Proximity-inducing compounds and methods

Abstract

The present invention relates to novel bifunctional molecules capable of undergoing bioorthogonal reactions and inducing proximity between two proteins, methods of inducing proximity between two proteins employing said novel molecules and bioorthogonal reactions and genetic code expansion, methods of evaluating protein-protein proximity interactions employing the novel bifunctional molecules, medical uses of the novel molecules, and methods of treatment of a disease involving post-translational modifications of a protein employing the novel bifunctional molecules.

Description

[0001] Novel proximity-inducing compounds and methods

[0002] Field

[0003] The present invention relates to novel molecules capable of inducing proximity between two proteins, methods of inducing proximity between two proteins employing said novel molecules, methods of evaluating protein-protein proximity interactions, medical uses of the novel molecules, and methods of treatment of a disease involving post-translational modifications of a protein.

[0004] Background

[0005] Induced proximity pharmacology is a promising therapeutic modality1. Its premise involves bringing two proteins together with a small molecule to elicit a therapeutic effect. The approach is exemplified by targeted protein degradation (TPD) where a ubiquitin E3 ligase enzyme causes the degradation of a neosubstrate. E3 ligases are writer enzymes that operate within an E1-E2-E3 enzyme cascade and are key components of the ubiquitin proteasome system (UPS)2. E3s catalyze the covalent transfer of the small protein modifier ubiquitin to substrates thereby marking them for degradation by a giant protease complex termed the proteasome. The degrader molecules employed in TPD can be bifunctional composed of distinct targeting ligands or smaller molecules that induce favourable protein-protein interactions with so-called molecular glue activity1.

[0006] For the -650 E3 ligases there is currently a small handful of specific ligands that have been elaborated into heterobifunctional degraders34. Those in clinical development typically leverage ligands for the E3s VHL and CRBN5. However, it is not conceivable these E3s will support the degradation of any target protein. Furthermore, VHL and CRBN are expressed throughout the body so there would be side effects from degrading neosubstrates with an essential role in healthy tissues. Hence, there is a need for ligands that bind new E3s including those with restricted tissue expression / activity profiles6.

[0007] Systematic exploration of an optimum TPD configuration for a single target of interest requires the development of specific ligands for all tractable E3 sites as each will furnish a distinct ternary complex that must be assessed for functionality. As there are -650 E3 ligases that can be selected from, each with multiple putative ligandable sites, the development of thousands of bifunctional ligands is necessary for comprehensive assessment. A parallel set of bifunctional compounds would then be needed for a new target. Furthermore, induced proximity involving other writer and eraser proteins also has great therapeutic potential7This chemistry burden prevents the potential of induced proximity research from being systematically assessed.

[0008] To help address these challenges, technologies have been developed that allow TPD, and other forms of induced proximity, to be assessed in the absence of a specific ligand. A platform known as Affinity-directed PRotein Missile (AdPROM) involves appending complementary protein tags (e.g. GFP and antiGFP-nanobody) to an E3 and a neosubstrate8. The advantages are that molecular biology is leveraged to achieve specificity and the ability to explore, in principle, any enzyme-neosubstrate combination. A caveat is the ternary complex is dictated by the binding of the protein tags, which are typically appended to the protein termini. Protein tags can also affect enzyme activity, leading to false negatives. Additionally, an observed phenotype may arise from a ternary complex that is strictly dependent on the protein tags and therefore not pharmacologically reproducible in a native context. Furthermore, as protein proximity is not pharmacologically induced, it precludes dose-response, pharmacokinetic and pharmacodynamic analyses. As a complement, systems involving conditional protein tags (HALOtag, dTAG, BromoTag) have been developed that come into proximity with their partner tag in the presence of a bifunctional probe molecule9-11. The size of these protein tags can cause perturbation and to mitigate this, smaller protein tags have been developed12. These, and others, can also be induced with small glue molecules allowing their improved bioavailability to be taken advantage of. However, these conditional systems still rely on terminal protein tags so the problems relating to restricted ternary complex formation and their potential effect on protein activity remain. To assess ligand tethering to internal protein sites COFFEE (covalent functionalization followed by E3 electroporation) technology has been developed13. Here a cysteine residue within a protein is modified in vitro with a bifunctional probe containing an electrophile. The adducted protein is then electroporated into cells. Whilst this enables the assessment of multiple pharmacologically relevant ternary complexes, the protein must be expressed and purified which may not always be feasible. Furthermore, to directly assess a specific target site, the protein must contain a single cysteine. Whilst the promise of induced proximity therapy is huge, new approaches to assess its potential and prioritise strategies in a pharmacologically informative and general manner are lacking.

[0009] There is a need for new technology which enables conditional tag-free induction of protein proximity. There is also a need for new technology which enables the comprehensive assessment of induced proximity across a spectrum of protein pairs to facilitate the broad interrogation of the therapeutic potential of induced proximity, and / or to help assign protein function thereby having utility in basic research. Summary

[0010] In a first aspect there is provided a proximity-inducing molecule of formula (I):

[0011] A-L-D

[0012] (I) for inducing proximity between a first protein (Pi) and a target protein (Pt), wherein:

[0013] A is a moiety capable of binding to the target protein or to a tag located on the target protein;

[0014] L is a linker; and

[0015] D is a moiety capable of undergoing a bioorthogonal reaction with a bioorthogonal reactive group.

[0016] Proximity between the first protein (Pi) and the target protein (Pt) may be induced in vivo, ex vivo or in vitro. Within the context of this disclosure the compounds are preferably configured to induced proximity in vitro.

[0017] The bioorthogonal reaction may be selected from: native chemical ligation, Staudinger ligation, copper-catalysed azide-alkyne cycloaddition, Copper-Free Azide-Alkyne Cycloaddition, strain-promoted [3 + 2] reaction, Inverse Electron Demand Diels-Alder (IEDDA) reaction, tetrazole ligation, oxime ligation, isocyanide click reaction, metal-catalysed coupling reaction, hydrazone ligation, photoinducible bioorthogonal reactions, optionally wherein the bioorthogonal reaction is an Inverse Electron Demand Diels-Alder (IEDDA) reaction.

[0018] A may be a ligand for the target protein. For example, the proximity-inducing molecule may comprise a known ligand for a target protein. The ligand A may be capable of binding the target protein. Binding may occur via a covalent bond or a non-covalent interaction.

[0019] A may be a ligand configured to bind to a target protein or polypeptide which is to be degraded by a ubiquitin ligase, optionally wherein A is linked to L through an amide bond.

[0020] A may be one or more of:

[0021] (i) a chemical moiety configured to bind to a protein within the bromo- and Extra-terminal (BET) family of proteins; or

[0022] (ii) a chemical moiety configured to bind to a protein within the bromo- and Extra-terminal (BET) family of proteins independently selected from: BRD2, BRD3 and BRD4; or (iii) a chemical moiety which is configured to selectively induce degradation of the BRD4 protein within the bromo- and Extra-terminal (BET) family of proteins;

[0023] (iv) an E3 ligase, optionally wherein the E3 ligase is selected from VHL and CRBN.

[0024] Optionally, A may be linked to L through an amide bond.

[0025] A may be a derivative of JQ1. Optionally, A may be linked to L through an amide bond. A may

[0026] A may be a moiety capable of binding to a protein tag. For example, in embodiments in which no suitable ligands for a target protein are available or known, the target protein may be tagged with a protein tag (e.g. HALOtag, bromoTag, dTAG). A may be a moiety capable of binding said protein tag.

[0027] By contacting the proximity-binding ligand with the first protein a complex comprising a covalent bond between the first protein and the proximity-inducing molecule may be formed. If this complex is in the vicinity of a target protein (e.g. inside a cell or in an extra-cellular environment), the moiety A may bind to the target protein (directly if A is a ligand for the target protein, or indirectly if A is a moiety capable of binding to a protein tag present on the target protein). The target protein and the first protein may then be brought into close proximity via the proximity-inducing molecule of formula (I) and a protein-protein interaction may take place.

[0028] D may be a moiety capable of undergoing a bioorthogonal reaction with a bioorthogonal reactive group. The bioorthogonal reactive group may be located in an unnatural amino acid (UAA) present on the first protein.

[0029] D may be a moiety comprising a diene, a dienophile or a heterodienophile. The diene, dienophile or heterodienophile may be capable of reacting in an Inverse Electron Demand Diels-Alder (I EDDA) reaction.

[0030] When D is a moiety comprising a diene, it may be configured to react with a dienophile present in the first protein. The diene may be any diene capable of reacting in a I EDDA reaction. D may be a moiety comprising a diene configured to react with a dienophile located in an unnatural amino acid (UAA) present in the first protein. D may be a moiety comprising an electron-poor diene. D may be a moiety comprising a nitrogen-containing diene. D may be a moiety comprising a diene present in an electron-deficient heterocycle. D may be a moiety comprising a cyclic azine. D may be a moiety comprising a diene selected from a diazine, a triazine, and a tetrazine. Preferably, D may be a moiety comprising a tetrazine. D may be , methyl or ethyl.

[0031] When D is a moiety comprising a dienophile or a heterodienophile, it may be configured to react with a diene present in the first protein. D may be a moiety comprising a dienophile or a heterodienophile configured to react with a diene located in an unnatural amino acid (UAA) present in the first protein. The dienophile or heterodienophile may be any suitable dienophile or heterodienophile capable of reacting in an Inverse Electron Demand Diels-Alder (I EDDA) reaction. D may be a moiety comprising an alkene or an alkyne. D may be a moiety comprising a strained dienophile or a strained heterodienophile. D may be a moiety comprising a strained olefin or an alkyne. D may be a moiety comprising a group selected from: cyclopropene, transcyclooctene (TCO), norbornene, and bicyclo [6.1.0] nonyne. D may be a moiety comprising a bicyclo [6.1.0] nonyne group.

[0032] D may be selected from transcyclooctene

[0033] R is H, methyl or ethyl.

[0034] L may be a linker configured to separate moieties A and D sufficiently to allow protein-protein interaction between the first protein and the target protein. L may be a linker comprising an aliphatic chain or a PEG group. L may be a linker of formula -(CH2-CH2-O)n-, wherein n is selected from 1 to 20, preferably wherein n is selected from 2 to 4, more preferably wherein n is selected from 2 and 3. In some embodiments, the compounds of Formula (I) may be selected from:

[0035] The compounds may also encompass a pharmaceutically acceptable, salt, enantiomer, stereoisomer, hydrate, solvate, or polymorph of the compounds of Formula (I) of the first aspect. Optionally, the compounds may encompass a pharmaceutically acceptable, salt, hydrate, or solvate of the compounds of Formula (I) of the first aspect.

[0036] In a second aspect there is provided a method of inducing proximity between a first protein (Pi) and a target protein (Pt), the method comprising: a) Preparing a recombinant protein (Pr) by genetic code expansion to encode into the first protein (Pi) an unnatural amino acid (UAA) comprising a bioorthogonal reactive group capable of reacting with a proximity-inducing molecule of formula (I) : A-L-D according to the first aspect; b) Forming a modified protein (Pm) by contacting the recombinant protein (Pr) with the proximity-inducing molecule to form a covalent bond between the unnatural amino acid (UAA) of the recombinant protein (Pr) and moiety D of the proximity-inducing molecule in a bioorthogonal reaction; c) Forming a ternary complex (Tc) by allowing the modified protein (Pm) to bind to the target protein (Pt).

[0037] Any one of steps a), b) and / or c) independently may take place in vivo, in vitro and ex vivo. Preferably, the method is performed in vitro or ex vivo. When the method is performed in vivo, it may be performed in a mammalian, such as a human or an animal. Preferably, the method may be performed in an animal, for example for an animal model or animal studies.

[0038] The target protein (Pt) may be present in a cell line into which the recombinant protein (Pr) is encoded and to which the proximity-inducing molecule of formula(l) is administered. In this case the method may take place inside a cell. The method may take place in vivo. The method may also take place in vitro. The method may take place in vivo by administering the compound of Formula (I) and recombinant protein (Pr) to a subject in need thereof. The subject in need thereof may be a mammalian. The subject in need thereof may be a human or an animal.

[0039] The bioorthogonal reaction may be selected from: native chemical ligation, Staudinger ligation, copper-catalysed azide-alkyne cycloaddition, Copper-Free Azide-Alkyne Cycloaddition, strain-promoted [3 + 2] reaction, Inverse Electron Demand Diels-Alder (IEDDA) reaction, tetrazole ligation, oxime ligation, isocyanide click reaction, metal-catalysed coupling reaction, hydrazone ligation, photoinducible bioorthogonal reactions, optionally wherein the bioorthogonal reaction is an Inverse Electron Demand Diels-Alder (IEDDA) reaction.

[0040] The Unnatural Amino Acid (UAA) of the recombinant protein (Pr) may comprise a group capable of reacting in an Inverse Electron Demand Diels-Alder (IEDDA) reaction. The unnatural amino acid (UAA) may be a non-catalytic amino acid present on the first protein. Moiety D of the proximity-inducing molecule may comprise a group capable of reacting in an Inverse Electron Demand Diels-Alder (IEDDA) reaction. The bioorthogonal reaction between the UAA of the recombinant protein (Pr) and moiety D of the proximity- inducing molecule may form a covalent bond between the proximity-inducing molecule and the UAA of the recombinant protein (Pr). The bioorthogonal reaction between the recombinant protein (Pr) and the proximity-inducing molecule may result in a cycloaddition between the UAA of the recombinant protein (Pr) and moiety D of the proximity-inducing molecule.

[0041] Moiety D of the proximity-inducing molecule may comprise a diene capable of reacting with a dienophile present in the recombinant protein (Pr) in an IEDDA reaction. D may be a moiety comprising a diene present in an electron-deficient heterocycle. D may be a moiety comprising a nitrogen-containing diene. D may be a moiety comprising a cyclic azine. The cyclic azine may be selected from a diazine, a triazine, and a tetrazine. The diene of D may be a tetrazine.

[0042] The unnatural amino acid (UAA) of the recombinant protein (Pr) may comprise a dienophile capable of reacting with a diene in moiety D of the proximity-inducing molecule in a IEDDA reaction. The unnatural amino acid (UAA) may be a non-catalytic amino acid present on the first protein. The UAA may comprise a moiety comprising an alkene or an alkyne. The UAA may be a moiety comprising a strained dienophile or a strained heterodienophile. The UAA may comprise a moiety comprising a strained olefin or an alkyne. The UAA may comprise a moiety comprising a group selected from: cyclopropene, transcyclooctene (TCO), norbornene, and bicyclo [6.1.0] nonyne. The UAA may comprise a moiety comprising a bicyclo [6.1.0] nonyne group. Alternatively, moiety D of the proximity-inducing molecule may comprise a dienophile capable of reacting with a diene present in the recombinant protein (Pr) in a I EDDA reaction. D may be a moiety comprising an alkene or an alkyne. D may be a moiety comprising a strained dienophile or a strained heterodienophile. D may be a moiety comprising a group selected from: cyclopropene, transcyclooctene (TCO), norbornene, and bicyclo [6.1.0] nonyne. D may be a moiety comprising a bicyclo [6.1.0] nonyne group.

[0043] The unnatural amino acid (UAA) of the recombinant protein (Pr) may comprise a diene configured to react with a dienophile in moiety D of the proximity-inducing molecule in a I EDDA reaction. The UAA of the recombinant protein (Pr) may comprise a moiety comprising a diene present in an electron-deficient heterocycle. The UAA of the recombinant protein (Pr) may comprise a moiety comprising a nitrogen-containing diene. The UAA of the recombinant protein (Pr) may comprise a moiety comprising a cyclic azine. The cyclic azine may be selected from a diazine, a triazine, and a tetrazine. The diene of the UAA of the recombinant protein (Pr) may be a tetrazine.

[0044] L may be a linker configured to separate moieties A and D sufficiently to allow protein-protein interaction between the first protein (Pi) and the target protein (Pt). The length of L may be customised to provide an ideal distance between the first protein (Pi) and the target protein (Pt) for a protein-protein interaction event to take place. L may be a linker comprising an aliphatic chain or a PEG group. L may be a linker of formula -(CH2-CH2-O)n-, wherein n is selected from 1 to 20, preferably wherein n is selected from 2 to 4, more preferably wherein n is selected from 2 and 3.

[0045] A may be a moiety capable of binding to the target protein (Pt). A may be a moiety capable of binding to a tag located on the target protein (Pt). Binding A to the target protein (directly or indirectly via a protein tag) may involve forming a covalent bond or a non-covalent interaction.

[0046] The first protein may be an enzyme. The first protein may be an enzyme involved in a post- translational modification of a protein (e.g. the target protein). The first protein may be an E3 ligase. The E3 ligase may be selected from VHL and CRBN. The first protein may be an alternative type of protein capable of proximity-induced destabilization such as an E2 conjugating enzyme.

[0047] The target protein (Pt) or the first protein may be any protein subject to a post-translational modification (PTM). The target protein (Pt) may be selected from:

[0048] (i) a protein within the bromo- and Extra-terminal (BET) family of proteins; or (ii) a protein of the bromo- and Extra-terminal (BET) family of proteins independently selected from: BRD2, BRD3 and BRD4; and

[0049] (iii) a BRD4 protein within the bromo- and Extra-terminal (BET) family of proteins.

[0050] The target protein (Pt) may be an enzyme. The target protein may be an enzyme involved in a post-translational modification of a protein (e.g. the target protein). The first protein may be an E3 ligase. The E3 ligase may be selected from VHL and CRBN.

[0051] The first protein may be any protein subject to a post-translational modification (PTM). The first protein (Pi) may be selected from:

[0052] (i) a protein within the bromo- and Extra-terminal (BET) family of proteins; or

[0053] (ii) a protein of the bromo- and Extra-terminal (BET) family of proteins independently selected from: BRD2, BRD3 and BRD4; and

[0054] (iii) a BRD4 protein within the bromo- and Extra-terminal (BET) family of proteins.

[0055] The method may further comprise attaching a protein tag to the target protein (Pt) prior to step c). In these embodiments, moiety A in the proximity-inducing molecule may be a ligand for the protein tag. The protein tag may be any suitable protein tag, such as a HALOtag, dTag or a bromoTag. In these embodiments, moiety A of the proximity-inducing molecule may be a ligand for the protein tag (e.g. a ligand for a HALOtag. dTag or a bromoTag).

[0056] In a third aspect there is provided a method of evaluating an induced proximity “interaction” between a first protein (Pi) and a target protein (Pt) in a cell, the method comprising: a) Providing a cell line; b) Genetically engineering the cell line to enable genetic code expansion to encode an unnatural amino acid (UAA) comprising a bioorthogonal reactive group into the first protein (Pi); c) Allowing the cell line to express a recombinant protein (Pr) of the first protein (Pi) comprising the unnatural amino acid (UAA); d) Providing to the cell line a proximity-inducing molecule according to the first aspect to form a ternary complex between the recombinant protein (Pr), the proximity-inducing molecule and the target protein (Pt); e) Testing the interaction between the recombinant protein (Pr) and the target protein (Pt) facilitated by the close proximity between them in the ternary complex. Genetically engineering the cell line by genetic code expansion to encode an unnatural amino acid (UAA) may involve transfecting the cell line, or stable integration, with an orthogonal tRNA-synthetase / tRNAstop codon pair (e.g. a tRNA-synthetase / tRNAcuA pair) configured to incorporate the unnatural amino acid (UAA) into the first protein (Pi) to form a recombinant protein (Pr) in response to an amber TAG stop codon. The tRNA-synthetase / tRNAcuA pair may be a mutant pyrrolysl-tRNA synthetase / tRNAcuA pair.

[0057] The tRNA-synthetase / tRNAstop codon pair may be supplied in an expression vector. The expression vector may be pBCNK-PylSTmam-

[0058] The method may further comprise cloning amber codon protein mutants with an epitope tag (e.g. a HA tag, a c-Myc tag, DYKDDDDK (FLAG®) tag, a GFP tag, V5 tag, a His Tag, a FITC tag, an APC tag and the like) into a separate expression vector. The epitope tag may enable ELISA, Western Blot, Surface plasmon resonance, immunoprecipitation, protein purification, imaging and the like. The epitope tag my be introduced at the C terminus of the amber codon protein. The separate expression vector may be the PiggyBac expression vector. The HA tag may enable detection of the recombinant protein (Pr) by Western Blotting.

[0059] The unnatural amino acid (UAA) may comprise a moiety comprising a dienophile capable of reacting with a diene in an Inverse Electron Demand Diels-Alder (IEDDA) reaction. The diene may be present in the proximity-inducing molecule of formula (I) (e.g. in moiety D). The unnatural amino acid (UAA) may comprise a moiety comprising an alkene or an alkyne. The unnatural amino acid (UAA) may comprise dienophile or a heterodienophile. The unnatural amino acid (UAA) may comprise a moiety comprising a strained dienophile or a strained heterodienophile. The unnatural amino acid (UAA) may comprise a moiety comprising a group selected from: cyclopropene, transcyclooctene TCO, norbornene, and bicyclo [6.1.0] nonyne- L-lysine (BCNK). The unnatural amino acid (UAA) may be selected from: cyclopropene- L- lysine, norbornene-L-lysine, and bicyclo [6.1.0] nonyne-L-lysine (BCNK).

[0060] The unnatural amino acid (UAA) may comprise a moiety comprising a diene capable of reacting with a dienophile in a IEDDA reaction. The dienophile may be present in the proximityinducing molecule of formula (I) (e.g. in moiety D). The unnatural amino acid (UAA) may comprise a moiety comprising a diene present in an electron-deficient heterocycle. The unnatural amino acid (UAA) may comprise a nitrogen-containing diene. The unnatural amino acid (UAA) may comprise a cyclic azine selected from a diazine, a triazine, and a tetrazine. The unnatural amino acid (UAA) may comprise a tetrazine. The unnatural amino acid (UAA) may comprise the moiety, where R= H, methyl or ethyl.

[0061] The induced proximity interaction may be a protein post-translational modification mediated by an enzyme. The post-translational modification may be selected from ubiquitination, acetylation, nitrosylation, methylations, hydroxylation, phosphorylation, SUMOylation, ISGylation, FATIOylation, UFMylation, glycosylation, AM Pylation, lipidation, proteolysis, redox post-translational modification, ADP-ribosylation and deamidation. The induced proximity may also apply to the reversal of natively installed modifications by phosphatases, deacetylases, deubiquitinating enzymes, lipid hydrolases and ubiquitin-like protein peptidases.

[0062] The proximity-inducing molecule of formula (I) may be according to any of the statements of the first aspect. Moiety A of the proximity-inducing molecule of formula (I) may be configured to bind to the target protein (Pt). Where a known ligand for the target protein (Pt) exists, moiety A may be a ligand for the target protein or a derivative thereof which retains the ligand function. Where a ligand for the target protein is not known or available (or does not exist), a tag may be attached to the target protein (Pt) and moiety A may comprise a ligand for the protein tag.

[0063] The method of the third aspect may further comprise attaching a protein tag to the target protein (Pt) prior to step d). In these embodiments, moiety A in the proximity-inducing molecule may be a ligand for the protein tag. The protein tag may be any suitable protein tag, such as a HALOtag, dTag or a bromoTag. In these embodiments, moiety A of the proximity-inducing molecule may be a ligand for the protein tag (e.g. a ligand for a HALOtag, dTag or a bromoTag).

[0064] The method may take place in vivo, in vitro, and ex vivo. Preferably, the method takes place in vitro or ex vivo. When the method is performed in vivo, it may be performed in a mammalian, such as a human or an animal. Preferably, the method may be performed in an animal, for example for an animal model or animal studies.

[0065] In a fourth aspect there is provided a method of treatment or prevention of a disease involving a post-translational modification of a target protein (Pt), the method comprising: a) Genetically engineering the cell line of a subject in need thereof by genetic code expansion to encode an unnatural amino acid (UAA) comprising a bioorthogonal reactive group into a first protein (Pi), wherein the first protein takes part in the post-translational modification of a target protein (Pt); b) Allowing the cell line to express a recombinant protein (Pr) of the first protein (Pi) comprising the unnatural amino acid (UAA); c) Administering to the subject in need thereof a therapeutically effective amount of a proximity-inducing molecule according to the first aspect to form a ternary complex between the recombinant protein (Pr), the proximityinducing molecule and the target protein (Pt); d) Allowing the posttranslational modification event to take place assisted by the proximity-inducing molecule.

[0066] In a fifth aspect there is provided a proximity- inducing molecule according to the first aspect for use in the treatment of a disease involving a post-translational modification of a target protein (Pt) and a first protein (Pi), wherein the proximity-inducing molecule is configured to bind (directly or indirectly) to the first protein (Pi) and to the target protein (Pt) to induce proteinprotein interaction.

[0067] The first protein may be an enzyme. The first protein may be an E3 ligase. The E3 ligase may be selected from VHL and CRBN.

[0068] The target protein (Pt). or the first protein may be any protein subject to a post-translational modification (PTM). The target protein (Pt) may be selected from:

[0069] (i) a protein within the bromo- and Extra-terminal (BET) family of proteins; or

[0070] (ii) a protein of the bromo- and Extra-terminal (BET) family of proteins independently selected from: BRD2, BRD3 and BRD4; and

[0071] (iii) a BRD4 protein within the bromo- and Extra-terminal (BET) family of proteins.

[0072] The target protein (Pt) may be an enzyme. The target protein may be an enzyme involved in a post-translational modification of a protein (e.g. the target protein). The first protein may be an E3 ligase. The E3 ligase may be selected from VHL and CRBN.

[0073] The first protein may be any protein subject to a post-translational modification (PTM). The first protein (Pi) may be selected from:

[0074] (i) a protein within the bromo- and Extra-terminal (BET) family of proteins; or

[0075] (ii) a protein of the bromo- and Extra-terminal (BET) family of proteins independently selected from: BRD2, BRD3 and BRD4; and

[0076] (iii) a BRD4 protein within the bromo- and Extra-terminal (BET) family of proteins. Brief description of the figures

[0077] Figure 1 shows a toolkit for induced bioorthogonal induced proximity in mammalian cells. A) BCNK (bicyclo [6.1.0] nonyne-L-lysine) is genetically encoded into proteins of interest in response to an in frame TAG stop codon. Two bioorthogonal proximity inducer (BPI) molecules (BPI-PEG3 and BPI-PEG2) were designed and synthesised. These are designed to enable BPI induced proximity between any protein containing BCNK and the BET family of bromodomain proteins. B) The target E3 ligases tested in this study were VHL and CRBN and were transiently expressed from piggybac plasmids pPB-EF1-VHL or pPB-EF1-CRBN, containing a TAG codon at the desired position. C) Wild-type or mutant Methanosarcina mazei pyrrolysyl tRNA synthetase / tRNAcuA pairs were expressed from plasmids pWT-PylSTmam or pBCNK-PylSTmam.

[0078] Figure 2 shows sites selected for unnatural amino acid (UAA) incorporation in the E3 ligases VHL and CRBN. A) 6 sites were selected for UAA incorporation in VHL that were in proximity of the VHL-1 ligand, which is the basis of the PROTAC molecule MZ118. B) 3 sites were selected for UAA incorporation in CRBN that were in proximity of the IMID ligand lenalidomide 18

[0079] Figure 3 shows Expressions assays of VHL and CRBN variants in HEK293 analysed by Western Blot. A) 6 VHL-HA variants were expressed bearing either BocK or BCNK at selected positions - this confirmed UAA incorporation into the 6 selected positions in VHL with variable efficiency. B) 3 CRBN-HA variants were expressed bearing either BocK or BCNK at selected positions - this confirmed UAA incorporation into 3 selected positions in CRBN. C) Experimental workflow for the initial assessment of UAA incorporation efficiency in mammalian cells. D) Cells were transfected with pBCNK-PylSTmam and pPB-EF1-VHLw / r or pPB-EF1- VHLTAGX (where TAGX denotes a variant where a sense codon is substituted to TAG at position X) variants and cultured in the presence or absence of BCNK. Untreated or BPI treated cells (1 h) were then subjected to in cellulo IEDDA labeling by treatment with TAMRA-tetrazine. Lysates were obtained which were then resolved by SDS-PAGE and imaged by in-gel fluorescence with a 560-580 nm emission filter.

[0080] Figure 4 shows degradation assays of long and short BRD4 isoforms, BRD3 and BRD2 mediated by VHL and CRBN variants in the presence of BPI-PEG3. A) BPI-induced degradation was tested in cells expressing 6 VHL variants and VHL wild-type. Different degradation profiles are observed dependent on the VHL site bearing BCNK and in presence of BPI-PEG3, whilst VHL wild-type is unable to degrade any substrate in the absence or presence of BPI. B) BPI-induced degradation was tested in cells expressing 3 CRBN variants and CRBN wild-type. Different degradation profiles are observed dependent on the CRBN sites bearing BCNK and in presence of BPI-PEG3, whilst CRBN wild-type is unable to degrade any substrate in the absence or presence of BPI. C) Degradation assay workflow.

[0081] Figure 5 shows images demonstrating targeted protein degradation through bioorthogonal proximity induction. A) Optimised transfection procedure that reduces the cell preparation time by 48 h and enables uniform transfection prior to experimentation. B) Selected VHL-HA variants sensitized at three positions with BCNK incorporation were treated with BPI for 1 hour. Lysates were analyzed by immunoblot using specified primary antibodies and IRDye® 680RD or IRDye® 800CW secondary antibodies. Immunoblotting for FLAG confirmed expression of the pyrrolysyl-tRNA-synthetase. Immunoblotting for HA (hemagglutinin) confirmed the expression of the E3 ligase containing the UAA. C) Quantitation of IRDye® fluorescence from immunoblot analysis. Cells were treated with BPI-PEG3-JQ1 (1 pM) and BET protein percentage was calculated for BCNK mutants relative to wild-type control. Column and error bars correspond to the mean and standard deviation, respectively (n = 3). Statistical analyses were carried out by multiple unpaired t-tests with Holm-Sidak correction and assumed single pooled variance (ns not significant, * P < 0.05, ** P < 0.01 , *** P < 0.001 , **** P < 0.0001). D) For HiBiT assays, successful transfection of the plasmids was determined by fluorescence microscopy. E) BPI-PEG3-JQ1 dose response analysis of HiBiT-BRD4 levels in cells expressing VHL with BocK at position 67. Statistical analyses were carried out by ordinary one-way ANOVA Dunnett’s multiple comparisons test relative to DMSO control, assuming single pooled variance. For all HiBiT experiments, n = 3. F) Dose-response with BPI-PEG3- JQ1 for cells expressing VHL with BCNK at position 67. G) Degradation of HiBiT-BRD4 is ablated upon treatment with MG132 (50 pM). H) Degradation of HiBiT-BRD4 is ablated upon treatment with MLN4924 (1 pM). I) Dmax is increased following FAC sorting for positively transfected cells. J) BPI-PEG3-JQ1 dose response analysis of HiBiT-BRD3 levels in cells expressing VHL with BCNK at position 67. K) BPI-PEG3-JQ1 dose response on HiBiT-BRD2 levels in cells expressing VHL with BCNK at position 67. L) Degradation of HiBiT-BRD4 with BCNK incorporation at multiple positions with VHL. Position is directed by the specified amber codon mutation. M) Structural model of (NEDD8)-CRL2VHL-UBE2R1~Ub complex. VHL-VHL- 1 ligand complex (PDB ID: 4W9H) was superposed onto VHL in the cryogenic electron microscopy (cryoEM) structure of the (NEDD8)-CRL2VHL-UBE2R1~Ub complex (PDB ID: 8RX0)52. For clarity, the BRD4 bromodomain has been omitted. To enable comprehensive surface generation of the E2 UBE2R1 (which was not possible using model 8RX0 due to unmodelled side chains) a high-resolution crystal structure (PDB ID: 2OB4) was superposed

[0082] 53 Figure 6 shows images of the BPI assessment of TPD with the E3 ligase CRBN. A) CRBN- HA variants that were sensitized at three positions with BCNK incorporation were treated with BPI for 1 hour. Lysates were analyzed by immunoblot using the specified primary antibodies and IRDye® 680RD or IRDye® 800CW secondary antibodies. B-E) Quantitation of IRDye® fluorescence from immunoblot analysis. Cells were treated with BPI-PEG3-JQ1 (1 pM) and BET protein percentage was calculated for BCNK mutants relative to wild-type control. Column and error bars correspond to the mean (n = 3) and standard deviation, respectively. Statistical analyses were carried out by multiple unpaired t-tests with Holm-Sidak correction and assumed single pooled variance (ns not significant, * P < 0.05, ** P < 0.01 , *** P < 0.001 , **** P < 0.0001). F) HiBiT-BRD4 is degraded in cells expressing CRBN with BCNK at position 353 at all tested BPI-PEG3-JQ1 concentrations. G) HiBiT-BRD4 degradation by BCNK containing CRBN is abated by MG132 treatment (50 pM). H) HiBiT-BRD4 degradation is also abated by MLN4924 treatment (1 pM). I) CRBN containing BCNK at position 353 indues significant HiBiT- BRD3 degradation at BPI concentrations above 10 nM. J) HiBiT-BRD2 is also significantly degraded at BPI concentrations above 10 nM. For all HiBiT assays, statistical analyses were carried out by ordinary one-way ANOVA Dunnett’s multiple comparisons test relative to DMSO control, assuming single pooled variance. Column and error bars correspond to the mean and standard deviation, respectively (n = 3).

[0083] Figure 7 shows images demonstrating BPI induced TPD with the non-E3 ligase effector LIBE2D1. A) The reactive closed E2~Ub was analyzed (PDB ID: 4AP4), and three sites were selected for sensitization with BCNK. B) Immunoblot analysis of all 3 positions in LIBE2D1 suggested variable levels of BPI-dependent BET protein degradation. C) The higher molecular weight bands diminished when lysates were treated with LISP2 (5 pM) for 30 minutes at 37 °C. D) With BCNK incorporation at position 90, significant HiBiT-BRD4 degradation was achieved at BPI concentrations above 250 nM. E) With BCNK incorporation at position 90, significant HiBiT-BRD3 degradation was achieved at BPI concentrations above 100 nM. F) With BCNK incorporation at position 90, significant HiBiT-BRD2 degradation was achieved at BPI concentrations above 50 nM. Above 250 nM, P values of < 0.001 and < 0.0001 were obtained and have been annotated collectively for clarity. G) With BCNK incorporation at position 111 , significant HiBiT-BRD4 degradation was achieved at BPI concentrations above 100 nM. Above 250 nM, P values of < 0.001 and < 0.0001 were obtained. H) With BCNK incorporation at position 111 , significant HiBiT-BRD3 degradation was achieved at BPI concentrations above 100 nM. Above 750 nM, P values of < 0.001 and < 0.0001 were obtained. I) With BCNK incorporation at position 128, degradation of HiBiT-BRD4 did not reach significance. J) Proteasome inhibition with MG132 did not significantly reduce HiBit-BRD4 degradation by LIBE2D1 containing BCNK at position 90. K) Inhibition of cullin ligase activation with MLN4924 reduced HiBiT-BRD4 degradation at a BPI-PEG3-JQ1 concentration of 250 nM. For J) and K), statistical analyses on the effect of BPI-PEG3-JQ1 were carried out by two-way ANOVA Dunnett’s multiple comparisons test relative to DMSO control, assuming single pooled variance.

[0084] Figure 8. A) Time course analysis after administering BPI probe to sensitized VHL. Cells (~6 x 106) were transfected with pPylSTAF (9 pg) and PPB-VHLTAG67 (9 pg), grown in the presence of BCNK and then seeded (~1 x 106) in parallel into multiple wells (6-well plate). Cells were washed and were then treated with BPI-PEG3-JQ1 or DMSO. Cells at the specified time points were harvested and analyzed by immunoblot. BRD4 and BRD2 degradation does not appreciably increase beyond 1 h, whereas improved BRD3 degradation is observed after prolonged incubation. The addition of the BPI probe to VHL causes as a slight reduction in electrophoretic mobility, suggestive of near quantitative I EDDA modification within 1 h. B) BPI dose response for cells transfected with pPB-VHLwr and pPylSTAF and cultured in the presence of BCNK. Pelleted cells were transfected with PPB-VHLTAG67 and pPylSTAF, incubated for 24 h and transferred into parallel wells in a 96-well culture plate. BPI-PEG3-JQ1 ligand was added at the specified concentrations for 1 h. Cells were directly treated with Nano- Glo® HiBiT Lytic Detection reagent according to the manufacturer’s instructions. Luminescence was then measured with a Pherastar FS plate reader (BMG Labtech). The mean is plotted, and error bars correspond to the standard deviation (n = 3). Statistical analyses were carried out by ordinary one-way ANOVA Dunnett’s multiple comparisons test, relative to DMSO control, assuming single pooled variance (ns not significant, * P < 0.05, ** P < 0.01 , *** P < 0.001 , **** P < 0.0001). C) Dose response with BPI-PEG2-JQ1 for cells expressing VHL containing BCNK at position 67. Pelleted cells were transfected with pPB- VHLTAG67 and pPylSTAF, incubated for 24 h and transferred into parallel wells in a 96-well culture plate. BPI-PEG2-JQ1 ligand was then added at the specified concentrations for 1 h. Cells were directly treated with Nano-Gio® HiBiT Lytic Detection reagent according to the manufacturer’s instructions. Luminescence was then then measured with Pherastar FS plate reader (BMG Labtech). The mean is plotted, and error bars correspond to the standard deviation (n = 3). Statistical analyses were carried out by ordinary one-way ANOVA Dunnett’s multiple comparisons test, relative to DMSO control, assuming single pooled variance (ns not significant, * P < 0.05, ** P < 0.01 , *** P < 0.001 , **** P < 0.0001).

[0085] Figure 9 Fluorescent activated cell sorting of cells expressing VHL containing BCNK at position 67. HiBiT-BRD4 HEK293 cells were identified based on forward scatter area (FSC- A) and side scatter area (SSC-A) properties and single cells were gated using proportionality of FSC-A against forward scatter width (FSC-W). HEK293 cells expressing double positively for GFP and mKate2 were collected; purity testing of this population showed 97.3% (98.32% x 98.83%) of cells collected were double positive single cells. A) Gating strategy and B) Purity assessment.

[0086] Figure 10 A) Time course analysis of BET protein degradation with UBE2D1 containing BCNK at position 90. Wild type HEK293 cells transfected with pPylSTAF and pPB- UBE2D1TAG9O were serially seeded and treated with DMSO or BPI-PEG3-JQ1 (500 nM). Samples were lysed at the specified time points and analyzed by immunoblot. Membranes were probed with HRP conjugated secondary antibodies and chemiluminescent detection. B) Gating Strategy and C) Purity assessment of fluorescent activated cell sorting of cells expressing UBE2D1 containing BCNK at position 90. HiBiT-BRD4 HEK293 cells were identified based on forward scatter area (FSC-A) and side scatter area (SSC-A) properties and single cells were gated using proportionality of FSC-A against forward scatter width (FSC- W). HEK293 cells expressing double positively for GFP and mKate2 were collected; purity testing of this population showed 96.3% (98.88% x 97.42%) of cells collected were double positive single cells. D) BPI dose response for FAC sorted cells expressing sensitized UBE2D1. No improvement in Dmax was observed, suggesting that transfection efficiency was not a limiting factor for degradation. The mean is plotted, and error bars correspond to the standard deviation (n = 3). Statistical analyses were carried out by ordinary one-way ANOVA Dunnett’s multiple comparisons test, relative to DMSO control, assuming single pooled variance (ns not significant, * P < 0.05, ** P < 0.01 , *** P < 0.001 , **** P < 0.0001).

[0087] Detailed description

[0088] The present invention aims to provide induced proximity of proteins mediated by novel proximity-inducing molecules and exploiting the potential of genetic code expansion and ultrafast bioorthogonal reactions. The inventors anticipate that this approach will have broad utility, enabling the therapeutic potential of induced proximity to be comprehensively assessed through the facile process of stop codon scanning, even when no known ligands for one or both of the proteins are available.

[0089] The inventors first envisaged employing genetic code expansion to evaluate proteins, for example in live mammalian cells (although this could also be performed extracellularly in vitro), for induced proximity applications in a tag-free manner. The inventors have devised a method of inducing proximity between a first protein (Pi) and a target protein (Pt), the method comprising: a) Preparing a recombinant protein (Pr) by genetic code expansion to encode into the first protein (Pi) an unnatural amino acid (UAA) comprising a bioorthogonal reactive group capable of reacting with a proximity-inducing molecule of formula (I) : A-L-D, wherein

[0090] A is a moiety capable of binding to the target protein (Pt) or to a tag located on the target protein;

[0091] L is a linker; and

[0092] D is a moiety capable of undergoing a bioorthogonal reaction with a bioorthogonal reactive group present in the first protein (Pi) to irreversibly form a covalent bond with the first protein (Pi); b) Forming a modified protein (Pm) by contacting the recombinant protein (Pr) with the proximity-inducing molecule of formula (I) to form a covalent bond between the unnatural amino acid (UAA) of the recombinant protein (Pr) and moiety D of the proximity-inducing molecule in a bioorthogonal reaction; c) Forming a ternary complex by allowing the modified protein (Pm) to bind to the target protein (Pt).

[0093] By supplementing cells with an orthogonal tRNA-synthetase / tRNAstoPcodon pair (e.g. tRNA- synthetase / tRNAcuA), genetic code expansion permits the incorporation of unnatural amino acids (UAAs) in response to an amber TAG stop codon. Proteins (e.g. “the first protein (Pi)” can be sensitised to bioorthogonal warheads (moieties comprising a functional group capable of reacting in a bioorthogonal reaction - in this case moiety D of the compounds of Formula (I)), by incorporating into a first protein (Pi) an UAA bearing a complementary bioorthogonal reactive group capable of reacting with moiety D of the proximity-inducing molecules of Formula (I) and which undergoes cellular labelling. In this proof of concept, and to ensure sufficient labelling of the first protein (Pi) at the low concentrations generated by cellular expression, the inventors designed proximity-inducing molecules bearing reactive groups capable of undergoing kinetically ultra-fast inverse electron demand Diels-Alder (I EDDA) reactions with UAAs, for example I EDDA reactions between electron-rich dienophiles and electron-poor dienes17. The inventors designed modified proteins (Pm) starting from first proteins (Pi) but including a range of suitable dienophiles. The UAAs can be incorporated into the modified proteins by any suitable tRNA-synthetase / tRNAstoPcodon pair (e.g. native and evolved pyrrolysl-tRNA synthetase / tRNAcuA pairs). The inventors reasoned that reaction rates between the dienophiles present in the UAAs of the modified proteins (Pm) and certain dienes (e.g. tetrazine dienes) are rapid (2= >103M'1s-1), enabling efficient labelling at cellular protein concentrations15. Likewise, the inventors reason that ultrafast I EDDA reactions can take place between a dienophile present in the proximity-inducing molecule of formula (I) (in moiety D) and a diene present in the UAA incorporated into a modified protein (Pm) starting from a first protein (Pi).

[0094] In order to establish a proof of concept which could subsequently be expanded to other proteins and proximity-inducing molecules, the inventors first designed and synthesised proof of concept proximity-inducing molecules (which they named bioorthogonal protein inducer (BPI) molecules and fall within the definition of compounds of Formula (I)) and assessed the biological effect of proximity induction by these molecules between any target protein (Pt) and a first protein (Pi) selected from the BET family bromodomain proteins.

[0095] The inventors designed the interaction between the proximity-inducing molecules (also referred to as BPIs) and the first protein (Pi) (e.g. BET family bromodomain proteins) to take place via bioorthogonal tethering to any site within the protein where the UAA is incorporated. The inventors initial proximity-inducing molecules (or BPIs) comprised a tetrazine moiety (moiety D, also referred to as bioorthogonal warhead) and the BET family ligand JQ1 (moiety A) with a linker therebetween (any suitable linker could be used ,such as an aliphatic chain or a PEG group, for example L can be a linker of formula -(CH2-CH2-O)n-, wherein n is selected from 1 to 20). In the proof of concept examples, the proximity-inducing molecules (BPI-PEG2 and BPI-PEG3) comprised PEG2 and PEG3 respectively as linkers (L) (Figure 1).

[0096] Therefore, the inventors designed a proximity-inducing molecule of formula (I):

[0097] A-L-D

[0098] (I) for inducing proximity between a first protein and a target protein, wherein:

[0099] A is a moiety capable of binding to the target protein or to a tag located on the target protein;

[0100] L is a linker; and

[0101] D is a moiety capable of undergoing a bioorthogonal reaction with a bioorthogonal reactive group present in the first protein to irreversibly form a covalent bond with the first protein. The specific proximity-inducing molecules of Formula (I) synthesised as proof of concept were

[0102] The inventors next established a cell line (in this particular case an engineered HEK293 cell line) that would express a first protein (Pi) that could be brought into proximity of a target protein (Pt) (in this case BET family bromodomain proteins) and elicit a measurable phenotype (proximity-induced event).

[0103] It is known that heterobifunctional degrader molecules (i.e. PROTACs) consisting of ligands specific for the E3 ligases VHL or CRBN, and which also comprise the ligand JQ1 , bind to BET proteins and facilitate BET protein degradation by the E3 ligases18 19. The inventors used these established targeted protein degradation (TPD) systems to inspire proof of concept and sought to incorporate the UAA bicyclo [6.1.0] nonyne-L-lysine (BCNK) into sites proximal to the established ligand binding pockets in E3 ligases VHL and CRBN (Figure 2a).

[0104] Although not essentially required, the inventors cloned amber codon VHL and CRBN mutants with a C-terminal HA tag (VHL-HA or CRBN-HA) into the PiggyBac expression vector to enable efficient mutant protein expression and detection by Western Blott. Solvent-exposed sites / side chains were chosen adjacent to the respective ligand and side chains proximal to proteinprotein interfaces (Figure 2B) were avoided - this ensured efficient bioorthogonal reaction, and reduced the chances of protein structural perturbation. The inventors also included an enhanced green fluorescent protein (EGFP) cassette to assess transfection efficiency and enable serial FAC sorting.

[0105] To mediate the incorporation of the UAA BCNK into the first protein (Pi - E3 ligases VHL and CRBN), the inventors supplied to the cell line a tRNA-synthetase / tRNAstoPcodon pair (in this case an evolved pyrrolysl-tRNA synthetase / tRNAcuA mutant) on a separate plasmid (pBCNK- PylSTmam) (Figure 1c). HEK293 cells were then cultured and supplemented with UAA (BCNK) up to 4 hours prior to transient transfection with both plasmids. Control cells were not supplemented with BCNK. As a benchmark, they also incorporated the widely used UAA Ne- Boc-L-lysine by supplying a plasmid harbouring the wild type pyrrolysl-tRNA synthetase / tRNAcuA pair (pWT-PylSTmam). For improved suppression efficiency, the tRNA gene could contain a U25C mutation. Depending on the site of amber codon mutation, optimized conditions led to efficient amber codon suppression in a UAA-dependent manner resulting in high levels of full-length VHL-HA and CRBN-HA proteins (Figure 3a and 3b). Plasmid pBCNK-PylSTmam contained Y306A and Y384F mutations in MmPylS to enable recognition of BCNK. MmPylS variants were cloned as bicistronic expression cassettes with mKate2 fluorescent protein to allow facile assessment of transfection efficiency by fluorescence-activated cell sorting (FACS). To improve cytoplasmic expression, a nuclear export sequence was appended to the N-terminus of the MmPylS variants.

[0106] Sensitized CRBN mediates BPI probe-dependent degradation

[0107] The inventors next studied BPI-dependent degradation of BET family proteins by recruiting CRBN. Using the immunoblot workflow with wild type cells, statistically significant degradation was only achieved for the long BRD4 isoform with BCNK incorporation at position Glu377 (Figure 6B-E). We next selected BCNK at site His353, which is expressed at a lower level, for further analysis with the HiBiT platform. Potent degradation of HiBiT-BRD4 was observed (DCso ~50 nM, Dmax 49%) and the characteristic hook effect manifested at elevated BPI concentrations (Figure 6F). Consistent with degradation being proteasome- and cullin E3- dependent, it was blocked by treatment with MG132 and MLN4924, respectively (Figure 6G, H). Significant degradation was also observed for HiBiT-BRD3 and HiBiT-BRD2 expressing cell lines (Dmax 30% and 25%, respectively) (Figure 6I, J). Hence, the BPI approach, in combination with HiBiT engineered reporter lines, provides a robust and general experimental platform that should allow interrogation of TPD enabled E3 ligases.

[0108] The inventors next explored whether the VHL variants (recombinant proteins Princorporating the UAAs) could induce targeted protein degradation (TPD) of BET family bromodomain proteins in the presence of one of their proximity-inducing molecules. To prevent the excess of UAA BCNK circulating in the media from quenching the proximity-inducing molecules administered to the cells, the inventors subjected cells to UAA washout. The proximityinducing molecule BPI-PEG3 (1 pM) was then added to the cultures and incubated for 16 h. Cells were then lysed and analysed by immunoblotting (Figure 4a). The tetrazine diene present in the warhead (moiety D) of the proximity-inducing molecule BPI-PEG3 underwent an Inverse Electron Demand Diels Alder (IEDDA) reaction with the dienophile present in UAA BCNK present in the recombinant VHL proteins and therefore the proximity-inducing molecule was incorporated into the recombinant protein (forming a modified protein) by the new covalent bond between the two species formed at the IEDDA reaction. HEK293 cells expressing VHL with BCNK at positions 67, 90 and 94 were also labelled in cellulo by treating with 3-(p- benzylamino)-1 ,2,4,5-tetrazine-5-carboxytetramethylrhodamine (TAMRA-tetrazine) (10 pM) for 1 h. Lysates were then resolved by SDS-PAGE and visualized by in-gel fluorescence. In addition to a prominent signal corresponding to VHL, other BCNK- dependent bands were observed, indicating a degree of off-target protein labelling (Figure 3D). Confirming that BCNK incorporated into VHL and the off-targets could rapidly react with BPI probe, the fluorescent signals were largely ablated by prior treatment of cells with BPI-PEG3-JQ1 (1 pM) for 1 hour (Figure 3D). Quantitative comparison of the bands corresponding to VHL with and without BPI treatment suggested ~95 % in cellulo conjugation of BPI-PEG3-JQ1 to all 3 VHL sites.

[0109] No observable degradation of any BET family bromodomain protein was observed in control cells expressing wild-type VHL-HA. However, cells expressing VHL-HA variants (recombinant proteins Prcomprising the BCNK UAA) having formed modified proteins (Pm) by incorporation of the proximity-inducing molecules at positions Asn67, Asn90, Glu94, ThrlOO, Thr105, and Ser111 exhibited degradation of BET family bromodomains to various extents. Marked degradation (-50%) was observed for the short isoform of BRD4 whereas degradation of the long BRD4 isoform was only observed with bioorthogonal tethering of BPI-PEG3 to position 67, 100, 105, and 111. In addition, degradation of BRD3 and BRD2 was also observed (Figure 4a).

[0110] Site-specifically sensitized VHL carries out BPI-dependent TPD

[0111] For an endogenous neosubstrate, the maximum level of degradation (Dmax) with the BPI approach would be governed by the efficiency of transgene delivery into the cell. To maximize this, the inventors developed an optimized transfection procedure where cells were uniformly transfected prior to seeding (Figure 5A). This also reduced the preparation time for degradation assays from 72 to 24 hours. Qualitative time course analysis for VHL containing BCNK at 67 confirmed efficient degradation efficiency, with maximum levels for BRD4 and BRD2 isoforms attainable after 1 h of BPI administration (Figure 8A). At prolonged time points, a modest improvement in BRD3 degradation was observed. In further support of rapid and efficient bioorthogonal labelling, the band corresponding to VHL-HA migrated with reduced electrophoretic mobility within 1 hour of BPI treatment (Figure 8A). As expected, following UAA washout - implemented to prevent BPI quenching - VHL-HA was destabilized, with a halflife of approximately 4 hours. The inventors next focused on sites 67, 90 and 94 for quantitation. Degradation of both BRD4 isoforms and BRD3 was significant only with BCNK incorporation at position 67 (Figure 5B, C). The inventors reasoned that the inability to achieve significance at other sites, and for BRD2, may have been due to experimental variability inherent to the immunoblot workflow, which compares cultures grown in the presence or absence of UAA. HiBiT quantitation of BET protein degradation

[0112] To characterize the system further, the inventors developed a higher throughput experimental platform based on a CRISPR / Cas9-engineered knock-in HEK293 cell line expressing HiBiT- tagged BRD4 (HiBiT-BRD4) 36. The HiBiT tag allows robust quantitative luminescence-based measurement of HiBiT-tagged protein levels by lysis and in situ detection using the Nano- Glo® HiBiT Lytic Detection System 37. To confirm successful transfection with pBCNK- PylSTmam / pWT-PylSTmam and PPB-EF-VHLTAGX plasmids, the inventors turned to fluorescence microscopy using EGFP and mKate2 channels (Figure 5D). As expected, no BRD4 degradation was observed relative to DM SO-treated controls when the unreactive control UAA BocK was incorporated at position 67 of VHL, demonstrating that BET protein degradation is not an artifact of general UAA incorporation (Figure 5E). Importantly, no degradation was observed across a broad range of BPI concentrations in cells transfected with pBCNK- PylSTmam and PPB-EF1-VHLWT, confirming the off-target proteins containing BCNK did not contribute to HiBiT-BRD4 degradation (Figure 8B).

[0113] The inventors next carried out a BPI-PEG3-JQ1 dose response on cells expressing VHL containing BCNK at position 67 (Figure 5F). Strikingly, a DC50 of ~ 1 nM (Dmax 69 % at 100 nM) was achieved, which is of comparable potency to optimized PROTACs containing VHL and JQ1 ligands (DC50 ~ 6 nM) 38. As expected for TPD with a bifunctional degrader, a hook effect manifested at higher BPI concentrations (Figure 5F). Comparable degradation efficiency was observed with BPI-PEG2-JQ1 containing the shorter PEG2 linker (Figure 8C). In support of HiBiT-BRD4 degradation being proteasome-dependent, it was ablated upon treatment with the inhibitor MG132 (Figure 5G). Consistent with degradation being mediated by VHL, it was also prevented by treatment with MLN4924, which blocks cullin ligase activation (Figure 5H).

[0114] The inventors next tested whether serial FAC sorting against EGFP and mKate2 fluorescence could enhance degradation efficiency and obtained 97.3 % double positive single cells (Figure 9A). As anticipated, a higher Dmax (92 %) was achieved (Figure 5I). Even without FACS, significant degradation of HiBiT-BRD3 and HiBiT-BRD2, in corresponding HEK293 cell lines, was observed (DC50 50 nM; Dmax 57 % and Dmax 40 %, respectively) albeit less efficiently than for BRD4 36. As VHL PROTACs degrade BRD4, BRD3, and BRD2 comparably, the inventors speculate that BPI-PEG3-JQ1 tethering to BCNK at position 67 results in an altered ternary complex, which accounts for the specificity profile and may resemble the profile of established PROTACs with a different linker in the BPI (Figure 5J, K) 38.

[0115] The inventors next assessed all our VHL sites in the HiBiT-BRD4 cell line. Significant HiBiT- BRD4 degradation was only observed when BCNK was incorporated in place of Asn67, Asn90, ThrlOO and Thr105 (Figure 5L). They considered the influence of expression level on the results. Although expression was amongst the lowest when BCNK was incorporated at His110, it was comparable to that of functional variants (Figure 4A). Furthermore, site Glu94 was non-functional, yet expression of this variant was amongst the highest. This suggests that sites His110 and Glu94 do not yield a productive ternary complex. These observations demonstrate that modest changes in the recruitment site can have a drastic impact on degradation efficiency. Interestingly, Asn67, the most efficient site, is proximal to the terminal VHL-1 acetyl group in ligand-bound VHL structures which has been selected as the exit vector for potent VHL-1 derived PROTACs (Figure 2A) 28,34. On the other hand, Asn90, ThrlOO and Thr105 are adjacent to distinct regions in VHL-1 , implying established PROTACs generate an optimal ternary complex but alternative exit vectors on VHL-1 could yield functional degraders (Figure 2A). The inventors next tested the outcome when recruiting to an unexplored site on the opposite side of the VHL binding pocket, by incorporating BCNK at sites Thr133 and Val137, but no significant degradation was observed (Figure 5L, M). Expression levels were comparable to functional mutants, suggesting that limited flexibility of the VHL receptor subcomplex prevents the formation of a productive ternary complex.

[0116] The inventors next applied the proximity-inducing molecule-dependent degradation of any BET family bromodomain protein by recruiting CRBN instead of VHL. Having confirmed cellular expression of recombinant protein CRBN-HA variants containing the UAA BCNK at positions Asn351 , His353, and Glu377, the inventors proceed to test for degradation of BET family bromodomains upon treatment with proximity-inducing molecules (BPI-PEG2 and BPI- PEG3). Significant degradation was observed for the short isoform of BRD4 for all tested positions. The degradation of the long BRD4 isoform was only observed with bioorthogonal tethering of BPI-PEG3 to position 353. BRD3 degradation was also observed for positions 353 and 377 (Figure 4b and 6A).

[0117] Using the immunoblot workflow with wild type cells, statistically significant degradation was only achieved for the long BRD4 isoform with BCNK incorporation at position Glu377 (Figure 5B-E). The inventors next selected BCNK at site His353, which is expressed at a lower level, for further analysis with the HiBiT platform. Potent degradation of HiBiT-BRD4 was observed (DCso ~50 nM, Dmax 49%) and the characteristic hook effect manifested at elevated BPI concentrations (Figure 5F). Consistent with degradation being proteasome- and cullin E3- dependent, it was blocked by treatment with MG132 and MLN4924, respectively (Figure 5G, H). Significant degradation was also observed for HiBiT-BRD3 and HiBiT-BRD2 expressing cell lines (Dmax 30% and 25%, respectively) (Figure 5I, J). Hence, the BPI approach, in combination with HiBiT engineered reporter lines, provides a robust and general experimental platform that should allow interrogation of TPD enabled E3 ligases. Assessment of a non-E3 effector

[0118] The inventors next tested the site-dependence on TPD for an unconventional effector. The E2 conjugating enzyme LIBE2D1 can facilitate lysine ubiquitination in vitro through E3- independent sampling of its reactive closed E2~Ub conformation39’40. Furthermore, a covalent tool molecule that targets cysteine 111 in LIBE2D1 and its isoforms has been elaborated into bifunctional BET protein degraders41. The inventors analysed the structure of the closed E2~Ub conformation and selected three sites hypothesized to support productive ternary complex formation (Figure 7A). They chose residues in proximity of the E2 catalytic site and residue Cys111. Residue Arg90, which lies adjacent to a shallow pocket, may be targetable with a ligand bearing a covalent warhead, as demonstrated for Cys42. Residue Lys128 has been shown to make contacts with substrates and is important for their ubiquitination40. The inventors also avoided residues known to play a role in stabilizing the closed E2~Ub conformation and E2 catalytic activity4344. Time-course analysis suggested optimal degradation after 2 h of BPI treatment (Figure 10A). LIBE2D1 also had reduced electrophoretic mobility after 1 h of BPI treatment consistent with efficient bioorthogonal conjugation (Figure 10A). Encouragingly, qualitative immunoblot analysis for all sites suggested BCNK-dependent BET protein degradation to different extents after 2 h treatment with BPI-PEG3-JQ1 (500 nM) (Figure 7B). This also allowed the expression levels of the BCNK variants to be deduced, revealing marked reduction for position 128.

[0119] Notably, with BCNK at site 90, BPI-PEG3-JQ1 treatment (250 nM) led to a discrete band with reduced electrophoretic mobility that approximated the molecular weight of a monoubiquitin adduct. This was observed for both BRD4 isoforms and BRD3 (Fig 7B). Consistent with monoubiquitination, the band was sensitive to lysate treatment with a promiscuous deubiquitinating enzyme (LISP2) (Figure 7C). When explored with the HiBiT platform, BCNK incorporation at sites 90 and 111 invoked significant BET protein degradation in all tested HiBiT reporter lines. For site 90, Dmax values of 13%, 24% and 17% were obtained for HiBiT-BRD4, HiBiT-BRD3 and HiBiT-BRD2, respectively (Figure 7D-F). For site 111 , Dmax values of 15 % and 21 % were achieved for HiBiT-BRD4, HiBiT-BRD3, respectively (Figure 7G,H). Site 128 did not yield significant degradation, although this might have been due to the weak expression of this mutant (Figure 7I and 7B). FAC sorting of cells (96.3 % double positive) expressing LIBE2D1 with BCNK at position 90 afforded no improvement in Dmax, suggesting that the modest degradation efficiency was not due to genetic mosacity in the unsorted cell population (Figure 10B,C,D).

[0120] Interestingly, a hook effect was only observed with BCNK incorporation at position 90 (Figure 7D & 10D). The inventors reasoned that this was due to recruitment at position Arg90 being less effective than Cys111 at forming an efficient ternary complex, potentially due to steric clash, misalignment or suboptimal lysine modification. Elevated I EDDA reactivity when BCNK is incorporated at position 90 might also account for this observation.

[0121] Curiously, MG132 did not block degradation of HiBiT-BRD4 in cells expressing LIBE2D1 containing BCNK at position 90 (Figure 7K). Degradation was modestly enhanced upon MLN4924 treatment, suggesting that cullin ligase activation is inhibitory (Figure 7L). The stability of the monoubiquitin adduct, coupled with the modest level of BET protein degradation (Dmax 13%), implied that the modification serves as an inefficient degradation signal - potentially due to the inability of LIBE2D1 to efficiently attach a polyubiquitin chain without robust cooperation with an E3 ligase. Interestingly, monoubiquitination has been linked to lysosomal degradation, which may explain why MG132 treatment failed to prevent degradation4546.

[0122] In summary, the BPI platform enables assessment of the outcomes of proximity induction at internal sites within effector proteins, extending beyond E3 ligases.

[0123] In summary, the inventors have developed a new methodology allowing the effect of induced proximity of two proteins to be assessed with no need for a specific ligand for at least one of the proteins (since the interaction between the first protein and the proximity-inducing ligand takes place via a bioorthogonal reaction between a UAA incorporated into the first protein and the proximity-inducing molecule). Therefore, specific ligand recruitment to the first protein is achieved with amino acid resolution by incorporating a UAA containing a bioorthogonal reactive group. The inventors demonstrated proof of concept with the established targeted protein degradation (TPD) systems harnessing the E3 ligases VHL and CRBN and preparing recombinant versions of VHL and CRBN incorporating the UAA BCNK. Modified VHL and CRBN proteins (Pm) are formed upon contacting the recombinant proteins (Pr) incorporating BCNK and proximity-inducing molecules comprising tetrazine dienes capable of reacting with BCNK and forming a covalent bond in an I EDDA reaction. The inventors also demonstrate E3- independent degradation by recruitment of an upstream E2 conjugating enzyme.

[0124] The case study presented herein is a proof of concept for the claimed methodology as it demonstrates the versatility of the claimed methods through the degradation of endogenous BET family proteins using the E3 ligases VHL and CRBN, as well as an E3-independent pathway via an E2 conjugating enzyme.

[0125] This methodology presents a significant advancement in targeted protein degradation (TPD) and induced proximity-based research by addressing a key bottleneck. By combining genetic code expansion with ultra-fast bioorthogonal chemistry, the approach enables site-specific sensitization of proteins to a heterobifunctional bioorthogonal proximity inducer (BPI) probe. When known ligands are available for the target protein (Pt), the proximity-inducing molecules to be used can be designed to contain said known ligands in moiety A of the molecules. When the modified proteins Pmencounter the target proteins (Pt), the ligand A binds (via a covalent or non-covalent interaction) to the target proteins (Pt), thus forming a ternary complex which induces the close proximity of the target protein and the first protein and enables proteinprotein interactions to take place.

[0126] However, the approach can be extended to target proteins for which no known ligand is available. In those cases protein tags (e.g. HALOtag, bromoTag, dTAG) can be appended to the target proteins and the proximity-inducing molecules are designed to comprise a ligand for the protein tag in moiety A.

[0127] The inventors envisage this approach can be extended to any protein pair, providing a specific ligand is available for the target protein or that a protein tag can be appended thereto. This methodology allows the effects of a plethora of induced proximity events to be rigorously assessed. For example, the methodology can be employed to trigger via induced proximity a protein post-translational modification mediated by an enzyme. The post-translational modification may be selected from ubiquitination, acetylation, nitrosylation, methylations, hydroxylation, phosphorylation, SUMOylation, ISGylation, FATIOylation, UFMylation, glycosylation, AMPylation, lipidation, proteolysis, redox post-translational modification, ADP- ribosylation and deamidation. The induced proximity may also apply to the reversal of natively installed modifications by phosphatases, deacetylases, deubiquitinating enzymes, lipid hydrolases and ubiquitin-like protein peptidases.

[0128] Importantly, this strategy facilitates precise and rapid protein degradation, closely mimicking drug-like behaviour without relying on artificial tags or purification steps. The platform holds significant potential for expanding TPD applications to previously undruggable targets, enabling systematic exploration of induced proximity modalities and accelerating the development of next-generation therapeutics for diseases driven by challenging protein targets.

[0129] As the BPI probe can be tethered to, in principle, any site in a protein with a single-residue resolution, it allows assessment of drug-like binding without the need for a specific ligand. In the context of induced proximity studies, this approach offers a valuable strategy for evaluating binding sites within effectors of interest before initiating ligand discovery efforts. The inventors anticipate that this methodology will be particularly useful for exploring pockets - and nucleophilic residues using covalent warheads - that could be targeted by novel ligands. Furthermore, by mapping sites that enable TPD, this strategy should facilitate more focused 1 small-molecule screening campaigns. This will be especially important for large or challenging effectors that are difficult to purify or express recombinantly in their full-length form, as well as for technologies such as DNA-encoded library (DEL) screening, where constructs focused on productive domains of interest can reduce detection of off-target binders. Importantly, the recruitment site is genetically defined, enabling assessment of induced proximity by ligand binding to, in principle, any internal site.

[0130] Induced proximity is a rapidly growing field with broad implications for medicine, basic research, and synthetic biology. While the potential design space is vast, success often hinges on ternary complex geometry. By enabling rapid assessment of recruitment site geometry with high site resolution, without the need for dedicated ligand development, the technology offers a powerful tool to accelerate the realization of induced proximity-based therapeutics and research applications.. By observing the phenotype of artificially induced proximity events, the inventor’s technology may also benefit basic research by delineating the function of poorly characterised proteins.

[0131] Methods

[0132] Cell culture

[0133] HEK293 cells were cultured (37 °C, 5% CO2) in media containing Dulbecco’s modified Eagle’s medium (DM EM) supplemented with 10% (v / v) foetal bovine serum (FBS), 2.0 mM L- glutamine and antibiotics (100 units ml-1 penicillin, 0.1 mg ml-1 streptomycin).

[0134] Transient transfections were performed using Lipofectamine2000 transfection reagent in a 3:1 Lipo:DNA ratio following the manufacturer's protocol.

[0135] Cells were rinsed with ice-cold PBS and extracted in RIPA buffer, complete EDTA-free protease inhibitor cocktail (Roche), and Benzonase® Nuclease (VWR). Lysates were then clarified by centrifugation at 4 °C for 10 min at 16,000 g. Supernatants (total cell extracts) were collected and protein concentration was determined by Bradford assay.

[0136] Transfection expression of VHL and CRBN variants

[0137] Expression assays were performed in 6-well plates. Routinely, 500,000 cells were seeded per well and grown overnight. The next day, media was exchanged, and the unnatural amino acid was added to the fresh growth media. 4 hours later, cells were transiently transfected using 1500 ng DNA per plasmid, 9pl of Lipofectamine2000 (1 mg / ml) and 500 pl reduced serum media (Opti-MEM, Gibco). Following 48 hours of growth, the supernatant was carefully removed, cells washed with PBS, and lysed in 100 l RIPA buffer (Sigma) with added Complete protease inhibitor (Roche) and Benzonase® Nuclease (VWR) while shaking.

[0138] Degradation assays

[0139] Before seeding cells, the 6-well plates were incubated with poly-L-Lysine for 30 min and washed twice with PBS. The expression of VHL and CRBN variants was performed as previously described. After 48h incubation, cells were washed with media twice, left in the incubator for 1 h, and washed again with media twice. Cells were treated with BPI to a final concentration of 1 pM. Following 16 hours of BPI treatment, the supernatant was carefully removed, cells washed with PBS, and lysed in 100 pl RIPA buffer (Sigma) with added Complete protease inhibitor (Roche) and Benzonase® Nuclease (VWR) while shaking.

[0140] Immunoblotting

[0141] Samples were mixed with NuPAGE LDS sample buffer (Thermofisher) and 5% BME, boiled at 90 C, and resolved by SDS-PAGE (4-12% NuPage gel, Thermofisher) with MES running buffer and transferred onto 0.45-pm PVDF membranes. Membranes were probed with the indicated antibodies in PBS-T or in TBS-T (tris buffered saline, 0.1 % Tween 20) containing 5% (w / v) bovine serum albumin (BSA) overnight at 4 °C. Detection was performed using horseradish peroxidase (HRP)-conjugated secondary antibodies in PBS-T for 1 h at 23 °C. ECL western blotting detection reagent (GE Life Sciences) was used for visualization according to the manufacturer’s protocol.

[0142] For chemiluminescent blots, detection was performed using horseradish peroxidase (HRP)- conjugated secondary antibodies in TBS-T for 1 h at 23 °C. ECL western blotting detection reagent (Thermo Fisher) was used for visualization according to the manufacturer’s protocol. Blots were visualized with a ChemiDoc imaging system (Bio-rad). Infrared fluorescent blots were visualized using a LI-COR Odyssey DLx system. Infrared fluorescent blots were carried out unless otherwise stated.

[0143] Antibodies

[0144] Expression of proteins was confirmed by immunoblotting with antibodies against HA tag (Roche, 11867423001 ), Actin(Proteintech, 66009-1-lg), FLAG (Sigma, F1804), BRD4 (Abeam, ab128874), BRD3 (Abeam, ab50818), BRD2 (Abeam, ab139690), and c-myc (CST, 5605S) and corresponding secondary HRP-linked antibodies (Anti-rat, CST, #7077; Anti-mouse, CST, #7076S; Anti-rabbit, CST, #7074S), or corresponding LI-COR infrared fluorescent secondary antibodies (Anti-rat IRDye 680RD #926-68076, Anti-mouse IRDye 800CW #926-32210, Antirabbit IRDye 680RD #926-68071).

[0145] Unnatural Amino acids

[0146] BocK was purchased from Bachem. CypK and BCNK were purchased from Sirius Fine Chemicals. UAA stock solutions were prepared for use in mammalian cell culture: (i) 100 mM BocK in 2N NaOH; (ii) 100 mM BCNK in 200 mM NaOH, 15% (v / v) DMSO. All solutions were stored at -20°C.

[0147] Immediately after adding to the cell culture medium, equal parts of HCI were added to neutralize pH. BocK was at a final concentration of 1 mM, while BCNK was at a final concentration of 0.25 mM.

[0148] Assessing incorporation efficiency of BCNK

[0149] Cell culture for assessing UAA incorporation into the corresponding target protein were performed in 6-well plates (30 °C, 5% CO2 incubator). Routinely, 500,000 cells were seeded per well and grown overnight. The next day, media was exchanged, and the unnatural amino acid was added to the fresh growth media. 4 hours later, cells were transiently transfected using 1.5 pg DNA per plasmid, 9 pl of Lipofectamine2000 (1 pg / pl) and 500 pl reduced serum media (Opti-MEM, Gibco). Following 48 hours of growth, the supernatant was carefully removed, cells washed with PBS, and lysed. Samples were prepared and utilised for Western Blot.

[0150] Assessment of in cellulo BPI conjugation efficiency with TA MRA -tetrazine

[0151] Transfection was performed with both plasmids, unless otherwise specified, following the optimised protocol. The next day, samples were washed twice with media, incubated for 1 h and washed twice with media again and were either treated with DMSO, or treated with BPI- PEG3-JQ1. Cells were then treated with 10 pM TAMRA-tetrazine for 1h. Lysates were prepared as described and concentrations were normalised to 2 mg / mL. Samples were resolved by SDS-PAGE and visualized with a with Typhoon Imager (Cytiva) using the 560-582 nm emission filter (Cy3).

[0152] Optimised transfection protocol

[0153] A transfection mix was prepared under the same conditions as before: 1.5 pg DNA per plasmid, 9 pL of Lipofectamine 2000 (1 pg / pL) in 500 pL reduced serum media (Opti-MEM, Gibco). ~1.5 x 106cells were then collected and media removed. The transfection mix was directly added into the cells in the absence of media and incubated for 15 min. Media containing 312.5 pM BCNK was added in a ratio of 4:1 with transfection mix / cells for a final cone, of 250 pM, and the sample was transferred to the 6-well plate.

[0154] Optimised VHL and CRBN degradation assays assessed by Western Blot

[0155] Transfection was performed with the optimised protocol. The next day, cells were washed with media twice, left in the incubator for 1 h, and washed again with media twice. Cells were treated with BPI at the specified concentrations. Following 1 hour of BPI treatment, the supernatant was carefully removed, cells were washed with PBS and lysed.

[0156] HiBiT degradation assay

[0157] HEK2393 cell lines had been genetically modified by CRISPR / Cas9 to constitutively express endogenous HiBiT-BRD4, HiBiT-BRD3 or HiBiT-BRD2. Routinely, 1 x 106cells were transfected with pBCNK-PylSTmam / pWT-PylSTmam and PPB-EF1-VHLTAGX, PPB-EF1-CRBNTAGX or PPB-EF1-UBE2D1TAGX plasmids (1.5 pg each) using the optimized protocol and left incubating overnight in media containing 250 pM BCNK. The next day, cells were washed with media twice, left in the incubator for 1 h, and washed again with media twice. Cells were then harvested, and 2 x 104cells were seeded into a 96-well plate per well (90 pL). Cells were treated with BPI-PEG2 or BPI-PEG3 (10 pL). After 1 h, each well was treated with Nano-Gio® HiBiT Lytic buffer (100 pL; Promega) as recommended by the manufacturer. Luminescence for each well was measured using Pherastar FS (BMG Labtech). For proteasome inhibition experiments, MG-132 (50 pM) was added after cell washing and incubated for 3 h. To inactivate cullin E3 ligases, the NEDD8 E1 inhibitor MLN4924 was added (1 pM) and incubated for 3 h. Cells were then washed a further two times. For control experiments incorporating BocK, the pWT-PylSTmam was used and 1 mM BocK was used instead of BCNK.

[0158] Fluorescence microscopy

[0159] Cells were transfected with PiggyBac and PylS plasmids following the optimised protocol. In parallel, a culture was set in the absence of the plasmids. The next day, cellular fluorescence was measured using a Zoe™ Fluorescent Cell Imager (Bio-rad). Images were recorded with a 100 mm resolution. Green channel was used to detect GFP expression. Red channel was used to detect mKate2 expression.

[0160] Fluorescence activated cell sorting

[0161] Using an MA900 cell sorter (Sony Biosciences), HEK293 cells were identified based on forward scatter area (FSC-A) and side scatter area (SSC-A) properties and single HEK293 cells were gated using proportionality of FSC-A against forward scatter width (FSC-W). HEK293 cells expressing double positively for GFP and mKate2 were collected; purity testing of this population showed the fraction of double positive single cells.

[0162] Deubiquitinating enzyme assay

[0163] Upon completion of the degradation assay, sample lysate concentrations were quantified using the Bradford assay. Lysates were then split and normalized to 2 mg / mL in a final volume of 25 pL. USP2, prepared as previously described2, was added to a final concentration of 5 pM, and its volume was accounted for within the total. Samples were incubated at 30 °C for 30 minutes at 300 rpm. Subsequently, lysates were resolved by reducing SDS-PAGE and visualized by Western blotting, as described previously.

[0164] Chemistry

[0165] Scheme 1- Synthesis of BPI-PEG3 (Compound of Formula I)

[0166] Scheme 2- Synthesis of BPI-PEG2 (Compound of Formula I)

[0167] BPI-PEG2

[0168] BPI-PEGX synthesis

[0169] All steps remained the same independently of which PEGX was used. Same concentrations, just changing the mass used for each PEGX according to their Molecular Weight. Below BPI- PEG3 as an example.

[0170] PEG3-JQ1

[0171] JQ1-acid (40 mg, 0.1 mmol) was dissolved in DCM (10 mL). NH2-PEG3 (22 mg, 0.15 mmol), HATLI (42 mg, 0.11 mmol) and DI PEA (34 u L, 0.2 mmol) were added to the solution. The mixture was left stirring at room temperature for 24h. The mixture was washed with HCI 2N to remove unreacted species. The organic phase was collected and concentrated in vacuo. The resulting crude was used for the next step. 4-NPCF coupling to PEG3-JQ1 followed by tetrazine coupling

[0172] PEG3-JQ1 crude was dissolved in 10 mL DCM. 4-NPCF (30 mg, 0.15 mmol) was added along with collidine (30 pL, 0.2 mmol). The reaction mixture was left to stir overnight. The mixture was then washed with 1 N HCI, and 2N NaOH four times. The organic phase was collected and concentrated in vacuo. The crude was used for the next step.

[0173] The crude was dissolved in 10 mL DCM. Collidine (30 uL, 0.23 mol) and DMAP (25 mg, 0.2 mmol) were added to the solution. Lastly, Tetrazine (30 mg, 0.11 mmol) was added. The reaction was left overnight. The product was purified by HPLC C18 column. It was obtained a final yield of 13.90 % for BPI-PEG3 and 35.91% for BPI-PEG2.

[0174] Citations

[0175] 1. Schreiber, S.L. (2024). Molecular glues and bifunctional compounds: Therapeutic modalities based on induced proximity. Cell Chem Biol 31, 1050-1063. 10.1016 / j.chembiol.2024.05.004.

[0176] 2. Zheng, N., and Shabek, N. (2017). Ubiguitin Ligases: Structure, Function, and Regulation. Annu Rev Biochem 86, 129-157. 10.1146 / annurev-biochem-060815- 014922.

[0177] 3. Li, W, Bengtson, M.H., Ulbrich, A., Matsuda, A., Reddy, V.A., Orth, A., Chanda, S.K., Batalov, S., and Joazeiro, C.A. (2008). Genome-wide and functional annotation of human E3 ubiguitin ligases identifies MULAN, a mitochondrial E3 that regulates the organelle's dynamics and signaling. PloS one 3, e1487.

[0178] 4. Bekes, M., Langley, D.R., and Crews, C.M. (2022). PROTAC targeted protein degraders: the past is prologue. Nature Reviews Drug Discovery 21, 181-200.

[0179] 5. Wang, S., He, F., Tian, C., and Sun, A. (2024). From PROTAC to TPD: Advances and Opportunities in Targeted Protein Degradation. Pharmaceuticals 17, 100.

[0180] 6. Guenette, R.G., Yang, S.W, Min, J., Pei, B., and Potts, PR. (2022). Target and tissue selectivity of PROTAC degraders. Chemical Society Reviews 51, 5740-5756. 7. Noblejas-Lopez, Tebar-Garcia, D., Lopez-Rosa, R., Alcaraz-Sanabria, A.,

[0181] Cristobal-Cueto, P., Pinedo-Serrano, A., Rivas-Garcia, L., and Galan-Moya, E.M. (2023). TACkling Cancer by Targeting Selective Protein Degradation. Pharmaceutics 15. 10.3390 / pharmaceutics 15102442.

[0182] 8. Macartney, T.J., Sapkota, G.P., and Fulcher, L.J. (2017). An Affinity-directed Protein Missile (AdPROM) System for Targeted Destruction of Endogenous Proteins. Bio Protoc 7, e2614. 10.21769 / BioProtoc.2614.

[0183] 9. Buckley, D.L., Raina, K., Darricarrere, N., Hines, J., Gustafson, J.L., Smith, I.E., Miah, A.H., Hading, J.D., and Crews, C.M. (2015). HaloPROTACS: Use of Small Molecule PROTACs to Induce Degradation of HaloTag Fusion Proteins. ACS Chem Biol 10, 1831-1837. 10.1021 / acschembio.5b00442.

[0184] 10. Nabet, B., Roberts, J.M., Buckley, D.L., Paulk, J., Dastjerdi, S., Yang, A., Leggett, A.L., Erb, M.A., Lawlor, M.A., Souza, A., et al. (2018). The dTAG system for immediate and target-specific protein degradation. Nat Chem Biol 14, 431-441. 10.1038 / s41589-018- 0021-8.

[0185] 11. Bond, A.G., Craigon, C., Chan, K.H., Testa, A., Karapetsas, A., Fasimoye, R., Macartney, T., Blow, J.J., Alessi, D.R., and Ciulli, A. (2021). Development of BromoTag: A "Bump-and-Hole"-PROTAC System to Induce Potent, Rapid, and Selective Degradation of Tagged Target Proteins. J Med Chem 64, 15477-15502. 10.1021 / acs.jmedchem.1 c01532.

[0186] 12. Mercer, J.A.M., DeCarlo, S.J., Roy Burman, S.S., Sreekanth, V., Nelson, A.T., Hunkeler, M., Chen, P.J., Donovan, K.A., Kokkonda, P., Tiwari, P.K., et al. (2024). Continuous evolution of compact protein degradation tags regulated by selective molecular glues. Science 383, eadk4422. 10.1126 / science.adk4422.

[0187] 13. Pinch, B.J., Buckley, D.L., Gleim, S., Brittain, S.M., Tandeske, L., D'Alessandro, P.L., Hauseman, Z.J., Lipps, J., Xu, L., Harvey, E.P., et al. (2022). A strategy to assess the cellular activity of E3 ligase components against neo-substrates using electrophilic probes. Cell Chem Biol 29, 57-66 e56. 10.1016 / j.chembiol.2021.08.007.

[0188] 14. Yi, H.B., Lee, S., Seo, K., Kim, H., Kim, M., and Lee, H.S. (2024). Cellular and Biophysical Applications of Genetic Code Expansion. Chem Rev 124, 7465-7530. 10.1021 / acs .chem rev.4c00112.

[0189] 15. Lang, K., Davis, L., Wallace, S., Mahesh, M., Cox, D.J., Blackman, M.L., Fox, J.M., and Chin, J.W (2012). Genetic encoding of bicyclononynes and trans-cyclooctenes for site-specific protein labeling in vitro and in live mammalian cells via rapid fluorogenic Diels-Alder reactions. Journal of the American Chemical Society 134, 10317-10320. Filippakopoulos, P, Qi, J., Picaud, S., Shen, Y, Smith, W.B., Fedorov, O., Morse, E.M., Keates, T, Hickman, T.T., Felletar, I., et al. (2010). Selective inhibition of BET bromodomains. Nature 468, 1067-1073. 10.1038 / nature09504. Lang, K., and Chin, J.W. (2014). Cellular incorporation of unnatural amino acids and bioorthogonal labeling of proteins. Chem Rev 114, 4764-4806. 10.1021 / cr400355w. Zengerle, M., Chan, K.H., and Ciulli, A. (2015). Selective Small Molecule Induced

[0190] Degradation of the BET Bromodomain Protein BRD4. ACS Chem Biol 10, 1770-1777. 10.1021 / acschembio.5b00216. Winter, G.E., Buckley, D.L., Paulk, J., Roberts, J.M., Souza, A., Dhe-Paganon, S., and Bradner, J.E. (2015). DRUG DEVELOPMENT. Phthalimide conjugation as a strategy for in vivo target protein degradation. Science 348, 1376-1381.

[0191] 10.1126 / science.aab1433.

Claims

Claims1- A method of inducing proximity between a first protein (Pi) and a target protein (Pt), the method comprising: a) Preparing a recombinant protein (Pr) by genetic code expansion to encode into the first protein (Pi) an unnatural amino acid (UAA) comprising a bioorthogonal reactive group capable of reacting with a proximity-inducing molecule of formula (I) :A-L-D(I) wherein:A is a moiety capable of binding to the target protein (Pt) or to a tag located on the target protein (Pt);L is a linker; andD is a moiety capable of undergoing a bioorthogonal reaction with the unnatural amino acid (UAA) present in the recombinant protein (Pr);b) Forming a modified protein (Pm) by contacting the recombinant protein (Pr) with the proximityinducing molecule of Formula (I) to irreversibly form a covalent bond between the unnatural amino acid (UAA) of the recombinant protein (Pr) and moiety D of the proximity-inducing molecule in a bioorthogonal reaction; c) Forming a ternary complex by allowing the modified protein (Pm) to bind to the target protein (Pt).2- The method of claim 1 , wherein the method is performed in vitro or ex vivo.3- The method of claim 1 or 2, wherein the bioorthogonal reaction is selected from: native chemical ligation, Staudinger ligation, copper-catalysed azide-alkyne cycloaddition, Copper-Free Azide-Alkyne Cycloaddition, strain-promoted [3 + 2] reaction, Inverse Electron Demand Diels-Alder (I EDDA) reaction, tetrazole ligation, oxime ligation, isocyanide click reaction, metal-catalysed coupling reaction, hydrazone ligation, photoinducible bioorthogonal reactions, optionally wherein the bioorthogonal reaction is an Inverse Electron Demand Diels-Alder (IEDDA) reaction.4- The method of any preceding claim, wherein D is a moiety comprising a diene or wherein D is a moiety comprising a dienophile or a heterodienophile.5- The method of any preceding claim wherein one of:(i) D is a moiety comprising a diene capable of reacting in a I EDDA reaction, optionally wherein D is a moiety comprising a diene present in an electron-deficient heterocycle, and / or nitrogen-containing diene, further optionally wherein D is a moiety comprising a cyclic azine selected from a diazine, a triazine, and a tetrazine, optionally wherein the diene of D is a tetrazine; further optionally wherein D is, where R is selected from H, methyl or ethyl; or(ii) D is a moiety comprising a dienophile capable of reacting with a diene in a I EDDA reaction, optionally wherein at least one of:D is a moiety comprising an alkene or an alkyne,D is a moiety comprising a strained dienophile,D is a moiety comprising a group selected from: cyclopropene, transcyclooctene TCO, norbornene, and bicyclo [6.1.0] nonyne,D is a moiety comprising a bicyclo [6.1.0] nonyne group.6- The method of any preceding claim, wherein L is a linker comprising an aliphatic chain or a PEG group, optionally wherein L is a linker of formula -(CH2-CH2-O)n-, wherein n is selected from 1 to 20, preferably wherein n is selected from 2 to 4, more preferably wherein n is selected from 2 and 3.7- The method of any preceding claim, wherein at least one of:(i) A is a ligand for the target protein (Pt) or A is a moiety capable of binding to a protein tag which is configured to be located on the target protein (Pt); and / or(ii) A is moiety configured to bind to the target protein (Pt) or polypeptide which is to be degraded by a ubiquitin ligase; and / or(iii) A is:(a) a chemical moiety configured to bind to a protein within the bromo- and Extra-terminal (BET) family of proteins; or(b) a chemical moiety configured to bind to a protein within the bromo- and Extra-terminal (BET) family of proteins independently selected from: BRD2, BRD3 and BRD4; or(c) a chemical moiety which is configured to selectively induce degradation of the BRD4 protein within the bromo- and Extra-terminal (BET) family of proteins;(d) an E3 ligase, optionally wherein the E3 ligase is selected from VHL and CRBN; or(iv) A is linked to L through an amide bond.8- The method of any preceding claim, wherein one of:(i) moiety D of the proximity-inducing molecule comprises a dienophile capable of reacting with a diene present in the recombinant protein (Pr) in a I EDDA reaction and wherein the unnatural amino acid (UAA) of the recombinant protein (Pr) comprises a diene capable of reacting with the dienophile in moiety D of the proximity-inducing molecule in a I EDDA reaction; or(ii) moiety D of the proximity-inducing molecule comprises a diene capable of reacting with a dienophile present in the recombinant protein (Pr) in a I EDDA reaction and wherein the unnatural amino acid (UAA) of the recombinant protein (Pr) comprises a dienophile capable of reacting with the diene in moiety D of the proximity-inducing molecule in a I EDDA reaction.9- The method of any preceding claim, wherein the method further comprises attaching a protein tag to the target protein (Pt) prior to step c) and wherein moiety A in the proximity-inducing molecule is a ligand for the protein tag, optionally wherein the tag is a dTag or a bromoTag.10- The method of any preceding claim, wherein one of the target protein (Pt) or the first protein (P1) is a protein subject to a post-translational modification (PTM) and the other of the target protein (Pt) or the first protein (P1) is an enzyme involved in the post- translational modification.11- A method of evaluating an induced proximity “interaction” between a first protein (P1) and a target protein (Pt) in a cell, the method comprising: a) Providing a cell line;b) Genetically engineering the cell line in vitro by genetic code expansion to encode an unnatural amino acid (UAA) comprising a bioorthogonal reactive group into the first protein (P1); c) Allowing the cell line to express a recombinant protein (Pr) of the first protein (P1) comprising the unnatural amino acid (UAA); d) Providing to the cell line a proximity-inducing molecule of formula (I) :A-L-D(I) wherein:A is a moiety capable of binding to the target protein (Pt) or to a tag located on the target protein (Pt);L is a linker; andD is a moiety capable of undergoing a bioorthogonal reaction with the unnatural amino acid (UAA) present in the recombinant protein (Pr); to form a ternary complex between the recombinant protein (Pr), the proximity-inducing molecule of Formula (I) and the target protein (Pt); and e) Testing the interaction between the recombinant protein (Pr) and the target protein (Pt) facilitated by the close proximity between them in the ternary complex.12- The method of claim 11 , wherein the method is performed in vitro or ex vivo.13- The method of claim 11 or 12, wherein genetically engineering the cell line by genetic code expansion to encode an unnatural amino acid (UAA) involves transfecting the cell line with an orthogonal tRNA-synthetase / tRNAstoPcodon pair configured to incorporate the unnatural amino acid (UAA) into the first protein (Pt) to form a recombinant protein (Pr) in response to an amber TAG stop codon, optionally wherein the tRNA-synthetase / tRNAstop codon pair is a tRNA-synthetase / tRNAcuA pair, further optionally wherein it is a mutant pyrrolysl-tRNA synthetase / tRNAcuA pair, further optionally wherein the tRNA-synthetase / tRNAcuA pair is supplied in an expression vector, optionally wherein the expression vector is pBCNK-PylSTmam.14- The method of any one of claims 11 to 13, further comprising cloning amber codon protein mutants with a C-terminal HA tag into a separate expression vector optionally wherein the separate expression vector is the PiggyBac expression vector.15- The method of any one of claims 11 to 14, wherein the unnatural amino acid (UAA) of the recombinant protein (Pr) comprises a moiety comprising a dienophile capable of reacting with a diene in an Inverse Electron Demand Diels-Alder (I EDDA) reaction, optionally wherein the UAA is configured to react with a diene present in the proximity-inducing molecule of formula (I) (e.g. in moiety D), optionally wherein one of:(i) the unnatural amino acid (UAA) of the recombinant protein (Pr) comprises a dienophile or a heterodienophile, optionally wherein the unnatural amino acid (UAA) comprises a moiety comprising a strained dienophile or strained heterodienophile, further optionally wherein the unnatural amino acid (UAA) comprises a moiety comprising a group selected from: cyclopropene, transcyclooctene TCO, norbornene, and bicyclo [6.1.0] nonyne, further optionally wherein the unnatural amino acid is bicyclo [6.1.0] nonyne-L-lysine (BCNK); or(ii) the unnatural amino acid (UAA) of the recombinant protein (Pr) comprises a moiety comprising diene capable of reacting with a dienophile in an Inverse Electron Demand Diels-Alder (I EDDA) reaction, optionally wherein the diene is configured to react with a dienophile present in the proximity-inducing molecule, optionally wherein the diene is present in an electron-deficient heterocycle, further optionally wherein at least one of:(a) the diene is a nitrogen-containing diene;(b) the UAA of the recombinant protein (Pr) comprises a moiety comprising a cyclic azine, optionally wherein the cyclic azine is selected from a diazine, a triazine, and a tetrazine;(c) the UAA comprises the moietywhere R is selected from H, methyl or ethyl.16- The method of any one of claims 11 to 15, wherein the induced proximity interaction is a protein post-translational modification mediated by an enzyme, optionally wherein the post-translational modification is selected from ubiquitination, acetylation, nitrosylation, methylation, hydroxylation, phosphorylation, SUMOylation, glycosylation, AMPylation, lipidation, proteolysis, redox post-translational modification and deamidation.17- A proximity-inducing molecule of formula (I):A-L-D(I) for inducing proximity between a first protein (Pi) and a target protein (Pt), wherein:A is a moiety capable of binding to the target protein (Pt) or to a tag located on the target protein;L is a linker; andD is a moiety capable of undergoing a bioorthogonal reaction with a bioorthogonal reactive group present in the first protein (Pi) to irreversibly form a covalent bond with the first protein (Pi), wherein D comprises a group selected from: transcyclooctene TCO, andwhere R is selected from H, methyl or ethyl.18- The proximity-inducing molecule of claim 17, wherein the bioorthogonal reaction is selected from: native chemical ligation, Staudinger ligation, copper-catalysed azidealkyne cycloaddition, Copper-Free Azide-Alkyne Cycloaddition, strain-promoted [3 + 2] reaction, Inverse Electron Demand Diels-Alder (I EDDA) reaction, tetrazole ligation, oxime ligation, isocyanide click reaction, metal-catalysed coupling reaction, hydrazone ligation, photoinducible bioorthogonal reactions, optionally wherein the bioorthogonal reaction is an Inverse Electron Demand Diels-Alder (I EDDA) reaction.19- The proximity-inducing molecule of claim 17 or 18, wherein L is a linker comprising an aliphatic chain or a PEG group, optionally wherein L is a linker of formula -(CH2-CH2- O)n-, wherein n is selected from 1 to 20, preferably wherein n is selected from 2 to 4, more preferably wherein n is selected from 2 and 3.20- The proximity-inducing molecule of any one of claims 17 to 19, wherein A is a ligand for the target protein; or wherein A is a moiety capable of binding to a protein tag which is configured to be located on the target protein.21- The proximity-inducing molecule of any one of claims 17 to 20, wherein at least one of:(i) A is ligand configured to bind to a target protein or polypeptide which is to be degraded by a ubiquitin ligase, optionally wherein A is linked to L through an amide bond(ii) A is a chemical moiety configured to bind to a protein within the bromo- and Extra-terminal (BET) family of proteins; or(iii) A is a chemical moiety configured to bind to a protein within the bromo- and Extra-terminal (BET) family of proteins independently selected from: BRD2, BRD3 and BRD4; or(iv) A is a chemical moiety which is configured to selectively induce degradation of the BRD4 protein within the bromo- and Extra-terminal (BET) family of proteins;(v) A is an E3 ligase, optionally wherein the E3 ligase is selected from VHL and CRBN; or(vi) A is a derivative of any one of (i) to (v) which retains the function,(vii) A is linked to L through an amide bond.22- The proximity-inducing molecule of any one of claims 17 to 21 , wherein the molecule of formula A is selected from selected from:pharmaceutically acceptable, salt, enantiomer, stereoisomer, hydrate, solvate, or polymorph.

Images