plotted using FigTree v1.4.4 software. The file containing the sequence-count- weighted top 100 sequences from r7 E focused medusa was used to produce the consensus sequence for the reads detected in this library.

[0153] DNA secondary structure prediction [0154] DNA secondary structure prediction was performed on the Nupack (https://nupack.org) web server using parameters that correspond to the binding and selection conditions: 25 oC folding temperature, 0.15 M NaCI and 0.001 M MgCI2. Example 6: Multivalent target-induced sensors

[0155] The inventors explored if multivalent target-induced sensors could be generated from a method according to the present invention.

[0156] The inventors hypothesized that if a multivalent assembly has multiple discrete conformational states, the conformation that facilitates the most effective ligand positioning for multivalent binding should be favoured upon the multivalent engagement of its binding units with the multimeric target protein. Crucially, this only holds true if the binding units can engage in synergistic binding with the multimeric target in cis. To test this hypothesis, the inventors devised a switchable MEDUSA variant (sMEDUSA) that can adopt two discrete conformational states: a sub-optimal open state for multivalent cis-binding, and a more optimal conformationally constrained closed state. In the open state, sMEDUSA resembles the Lin structure, while in the closed state it is akin to the original MEDUSA assembly (see previous examples) with the base 2T scaffold. The structural basis for the switchable behaviour of sMEDUSA arises from its scaffold, which contains short (6 nt) complementary regions at the 5’ and 3’ termini flanking three FNAP- binding sites. Upon the hybridization of this scaffold with three binding units, a molecular switchable system with dynamic equilibrium between open and closed states is obtained. By introducing a FRET pair at the root of the dynamic hairpin, the proportion of closed-state molecules can be monitored in bulk by ratiometric readout of fluorescence intensities at 670 nm (indicative of the closed conformation) and 560 nm (indicative of the open conformation) (Figure 8a). The inventors successfully validated the system by modulating the equilibrium between open and closed states using temperature and Mg2+ concentration. Both a decrease in temperature and an increase in Mg2+ concentration led to the stabilization of the hairpin and a subsequent increase in the closed-state population (Figure 8b).

[0157] To explore the dynamic range of this system and to test whether the same shift in the equilibrium between the open and closed conformations of sMEDUSA could be achieved by cooperative multivalent binding in cis, the inventors conducted a sandbox experiment with sMEDUSAs featuring short DNA sequences with different GC content in place of FNAP binding units (Figure 9a; top). On the opposite side, the spike protein target was replaced with 2T, Flex, and Lin “host” MEDUSAs displaying complementary DNA sequences (Figure 9a; bottom). We observed that the conformationally constrained 2T MEDUSA target was the most effective in mediating the FRET increase of the sMEDUSAs, and the increase in GC content of the binding units correlated with the increase in FRET (Figure 9e). Next, the inventors tested whether the conformational dynamics of sMEDUSA could be altered by the spike protein. We compared ml and m2 sMEDUSAs, as well as sMEDUSAs with scrambled non-modified control sequences, in the FRET assay with increasing concentrations of spike protein. Excitingly, the addition of the protein to the m2 sMEDUSA led to a more substantial increase in FRET compared to ml sMEDUSA (Figure 8c,d,e), indicating effective multivalent cis-binding to the spike protein and consequently a shift of the conformational state equilibrium towards the closed conformation. This data shows the potential of the present invention in molecular sensing applications.

[0158] Methods:

[0159] FRET assays

The sMEDUSAs were folded in 1x DPBS with 1 mM MgCI₂ at a concentration of 0.2 pM. Upon completion of the thermocycler program, a working solution of 0.1 pM assemblies was prepared by diluting the folding mixture twofold with RF water. Dilutions of Wuhan WT spike protein (ExcellGene) were prepared from a 1 mg/mL stock solution in 1x DPBS with 1 mM MgCI₂. The MgCI₂ concentration was adjusted proportionally to the volume of spike protein stock solution using a 10 mM MgCI₂ solution. The assay was performed in a 96-well microplate (PS, F-bottom with chimney well, pCLEAR®, black, non-binding, Greiner Bio-One) by mixing 65 pL of 100 nM sMEDUSA with 65 pL of spike protein solutions at final concentrations of 10, 20, 40, and 100 pg/mL. The mixture was incubated at room temperature for 20 minutes, followed by a 10-minute temperature equilibration at 37°C. Spectral scans were acquired at 37°C using a BioTek Cytation 5 Plate Imager under the following parameters: excitation at 540/10 nm, emission range from 560/10 to 700 nm, emission step size of 4 nm, gain set to 100, optics positioned at the top, and a read height of 7 mm.

Claims

1. Method for identifying nucleic acid-based binders having binding affinities for a user- defined target, where said target provides a target-tailored geometry-and-valency selection constraint present during a Systematic Evolution of Ligands by Exponential enrichment (SELEX) procedure, the method comprising:

Preparation of a random or combinatorial library of nucleic acid polymers; Preparation of a target-specific multimerization scaffold where geometry, dimension and valency match the desired target binding-site(s);

Hybridizing said random or combinatorial library with said scaffold to create a multimeric combinatorial library; and

Performing iterative rounds of affinity-based selections according to the SELEX procedure between said multimeric combinatorial library of nucleic acid polymers and the target to allow to enrich for nucleic acid-based binders capable of engaging in multivalent interactions with the target;

Determination of the selected binder sequences by NGS (Next Generation Sequencing).

2. The method according to claim 1, wherein the type of nucleic acid polymers selected for the random or combinatorial library includes DNA, RNA, base- or backbone-modified DNA, base- or backbone-modified RNA, peptide nucleic acids (PNAs), or a combination thereof.

3. The method according to any one of claims 1 to 2, wherein the target-specific multimerization scaffold is made of DNA (cyclic single stranded DNA or linear single stranded DNA), RNA (cyclic single stranded RNA or linear single stranded RNA) or XNA including: LNA, PNA, TNA, FANA, GNA, CeNA and HNA, peptides, PEG dendrimers, synthetic polymers, or combinations thereof.

4. The method according to any one of claims 1 to 3, wherein the target is selected from the group consisting of proteins, enzymes, receptors, transcription factors, structural proteins, transport proteins, DNA, RNA (including mRNA, rRNA, tRNA, and miRNA), multimeric protein complexes, protein-nucleic acid complexes, cell surface complexes, hormones, metabolites, drugs, viruses, bacteria, parasites, fungi, cancer cells, stem cells, and specific tissue types, and combination thereof.

5. The method according to any one of claims 1 to 4, wherein the identification of the target-specific multimerization scaffold's geometry and valency is based on artificial intelligence predictions, computational modeling, or on available crystal structure data.

6. The method according to any one of claims 1 to 5, wherein the target-specific multimerization scaffold is designed in a specific geometric arrangement selected in the group consisting of points, lines, rays, line segments, triangles, which can be classified as equilateral, isosceles, scalene, right, acute, or obtuse, quadrilaterals that encompass squares, rectangles, parallelograms, rhombuses, trapezoids, and kites, other polygons including pentagons, hexagons, heptagons, octagons, nonagons, decagons, hendecagons, and dodecagons.

7. The method according to any one of claims 1 to 6, wherein the valency of the targetspecific multimerization scaffold is defined to be between 2 and 10, preferably between 2 and 7.

8. The method according to any one of claims 1 to 7, wherein the SELEX procedure is selected from the group consisting of Counter-SELEX, Cell-SELEX, or Photo-SELEX.

9. The method according to any one of claims 1 to 8, wherein the target is a trimeric viral capsid protein, preferably the SARS-CoV2 Spike trimer or the HIV Env complex.

10. The method according to any one of claims 1 to 9, wherein the library is constructed from FNAPs.

11. The method according to any one of claims 1 to 10, wherein the scaffold is a triangle with vertices of between 5 to 15 nm, preferably 7 to 12 nm.

12. Use of the Systematic Evolution of Ligands by Exponential enrichment (SELEX) procedure to identify nucleic acid-based binders having binding affinities for a user- defined target, wherein a random or combinatorial library of nucleic acid polymers is hybridized with a target-specific multimerization scaffold where geometry, dimension and valency match the desired target binding-site(s) before performing at least two iterative rounds of affinity- based selections according to the SELEX procedure.

13. Nucleic acid-based binders capable of engaging in multivalent interactions with a target, wherein said nucleic acid-based binders are selected according to the method according to claims 1 to 11.

14. Multivalent binding complex comprising a cyclic single-stranded DNA (cssDNA) multimerization scaffold and three binding units that sequence-specifically hybridize to said multimerization scaffold, preferably for use in the method according to claims 1 to 11.