SYSTEM FOR COVALENTLY LINKING PROTEINS

Abstract

The present invention relates to a system for generating intermolecular covalent bonds (e.g. amide, e.g. isopeptide bonds) between polypeptides. In particular, it provides the use of a chimeric protein to generate an anhydride group on a polypeptide for the formation of a covalent bond, wherein the chimeric protein comprises (i) a domain comprising the polypeptide and (ii) a domain comprising a self-processing module that contains an N-terminal dipeptide of aspartate or glutamate and proline (D/E-P), wherein (i) and (ii) are linked by a peptide bond between the aspartate or glutamate residue at the N-terminus of (ii) and the amino acid at the C-terminus of (i) and wherein the self-processing module cleaves the peptide bond between the proline residue and the aspartate or glutamate residue in the self-processing module to release the polypeptide and generate the anhydride group on the aspartate or glutamate residue.

Claims

1. Use of a chimeric protein to generate an anhydride group on a polypeptide for the formation of a covalent bond, wherein the chimeric protein comprises: (i) a domain comprising the polypeptide; and (ii) a domain comprising a self-processing module that contains an N-terminal dipeptide of aspartate or glutamate and proline (D/E-P), wherein (i) and (ii) are linked by a peptide bond between the aspartate or glutamate residue at the N-terminus of (ii) and the amino acid at the C-terminus of (i) and wherein the self-processing module cleaves the peptide bond between the proline residue and the aspartate or glutamate residue in the self-processing module to release the polypeptide and generate the anhydride group on the aspartate or glutamate residue.

2. The use of claim 1 further comprising using the anhydride group on the polypeptide to: (i) form an intramolecular covalent bond in the polypeptide; or (ii) conjugate the polypeptide to a second polypeptide via a covalent bond.

3. A method of producing an anhydride group on a polypeptide for use in directing the formation of a covalent bond comprising: (a) providing a chimeric protein comprising: (i) a domain comprising the polypeptide; and (ii) a domain comprising a self-processing module that contains an N-terminal dipeptide of aspartate or glutamate and proline (D/E-P), wherein (i) and (ii) are linked by a peptide bond between the aspartate or glutamate residue at the N-terminus of (ii) and the amino acid at the C-terminus of (i) and wherein the self-processing module cleaves the peptide bond between the proline residue and the aspartate or glutamate residue under suitable conditions; (b) inducing the self-processing module to cleave the peptide bond between the proline residue and the aspartate or glutamate residue to release the polypeptide and generate the anhydride group on the aspartate or glutamate residue, thereby producing a polypeptide comprising an anhydride group.

4. The method of claim 3 further comprising a step of isolating the polypeptide comprising an anhydride group and/or storing the polypeptide comprising an anhydride group under conditions in which the anhydride group is stable.

5. The method of claim 3, being a method of forming an intramolecular covalent bond in a polypeptide (e.g. a method of cyclizing a polypeptide) comprising: (a) providing a chimeric protein comprising: (i) a domain comprising the polypeptide; and (ii) a domain comprising a self-processing module that contains an N-terminal dipeptide of aspartate or glutamate and proline (D/E-P), wherein (i) and (ii) are linked by a peptide bond between the aspartate or glutamate residue at the N-terminus of (ii) and the amino acid at the C-terminus of (i) and wherein the self-processing module cleaves the peptide bond between the proline residue and the aspartate or glutamate residue under suitable conditions; and (b) inducing the self-processing module to cleave the peptide bond between the proline residue and the aspartate or glutamate residue to release the polypeptide and generate an anhydride group on the aspartate or glutamate residue that reacts with a functional group in the polypeptide to form a covalent bond, thereby forming an intramolecular covalent bond in the polypeptide (e.g. thereby cyclizing the polypeptide).

6. Use of claim 1 or 2, being the use of a chimeric protein to conjugate a first polypeptide to a second polypeptide via an isopeptide bond, wherein the chimeric protein comprises: (i) a domain comprising the first polypeptide; and (ii) a domain comprising a self-processing module that contains an N-terminal dipeptide of aspartate or glutamate and proline (D/E-P), wherein (i) and (ii) are linked by a peptide bond between the aspartate or glutamate residue at the N-terminus of (ii) and the amino acid at the C-terminus of (i) and wherein the self-processing module cleaves the peptide bond between the proline residue and the aspartate or glutamate residue in the self-processing module to release the first polypeptide and generate an anhydride group on the aspartate or glutamate residue at the C-terminus of the first polypeptide that reacts with a functional group on the second polypeptide to form the covalent bond.

7. The use of claim 6, wherein the second polypeptide binds non-covalently to the chimeric polypeptide via an interaction with the domain comprising the first polypeptide.

8. The method of claim 3, being a method of conjugating a first polypeptide to a second polypeptide via a covalent bond comprising: (a) providing a chimeric protein comprising: (i) a domain comprising the first polypeptide; and (ii) a domain comprising a self-processing module that contains an N-terminal dipeptide of aspartate or glutamate and proline (D/E-P), wherein (i) and (ii) are linked by a peptide bond between the aspartate or glutamate residue at the N-terminus of (ii) and the amino acid at the C-terminus of (i) and wherein the self-processing module cleaves the peptide bond between the proline residue and the aspartate or glutamate residue under suitable conditions; (b) contacting the chimeric protein of (a) with the second polypeptide, wherein the second polypeptide binds non-covalently to (i); (c) inducing the self-processing module to cleave the peptide bond between the proline residue and the aspartate or glutamate residue to release the first polypeptide and generate an anhydride group on the aspartate or glutamate residue that reacts with a functional group on the second polypeptide to form the covalent bond, thereby conjugating the first and second polypeptides.

9. The use or method of any preceding claim, wherein the covalent bond is an amide bond.

10. The use or method of any preceding claim, wherein the functional group is an amine.

11. The use of any one of claim 6, 7, 9 or 10, or the method of any one of claims 8 to 10, wherein the second polypeptide is attached to the surface of a cell or is in the extracellular matrix.

12. The use or method of claim 11, wherein the cell is located in a subject or the extracellular matrix is located in an organ and/or subject.

13. The use of any one of claim 6, 7, 9 or 10 or method of any one of claims 8 to 10, wherein the second polypeptide is attached to an exosome, virus, virus-like particle, nanoparticle or solid support.

14. The use or method of any preceding claim, wherein the self-processing module comprises: (1) an amino acid sequence as set forth in SEQ ID NO: 1; (2) a portion of (1) comprising an amino acid sequence as set forth in SEQ ID NO: 5; (3) an amino acid sequence with at least 80% sequence identity to an amino acid sequence as set forth in SEQ ID NO: 1 or 2; or (4) a portion of (3) comprising an amino acid sequence with at least 80% sequence identity to an amino acid sequence as set forth in SEQ ID NO: 5 or 6, wherein the amino acid sequence comprises aspartate or glutamate at position 1, proline at position 2 and one or more of the following: 1) alanine at position 17; 2) alanine at position 23; 3) arginine at position 28; 4) glutamine at position 30; and wherein the self-processing module cleaves the peptide bond between the first and second amino acids of the domain comprising a self-processing module under suitable conditions.

15. The use or method of any preceding claim, wherein the self-processing module comprises: (1) an amino acid sequence as set forth in SEQ ID NO: 1; (2) a portion of (1) comprising an amino acid sequence as set forth in SEQ ID NO: 5; (3) an amino acid sequence with at least 99% sequence identity to an amino acid sequence as set forth in SEQ ID NO: 1; or (4) a portion of (3) comprising an amino acid sequence with at least 99% sequence identity to an amino acid sequence as set forth in SEQ ID NO: 5, wherein the amino acid sequence comprises aspartate or glutamate at position 1 and proline at position 2; and wherein the self-processing module cleaves the peptide bond between the first and second amino acids of the domain comprising a self-processing module under suitable conditions.

16. A composition comprising: (i) a polypeptide having an anhydride group on a C-terminal aspartate or glutamate residue, wherein the aspartate or glutamate residue in the polypeptide is not present at the equivalent position in the amino acid sequence of the corresponding endogenous polypeptide or portion thereof; and (ii) a solvent that prevents hydrolysis or reaction of the anhydride group.

17. A polypeptide (e.g. a cyclized polypeptide) comprising an intramolecular covalent bond formed between an aspartate or glutamate residue and a functional group (e.g. an amine on a lysine residue or at the N-terminus), wherein: (i) the aspartate or glutamate residue in the polypeptide is not present in the amino acid sequence of the corresponding endogenous polypeptide or portion thereof; and (ii) the functional group in the polypeptide is present at an equivalent position in the corresponding endogenous polypeptide or portion thereof.

18. A product comprising a first polypeptide conjugated to a second polypeptide via a covalent bond between an aspartate or glutamate residue in the first polypeptide and a functional group in the second polypeptide, wherein: (i) the aspartate or glutamate residue in the first polypeptide is not present at the equivalent position in the amino acid sequence of the corresponding endogenous polypeptide or portion thereof; and (ii) the functional group in the second polypeptide is present at the equivalent position in the amino acid sequence of the corresponding endogenous polypeptide.

19. The product of claim 18, wherein: (i) the first polypeptide comprises an amino acid sequence that corresponds to the amino acid sequence of an endogenous polypeptide or a portion thereof except that the endogenous polypeptide or portion thereof does not contain an aspartate or glutamate residue at its C-terminus; and (ii) the second polypeptide comprises an amino acid sequence that corresponds to the amino acid sequence of an endogenous polypeptide or a portion thereof which contains a functional group at an equivalent position to the functional group in the second polypeptide.

20. The polypeptide of claim 17 or product of claim 18 or 19, wherein the covalent bond is an amide bond.

21. The polypeptide of claim 17 or 20 or product of any one of claims 18 to 20, wherein the functional group is an amine.

22. A pharmaceutical composition comprising: (a)(1) a chimeric protein comprising: (i) a domain comprising the first polypeptide; and (ii) a domain comprising a self-processing module that contains an N-terminal dipeptide of aspartate or glutamate and proline (D/E-P), wherein (i) and (ii) are linked by a peptide bond between the aspartate or glutamate residue at the N-terminus of (ii) and the amino acid at the C-terminus of (i) and wherein the self-processing module cleaves the peptide bond between the proline residue and the aspartate or glutamate residue under suitable conditions; (2) a polypeptide comprising an anhydride group on a C-terminal aspartate or glutamate residue, wherein the aspartate or glutamate residue in the polypeptide is not present at the equivalent position in the amino acid sequence of the corresponding endogenous polypeptide or portion thereof (e.g. obtained by the method of any one of claim 3, 4, 9 or 10); (3) a composition as defined in claim 16; (4) a polypeptide as defined in claim 17, 20, or 21; or (5) a product as defined in any one of claims 18 to 21; and (b) one or more pharmaceutically acceptable excipients and/or diluents.

23. A pharmaceutical composition as defined in claim 22 for use in therapy or diagnosis.

24. A method of treating a disease in a subject comprising administering to a subject in need thereof a therapeutically effective amount of a pharmaceutical composition of claim 22, thereby treating the disease.

25. The use, method or pharmaceutical composition of any preceding claim, wherein the chimeric protein comprises N-terminus to C-terminus: (i) a domain comprising a polypeptide; and (ii) a domain comprising a self-processing module comprising: (1) an amino acid sequence as set forth in any one of SEQ ID NOs: 1-4; (2) a portion of (1) comprising an amino acid sequence as set forth in any one of SEQ ID NOs: 5-8; (3) an amino acid sequence with at least 60% sequence identity to an amino acid sequence as set forth in any one of SEQ ID NOs: 1-4; or (4) a portion of (3) comprising an amino acid sequence with at least 60% sequence identity to an amino acid sequence as set forth in any one of SEQ ID NOs: 5-8, wherein the first (N-terminal) amino acid of the domain comprising a self-processing module is an aspartate or glutamate and the second amino acid of the domain comprising a self-processing module is proline; and wherein the self-processing module cleaves the peptide bond between the first and second amino acids of the domain comprising a self-processing module under suitable conditions.

26. The use, method or pharmaceutical composition of claim 25, wherein the chimeric protein further comprises a linker between (i) and (ii), preferably wherein the linker comprises the motif X.sub.1X.sub.2X.sub.3, wherein: (a) X.sub.1 and X.sub.2 are independently selected from any amino acid, preferably G and S; and (b) X.sub.3 is selected from R, N, Q, F, V, H, Y or W, preferably V, H, Y or W.

27. A chimeric protein comprising N-terminus to C-terminus: (i) a domain comprising a polypeptide; (ii) a domain comprising a linker; and (iii) a domain comprising a self-processing module comprising: (1) an amino acid sequence as set forth in any one of SEQ ID NOs: 1-4; (2) a portion of (1) comprising an amino acid sequence as set forth in any one of SEQ ID NOs: 5-8; (3) an amino acid sequence with at least 60% sequence identity to an amino acid sequence as set forth in any one of SEQ ID NOs: 1-4; or (4) a portion of (3) comprising an amino acid sequence with at least 60% sequence identity to an amino acid sequence as set forth in any one of SEQ ID NOs: 5-8, wherein the first (N-terminal) amino acid of the domain comprising a self-processing module is an aspartate or glutamate and the second amino acid of the domain comprising a self-processing module is proline; and wherein the self-processing module cleaves the peptide bond between the first and second amino acids of the domain comprising a self-processing module under suitable conditions.

28. The chimeric protein of claim 27, wherein the linker comprises the motif X.sub.1X.sub.2X.sub.3, wherein: (a) X.sub.1 and X.sub.2 are independently selected from any amino acid, preferably G and S; and (b) X.sub.3 is selected from R, N, Q, F, V, H, Y or W, preferably V, H, Y or W.

29. The chimeric protein of claim 27 or 28, wherein the self-processing module comprises: (1) an amino acid sequence as set forth in SEQ ID NO: 1; (2) a portion of (1) comprising an amino acid sequence as set forth in SEQ ID NO: 5; (3) an amino acid sequence with at least 80% sequence identity to an amino acid sequence as set forth in SEQ ID NO: 1 or 2; or (4) a portion of (3) comprising an amino acid sequence with at least 80% sequence identity to an amino acid sequence as set forth in SEQ ID NO: 5 or 6, wherein the amino acid sequence comprises aspartate or glutamate at position 1, proline at position 2 and one or more of the following: (1) alanine at position 17; (2) alanine at position 23; (3) arginine at position 28; (4)glutamine at position 30; and wherein the self-processing module cleaves the peptide bond between the first and second amino acids of the domain comprising a self-processing module under suitable conditions.

30. The chimeric protein of any one of claims 27 to 29, wherein the self-processing module comprises: (1) an amino acid sequence as set forth in SEQ ID NO: 1; (2) a portion of (1) comprising an amino acid sequence as set forth in SEQ ID NO: 5; (3) an amino acid sequence with at least 99% sequence identity to an amino acid sequence as set forth in SEQ ID NO: 1; or (4) a portion of (3) comprising an amino acid sequence with at least 99% sequence identity to an amino acid sequence as set forth in SEQ ID NO: 5, wherein the amino acid sequence comprises aspartate or glutamate at position 1 and proline at position 2; and wherein the self-processing module cleaves the peptide bond between the first and second amino acids of the domain comprising a self-processing module under suitable conditions.

31. Use of a chimeric protein as defined in any one of claims 27 to 30 to isolate (e.g. purify) a desired polypeptide, wherein the polypeptide in domain (i) of the chimeric protein is the desired polypeptide.

32. A method of isolating (e.g. purifying) a desired polypeptide comprising: a) providing a sample comprising a chimeric protein of any one of claims 27 to 30, wherein the polypeptide in domain (i) of the chimeric protein is the desired polypeptide; b) contacting the sample of a) with a solid support under conditions that enable said chimeric protein to selectively bind to said solid support, thereby forming a non-covalent complex between said chimeric protein and the solid support; c) washing the solid support with a buffer; d) inducing the self-processing module to cleave the peptide bond between the proline residue and the aspartate or glutamate residue (i.e. between residues 1 and 2) to release the desired polypeptide; and e) separating the desired polypeptide from the solid support.

33. The chimeric protein of any one of claims 27 to 30, wherein the polypeptide in domain (i) of the chimeric protein is a growth factor, cytokine, chemokine or a portion or derivative thereof.

34. The chimeric protein of claim 33, wherein the growth factor, cytokine or chemokine is selected from any one of TGFα, epigen, epiregulin, EGF, HB-EGF, TGFβ, TNFα, IL1RA, IL-1β, IL-2, IL-4, IL-5, IL-6, IL-7, IL-8 (CXCL8), IL-9, IL-10, IL-12, IL-13, IL-15, IL-17, CCL11, BasicFGF, G-CSF, GM-CSF, INFα, INFγ, CXCL10, CCL2, CCL3, CCL4, PDGF-β, CCL5, VEGF or a functional portion or derivative thereof, preferably TGFα or a functional portion or derivative thereof.

35. The chimeric protein of claim 33 or 34, wherein the polypeptide in domain (i) of the chimeric protein comprises an amino acid sequence as set forth in SEQ ID NO: 17.

36. The chimeric protein of any one of claims 33 to 35, wherein the chimeric protein comprises an amino acid sequence as set forth in SEQ ID NO: 16.

37. A nucleic acid molecule encoding the chimeric protein of any one of claims 27 to 30 or 33 to 36.

Description

[0306] The invention will now be described in more detail in the following non-limiting Examples with reference to the following drawings:

[0307] FIG. 1 shows: (a) a schematic of the FrpC self-processing module (SPM), which catalyzes autoproteolytic cleavage at an Asp-Pro bond, induced by calcium. The resultant anhydride enables protein-protein crosslinking via reaction with nucleophilic side-chains; (b) a schematic of the chimeric protein (NeissLock probe) and its utility to conjugate two polypeptides. The SPM is recombinantly fused to a binding protein which docks with the target protein. Adding calcium promotes generation of the anhydride and the binding protein then can form a covalent bond to the target protein; (c) a photograph of an SDS-PAGE gel with Coomassie staining showing a time-course of SPM cleavage with Ala preceding Asp-Pro; and (d) a histogram of SPM cleavage rate with each residue before Asp-Pro, moving from the least cleaved residue at 60 min on the left to the most cleaved residue on the right (mean of triplicate±1 s.d.; some error bars are too small to be visible).

[0308] FIG. 2 shows (a) a diagrammatic representation of the considerations for binder/target complex selection. The target protein should have a lysine or N-terminal amine in proximity and sterically accessible to the C-terminus of the binder protein, to enable reaction with the anhydride formed during activation. To avoid quenching by self-reaction, the binder protein should not feature a lysine close to its own C-terminus; and (b) a flow chart of the disCrawl distance database pipeline, i.e. the computer implemented method of selecting polypeptides for use in the method of the invention.

[0309] FIG. 3 shows (a) a photograph of an SDS-PAGE gel with Coomassie staining showing Ornithine Decarboxylase (ODC) reacted covalently with Ornithine decarboxylase antizyme (OAZ). ODC and OAZ-Y-SPM (with a Tyr before the SPM) were incubated at each 10 μM for 16 h with or without calcium, boiled in SDS loading buffer; (b) intact protein electrospray ionization MS confirms covalent coupling of OAZ-Y to ODC, with a loss of water (−18) indicating isopeptide formation and (c) a photograph of an SDS-PAGE gel with Coomassie staining showing specific reaction by NeissLock. OAZ-GSY-SPM was incubated overnight with each protein at 10 μM with the cognate partner ODC or non-cognate DogTag-MBP or SpyTag003-sfGFP. All lanes are in the presence of calcium. Samples were analyzed by SDS-PAGE with Coomassie staining.

[0310] FIG. 4 shows (a) a photograph of an SDS-PAGE gel with Coomassie staining showing a time-course for OAZ-Y-SPM coupling. OAZ-Y-SPM was incubated with ODC for the indicated time in the presence of Ca.sup.2+; (b) a spacer increases cleavage efficiency. ODC was incubated with OAZ-Y-SPM or OAZ-GSY-SPM for the indicated time in the presence of Ca.sup.2+ and the extent of cleavage was determined by SDS-PAGE with Coomassie staining (mean of triplicate±1 s.d.; some error bars are too small to be visible); and (c) pH-dependence of cleavage. OAZ-GSY-SPM was incubated with Ca.sup.2+ for the indicated time at the indicated pH and cleavage of SPM was determined (mean of triplicate±1 s.d.).

[0311] FIG. 5 shows (a) a photograph of an SDS-PAGE gel with Coomassie staining showing disruption of ODC/OAZ affinity blocked conjugation. OAZ-GSY-SPM or the non-binding OAZ-GSY-SPM was incubated with ODC along with Ca.sup.2+ with each protein at 0.5 μM for 0 or 60 min, before SDS-PAGE with Coomassie staining; and (b) a photograph of an SDS-PAGE gel with Coomassie staining showing different sites on ODC can react with OAZ. OAZ-GSY-SPM was incubated with the indicated ODC mutant overnight at 37° C. before SDS-PAGE with Coomassie staining.

[0312] FIG. 6 shows (a) a photograph of an SDS-PAGE gel with Coomassie staining showing NeissLock reaction to soluble epidermal growth factor receptor (EGFR). TGFα-GSY-SPM was incubated with sEGFR with or without Ca.sup.2+ for 90 minutes at 37° C. Subsequently, samples were deglycosylated with PNGase F Kit (NEB), i.e. denatured with Glycoprotein Denaturing Buffer and digested at 37° C. with PNGase F before SDS-PAGE with Coomassie staining; (b) a photograph of a Western blot showing NeissLock conjugation to EGFR on cells. A431 cells were incubated with TGFα-GSY-SPM for 5 min at 37° C. or 30 min at 4° C. according to the indicated times. Samples were washed and optionally incubated with Ca.sup.2+ for 15 min at 37° C. or 30 min at 4° C. according to the indicated times. Non-processing TGFα-GSY-[DA]SPM or non-binding TGFα[R42A]-GSY-SPM controls were tested. Cells were lysed and Western blot was performed against Transforming Growth

[0313] Factor-alpha (TGFα); and (c) a photograph of a Western blot showing condition-dependence of reaction with EGFR. A431 cells were incubated with TGFα-GSY-SPM for varying times at different temperatures, before Western blot against TGFα. 1,2: Dynasore treated, 5 min binding to TGFα-GSY-SPM at 37° C., washed, with or without calcium for 15 min at 37° C. 3,4: As 2,1, respectively without prior dynasore treatment. 5: As 4, but cells were co-incubated with TGFα-GSY-SPM and calcium at the same time. 6,7: Cells were incubated with TGFα-GSY-SPM at 4° C. for 30 min, then washed, then without or with 30 min calcium incubation. 8: As 7, but cells were not washed before adding calcium. 9: As 5, but co-incubation for 30 min at 4 ° C. C: Control without TGFα-GSY-SPM.

[0314] FIG. 7 shows (a) introduction of C157A in OAZ decreased protein aggregation and improved cleavage rate. OAZ-GSY-SPM or the C157A mutant was incubated with Ca.sup.2+ for the indicated time at 37° C. and cleavage was analyzed by SDS-PAGE with Coomassie staining; and (b) the effect of SPM truncations on cleavage. OAZ-GSY-SPM or various modifications (illustrated on right) were incubated at 37° C. with Ca.sup.2+ for the indicated time, before analysis of cleavage by SDS-PAGE with Coomassie staining. Data represent mean of triplicate±1 s.d. (some error bars are too small to be visible).

[0315] FIG. 8 shows a photograph of an SDS-PAGE gel with Coomassie staining showing the necessity of D414 in SPM for cleavage and coupling. ODC was incubated with OAZ-GSY-SPM with or without D414A mutation for the indicated time with Ca.sup.2+.

[0316] FIG. 9 shows the results of an investigation of aspartyl anhydride chemical reactivity. (a) After SPM activation by Ca.sup.2+, the released affibody features an aspartic anhydride. The anhydride then reacts with free nucleophiles or nucleophiles within the affibody (resulting in cyclization). Various nucleophiles were chosen: [1] N-terminal amine minic, [2] Lysine side-chain mimic, [3/4] thiols, and [5] Tyrosine side-chain mimic. [3] forms a labile thioester, whereas [4] may undergo S,N-acyl shift to yield an amide; (b) Quantification of nucleophile reaction with anhydride. Affibody-SPM was incubated with Ca.sup.2+ for 60 min at 37° C. in the presence of 1 or 10 mM of the indicated nucleophile. Products were analyzed by SDS-PAGE with Coomassie staining. Reaction with nucleophile in solution was quantified by the decrease in the level of cyclization. The ratio of linear to cyclized affibody is plotted at the right. (c, d) Anhydride lifetime. Generation of anhydride from affibody-SPM was initiated by adding Ca.sup.2+. At the indicated time-point, cleavage was stopped with EDTA and anhydrides were quenched with free cysteine. The abundance of each species was determined by SDS-PAGE with Coomassie staining. The different kinetics of SPM appearance and affibody cyclization are indicative of the life-time of the anhydride.

[0317] FIG. 10 shows the results of experiments to identify crosslinking sites for ODC reaction: (a) SDS-PAGE with Coomassie staining for OAZ-Y-SPM coupling to wt or K92R ODC. The position of K92 and K121 in the ODC/OAZ complex is shown (PDB 4ZGY); (b) Truncation of first 9 amino acids and removal of N-terminal His-tag (ΔH6Δ1-9), together with introduction of K92R, K12R, K74R and K78R (4KR) reduced conjugation of OAZ-GSY-SPM (SDS-PAGE with Coomassie staining). Re-insertion of the original N-terminus or re-introduction of K92 or K121 rescued coupling. Time where Ca.sup.2+ was present is indicated.

[0318] FIG. 11 shows a photograph of an SDS-PAGE gel with Coomassie staining showing changes in conjugation pattern for OAZ-Y-SPM to ODC K92R and ODC K92R double mutants. OAZ-Y-SPM was incubated with wt ODC or the indicated mutants with or without Ca2+.

[0319] FIG. 12 shows cleavage and crosslinking activities of SPM homologues: (a) SPM homologues, including the −1 and −2 positions relative to the cleavable D-P bond, were fused to OAZ. The Coomasie-stained gel shows formation of cleavage and crosslinked products after 10 μM OAZ-SPM was incubated overnight with 10 μM ODC at 37° C., pH 7.4, and in the presence of 10 mM CaCl.sub.2; (b) Different SPM homologues showed varying cleavage rates. Time course showing relative cleavage rates of different SPM homologues after incubation of 10 μM OAZ-SPM with 10 μM ODC at 37° C., pH 7.4, and with addition of 10 mM CaCl.sub.2. Mean±1 s.d., n=3; and (c) FrpA cleaved more rapidly than FrpC even at 25° C. and with low calcium concentration. Time course showing cleavage rates of FrpA and FrpC. 10 μM OAZ-SPM was incubated with 10 μM ODC at 25° C., pH 7.4 and with addition of 1 mM CaCl.sub.2. Mean±1 s.d., n=3.

EXAMPLES

Example 1—Characterisation of a Chimeric Protein Comprising a Self-Processing Module (SPM) From the FrpC Protein of Neisseria Meningitidis

[0320] To determine whether the cleavage efficiency of the SPM from the FrpC protein of Neisseria meningitidis could be maintained when coupled to other proteins, SPM (SEQ ID NO: 2) was fused to the unstructured SpyTag peptide and substituting each of the 20 amino acids in front of the reactive Asp-Pro was tested (SpyTag-X-SPM) (FIG. 1c). Protein constructs were all expressed in Escherichia coli and purified using Ni-NTA. Each purified chimeric protein was then incubated with 10 mM Ca.sup.2+ in the presence of 10 mM cysteine at 37° C. for 5, 15 or 60 min. The reaction was stopped by addition of EDTA and boiling in sodium dodecyl sulfate, ahead of SDS-PAGE with Coomassie staining (FIG. 1c). When X was D, G or P, there was minimal cleavage even after 60 min (FIG. 1d). H, Y, and W were most efficient at 60 min and Y is the residue here in the native FrpC. Therefore, Y was selected as the amino acid residue preceding the D-P dipeptide in the chimeric protein in further experiments for the development of “NeissLock”.

Example 2—Semi-Automated Analysis of Protein Structures to Identify NeissLock Candidates

[0321] It was hypothesised that key features for the NeissLock strategy are likely to be the distance from the C-terminal anhydride to the nearest nucleophile on the binding protein, steric constraints so that the presence of SPM would disrupt complex formation, and the possibility of an own-goal (where a nucleophile on the target protein rather than the binding protein reacts with the anhydride) (FIG. 2a). A computational approach was developed to search the Protein Data Bank (PDB) to assess how complexes matched these criteria. The steps used by the computational approach are set out in FIG. 2b. First, a database with distances from the most distal resolved residue in a given polypeptide to nucleophilic residues in the same structure was generated from entries in the PDB. This database was then sorted and filtered, and structures were shortlisted after visualization and inspection in PyMOL (see Table 1 above). Due to promising structural characteristics, in combination with expression from E. coli, the complex between Ornithine Decarboxylase (ODC) and Antizyme (OAZ) (PDB 4ZGY) was selected as a model system. In addition, Epidermal Growth Factor Receptor/Transforming Growth Factor alpha (EGFR/TGFα, PDB ID 1MOX) was chosen for further study due to the biological importance of these proteins in cancer and cell survival.

Example 3—Establishing the NeissLock Principle

[0322] In the ODC/OAZ crystal structure, the last resolved residue (E219) of OAZ is 3.5 A from K92 on ODC. Furthermore, E219 appeared to be sterically accessible and far from nucleophiles on OAZ itself. As further truncations in OAZ have previously been described, OAZ was truncated to E219 (hereafter referred to as “OAZ”) and Tyr was introduced as a spacer for SPM fusion (see above) to yield OAZ-Y-SPM as a chimeric protein comprising a binding polypeptide (i.e. a NeissLock-probe).

[0323] The boundaries of the SPM within FrpC are defined as 414-657. A stepwise truncation according to predicted secondary structure revealed that shortened forms of SPM (414-591, 414-613 and 414-635), while functional, were lower yielding and less pure than 414-657 after standard purification from E. coli expression. In addition, the shortened form of SPM (414-591) showed reduced cleavage rate (FIG. 7b). Thus, the full length “long” SPM (comprising amino acids 414-657 of FrpC, SEQ ID NO: 14) was selected for use in further experiments.

[0324] Upon addition of calcium, OAZ-Y-SPM undergoes self-processing to yield SPM and two OAZ species of differing mobility (FIG. 3a). Based on electrospray ionization-MS, these correspond to a linear OAZ species from hydrolysis and a cyclized species from self-reaction of a nucleophile on OAZ with its own anhydride. The formation of higher-molecular weight products indicative of self-conjugation of OAZ was observed in trace amounts (FIG. 3a). However, when ODC was mixed with OAZ-Y-SPM, no such higher-molecular weight products were observed. Instead OAZ nearly quantitatively conjugated to ODC (FIG. 3a). This covalent conjugation was validated by intact protein electrospray ionization MS. After OAZ-Y-SPM self-processing, masses corresponding to SPM (calculated: 26,414.80, observed: 26,415.61), OAZ-Y-SPM (calculated: 42,024.19, observed: 42,025.87), ODC (calculated: 52,929.42, observed: 52933.81) and ODC:OAZ-YD conjugate (calculated: 68,538.80, observed: 68,539.76) were identified (FIG. 3b). Mutation of D414 in SPM to alanine abolished calcium-induced cleavage of OAZ-GSY-SPM and covalent conjugation to ODC (FIG. 8), supporting the key role of this residue for reaction.

[0325] The parameters determining cleavage of the chimeric protein and conjugation of the binding and target polypeptide were explored using the ODC/OAZ model system. The OAZ-Y-SPM displayed reduced cleavage rate (FIG. 4a) compared to SpyTag-Y-SPM (FIG. 1c) or Affibody-Y-SPM (FIG. 9). Steric hindrance was proposed as the reason for reduced cleavage rate in SPM fusion proteins. Accordingly, a GS-linker was introduced into OAZ-Y-SPM to produce OAZ-GSY-SPM and its effect on cleavage rate and conjugation efficiency was tested. A significant increase in cleavage rate was observed in OAZ-GSY-SPM compared to OAZ-Y-SPM (FIG. 4b).

[0326] The pH-dependence of cleavage and conjugation was also tested using the ODC/OAZ model system. Since reaction is proposed to be principally from nucleophilic attack by the ε-amine of Lys, with a typical pK.sub.a of 10, it was important to test if the NeissLock approach was feasible at neutral pH (e.g. between pH 6.5 and 8.5). It was surprisingly found that cleavage was most efficient at pH 6.5 or 7.0 and but still readily occurred up to pH 8.5 (FIG. 4c). Similar to the cleavage rate, the rate of formation of the cross-linked product was highest at pH 6.5 and decreased with higher pH. However, significant changes in conjugation efficiency were not observed at different pH values within the tested range (calculated from the ratio of crosslinked product to cleavage product, SPM).

[0327] To investigate the specificity of NeissLock reaction, an AP-tag (Acceptor Peptide for site-specific biotinylation) was introduced to OAZ-GSY-SPM to enable SPR affinity measurements (AP-OAZ-GSY-SPM). Residue 175 in OAZ was changed from C to A (C175A) to produce AP-OAZ.sup.c175A-GSY-SPM, in order to reduce aggregation (FIG. 7a). Conjugation of AP-OAZ.sup.c175A-GSY-SPM to irrelevant (non-binding) proteins, maltose binding protein (MBP) or superfolder green fluorescent protein (sfGFP) was tested. At 10 μM concentration of both AP-OAZ.sup.c175A-GSY-SPM and added protein, conjugation of 63% ODC to ODC:AP-OAZ.sup.c175A-GSY-D was observed (FIG. 3c). For MBP, SDS-PAGE suggested trace levels of conjugation (1-2%), whereas no trace of conjugation to SpyTag003-sfGFP was identified (FIG. 3c).

[0328] The affinity-dependence of NeissLock was assessed. Two mutations reported to reduce binding in mouse OAZ/ODC (K153E and V198A) as well as a third mutation (charge inversion via R188E) were introduced into OAZ, to design the low affinity binder OAZ[K153E, R188E, V198A]-GSY-SPM. SPR was used to determine the KD of binding of AP-OAZ[K153E, R188E, V198A]-GSY-SPM to ODC and was found to be unmeasurable by SPR (indicating Kd>100 μM). For wild type AP-OAZ-GSY-SPM binding to ODC, a K.sub.d of 0.12 μM was measured. Upon addition to ODC, no detectable cross-linked product was observed for AP-OAZ[K153E, R188E, V198A]-GSY-SPM after overnight incubation (FIG. 5a). This shows that the chimeric protein OAZ-Y-SPM and derivatives thereof may be used to selectively conjugate OAZ to ODC in an affinity-dependent manner, providing proof-of-concept for NeissLock conjugation.

Example 4—Conjugation of OAZ to Other Nucleophiles in the Target Polypeptide, ODC

[0329] As discussed above, OAZ was identified as a suitable NeissLock probe (i.e. a binding polypeptide in the chimeric protein) based on the proximity of the distal resolved residue E219 to ODC K92 and it was hypothesized that crosslinking primarily occurred at ODC K92. Tryptic liquid chromatography-mass spectrometry/mass spectrometry (LC-MS/MS) was used to characterise OAZ-ODC conjugate produced from the OAZ-Y-SPM chimeric protein and crosslinked peptides at K92 were identified. When the OAZ-Y-SPM chimeric protein (NeissLock probe) was reacted with K92R ODC, it was surprisingly found that high amounts of covalent conjugation were observed (FIG. 5b and FIG. 10a). Optimized gel conditions were used to resolve at least two distinct conjugated species, suggesting that K92 is a primary crosslink site with alternative crosslink sites on ODC. The slower and faster migrating products of OAZ-Y-SPM conjugated to ODC K92R were subjected to tryptic LC-MS/MS. This resulted in the identification of another crosslink site at K121 in the slowly migrating product. The third crosslink site in either the slower or the faster migrating band was not identified by tryptic LC-MS/MS.

[0330] Comparison of OAZ-GSY-SPM to OAZ-Y-SPM resolved under optimized conditions revealed that OAZ-GSY-SPM showed traces of a distinct conjugation product even during conjugation with wild type ODC (FIG. 5b), whereas no such trace was observed for OAZ-Y-SPM (FIG. 11). This indicates that the GS spacer altered the availability of nucleophiles in the target polypeptide (ODC), potentially by introducing increased range and flexibility. To further explore the spatial requirements for crosslinking, attempts were made to rescue the wild type-like banding pattern for OAZ-Y-SPM or OAZ-GSY-SPM conjugation to ODC K92R by reintroducing lysine residues in proximity to K92R. Along the α-helix on which K92 is positioned, mutations T93K, Q96K or S100K were introduced into the ODC K92R background (FIG. 5b). Furthermore, ODC K92R T396K, which is opposite this α-helix, was tested (FIG. 5b). A wild type-like banding pattern was observed in ODC K92R Q96K (FIG. 4e), which has a similar orientation to K92—facing towards K121—and, especially accounting for the additional residues introduced after E219, i.e. -YD or -GSYD, is likely at a comparable distance to the aspartic anhydride.

[0331] Finally, additional crosslinking sites on ODC were located by rational mutagenesis. From tryptic LC-MS/MS, K92 and K121 were already identified as crosslinking sites. The faster migration of one of the product bands indicated that the crosslinking products would be less branched than for crosslinking to ODC K121, i.e. closer to the terminus. First, lysines in proximity to OAZ E219 were mutated to make ODC ‘4KR’ (ODC K92R K121R K74R K78R) and ODC ‘8KR’ (ODC 4KR with additional K141, K69, K148 and K150). Compared to wild-type ODC, conjugation of OAZ-GSY-SPM to ODC 4KR or ODC 8KR showed similarly reduced efficiency, however, significant amounts of product formation were still observed. It was hypothesised that the unresolved N-terminal region of ODC—which further harbours flexible tags—could be another crosslinking site, especially considering the good reactivity of the N-terminal amine (FIG. 9). Although the proximal resolved residue of ODC in the 4ZGY structure faces outwards, far away from the ODC/OAZ interface, alignment of a structure of the ODC homodimer (PDB ID 1D7K) indicated that the N-terminus of ODC might loop back towards the binding interface.

[0332] Accordingly, the mutations in ODC 4KR were combined with removal of the N-terminal His-tag as well as further truncation of unresolved or flexible residues based on PDB 1D7K to make ‘ΔH6Δ1-9 4KR’. Although removal of the His-tag alone did not appear to significantly reduce conjugation efficiency, conjugation of OAZ-GSY-SPM to ΔH6Δ1-9 4KR yielded only low amounts of conjugation. The amount of crosslinked product was 6.7% that of wild type ODC (FIG. 10b). Subsequent reintroduction of R74K and R78K into ODC 4KR appeared to rescue none to small amounts of conjugation (FIG. 10b), consistent with the observations made for spatial preference in ODC K92R Q96K compared to ODC K92R T93K/S100K (FIG. 5b). However, reintroduction of either R92K or R121K rescued high levels of conjugation, notably yielding a slow migrating species as the main product for ODC 4KR R121K, consistent with previous observations. These observations confirm a third major crosslinking site within the N-terminal region, likely at the N-terminal amine as this region contains no lysines. Overall, conjugation appears to be most efficient for K92 and, according to the ratio of slow and fast migrating bands in ODC K92R (FIG. 5b), similarly efficient for K121 and the N-terminal amine, wherein these ratios are influenced by linker length (FIG. 5b and FIG. 11). In summary, the OAZ anhydride conjugates efficiently with multiple different crosslinking sites on ODC.

[0333] To understand the solution behaviour of OAZ-GSY-SPM, size exclusion chromatography with multi-angle light scattering (SEC-MALS) was performed. This analysis gave a close correspondence between the predicted and observed M.sub.W for a monomeric protein.

Example 5—Use of a Chimeric Protein (TGFα-GSY-SPM) to Conjugate a Polypeptide to Cells

[0334] The TGFα/EGFR complex was identified as a promising candidate for use in the method of the invention. This was validated by testing conjugation of TGFα-GSY-SPM to the soluble ectodomain fragment of EGFR, sEGFR501 in vitro. The complex glycosylation of sEGFR501 expressed in 293Expi cells led to heterogeneous gel mobility. Therefore this construct was expressed with the mannosidase inhibitor kifunensine and treated with PNGase F before resolving it on SDS-PAGE, which resulted in a single sharp band. Co-incubation of 10 μM sEGFR with 100 μM TGFα-GSY-SPM in the presence of Ca.sup.2+ led to the formation of a new species, a covalent complex between sEGFR and TGFα, which is not present from autoproteolysis of TGFα-GSY-SPM alone (FIG. 6a). Under these conditions, ˜50% of sEGFR was conjugated.

[0335] To test cellular interaction of SPM fusion, the interaction of TGFα-GSY-SPM at the mammalian cell surface was assessed. The A431 cell line, which displays high levels of EGFR, was used. MCF-7 was used as a negative control since it has low levels of EGFR. AlexaFluor-488 conjugated anti-EGFR affibody was used as a positive control. His.sub.6-TGFα-SPM detected with anti-His-phycoerythrin (PE) resulted in clear visualization of A431 cellular membranes, which was not the case for MCF-7, supporting specific receptor binding. Covalent reaction of TGFα-GSY-SPM to EGFR on cells was then tested. A431 cells incubated with TGFα-GSY-SPM showed conjugation of TGFαto EGFR as determined by Western blot (FIG. 6b). Importantly, incubation with either TGFα-GSY-[DA]-SPM (non-cleaving) or TGFα[R42A]-GSY-SPM, a low-binding mutant of TGFα, blocked reaction, indicating that conjugation was dependent on both SPM-processing and EGFR-binding (FIG. 6b). Subsequent testing of different cleavage conditions showed that both co-incubation of TGFα-GSY-SPM with calcium as well as inhibition of endocytosis with dynasore further improved coupling yield (FIG. 6c).

Example 6—Characterising Other Self-Processing Modules and Their Use in the Chimeric Protein

[0336] To verify the utility of other SPMs in the methods and uses of the invention, SPMs with homology to the SPM from FrpC protein from Neisseria meningitidis (SEQ ID NO: 2) were identified. In particular, an SPM was identified in: the FrpA protein from Neisseria meningitidis (SEQ ID NO: 1), which shows 98.37% sequence identity to SEQ ID NO: 2; the haemolysin-type calcium binding protein related domain-containing protein from Alysiella filiformis (SEQ ID NO: 3), which shows 71.95% sequence identity to SEQ ID NO: 2; and the bifunctional haemolysin/adenylate cyclase precursor protein from Kingella negevensis (SEQ ID NO: 4), which shows 60.41% sequence identity to SEQ ID NO: 2.

[0337] Each of the SPMs was used to produce a chimeric protein containing a domain (i) sequence containing AP-GSS-His6-OAZ (SEQ ID NO: 13); a linker domain comprising GVY, GIV or GGY, and the SPM sequence set out above. The sequences of the chimeric proteins are set out in SEQ ID NOs: 9-12 (i.e. comprising SEQ ID NOs: 1-4, respectively).

[0338] The chimeric proteins were assessed for their ability to promote the proximity-dependent conjugation of OAZ to ODC as described in Example 3. As shown in FIG. 12a, all of the chimeric proteins were able to promote the proximity-dependent conjugation of OAZ to ODC. Moreover, it was surprisingly determined that the SPM from FrpA (SEQ ID NO: 1) displayed a substantially faster rate of autoproteolytic cleavage and a higher yield of cleavage compared to the other SPMs (FIGS. 12b and 12c). Notably, SEQ ID NO: 1 differs from SEQ ID NO: 2 at positions 17 (A vs T), 23 (A vs S), 28 (R vs T) and 30 (Q vs N) (using the numbering of SEQ ID NOs: 1 and 2). It is hypothesised that one or all of these differences results in the improved activity of the SPM from the FrpA protein.

METHODS

Plasmids and Cloning

[0339] For cloning of constructs, Q5 High-Fidelity Polymerase (NEB) or KOD Hot Start DNA Polymerase (EMD Millipore) was used for PCR followed by Gibson assembly. Residue numbers for SPM derive from FrpC of N. meningitidis serogroup B (strain MC58) (UniProt Q9JYV5). The SPM sequence was based on residues 414-657 of FrpC. SpyTag-A-SPM has the following organization: N-terminal (M)GSS-linker, His.sub.6-tag, SSG-linker, thrombin cleavage site, Ndel restriction site, G-spacer, SpyTag, alanine, SPM, GSG-linker, C-tag. Residue numbers for OAZ and ODC were based on the crystal structure of the OAZ:ODC complex (PDB 4zgy). Residues 95-219 of human OAZ (UniProt P54368) were used for pET28a-His.sub.6-OAZ-SPM-Ctag. The truncation of OAZ1 corresponds to the region modelled in

[0340] PDB 4zgy. pET28a-His.sub.6-OAZ-SPM-Ctag has the following organization: N-terminal (M)GSS-linker, His.sub.6-tag, OAZ, SPM, GSG-linker, C-tag. Human ODC1 (UniProt P11926) was cloned into pET28a-His.sub.6-ODC-Ctag to give the following organization: N-terminal (M)GSS-linker, His.sub.6tag, SSG-linker, ODC1, GSG-linker, C-tag. pET28a-TGFα-GSY-SPM-His.sub.6-Ctag includes mature TGFαsequence that was taken from residues 40-89 of human protransforming growth factor alpha (UniProt P01135). His.sub.6-TGFα-SPM has the following organization: N-terminal (M)GSS-linker, His.sub.6-tag, SSG-linker, TGFα, SPM, GSG-linker, C-tag. DNA primers and gene fragments codon optimized for E. coli expression were ordered from Integrated DNA Technologies before cloning into the pET28a backbone. All constructs were validated by Sanger sequencing.

Mammalian Protein Expression

[0341] Expression of the ectodomain of human EGFR was carried out using pENTR4-sEGFR501-His.sub.6 that has the organization: tissue plasminogen activator (tPA) secretion leader sequence, soluble fragment of extracellular domain of human EGFR (UniProt P00533, residues 25-525), GSGESG (SEQ ID NO:15), His.sub.6s. pENTR4-sEGFR501-His.sub.6was transfected into the Expi293 Expression System (ThermoFisher) using the ExpiFectamine 293 Transfection Kit (ThermoFisher).

[0342] Secreted sEGFR501 was recovered from the cell supernatant using Ni-NTA affinity purification.

Database Search for Model Protein Complex

[0343] To identify candidate complexes for covalent fusion by C-terminal activation, protein structures were screened for the distance of the C-terminal resolved residue to Lys ε-amino groups (CTε). First, protein structures were retrieved from the worldwide protein data bank (wwPDB, www.wwpdb.org). Initial analysis was performed using the programming language Python (Python Software Foundation, www.python.org); in particular, the Biopython PDB module was used to interpret structural data. A set of protein structures was pre-selected based on inter- and intra-chain CTε, chain count, and other metadata. Preselected structures were visually inspected in PyMOL (version 2.0) and a final selection was made, taking into account the biological relevance of the complex and experimental data such as ease of purification and complex K.sub.d.

Bacterial Protein Expression and Purification

[0344] For pET28a-His.sub.6-OAZ-SPM-Ctag, pET28-His.sub.6-ODC1-Ctag or related plasmids, the plasmids were transformed into chemically-competent E. coli BL21 (DE3) RI PL (Agilent Technologies). Cells were then plated on LB agar with 50 pg/mL kanamycin and incubated overnight at 37° C. Single colonies were picked to inoculate 11 mL of LB with 50 μg/mL kanamycin and 34 μg/mL chloramphenicol before 16-20 hours of incubation at 37° C. with shaking at 200 rpm. 10 mL of the overnight culture was used to inoculate 1 L of LB with 50 μg/mL kanamycin and 34 μg/mL chloramphenicol in a baffled flask. Cultures were incubated at 37° C. with shaking at 200 rpm until OD.sub.600 reached ˜0.6, upon which cultures were induced using with 0.42 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) (Fluorochem, UK) and incubated at 25° C. with shaking at 200 rpm for 16-18 h. Cells were harvested from the culture medium with a JLA8.1 rotor at 4° C., washed and pelleted. Cell pellets were immediately processed or stored at −80° C. until further use. For constructs containing TGFα-SPM and variants thereof, the same protocol was used except that protein was induced at 18° C. instead of 25° C. For His.sub.6-TGFα-SPM, the protein was induced from the Rosetta-Gami 2 (DE3) strain instead of BL21 (DE3) RIPL.

ODC and OAZ-SPM Protein Purification

[0345] For variants of ODC and OAZ-SPM, cells were harvested and lysed by sonication in lysis buffer [30 mM Tris-HCl, 200 mM NaCl, 5% (v/v) Glycerol, 15 mM imidazole, pH 7.5] supplemented with mixed protease inhibitors (cOmplete mini EDTA-free protease inhibitor cocktail, Roche), 1 mM phenylmethylsulfonyl fluoride (PMSF), 1 mg/mL lysozyme (Sigma-Aldrich), 2 U/mL benzonase (Sigma-Aldrich) and 5 mM 2-Mercaptoethanol (Sigma-Aldrich). While kept on ice, the lysate was sonicated thrice for 1 min at 50% duty cycle with 1 min rest period in between. The cell lysate was then centrifuged at 16,900 g for 10-20 min at 4° C. The clarified lysate was then added to Ni-NTA resin (Qiagen). After addition to a Polyprep gravity column, the Ni-NTA resin was washed twice with 5 packed resin volumes of Ni-NTA buffer (50 mM Tris-HCl, 300 mM NaCl, pH 7.8) with 10 mM imidazole and 5 mM 2-Mercaptoethanol (Sigma-Aldrich). This was followed by another two washes with 5 packed resin volumes of Ni-NTA buffer with 30 mM imidazole and 5 mM 2-Mercaptoethanol (Sigma-Aldrich). The protein was eluted from the Ni-NTA resin using Ni-NTA buffer with 200 mM imidazole and 5 mM 2-Mercaptoethanol (Sigma-Aldrich). The protein was concentrated using a Vivaspin centrifugal concentrator with 10 or 30 kDa cut-off (GE Healthcare) before loading onto a pre-equilibrated HiLoad 16/600 Superdex 200 pg size exclusion chromatography column (GE Healthcare) connected to an AKTA Pure 25 (GE Healthcare) fast protein liquid chromatography (FPLC) machine at 4° C. 50 mM HEPES, 150 mM NaCl, 2 mM TCEP, pH 7.4 buffer was used for gel filtration. An additional 0.02 mM pyridoxal phosphate (PLP) was added to the gel-filtration buffer when purifying ODC. Fractions were collected according to the A.sub.280 peak and verified by SDS-PAGE, before another round of concentration using Vivaspin centrifugal concentrator with 10 or 30 kDa cut-off (GE Healthcare). For ODC variants without an N-terminal His-tag, the clarified lysate was added to CaptureSelect™ C-tagXL Affinity Matrix (ThermoFisher) instead of Ni-NTA. After addition to a Polyprep gravity column, the resin was washed four times with 5 packed resin volumes of wash buffer (20 mM Tris-HCl, 5 mM 2-Mercaptoethanol, pH 7.4). The protein was eluted from the C-tagXL resin using 50 mM HEPES, 5 mM 2-Mercaptoethanol and 2M MgCl.sub.2, pH 7.8.

Protein Analysis

[0346] Protein concentrations were estimated using a NanoDrop spectrophotometer, with extinction coefficients estimated using the ExPASy server. SDS-PAGE was done using 10%, 16% or 18% polyacrylamide gels in an XCell SureLock system (ThermoFisher) run at 180V or 200V. SDS-PAGE gels were stained using InstantBlue (Expedeon) and destained with water before imaging with a ChemiDoc XRS imager. Quantification was carried out using Image Lab software (version 5.2.1).

Cleavage and Coupling Assays

[0347] Reactions were carried out in the reaction buffer (50 mM HEPES, 150 mM NaCl, 2 mM TCEP, pH 7.4) at 37° C. When measuring the pH-dependence of the reaction, an additional 50 mM 2-(N-morpholino)ethanesulfonic acid (MES) was added for proper buffering over the pH range tested. For reactions analyzing the effect of the −1 position on cleavage rate, 10 μM of SpyTag-X-SPM was used. For reactions for analysing the speed and pH-dependence of coupling, OAZ-SPM was reacted with ODC at a 1:1 ratio with each protein at 10 μM or at the indicated concentrations. The cleavage of SPM was induced by addition of the HEPES reaction buffer, pre-equilibrated to 37° C., containing calcium chloride at a final concentration of 10 mM. After the indicated time, the reaction was stopped by addition of 5×SDS-loading buffer [0.19 M Tris-HCl pH 6.8, 20% (v/v) glycerol, 100 μM bromophenol blue, 0.19 M SDS] containing EDTA added to a final concentration of 15 mM in the reaction mixture. Protein samples were then heated on a Bio-Rad C1000 thermal cycler at 95° C. for 3 min. For time courses, the 0 h time point was taken by addition of the stop buffer to the reaction before addition of the start buffer. Finally, cleavage and coupling reactions were analyzed by gel densitometry of 10%, 16% or 18% polyacrylamide gels. The percentage cleavage of SPM was determined from the reduction in intensity of SpyTag-X-SPM or OAZ-SPM from the 0 h time point.

Anhydride Reactivity Test

[0348] 20 μM Affibody-SPM was incubated with 10 mM CaCl.sub.2 in 50 mM HEPES, 150 mM NaCl, pH 7.4 (HBS) with 1 mM or 10 mM of the indicated nucleophiles at 37° C. for 1 h, before inhibiting the reaction with 75 mM EDTA in 5×SDS loading buffer. Samples were resolved on 18% SDS-PAGE without prior boiling. For anhydride lifetime tests, 7.5 μM Affibody-SPM was incubated for the indicated amount of time with 10 mM CaCl.sub.2 in 50 mM HEPES, 150 mM NaCl, pH 7.4. Samples were then quenched with 5 μL 100 mM EDTA and 100 mM Cysteine in HBS. Samples were boiled in SDS loading buffer before resolving on SDS-PAGE.

SEC-MALS

[0349] OAZ-SPM was prepared at 2 mg/mL in 100 μL of buffer containing 50 mM HEPES, 150 mM NaCl, 2 mM TCEP, 0.02 mM PLP, pH 7.4 before injection into a Superdex 200 HR 10/30 column (GE Healthcare) connected to a Shimadzu HPLC system with an attached Wyatt Dawn HELEOS-II 8-angle light scattering detector and Wyatt Optilab rEX refractive index monitor. SEC-MALS was carried out at room temperature with 50 mM HEPES, 150 mM NaCl 2 mM TCEP, 0.02 mM PLP, pH 7.4 running buffer.

Surface Plasmon Resonance

[0350] Surface plasmon resonance was carried out using a Biacore T200 (GE Healthcare). AP-OAZ-GSY-SPM was biotinylated using GST-BirA. Biotinylated AP-OAZ-GSY-SPM was immobilized onto the sensor chip using the Biotin CAPture reagent from the Biotin CAPture Kit, Series S (GE Healthcare) and following each run, the chip was regenerated using the provided solutions, following the manufacturer's protocol. Serial dilutions of ODC were tested when measuring the K.sub.d of OAZ binding to ODC. 1.25 μM ODC was diluted down to 78.1 nM for wild-type OAZ and for binding mutants of OAZ, 97 μM of ODC was diluted down to 1.51 μM.

Mass Spectrometry

[0351] For intact protein mass spectrometry, a RapidFire 365 platform (Agilent) comprising a jet-stream electrospray ionization source coupled to a 6550 Accurate-Mass Quadrupole Time-of-Flight (Q-TOF) (Agilent) detector was used. With the RapidFire platform, protein samples prepared at 10 μM in 70 μL were acidified to 1% (v/v) formic acid before aspiration under vacuum for 0.3 s and loading onto a C4 solid-phase extraction cartridge. Washes using 0.1% (v/v) formic acid in water was carried out for 5.5 s before sample elution onto the Q-TOF detector for 5.5 s.

Tryptic LC-MS/MS

[0352] Conjugated OAZ-Y-SPM/ODC or OAZ-Y-SPM/ODC K92R were resolved on 18% SDS-PAGE at 180 V for 100 min to separate different conjugate species. Bands were cut from the gel, in particular higher and lower conjugate bands, and submitted to the Oxford Biochemistry Proteomics facility for further processing.

Cell Staining with TGFα-SPM

[0353] A431 and MCF-7 were cultured in Dulbecco's Modified Eagle Medium supplemented with 10% fetal bovine serum, 1% penicillin, 1% streptomycin, and 1% GlutaMAX at 37° C., 5% CO.sub.2. Before cell staining, A431 and MCF-7 was seeded onto glass-bottom petri dishes. The glass dishes were transferred to 4° C. to prevent receptor internalization, the medium was removed, and cells were washed twice with 1 mL PBS +5 mM MgCl.sub.2 (PBS-M). Then, cells were incubated with PBS-M with 1% (w/v) bovine serum albumin (BSA) and 1.5 μM anti-EGFR-Affibody conjugated to AlexaFluor-488 or 3 μM His.sub.6-TGFα-SPM (from Rosetta-Gami 2) as indicated. After 30 min, cells were washed twice with PBS-M +1% BSA. Samples not incubated with affibody were incubated with 450 μL Anti-His-Phycoerythrin at 1:200 in PBS-M +1% BSA; affibody samples were incubated with only PBS-M +1% BSA instead. After 15 min, cells were washed twice and then covered with 1 mL PBS-M. Samples were imaged with a DV core inverted microscope (Micron Oxford), using a FITC (green false colour) or TRITC filter (red false colour).

Cell Conjugation with TGFα-GSY-SPM

[0354] A431 cells were seeded into 25 cm.sup.2 flasks and grown overnight. Before cell conjugation, cells were starved in Dulbecco's Modified Eagle Medium. For cell conjugation, TG Fα-GSY-SPM, TGFα-GSY-[DA] SPM or TGFα[R42A]-GSY-SPM diluted in HEPES-buffered saline (50 mM HEPES, 150 mM NaCl, pH 7.4) supplemented with 5 mM MgCl.sub.2 (HBS-M) were added to cells. Cells were either incubated for the indicated time at indicated temperature before washing with HBS-M. Subsequently, 2 mM CaCl.sub.2 diluted in HBS-M was added to the cells. Alternatively, CaCl2 diluted in HBS-M was added immediately after addition to the protein solution without washing (co-incubation) or added after the indicated amount of time without washing (directly). After protein conjugation, cells were placed on ice and washed with HBS-M. Optionally, cell flasks were frozen at −80° C. before further processing. Cells were lysed by addition of hot SDS lysis buffer (1% SDS in 10 mM Tris-HCl, 1 mM EDTA, pH 8.0), followed by sonication, heating and centrifugation.

Western Blot

[0355] Cell lysates were diluted in reducing SDS loading buffer and resolved on SDS-PAGE as described above. Proteins were transferred overnight at 30 V, 4° C. to methanol-activated Polyvinylidene fluoride (PVDF) membrane in transfer buffer (7.2 g/L glycine, 1.44 g/L Tris base in 20% methanol). Membranes were blocked with 5% (w/v) skim milk in PBS pH 7.4 with 0.05% (v/v) Tween-20 (PBS-T). Subsequently, membranes were incubated with primary antibodies at 1:1000 dilution in 5% (w/v) skim milk in PBS-T, i.e. mouse anti-TGFα (MF9, Novus Biologicals) or mouse anti-EGFR (LA22, Merck). Membranes were washed 3-4 times with PBS-T before addition of secondary goat anti-mouse horseradish peroxidase HRP antibody (Sigma-Aldrich A4416) at 1:5000 dilution in 5% (w/v) skim milk with PBS-T. After additional washes with PBS-T, membranes were incubated with SuperSignal™ West Pico PLUS Chemiluminescent Substrate before measuring chemiluminescence on a ChemiDoc XRS imager.

Graphics/ Structure Visualization

[0356] The structure of OAZ/ODC was obtained from PDB 4zgy and TGFα/EGFR from PDB 1mox, respectively. Structures were visualized using PyMOL (version 2.0). Figures were prepared using the FIJI distribution of ImageJ and the open-source graphics editor inkscape (inkscape.org).