Funding

The work leading to this invention has received funding from the European Research Council under the European Union's Seventh Framework Programme (FP7/2007-2013) / ERC grant agreement n° 339367.

Field

The present invention relates to methods for the synthesis and enzymatic modification of peptides.

Background

Cyclopeptides display wide-ranging applications in the fields of biomedicine and nanotechnology. ¹ ’ ² ’ ³ ’ ⁴ Notably, many natural cyclopeptides show a diverse range of potent biological activities and several are used in clinical practice (e.g. cyclosporin A, vancomycin). ⁵ A cyclic peptide was recently approved for the treatment of multiple sclerosis. ⁶

In theory, peptide-based macrocycles represent a particularly valuable class of molecules for exploitation in therapeutic design. Like peptides, they can be exquisitely tailored to engage the extended interfaces that are characteristic of many protein-protein interfaces. ⁷ ’ ⁸ Yet unlike peptides, macrocyclic peptides are resistant to protease degradation and are capable of passive diffusion across cell membranes. The propensity for peptide (and other) macrocycles to penetrate the cell membrane has resulted in the coining of the term‘beyond the rule of five’. ⁹ Particular classes of non-natural peptide based macrocycles such as bicyclic peptides and stapled peptides have attracted a lot of attention. ¹⁰ From a chemical viewpoint, ring closure of macrocycles by peptide bond formation is not routine. Different synthetic strategies have been reported both in solution and on solid phase ¹¹ ’ ^{12 13} (e. g. head to tail cyclization, side chain to side chain cyclization, lactamization, azide-alkyne cyclization, ring closing metathesis, disulfide bridges). All have advantages as well as drawbacks; thus far the use of enzymes on resin-bound peptides has had limited success. ^{14 15}

The cyanobactin family of ribosomally synthesized and post-translationally modified peptides (RiPPs), which includes patellamides, ulithiacyclamides and trunkamides, are peptide macrocycles that are highly diverse in size (residue number), sequence and residue tailoring (e. g. heterocyclization, oxidation, prenylation). ¹⁶ Given this enormous chemical diversity it is unsurprising that the biological properties are equally diverse, including P-gp inhibition, cytotoxicity, immunomodulation, antifungal, antibacterial and antiviral properties. ¹⁶ Patellamides and ulithiacyclamides are cyclooctapeptides and trunkamide is a prenylated cycloheptapeptide, produced by P. didemni, an obligate, uncultured symbiont of L patella . ^{17 18 19} Patellamide D’s structure has been assigned by X-ray crystallography ²⁰ and it adopts a closed‘figure of eight’ conformation (II) in solution ²¹ (Figure 1 B). This conformation is stabilized by intramolecular hydrogen bondings and tt-p stacking. Nature has evolved multi enzyme pathways to produce these molecules. The most well studied, the patellamide pathway, has seven genes, one of which encodes PatE the seventy-one residue precursor peptide, which contains two, eight residue cassettes (denoted core peptides) that are modified by the other proteins to yield the final product. ²² The final product contains oxazolines (derived from the heterocyclization of serine/threonine) and thiazoles (oxidized form of thiazoline which is derived from the heterocyclization of cysteine). The two residues adjacent to the thiazoles are epimerized from an L-configuration to a D-configuration. Remarkably, the sequence of the PatE core peptides are highly variable since catalysis and recognition occur on separate regions of PatE. The only requirement within the core is a five membered heterocyclic ring (thiazoline, oxazoline, c/s-proline) at the C-terminal position to facilitate macrocyclization by PatG; an otherwise extremely promiscuous enzyme ²³ . This promiscuity has previously been exploited to insert non- proteinaceous groups into macrocycles. ²⁴ _’ ²⁵ Many of the enzymatic transformations require motifs outside the core peptide and this has been exploited to engineer a heterocyclase that can process synthetic core peptides (without any leader). ²⁶ Combining this with additional enzymes in vitro has afforded non-natural cyclooctapeptides. ²⁷ ’ ²⁸ The use of enzymes with purified peptide substrates allows enormous flexibility in the design of the final product, however it is not compatible with the generation of large diverse libraries. Entirely in vivo production using bacteria engineered to contain the pathway ²⁹ , or by use of interns ³⁰ · ³¹ can generate very large arrays of peptide macrocycles but as the building blocks are limited to amino acids, diversity is inherently limited. Additionally such approaches are difficult to make compatible with large- scale defined compound library production. Total synthesis of cyanobactins has been reported but requires at least 14 steps. Ulicyclamide, ulithiacyclamide, patellamide A and patellamide B have been synthesized by total synthesis, ³² · ³³ _’ ³⁴ _’ ³⁵ while trunkamide A has been synthesized by a solid-phase total synthesis approach. ³⁶ The inability to generate highly chemically diverse (diversity beyond variation of amino acids) libraries of cyanobactins on a useful scale has hindered their development as potential drug lead molecules.

Two chemoenzymatic in vitro syntheses of cyanobactin analogues have been described previously.

Both the approaches have been performed on a very small scale and the final cycles characterized only by ESI-MS. The first, developed by Houssen et al. relied on recombinant peptide substrates and so suffers the same diversity limitations as the in vivo methods of cyanobactin production described previously. ³⁷ Later, Sadar et al, described a one-pot synthesis of cyanobactins, using purified synthetic substrates ³⁸ . The synthetic substrates incorporated a minimal leader peptide, making them 25 residues long ³⁸ .

Summary

The present inventors have developed a chemoenzymatic route for the production of modified peptides, including macrocycles, in which peptides generated by solid-phase synthesis are directly modified by enzymes in a“one-pot” manner without intermediate purification or re-dissolution. This may be useful for example in generating libraries of modified peptides.

An aspect of the invention provides a method of producing a modified peptide comprising;

providing a peptide covalently linked to a solid support by a benzyl linker,

cleaving the benzyl linker to release the peptide from the support and produce a peptide solution comprising the released peptide, and

treating the peptide in the peptide solution with one or more modifying enzymes to produce a modified peptide. Another aspect of the invention provides a method of producing a population of modified peptides comprising;

providing a population of diverse peptides, each peptide being attached to a solid support through a benzyl linker,

cleaving the benzyl linker to release the population from the support and produce a peptide solution comprising the released population, and

treating the released population of peptides in the peptide solution with one or more modifying enzymes to produce a modified population of diverse peptides.

Preferably, the benzyl linker is a hydroxymethylbenzoic acid (HMBA) linker.

Other aspects and embodiments of the invention are described in more detail below.

Brief Description of the Figures

Figure 1 A shows the structure of cyanobactin metabolites. Figure I B shows a representation of X-ray crystal structure (C = grey, O = red, S = yellow, N = blue) and the chemical structure of patellamide D with key hydrogen bonds (red lines), tt-p stacking interactions (black line) and protons useful for the NMR analysis highlighted. Figure 1 C shows an overview of the planned one-pot enzymatic synthesis of five patellamide-like macrocycles from one resin-bound peptide sequence that was prepared using automated peptide synthesis technology. After cleavage from the resin, it was planned that the resulting unprotected linear peptide would be processed by different combinations of enzymes leading to one pot syntheses of different complex cyclic peptides. A key challenge in this approach is the proof of structure of the resulting cyclic peptides hence no stereochemistry is shown. See below for a detailed discussion on structure assignment. Enzymes: plain pentagon = LynDfusion (cysteine heterocyclase), striped pentagon = MicDfusion (cysteine, threonine, serine heterocyclase), Square = PatGmac (macrocyclase), light circle = ArtGox (thiazoline oxidase), dark circle = ArtGox (thiazoline, oxazoline oxidase).

Figure 2 shows an overview of the method of the invention that employs a base cleavable linker (HMBA) enabling release of a crude linear peptide in solution for subsequent enzyme-catalyzed modification. Protected amino acids are indicated with *.

Figure 3A shows proof of concept studies demonstrating that the use of the HMBA-Doc linker system enabled enzyme mediated conversion of the peptide sequence VGAGIGWCAYD-Doc to cyclic peptides 4 and 5 on reaction with LynDfusion, PatGmac. 4 and 5 were isolated as single diastereomers but assignment of the absolute stereochemistry was a key challenge addressed in this work. Figure 3B shows one-pot enzymatic synthesis used to prepare cyclic peptide 8 (single diastereomer, absolute configuration unassigned) that was expected to be Patellamide C via linear peptide 7 using MicDfusion, PatGmac, ArtGox.

Figure 4 shows examples of reaction schemes for one-pot enzymatic syntheses as described herein. Figure 4A shows a one-pot enzymatic synthesis carried out with the peptide sequence VGAGIGWPAYD- Doc. Figure 4B shows one-pot enzymatic synthesis of cycle 4, which is as a 4:1 mixture of diastereomers (4A/4B) by HPLC. Figure 4C shows ArtGox oxidation of cycle 4A to 5. Figure 4D shows one-pot enzymatic synthesis of cycles S19 and S20 employing the resin bound peptide S16. Figure 4E shows one-pot enzymatic synthesis of cycle S22 employing the resin bound peptide S16. Figure 4F shows one- pot enzymatic synthesis of cycle S24 employing the resin bound peptide 6. Figure 4G shows one-pot enzymatic synthesis of cycle S25 employing the resin bound peptide S16. Figure 4H shows one-pot enzymatic synthesis of cycle S29 employing the resin bound peptide S26. Figure 4I shows one-pot enzymatic synthesis of cycle 8 (epi epi patellamide C) employing the resin bound peptide 6. Figure 4J shows one-pot enzymatic synthesis of cycle S33 (epi epi patellamide D) employing the resin bound peptide S30. Figure 4K shows one-pot enzymatic synthesis of cycle S34 employing the resin bound peptide S16. Figure 4L shows one-pot enzymatic synthesis of cycle S38 employing the resin bound peptide S35. Figure 4M shows one-pot enzymatic synthesis of cycle S42 employing the resin bound peptide S39. Figure 4N shows one-pot enzymatic synthesis of cycle 13 _e employing the resin bound peptide 11.

Figure 5A shows selected ¹ H NMR chemical shifts of authentic patellamide C and epi epi patellamide C (8). Figure 5B shows selected ¹ H NMR signals of authentic patellamide D, epi epi patellamide D and a mixture of the two species. * = unassigned peaks.

Figure 6A shows mechanism of the enzymatic preparation of the OxZ unit ⁴⁶ . Figure 6B shows the structures of syn , anti OxZ model dipeptides (9, 10). Figure 6C shows experimental and DFT calculated J (a-CH, b-CH) (OxZ) in Hz listed for 9, 10, epi epi patellamide C (8), patellamide C. The DFT calculation has been done assuming 8 as an L-Phe, L-Ala epimer of patellamide Cu).

Figure 7 A shows 4 possible epimers for epi epi patellamide C (8). Figure 7B shows cycle l-OxH-A-ThH- 1- OxH-A-ThH is symmetric only if both a-CH (Ala) stereocentres are either D(R) or L(S). Figure 7C shows one-pot enzymatic synthesis used to prepare cyclic peptide 13 _e (single diastereomer) using MicDfusion, PatGmac, ArtGox. Figure 7D shows selected ¹ H NMR signals of 13 _e , a mixture 13 _e and 14 and a mixture of 13 _e , 13s and 14.

Figure 8 shows a schematic for synthesis of a library of linear patellamide analogues.

Detailed Description

This invention relates to methods for the“one-pot” release of peptides from solid phase after synthesis and their subsequent enzymatic modification. A peptide covalently linked to a solid support by a benzyl linker, preferably a hydroxymethylbenzoic acid (HMBA) linker, is introduced to a solution under conditions that cleave the linker. The peptide is released from the support into the solution and treated in the solution with one or more modifying enzymes to produce a modified peptide. This avoids the need for the purification and re-dissolution of peptides in aqueous solution, which is resource-intensive and reduces yield, and, for example, allows peptide libraries generated by split and pool synthesis to be subjected to enzymatic modification to generate highly diverse molecular libraries. A peptide as described herein is a linear sequence of amino acid residues linked by peptidyl bonds. A suitable peptide may have from 4 to 30 residues, preferably 4 to 25 residues, more preferably 6 to 23 residues, 6 to 20 residues or 6 to 11 residues.

A peptide as described herein is produced by solid phase synthesis and is covalently attached to a solid support when introduced to the solution. Conveniently, the peptide may be synthesized on the solid support (i.e. solid phase synthesis).

A solid support is an insoluble, non-gelatinous body which presents a surface on which the nascent peptide can be attached during peptide synthesis. Examples of suitable supports include beads.

Peptides are typically synthesized by solid phase synthesis in a stepwise fashion from the C terminus to the N terminus. In an initial step, an N protected amino acid is covalently attached to an insoluble solid support via its carbonyl group. Suitable groups for N protecting the amino acid include 9- fluorenylmethyloxycarbonyl group (Fmoc) and t-butyloxycarbonyl (Boc). Following covalent attachment of the N protected amino acid, the N protecting group is removed and the deprotected NP group of the attached amino acid is reacted with the carboxylic acid group of the next N protected amino acid to generate a nascent peptide comprising 2 amino acids that is covalently attached to the solid phase. This process is repeated until the complete peptide sequence is built up on the solid phase. In some embodiments, protecting groups may be employed to prevent functional groups in the side chains of amino acids from reacting with an incoming N protected amino acids. These side chain protecting groups may be present throughout the synthesis of the peptide and may be removed in a final deprotection step. Methods of solid phase peptide synthesis are well-established in the art (see for example Coin et al Nature Protocols 2, 3247-3256 (2007) Stawikowski (2002) Curr Protoc Protein Sci. 2002 Unit-18.1. oi:10.1002/0471 140864. ps1801 s26; Chan and White; Fmoc Solid Phase Peptide Synthesis - A Practical Approach. Oxford University Press, 2000; Stewart, J. M.; Young, J. D. Solid-Phase Peptide Synthesis (2nd ed.), Pierce Chemical Co., Rockford, IL, 1984; Atherton, E.; Sheppard, R. C., Solid-Phase Peptide Synthesis: A Practical Approach. Oxford University Press: New York City, 1989; M. Bodanzsky and A. Bodanzsky, The Practice of Peptide Synthesis, Springer Verlag, New York (1984); J. H. Jones, The Chemical Synthesis of Peptides. Oxford University Press, Oxford 1991 ; in Applied Biosystems 430A User’s Manual, ABI Inc., Foster City, California; G. A. Grant, (Ed.) Synthetic Peptides, A User’s Guide.

W. FI. Freeman & Co., New York 1992, and G.B. Fields, (Ed.) Solid-Phase Peptide Synthesis (Methods in Enzymology Vol. 289). Academic Press, New York and London 1997).

A peptide for modification as described herein may comprise any amino acid sequence of interest.

An amino acid in a peptide described herein may comprise a carboxylic acid group and amino group. The peptide may include modified amino acids, unmodified amino acids, heterocyclic amino acids, non- heterocyclic amino acids, a-amino acids, b-amino acids, g-amino acids, naturally occurring amino acids and/or non-naturally occurring amino acids.

In some preferred embodiments, a peptide may comprise one or more amino acid residues which may be altered or alterable by enzymatic treatment (i.e. modifiable residues). For example, the peptide may comprise one or more amino acids that are converted into heterocyclic molecules by heterocyclases. Heterocyclisable molecules include cysteine, selenocysteine, tellurocysteine, threonine, serine and 2, 3- diaminopropanoic acid. In some preferred embodiments, a peptide may comprise one or more residues that undergo prenylation by a prenyltransferase. Suitable residues depend on the prenyltransferase that is employed and may include Ser/Thr (e.g. TruF1), Tyr (e.g. PagF, LynF) and Trp (e.g. KgpF).

A peptide may comprise one or more motifs or signals which facilitate the structural alteration of the peptide by enzymatic treatment. In preferred embodiments, the one or more motifs or signals may facilitate the macrocyclization of the peptide by a macrocyclase. Suitable motifs or signals include a cyclization signal. A cyclization signal is a recognition site for a macrocyclase. The peptide may comprise a cyclization signal at its C terminus. The cyclization signal may be heterologous i.e. a sequence that is not naturally associated with the remainder of the peptide. Typically, a cyclization signal will comprise the sequence; small residue - bulky residue - acidic residue. Suitable cyclization signals include AYD, AYE, SYD, AFD and FAG, preferably AYD. The sequence of the cyclization signal in a peptide may depend on the cyanobacterial macrocyclase that is being used for macrocyclization. In some preferred

embodiments, the macrocyclase may be a PatG macrocyclase (PatGmac) and the cyclization signal may be AYD. A cyclic residue (e.g. proline or thiazoline) may be present next to the cyclization signal in the peptide. Macrocyclase treatment may be useful, for example, in generating macrocyclic peptides. The cyclization signal at the C terminus of the peptide is not incorporated into the macrocyclic peptide and, in the presence of the cyclization signal, the remainder of the peptide, including the cyclic residue, undergoes cyclization to form a macrocyclic peptide (Bras et al Chemistry. 2016 Sep 5; 22(37): 13089- 13097; McIntosh JA, et al. Journal of the American Chemical Society. 2010; 132: 15499-15501).

In some embodiments, one or more residues in the peptide may comprise a reactive functionality which allows further chemical modification. Suitable residues may contain side chains with linking groups, such as -NH , -COOH, -OH and -SH.

The peptide is connected to the solid support via a benzyl linker. The benzyl linker is selected from moieties of formula (I)

'*' and‘**’ denote positions of attachment to the solid support and the peptide or spacer if present. Preferably‘*’ is the position of attachment to the peptide or spacer if present and‘**’ is the position of attachment to the solid support.

X is -O- or -N(R’)-. Preferably, X is -N(R’)-.

Each R ¹ is independently selected from Ci- -alkyl, C -io-alkenyl, C -io-alkynyl, Ci- -alkoxy,

Ci-io-halo-alkyl, -(C=0)R’, -(C=0)0R’, -(C=0)NR’R”, NO , halo or -CN. n is an integer from 1 to 4. Each R’ or R” is independently selected from hydrogen, Ci-io-alkyl, C2-io-alkenyl, C2-io-alkynyl,

Ci-io-alkoxy, Ci-io-halo-alkyl. Preferably, the benzyl linker is a hydroxymethylbenzoic acid (HMBA) linker, such as

4-(hydroxymethyl)benzoic acid, 2-(hydroxymethylbenzoic acid or a mixture thereof. In some preferred embodiments, the benzyl linker is 4-(hydroxymethyl)benzoic acid.

The term hydroxymethylbenzoic acid linker refers to a linker having the following structure:

The term hydroxymethylbenzoic acid linker refers to a linker which is formally a diradical of

hydroxymethylbenzoic acid. X may be -O- or -N(R’)-, preferably X is -N(R’)-. That is, the X moiety may be derived from the carboxylic acid of the hydroxymethylbenzoic acid or from the group to which the linker is attached.

The peptide may be connected directly to the linker on the solid support without a spacer or more preferably the peptide may be connected to the linker via a spacer. Suitable spacers include molecules with terminal amine and carboxyl groups such as a spacer of formula (II)

'*' and‘**’ denotes positions of attachment to the peptide and the linker. Preferably,‘*’ is the position of attachment to the peptide and‘**’ is the position of attachment to the benzyl linker.

R ² is selected from hydrogen, Ci e alkyl. Preferably R ² is hydrogen.

L ¹ is selected from C-MS alkylene, C2-18 alkenylene, C2-18 alkyneylene and -CH2(-CH2-0-) _n -CH2- wherein n is an integer from 1 to 5. Preferably L ¹ is-CH ₂ (-CH ₂ -0-) _m -CH ₂ - wherein m is an integer from 1 to 5.

Preferably, the spacer is an 8-amino-3, 6-dioxaoctanoic acid (Doc) spacer.

The term“8-amino-3, 6-dioxaoctanoic acid spacer” refers to a spacer having the following structure:

That is the term 8-amino-3, 6-dioxaoctanoic acid spacer refers to a spacer which is formally a diradical of 8-amino-3, 6-dioxaoctanoic acid. For example, the peptide-spacer-linker-solid support structure may be as follows:

Peptide -M-IH

J.2

R T

o

Wherein R ² , L and X are as defined above.

The term C _x-xx refers to the number of carbon atoms in a functional group. For example, Ci -12-alkyl refers to an alkyl group have from 1 to 12 carbon atoms.

The term alkyl refers to a monovalent moiety obtained by removing a hydrogen atom from a carbon atom of a linear or branched saturated hydrocarbon compound.

The term alkylene refers to a divalent moiety obtained by removing two hydrogen atoms from carbon atom(s) of a linear or branched saturated hydrocarbon compound.

The term alkenyl refers to a monovalent moiety obtained by removing a hydrogen atom from a carbon atom of a linear or branched hydrocarbon compound having one or more carbon-carbon double bonds.

The term alkenylene refers to a divalent moiety obtained by removing two hydrogen atoms from carbon atom(s) of a linear or branched hydrocarbon compound having one or more carbon-carbon double bonds.

The term alkynyl refers to a monovalent moiety obtained by removing a hydrogen atom from a carbon atom of a linear or branched hydrocarbon compound having one or more carbon-carbon triple bonds

The term alkynylene refers to a divalent moiety obtained by removing two hydrogen atoms from carbon atom(s) of a linear or branched hydrocarbon compound having one or more carbon-carbon triple bonds

The term alkoxy refers to a group -OR wherein R is an alkyl group as defined herein.

The term halo refers to a halogen substituent, such as fluoro, chloro, bromo or iodo.

The term halo-alkyl refers to an alkyl groups as defined herein wherein one or more hydrogen atoms have been replaced with halo groups.

HMBA linkers are resistant to the acidic conditions, such as the presence of trifluoroacetic acid (TFA), and may therefore be used to immobilize the nascent peptide on the support during solid phase synthesis. The attachment of molecules to a solid phase using HMBA linkers is established in the art (see for example Hansen et al Org Biomol Chem. (2016) Mar 28; 14 (12):3238-45; Sheppard, R. C. et al. Int. J. Pept. Protein Res. 1982, 20, 451). HMBA linkers are cleavable for example under strongly basic conditions.

Suitable solid supports for solid-phase peptide synthesis are well-known in the art and may include resins and polymers such as polystyrene-polyethylene glycol (PEG) composites, PEG and poly-a-lysine (e-PL) (see for example Albericio F (2000). Solid-Phase Synthesis: A Practical Guide. Boca Raton: CRC Press). Conveniently, the support may be in form of beads. Solid supports comprising benzyl linkers, such as HMBA linkers, may be produced using standard techniques or obtained from commercial suppliers (e.g. HMBA-ChemMatrix®, PCAS BioMatrix Inc CA; Hypogel®-HMBA, Sigma Aldrich Inc, USA, and HMBA Tentagel®, Rapp Polymere GmbH, DE).

Following the synthesis of a peptide on a solid support, the peptide-bound support may be washed and dried. For example, the peptide-bound support may be washed with one or more organic solvents, such as dichloromethane (DCM), dimethylformamide (DMF), and/or diisopropylethylamine (DIEA), and dried in vacuo.

Following synthesis and optional washing and drying, the peptide is separated from the support through cleavage of the benzyl linker and released into the peptide solution. The peptide remains in solution during the release and modification steps described herein and is not subjected to intermediate purification, precipitation or redissolution. For example, the peptide is not purified or isolated before treatment with the one or more modifying enzymes. This reduces losses due to inefficient recovery and allows small quantities of peptide to be modified as described herein.

Between the steps of the method, for example between cleavage and enzyme treatment and between treatments with different enzymes, the pH, ionic strength and/or other conditions of the peptide solution may be adjusted in order to provide suitable reaction conditions.

The peptide may be released from the solid support by exposure to conditions that cause cleavage of the HMBA or other benzyl linker. Suitable conditions for cleavage of HMBA and other benzyl linkers are known in the art. For example, the peptide-bound support may be exposed to basic conditions, for example by introduction to a base solution. The peptide-bound support may for example be exposed to the basic conditions for 20 minutes at room temperature to release the peptide. The peptide may be released into the base solution to produce the peptide solution.

A base solution is an aqueous solution comprising a nucleophile, such as ammonia, a primary amine, hydrazine, an alkoxide or alcohol, methanol/triethylamine, or hydroxide, such as NaOH or KOH. For example, the peptide-bound support may be introduced to a base solution comprising 0.01 to 1 M NaOH, for example 0.1 M NaOH (see for example J. Hansen, F. Diness, M. Meldal Org. Biomol. Chem. 2016, 14, 3238).

In some embodiments, the base solution may further comprise a reducing agent, such as dithiothreitol (DTT) e.g. at 5mM. This may be useful for example in preventing oxidation of thiol groups in the peptide. The support-bound peptide may be exposed to the cleavage conditions for a time that is sufficient to cleave the HMBA or other benzyl linker. For example, the support-bound peptide may be exposed to the basic conditions for 15 to 60 mins, preferably 30 mins.

Following release of the peptide from the solid support, the base conditions of the peptide solution may be quenched in order to facilitate enzymatic modification. For example, the pH of the peptide solution may be reduced, for example from strong basic conditions, such as pH 1 1 or above, to neutral or mild basic conditions, such as pH 7.0-9.0.

In some embodiments, an acid, such as HCI, H2SO4, HNO3, or H3PO4, may be added to the peptide solution. For example, an equal volume of 0.1 M to 2M HCI may be added to the peptide solution to reduce the pH.

Following cleavage of the HMBA or other benzyl linker, the released peptide in the peptide solution is modified by treatment with one or more modifying enzymes. For example, a modifying enzyme may be added to the peptide solution in order to modify the peptide. Modifying enzymes alter the structure of the peptide, for example by modifying one or more amino acids within the peptide, cyclizing the peptide or adding one or more additional groups to the peptide.

The peptide may be treated sequentially with the one or more modifying enzymes in a series of steps; or may be treated simultaneously.

The one or more modifying enzymes may be enzymes from ribosomally synthesized and post- translationally modified peptide (RiPP) biosynthesis pathways. For example, the one or more modifying enzymes may be cyanobacterial enzymes, such as patellamide, trunkamide or microcyclamide biosynthesis pathway enzymes.

Suitable modifying enzymes include macrocyclases, heterocyclases, oxidases and/or prenylases (see for example, Czekster et al Curr Opin Chem Biol. (2016) 35: 80-88). For example, the peptide in the peptide solution may be treated with;

(i) a heterocyclase and (ii) a macrocylase;

(i) a heterocyclase, (ii) a macrocylase and (iii) an oxidase,

(i) a macrocyclase and (ii) an oxidase

In some embodiments, where a macrocyclic peptide is not required, the peptide in the peptide solution may be treated with;

(i) a heterocyclase and (ii) an oxidase.

Preferably the peptide is treated with the modifying enzymes sequentially in the order set out above. In some embodiments, the peptide solution may be split and different portions may be treated with different sets of one or more modifying enzymes. Optionally, the portions may be pooled and re-split into portions between enzyme treatments to generate a diverse population of modified peptides with different combinations of modifications.

In some preferred embodiments, the peptide solution is treated with a macrocyclase. The macrocyclase catalyzes the cyclization of peptides in the peptide solution which contain a cyclization signal to form macrocyclic rings through the formation of a peptide bond between the N-terminal amine and the amino acid next to the cyclization signal. Suitable macrocyclases include PatG macrocyclase (AAY21 156.1 Gl:62910843; SEQ ID NO: 1) and TruG (gi|167859101 |gb|ACA04494.1) from Prochloron and macrocylases from Anabaena spp, such as ADA00395.1 Gl:280987232; ACK37889.1 Gl:217316956 and AED99446.1 Gl:332002633; Oscillatoria sp, such as Gl:300866529 ZP_0711 1219.1 ; Microcystis spp such as Gl:389832527 CCI23764.1 , Gl:158934376 CAO82089.1 , Gl:389788443 CCI15902.1 , Gl:389678154 CCH92964.1 , Gl:389802072 CCI18832.1 , Gl:389882395 CCI37144.1 , Gl:389826370 0012311 1.1 ; Gl:389731219 CCI04703.1 , Gl:389716328 CCH99432.1 , Gl:389831597 CCI25524.1 and Gl:159027550 CAO86920.1 ; Nostoc spongiaeforme spp, such as TenG (Gl:167859092 ACA04486.1); lyngya spp, such as Gl:1 19492374 ZP_01623710.1 ; Nodularia spp, such as Gl :1 19512474

ZP_01631555.1 ; Anabaena spp, such as AcyG (Gl:280987232 ADA00395.1) Planktothrix spp, such as Gl:332002633 AED99446.1 , Trichodesmium spp, such as Gl:1 13475997 YP_722058.1 ; Arthrospira spp, such as ZP_06384654.1 Gl:284054444, Gl:284054071 ZP_06384281 .1 , Gl:291571075 BAI93347.1 , Gl:284054444 ZP_06384654.1 , and Gl:376002294 ZP_09780130.1 , and plant spp, including Saponaria vaccaria (Barber et al J Biol Chem 2013, 288 (18), 12500-10, such as AGL51088.1.

Other suitable macrocyclases are available in the art (Lee, S. W. et al (2008) PNAS 105(15), 5879-5884). Preferred macrocyclases include PatGmac

A macrocyclase for use as described herein may comprise the amino acid sequence of any one of the reference macrocyclase sequences listed above or may be a fragment or variant thereof. In some preferred embodiments, a macrocyclase may be a PatG macrocyclase which comprises the amino acid sequence of SEQ ID NO: 1 or a fragment or variant thereof.

The peptide solution may be treated with the macrocyclase under suitable conditions for the cyclization of peptide. Suitable conditions would be apparent to those skilled in the art and examples of suitable reaction conditions are described in section 2.10 below. In some preferred embodiments, conditions may include 500mM NaCI and/or pH 9. For example, the linear peptide substrate may be treated with the macrocyclase in 100 mM bicine, 500 mM NaCI and 5 % DMSO at pH 9.

In some embodiments, the peptide solution is treated with a heterocyclase (also called a

cyclodehydratase). The heterocyclase converts heterocyclizable residues in peptides in the peptide solution into heterocycles. Heterocyclizable amino acids include cysteine, selenocysteine, tellurocysteine, threonine, serine, 2, 3-diaminopropanoic acid and synthetic derivatives thereof with additional R groups at the alpha and beta position. The heterocyclase may convert the cysteine residues in the peptide into thiazolines; threonine/serine residues into oxazolines; selenocysteines into selenazolines;

tellurocysteines into tellurazolines, 2, 3 diaminopropanoic acids into imadazolines and/or aminoalanines into imidazolines. Homocysteine, homoserine, 2, 4-diaminobutanoic acid and alpha/beta/gamma substituted analogues thereof may be converted into 5, 6-dihydro-4H-1 ,3-thiazine, 5,6-dihydro-4H-1 ,3- oxazine and 5,6-dihydro-4H-1A2-pyrimidine respectively.

In some embodiments, the heterocyclase may be a cysteine heterocyclase which converts cysteine residues into thiazolines. Suitable cysteine heterocyclases include LynD, TruD (ACA04490.1

Gl:167859097) and TenD (ACA04483.1 Gl:16785908) and fusion variants thereof. In other

embodiments, the heterocyclase may be a cysteine/serine/threonine heterocyclase which converts cysteine residues into thiazolines and threonine/serine residues into oxazolines. Suitable

cysteine/serine/threonine heterocyclases include MicD and PatD (AAY21153.1 Gl:6291084) and fusion variants thereof. Other suitable heterocyclases are available in the art (Lee, S. W. et al (2008). PNAS 105(15), 5879-5884; Koehnke et al (2013) Angewandte Chemie 125 52 14241 -14246).

The heterocyclase may be added to the peptide solution in the presence of a leader peptide comprising a heterocyclase substrate leader sequence

More preferably, the heterocyclase may be a modified heterocyclase that comprises a heterocyclase substrate leader sequence fused to the catalytic domain of the enzyme. The heterocyclase substrate leader sequence, whether in cis or trans, activates the heterocyclase and allows heterocyclic groups to be introduced into peptides which completely lack a heterocyclase substrate leader sequence (Koehnke J, et al. Nat Chem Biol. 2015; 1 1 :558-563; Czekster et al Curr Opin Chem Biol. (2016) 35: 80-88).

The heterocyclase substrate leader sequence may be from the same source (e.g. the same bacteria species) as the catalytic domain or may be from a different source. For example, the heterocyclase substrate leader sequence may be a fragment of the pre-pro-peptide that forms the natural substrate for the heterocyclase. The heterocyclase substrate leader sequence may be directly linked to the heterocyclase sequence or via a linker. Suitable heterocyclase substrate leader sequences and linkers are well known in the art.

For example, a heterocyclase substrate leader sequence may comprise the sequence LAELX ₁ EEX ₂ X3 where Xi is S or T, preferably S, X ₂ is A, V, T or N and X3 is L or I. Suitable heterocyclase substrate leader sequences may comprise LAELSEEAL, LAELSEETL or LAELSEEAI or a variant of any one of these sequences. Preferred heterocyclase substrate leader sequences may consist of the sequence

X4X5X6X7X8 LAEL X1EEX2X3LX9X10X11X12 (SEQ ID NO: 29) where Xi is S or T, preferably S; X is A, V, T or N; X3 is L or I or optionally absent; X is T, Q or K; X5 is Q, L or K; Xe is A, P or S; X7 is A, D or S; Xe is E, L, A, H or Q or Y; Xg is G or A, preferably G; X ₁₀ is S, D, G, or absent; Xn is T, L, N, A or V or absent; and X ₁₂ is T, P, A, E, G, D or absent. An example of a preferred heterocyclase substrate leader sequence consists of the amino acid sequence; X ₁ QLSSQLAELSEEALGDAG where Xi is absent, G, AG, TAG, LTAG or RLTAG. For example, the heterocyclase substrate leader sequence may consist of the sequence QLSSQLAELSEEALGDAG or a variant thereof. Suitable modified heterocyclases include LynDfusion (SEQ ID NO: 2 or a fragment or variant thereof) and MicDfusion (SEQ ID NO: 3 or a fragment or variant thereof). Other modified heterocyclases are available in the art (see for example WO2016/071422).

The released peptide may be treated with the heterocyclase under suitable conditions to heterocyclize one or more heterocyclizable residues therein. For example, the released peptide may be treated with the heterocyclase in aqueous solution at ambient temperature in the presence of Mg2+ and ATP at pH 8 to 10, preferably 9. For example, before addition of the heterocyclase, the peptide solution may be adjusted to comprise 100 mM bicine, 150 mM NaCI, pH = 9.0. The rate of heterocyclization may be altered by changing the conditions, for example altering the Mg2+ and ATP concentrations.

The highest temperature tolerated by the heterocyclase is generally preferred as this leads to increased reaction rates. The optimal temperature for reaction under a defined set of conditions may be determined experimentally.

In some embodiments, the peptide solution is treated with an oxidase, for example a flavin

mononucleotide (FMN)-dependent dehydrogenase, following treatment with heterocyclase. The oxidase catalyzes the oxidation of one or more heterocyclic residues in peptides in the peptide solution, such as thiazolines, oxazolines, selenazolines, tellurazolines and imidazolines. Thiazoline (Thn) residues in the peptide may be oxidized into thiazoles (Thz); oxazoline residues (Oxn) in the peptide may be oxidized into oxazoles (Oxz); selenazolines (Sen) in the peptide may be oxidized into selenazoles (Sez);

tellurazolines (Ten) in the peptide may be oxidized into tellurazoles (Tez) and imidazolines (Imn) in the peptide may be oxidized into imidazoles (Imz). An oxidase may oxidize all the heterocyclic residues described herein or combinations thereof, for example oxazolines and thiazolines; or only thiazolines.

Oxidase treatment may occur directly after heterocyclization to oxidize one or more heterocyclic residues in the peptide or oxidization may occur at a different stage, for example, the cyclic peptide may be treated with the oxidase after macrocyclization. Suitable oxidases include thiazoline oxidases and thiazoline, oxazoline oxidases, such as ArtGox (SEQ ID NO: 4 or a fragment or variant thereof), PatGox, ThcOx, TriOx, McaOx, ArtGox, LynGox or TenGox. Before treatment with the oxidase, the peptide solution may be adjusted to comprise 100 mM bicine, 150 mM NaCI, pH = 9.0 buffer (5% DMSO) and FMN cofactor (1 to 500mM, for example 200 mM) and ArtGox (1 to 50 pM for example 5 pM).

In some embodiments, the peptide solution may be treated with a prenylase to O-prenylate serine, threonine and/or tyrosine residues in peptides in the solution. Preferably, the peptide solution may be treated with a prenylase after treatment with the other modifying enzymes i.e. the prenylation is the final enzymatic modification step in the method.

Suitable prenylases include LynF (Mcintosh et al J Am Chem Soc 2011 133(34) 13698-13705; SEQ ID NO: 5), PatF (Gl: 62910842 AAY21 155.1 ), Gl: 167859100 ACA04493.1 (TruF2), and Gl: 167859099 ACA04492.1 (TruF1) from Prochloron spp; Gl: 159027547 CAQ86917.1 , Gl: 158934373 CAQ82086.1 , Gl: 389788445 CCI15906.1 , Gl: 389678155 CCH92965.1 (TenF), Gl: 166362791 YP_001655064.1 , Gl:389831610 CCI25499.1 , Gl:389826377 CCI23120.1 , Gl: 389826383 CCI23131 .1 , Gl: 389832530 CCI23767.1 , Gl:389716343 CCH99420.1 , Gl:389882386 CCI37135.1 , Gl:389720299 CCH95988.1 , Gl:389732896 CCI03253.1 , Gl:389734240 CCI02071.1 , Gl:389801748 CCI19127.1 and Gl: 389802082 CCI18842.1 from Microcystis spp; Gl:167859091 ACA04485.1 (TenF) from Nostoc spongiaeforme spp; Gl:1 19492371 ZP_01623707.1 from Lyngbya spp; Gl:280987227 ADA00390.1 (AcyF) from Anabaena sp; Gl:376002283 ZP_097801 19.1 , Gl:284054206 ZP_06384416.1 from Arthrospira sp; Gl:332002616 AED99429.1 from Planktothrix spp; Gl:300866527 ZP_07111217.1 from Oscillatoria spp.; and

Gl:220905949 YP_002481260.1 from Cyanothece spp. Other suitable prenylases are known in the art (McIntosh JA, et al. J Am Chem Soc. 2011 ; 133:13698-13705).

As described above, modifying enzymes, such as macrocyclases, oxidases and heterocyclases may comprise an amino acid sequence which is a variant or fragment of a reference amino acid sequence.

A variant of a reference amino acid sequence may have an amino acid sequence having at least 60%, at least 70%, at least 75%, at least 80%, at least 85%, at least 90%, at least 95% or at least 98% sequence identity to the reference amino acid sequence of a modifying enzyme. Suitable reference amino acid sequences for modifying enzymes are provided above.

Amino acid sequence identity is generally defined with reference to the algorithm GAP (GCG Wisconsin Package™, Accelrys, San Diego CA). GAP uses the Needleman & Wunsch algorithm (J. Mol. Biol. (48): 444-453 (1970)) to align two complete sequences that maximizes the number of matches and minimizes the number of gaps. Generally, the default parameters are used, with a gap creation penalty = 12 and gap extension penalty = 4. Use of GAP may be preferred but other algorithms may be used, e.g. BLAST or TBLASTN (which use the method of Altschul et al. (1990) J. Mol. Biol. 215: 405-410), FASTA (which uses the method of Pearson and Lipman (1988) PNAS USA 85: 2444-2448), or the Smith-Waterman algorithm (Smith and Waterman (1981) J. Mol Biol. 147: 195-197), generally employing default parameters.

Particular amino acid sequence variants may differ from that in a given sequence by insertion, addition, substitution or deletion of 1 amino acid, 2, 3, 4, 5-10, 10-20 or 20-30 amino acids. In some embodiments, a variant sequence may comprise the reference sequence with 1 , 2, 3, 4, 5, 6, 7, 8, 9, 10 or more residues inserted, deleted or substituted. For example, up to 15, up to 20, up to 30, up to 40, up to 50 or up to 60 residues may be inserted, deleted or substituted.

A fragment is a truncated protein which contains less than the full-length amino acid sequence but which retains the activity of the full-length protein sequence. A fragment may comprise at least 100 amino acids, at least 200 amino acids or at least 300 contiguous amino acids from the full-length sequence.

In some embodiments, a buffering solution and, optionally one or more additional co-factors, may be added to the peptide solution to provide suitable conditions for the activity of a modifying enzyme before or at the same time as the modifying enzyme is added. Suitable co-factors may include ATP and MgCL for heterocyclase activity and FMN for oxidase activity. Preferably, co-factors remain in the peptide solution during subsequent modification steps.

Suitable conditions for specific enzymes are well known in the art or may be determined using standard techniques. The buffering solution may produce any convenient pH concentration upon addition to the peptide solution, for example pH 8 to 10, preferably pH 9.0. The buffering solution may comprise a salt, for example NaCI, concentration of 10-1000mM, preferably 100-500mM. The peptide solution may, for example, comprise 100 mM bicine, 150 mM NaCI at pH 9.0.

The modifying enzyme may be present in the buffering solution or may be added separately. The concentration of modifying enzyme in the peptide solution may be sufficient to modify the peptide in the solution. Suitable enzyme concentrations might include 1 to 100pM, preferably 10 to 50 mM

The peptide may be exposed to the modifying enzyme in the peptide solution for sufficient time for the peptide to be modified by the enzyme. For example, the peptide may be exposed to the modifying enzyme in each step for 6 to 48 hours, preferably about 12 hours

The peptide may be exposed to the modifying enzyme in the peptide solution at a suitable temperature for the peptide to be modified by the enzyme. The highest temperature tolerated by the modifying enzyme is generally preferred as this leads to increased reaction rates. The optimal temperature for reaction under a defined set of conditions may be determined experimentally. For example, the peptide may be exposed to the modifying enzyme for 25°C to 35°C, preferably about 29°C.

A modified peptide produced by a method described herein comprises a sequence of amino acids with one or more enzymatic modifications. For example, the modified peptide may comprise one or more heterocycles or one or more oxidized heterocycles; and/or may be macrocyclic

After treatment with the one or more modifying enzymes, the modified peptide may be isolated and/or purified, for example by HPLC, gel filtration, or size exclusion chromatography. Conveniently, the peptide solution may be extracted three times with BuOH and the fractions combined, dried and solubilized in HSO/MQOH and then purified by HPLC.

Following purification, the modified peptides may be further characterized, for example by liquid chromatography-mass spectrometry (LC-MS), tandem mass spectrometry (MS-MS), circular dichroism (CD) and/or nuclear magnetic resonance spectroscopy (NMR).

A modified peptide produced by a method described herein may be subjected to further chemical modification. Suitable modifications include derivatization with a heterologous moiety, for example, a moiety containing a natural side group such as OH, NH2, COOH, SH, or an unnatural side group suitable for coupling reactions and click chemistry. Click-chemistry involves the Cu(l)-catalyzed coupling between two components, one containing an azido group and the other a terminal acetylene group, to form a triazole ring. Since azido and alkyne groups are inert to the conditions of other coupling procedures and other functional groups found in peptides are inert to click chemistry conditions, click-chemistry allows the controlled attachment of almost any linker to the cyclic peptide under mild conditions. For example, non-cyclized cysteine residues of the cyclic peptide may be reacted with a bifunctional reagent containing a thiol-specific reactive group at one end (e.g. iodoacetamide, maleimide or phenylthiosulfonate) and an azide or acetylene at the other end. Label groups may be attached to the terminal azide or acetylene using click-chemistry. For example, a second linker with either an acetylene or azide group on one end of a linker and a chelate (for metal isotopes) or leaving group (for halogen labelling) on the other end (Baskin, J. (2007) PNAS 104(43)16793-97) may be employed.

In some embodiments, a modified peptide may be labelled with a detectable label. The detectable label may be any molecule, atom, ion or group which is detectable in vivo by a molecular imaging modality. Suitable detectable labels may include metals, radioactive isotopes and radio-opaque agents (e.g.

gallium, technetium, indium, strontium, iodine, barium, bromine and phosphorus-containing compounds), radiolucent agents, contrast agents and fluorescent dyes.

In some embodiments, a modified peptide may be attached to an antibody molecule, such as an antibody or antibody fragment or derivative, for example for use in antibody-directed drug therapies. Suitable techniques for the conjugation of modified peptides and antibodies are well known in the art.

The methods described herein may be particularly useful in producing libraries of modified peptides. A method of producing a population of modified peptides may comprise;

providing a population of diverse peptides attached to a solid support through a benzyl linker, preferably an hydroxymethylbenzoic acid (HMBA) linker, , and optionally a spacer, such as 8-amino-3, 6- dioxaoctanoic acid (Doc),

cleaving the benzyl linker to release the population from the support and produce a peptide solution comprising the population,

treating the released population in the peptide solution medium with one or more modifying enzymes to produce a modified population of diverse peptides.

The cleavage of HMBA linkers and other benzyl linkers, and treatment of released peptide with one or more modifying enzymes is described in more detail above.

The peptides in the population may comprise one or more diverse residues i.e. the residue at one or more positions in the peptide sequence may be different in different peptides in the population. For example, the residue at two, three, four or more, or all of the positions in the peptide sequence may be different in different peptides in the population.

Suitable methods of producing diverse peptide populations are well-known in the art (see for example Furka et al (1991) Int J Pept Protein Res 37(6) 487-493). In some embodiments, the peptides in the population comprise;

a first amino acid sequence that is the same in all the peptides in the population, the first amino acid being located at the C terminal end of the peptides and attached to the linker, optionally via a spacer, a first diverse amino acid residue linked to the N terminal of the first amino acid sequence, wherein the first diverse residue is different in different peptides in the population, and

a second amino acid sequence that is linked to the N terminal of the diverse amino acid residue.

In some preferred embodiments, the first amino acid sequence may comprise a cyclization signal.

Suitable cyclization signals may include AYD, AYE, SYD, AFD and FAG, preferably AYD, and are described in more detail above.

The first amino acid sequence may further comprise an amino acid directly N terminal to the cyclization signal that is a heterocyclic amino acid or a heterocyclizable amino acid. Heterocyclic amino acids may include thiazoline, thiazole, oxazoline, oxazole, or proline. Heterocyclisable amino acids are described above and may include cysteine, selenocysteine, tellurocysteine, threonine, or serine, 2,3- diaminopropanoic acid. Heterocyclisable amino acids may be converted into heterocyclic groups using a heterocyclase, as described herein.

In some embodiments, the second amino acid sequence may be the same in all of the peptides in the population. The sequences of the peptides in the population may be the same at all positions except the first diverse amino acid residue.

In other embodiments, the second amino acid sequence may be different in different peptides in the population. For example, the amino acids at 1 , 2, 3, 4, 5 or more positions in the second amino acid may be different in different peptides in the population.

Suitable methods of synthesizing a population of diverse peptides are established in the art. For example, a population of diverse peptides may be produced by split-and-pool synthesis. For example, the population may be produced by multiple cycles of;

(i) providing a population of nascent peptides having a free end and an end attached to a solid support through a benzyl linker, preferably an HMBA linker,

(ii) splitting the population into portions,

(iii) adding a different amino acid residue to the free ends of the nascent peptides of each portion,

(iv) combining the portions to produce a population of nascent peptides, said population comprising diverse amino acid residues at the free end.

In some embodiments, the population of nascent peptides may comprise a first amino acid sequence that is common to all the peptides in the population. The first amino acid sequence may comprise or consist of a cyclization signal at the C terminal end and a heterocyclic or heterocyclizable residue. For example, a population of diverse peptides may be provided by a method comprising;

coupling a benzyl linker, preferably an HMBA linker, to a plurality of solid supports, optionally coupling a spacer to the coupled benzyl linker,

covalently linking attaching in series the amino acid residues of the first amino acid sequence to the coupled linker or optional spacer, thereby producing a population of nascent peptides consisting of the first amino acid sequence, said nascent peptides having a free end and an end attached to the plurality of solid supports through the linker and optional spacer,

splitting the plurality of solid supports into two or more sub-pluralities, each sub-plurality having a sub-population of nascent peptides attached thereto,

covalently linking a diverse amino acid residue to N terminal of the first amino acid sequence of the nascent peptides, wherein the diverse amino acid residue is different in different sub-populations of nascent peptides,

covalently linking in series the amino acid residues of the second amino acid sequence to the N terminal of the diverse residue in each of said sub-population of nascent peptides, to product sub populations of peptides attached to sub-pluralities of solid supports, wherein each sub-population of peptides has a different amino acid sequence, and

combining the sub-populations to produce a population of peptides, the peptides in the population having different amino acid sequences being attached to solid supports by the benzyl linker and optional spacer.

The benzyl linker may then be cleaved to release the diverse population from the solid supports and produce a peptide solution comprising the population. The released population in the peptide solution medium may then be treated with one or more modifying enzymes to produce a modified population of diverse peptides, as described herein.

Conveniently, the solid support may be modular, such as beads, or polymeric entities, such as

SynPhase™ lanterns ( imotopes Pty Ltd, AU), for example, to facilitate splitting and pooling during synthesis. The peptides in the population may be attached to multiple beads or other modular supports, and the peptide molecules that are attached to each single bead or other modular support may have the same amino acid sequence.

Following treatment with the one or more modifying enzymes, the modified population of diverse peptides may be isolated and/or purified. For example, the population may be alcohol extracted and purified by RT-HPLC.

Following production and optional purification, the modified population of diverse peptides may be screened for binding to a target molecule and/or for a target activity.

Suitable methods for determining binding of a modified peptide to a target molecule are well known in the art and include ELISA, bead-based binding assays (e.g. using streptavidin-coated beads in conjunction with biotinylated target molecules, surface plasmon resonance, flow cytometry, Western blotting, immunocytochemistry, immunoprecipitation, and affinity chromatography. Alternatively, biochemical or cell-based assays, such as an HT-1080 cell migration assay or fluorescence-based or luminescence- based reporter assays may be employed. Other aspects and embodiments of the invention provide the aspects and embodiments described above with the term“comprising” replaced by the term“consisting of and the aspects and embodiments described above with the term“comprising” replaced by the term” consisting essentially of.

It is to be understood that the application discloses all combinations of any of the above aspects and embodiments described above with each other, unless the context demands otherwise. Similarly, the application discloses all combinations of the preferred and/or optional features either singly or together with any of the other aspects, unless the context demands otherwise.

Modifications of the above embodiments, further embodiments and modifications thereof will be apparent to the skilled person on reading this disclosure, and as such, these are within the scope of the present invention.

All documents and sequence database entries mentioned in this specification are incorporated herein by reference in their entirety for all purposes.

“and/or” where used herein is to be taken as specific disclosure of each of the two specified features or components with or without the other. For example“A and/or B” is to be taken as specific disclosure of each of (i) A, (ii) B and (iii) A and B, just as if each is set out individually herein.

Experimental

f _ Materials

The HPLC grade acetonitrile (MeCN) was purchased from VWR. Aqueous buffers and aqueous mobile- phases for HPLC were prepared using water purified with an Elga ^® Purelab Milli-Q water purification system (purified to 18.2 MΏ.ati). Other chemicals and reagents were purchased from Sigma and used without any further purification. NMR spectra ( ¹ H, 2D) were recorded on a Bruker Ultrashield 700 spectrometer {5H (700 MHz, 5C (175 MHz)} at room temperature in the deuterated solvent stated.

Chemical shifts are expressed in parts per million (ppm) from DMSO-d6 (5H=2.50) (Fulmer, G. R. et. al. Organometallics 29, 2176-2180 (2010)). Multiplicities are described as s (singlet), d (doublet), dd (doublet of doublets), t (triplet), m (multiplet), br (broad). Coupling constants J are quoted in Hertz (Hz) to the nearest 0.1 Hz. Signals of protons and carbons were assigned, as far as possible, by using the following two dimensional NMR spectroscopy techniques: [ ¹ H, ¹ H] COSY (Correlation Spectroscopy), [ ¹ H, ¹ H] TOCSY (Total Correlation Spectroscopy), HSQC (Heteronuclear Single Quantum Coherence), [ ¹ H, ¹ H] ROESY (Rotating-frame NOE Spectroscopy) and long range [ ¹ H, ¹³ C] HMBC (Heteronuclear Multiple Bond Connectivity). EXSY (Exchange Spectroscopy) experiment was used to identify equilibrium chemical exchange either at room temperature (rt) or with heating.

Low-resolution mass spectra were obtained with an Agilent 6130 single quad apparatus equipped with an electrospray ionization source. High-resolution mass spectra (HRMS) were obtained with a Thermo Exactive Orbitrap mass spectrometer. Reactions performed in the enzymatic media were monitored using MALDI-MS acquired using a 4800 MALDI TOF/TOF Analyzer (ABSciex, Foster City, CA) equipped with a Nd:YAG 355 nm laser and calibrated using a mixture of peptides. The spot was analyzed in positive MS mode between 800 and 4000 m/z, by averaging 1000 laser spots. The samples, diluted in water to reduce the buffer concentration, (0.5 ml) were applied to the MALDI target along with alpha-cyano-4- hydroxycinnamic acid matrix (0.5 ml, 10 mg/ml in 50:50 acetonitriie:0.1 % TFA) and allowed to dry. When analyzing bead bound peptides linked via a photocleavable linker analogous conditions to those reported by Semmler were used (Semmler, A. et al J. Am. Soc. Mass Spectrom. 21 , 215-219 (2010)) MSMS data were acquired using a TripleTOF 5600+. The sample was subjected to chromatography on an Acclaim PepMap 100 C18 trap and an Acclaim PepMap RSLC C18 column (ThermoFisher Scientific), using a nano-LC Ultra 2D plus loading pump and autosampler (Eksigent). The sample was injected at neutral pH to avoid acid catalyzed ring opening. The trap was washed with 2% acetonitrile, 0.05% trifluoroacetic acid, and the patellamide was then eluted with a gradient of increasing acetonitrile, containing 0.1 % formic acid (15-40% acetonitrile in 5 min, 40-95 % in a further 1 min, followed by 95% acetonitrile to clean the column, before re-equilibration to 15% acetonitrile). The eluent was sprayed into a TripleTOF 5600+ electrospray tandem mass spectrometer (Sciex) operating with standard nanospray conditions, an analyzed in Product Ion Scan mode isolating the m/z of interest. The collision energy was adjusted to give optimal fragmentation. The MSMS fragmentation pattern was interrogated for diagnostic peaks.

Analytical RP-HPLC was performed on an Agilent infinity 1260 series equipped with a VWD detector and a single quadrupole MS using a Macherey-Nagel Nucleodur C18 column (10 pm ^c 4.6 ^c 250 mm).

Several chromatographic systems were used; System A1 : 1 ml/min flow rate with MeCN and 0.1 % aqueous TFA [95% TFA (5 min), linear gradient from 5 to 100% of MeCN (25 min)] and UV detection at 338 nm. System A2: 1 ml/min flow rate with MeCN and 0.1 % aqueous TFA [95% TFA (3 min), linear gradient from 5 to 95% of MeCN (15 min)] and UV detection at 220nm, 254 nm, 280 nm. System A3: 1 ml/min flow rate with MeCN and 0.1 % aqueous TFA [95% TFA (3 min), linear gradient from 5 to 95% of MeCN (27 min)] and UV detection at 220nm, 254 nm, 280 nm. System A4: 1 ml/min flow rate with MeCN and 0.1 % aqueous TFA [95% TFA (3 min), linear gradient from 5 to 95% of MeCN (40 min)] and UV detection at 220 nm, 254 nm, 280 nm. System A5: 1 ml/min flow rate with MeCN and 5 mM aqueous ammonium carbonate (AmC) [95% AmC (3 min), linear gradient from 5 to 95% of MeCN (40 min)] and UV detection at 220 nm, 254 nm, 280 nm. Semi-preparative HPLC was performed on an Agilent Infinity preparative scale purification 1260 series equipped with a VWD detector using a Macherey-Nagel Nucleodur C18 column (10 pm ^c 16 ^c 250 mm). The chromatographic system used was System P1 : 10 ml/min flow rate with MeCN and 0.1 % aqueous TFA [95% TFA (5 min), linear gradient from 5 to 35% of MeCN (5 min), linear gradient from 35 to 37% of MeCN (20min)] and UV detection at 220 nm, 254 nm,

280 nm.

System P2: 10 ml/min flow rate with MeCN and 0.1 % aqueous TFA [95% TFA (5 min), linear gradient from 5 to 50% of MeCN (3 min), linear gradient from 50 to 80% of MeCN (60 min)] and UV detection at 220 nm, 254 nm, 280 nm. System P3: 10 ml/min flow rate with MeCN and 5 mM aqueous ammonium carbonate (AmC) [95% Amc (5 min), linear gradient from 5 to 35% of MeCN (5 min), linear gradient from 35 to 37% of MeCN (20min)] and UV detection at 220 nm, 254 nm, 280 nm. System P4: 10 ml/min flow rate with MeCN and 5 mM aqueous ammonium carbonate (AmC) [95% AmC (5 min), linear gradient from 5 to 50% of MeCN (3 min), linear gradient from 50 to 80% of MeCN (30 min)] and UV detection at 220 nm, 254 nm, 280 nm.

CD measurements were performed at room temperature in MeOH in a 1 mm path length cell with a Bio- Logic MOS-500 spectrometer.

2. _ Methods

2. 1 Cloning, Expression and Purification of Biosynthetic Enzymes: AcLynD, AcMicD, PatGmac, and ArtGox.

PatGmac was cloned from genomic DNA ( Prochlon sp.) into the pHISTEV vector. LynDfusion was cloned into pJexpress 411 plasmid (DNA 2.0). MicDfusion a variant of MicD based on the LynDfusion construct, which is able to processes serine/threonine residues in addition to cysteine residues in a leaderless peptide substrate, was cloned into pJexpress 411 plasmid (DNA 2.0). ArtGox was cloned into

PHISSUMOTEV. All enzymes were expressed in Eschericia coli BL21 (DE3) cells grown in auto-induction media as previously described (J. Koehnke et al. Nat. Struct. Mol. Biol. 19, 767-772 (2012)). For ArtGox the auto-induction media was supplemented with 5 mg L ^_1 riboflavin.

All enzymes were purified as follows: Cell pellets were resuspended in lysis buffer (20 mM Tris pH 8.0, 500 mM NaCI, 20 mM Imidazole, 3 mM BME), supplemented with 0.4 mg DNAse (SIGMA) per gram of wet cell pellets and complete EDTA-free protease inhibitor tablets (Roche; 1 per 50 mL resuspension). The cells were lysed via passage through a cell disruptor at 207 MPA (Constant Systems) and clarified by centrifugation (40, 000 g, 4 °C, 20 min). The supernatant was passed through a Ni-NTA-sepharose 6 Fast Flow column (GE Healthcare) equilibrated in lysis buffer. The bound proteins were washed with lysis buffer and eluted with elution buffer (20 mM Tris pH 8.0, 500 mM NaCI, 250 mM Imidazole, 3 mM BME). The eluted proteins were dialyzed (3 x 300 mL, 4 °C) into storage buffer (10 mM Bicine pH 8.5, 150 mM NaCI, 1 mM TCEP). ArtGox lysis and elution buffers were supplemented with 50 mM FMN.

2.2 Solid-phase peptide synthesis of S1

Amino-PEGAi9oo resin (0.20 mmol/g) from Agilent Technologies was employed. Fmoc-photo linkerwas purchased from Iris Biotech. Synthesis was performed manually on a 0.020 mmol scale. Amino acids were coupled using a 5-fold excess (0.10 mmol). A solution of HBTU (0.10 mmol, 38 mg) in DMF (1 ml) was added to the Fmoc protected amino acid, and DIEA (0.10 mmol, 17 pL) was then added and the amino acid was activated for 30 s prior to coupling. Derivatization of the resin with Fmoc photo-linker was attempted using a triple treatment of 90 mins. A single treatment of 75 mins was used for all the amino acids. Prior to adding the next amino acid, the peptide-bound resin was Fmoc deprotected with 20% piperidine/DMF (v/v, 1 mL) for 10 mins. Each coupling step was checked using the ninhydrin test. After coupling, the resin was washed with DMF (6x2 mL). The final peptide-bound resin was dried under vacuum and the amino acid side-chain deprotection was performed with the reagent cocktail: 94% trifluoroacetic acid (TFA), 4% H2O, 2% triisopropylsilane (TIS) (2 mL) for 2hrs. Finally, the peptide-bound resin was washed with DCM (6x2 mL) and DMF (6x2 mL), neutralized with 10% DIEA/DMF (10x2 mL), washed with DMF (6x2 mL) and DCM (6x2 mL) and dried. The dry resin-bound peptide was used for the enzymatic reaction. 2.3 Solid-phase peptide synthesis of S3

Preloaded G ly-2- ch I o rot rity I chloride resin (0.75 mmol/g) from Bachem was employed. Synthesis was performed manually on a 0.15 mmol scale. Amino acids were coupled using a 5-fold excess (0.75 mmol). A solution of HBTU (0.75 mmol, 288 mg) in DMF (2 ml) was added to the Fmoc protected amino acid, and DIEA (0.75 mmol, 125 pL) was then added and the amino acid was activated for 30 s prior to coupling. A single treatment of 1 hr was used for all the amino acids. Prior to adding the next amino acid, the peptide-bound resin was Fmoc deprotected with 20% piperidine/DMF (v/v, 2 ml_) for 10 mins. Each coupling step was checked using the ninhydrin test. After coupling, the resin was washed with DMF (6x2 ml_). The final peptide-bound resin was dried under vacuum and the peptide was cleaved from the resin with the reagent cocktail: 20% HFIP (hexafluoro isopropanol) in DCM (v/v) (2 ml_) for 30 mins. The resin was filtered away and the filtrate was concentrated in vacuo to afford a white powder (69 mg, yield =

64%).

2.4 Solid-phase peptide synthesis of S4-S6, S11

Amino-PEGAi9oo resin (0.20 mmol/g) from Agilent Technologies was employed. Fmoc-Glu (EDANS) and Fmoc-Asp(OMpe) were purchased from Merck Millipore. Fmoc-Asp(OMpe) was employed instead of Fmoc-Asp(Ot-Bu) in order to prevent aspartimide formation. Synthesis was performed on an automatic peptide synthesizer (Biotage Syrowave) on a 0.020 mmol scale. Amino acids were double coupled using a 4-fold excess (0.08 mmol) for 1 hr at RT making use of DIC (0.08 mmol, 13 pL, 0.5 M in DMF), Oxyma (0.08 mmol, 1 1 mg, 1 M in DMF) for the first coupling and HBTU (0.08 mmol, 31 mg, 0.5 M in DMF), DIEA (0.08 mmol, 14 pL, 2 M in NMP) for the second coupling. Derivatization of the resin with Fmoc photo linker was attempted manually using a triple treatment of 90 mins. Prior to adding the next amino acid, the peptide-bound resin was Fmoc deprotected with 20% piperidine/DMF (v/v, 1 ml_) for 10 mins. The final peptide-bound resin was dried under vacuum and the side-chain deprotection was performed with the reagent cocktail: 94% trifluoroacetic acid (TFA), 4% H O, 2% triisopropylsilane (TIS) (2 mL) for 2hrs. Finally, the peptide-bound resin was washed with DCM (6x2 mL) and DMF (6x2 mL), neutralized with 10% DIEA/DMF (10x2 mL), washed with DMF (6x2 mL) and DCM (6x2 mL) and dried. The dry resin- bound peptide was used for the enzymatic reaction.

2.5 Solid-phase peptide synthesis of S13

HMBA-tentagel resin (0.21 mmol/g) from Rapp Polymere was employed. Synthesis was performed on an automatic peptide synthesizer (Biotage Syrowave) on a 0.021 mmol scale. Amino acids were double coupled using a 4-fold excess (0.084 mmol) for 20 mins at 75 °C employing DIC (0.084 mmol, 13 pL, 0.5 M in DMF), Oxyma (0.084 mmol, 12 mg, 1 M in DMF) for the first coupling and HBTU (0.084 mmol, 32 mg, 0.5 M in DMF), DIEA (0.084 mmol, 14 pL, 2 M in NMP) for the second coupling. After the coupling of the first amino acid, a capping procedure was performed with acetic anhydride:DIEA:DMF 1 :1 :3 (v/v/v, 2 mL) for 5 mins at RT. Prior to adding the next amino acid, the peptide-bound resin was Fmoc deprotected with 20% piperidine/DMF (v/v, 1 mL) for 10 mins. The Fmoc deprotection on the C-terminal amino acid was carried out manually in order to estimate the loading by UV determination of the dibenzo-fulvene at 301 nm (loading = 0.084 mmol/g). The final peptide-bound resin was dried under vacuum and the amino acid side-chain deprotection was performed with the reagent cocktail: 94% trifluoroacetic acid (TFA), 4% H2O, 2% triisopropylsilane (TIS) (2 ml.) for 2hrs. Finally, the peptide-bound resin was washed with DCM (6x2 mL) and DMF (6x2 mL), neutralized with 10% DIEA/DMF (10x2 mL), washed with DMF (6x2 mL) and DCM (6x2 mL) and dried. The dry resin-bound peptide was used for the enzymatic reaction.

2.6 Solid-phase peptide synthesis of 1, 6, S16, S26, S30, S35, S39, S43

HMBA-chemmatrix resin (0.53 mmol/g) from PCAS Biomatrix was employed. Synthesis was performed on an automatic peptide synthesizer (Biotage Syrowave) on a 0.053 mmol scale. Amino acids were double coupled using a 4-fold excess (0.21 mmol) for 20 min at 75 °C making use of DIC (0.21 mmol, 32 pL, 0.5 M in DMF), Oxyma (0.21 mmol, 30 mg, 1 M in DMF) and HBTU (0.21 mmol, 81 mg, 0.5 M in DMF), DIEA (0.21 mmol, 35 pL, 2 M in NMP). The last two amino acids of the sequence were triple coupled for 20 min at 75 °C performing a third coupling with DIC (0.21 mmol, 32 pL, 0.5 M in DMF), Oxyma (0.21 mmol, 30 mg, 1 M in DMF). After the coupling of the first amino acid, a capping procedure was performed with acetic anhydride:DIEA:DMF 1 :1 :3 ( v/vA /, 2 mL) for 5 mins at RT. Prior to adding the next amino acid, the peptide-bound resin was Fmoc deprotected with 20% piperidine/DMF (v/v, 1 mL) for 10 mins. The Fmoc deprotection on the C-terminal amino acid was carried out manually in order to estimate the loading by UV determination of the dibenzo-fulvene at 301 nm (loading = 0.24 mmol/g (1), 0.20 mmol/g (6), 0.21 mmol/g (S26), 0.23 mmol/g (S30), 0.24 mmol/g (S35), 0.19 mmol/g (S39), 0.25 mmol/g (S43)). The peptidyl-resin was dried under vacuum and the side-chain deprotection was performed with the reagent cocktail: 94% trifluoroacetic acid (TFA), 2% H2O, 2% triisopropylsilane (TIS), 2% 3,6-dioxa-1 ,8-octane-dithiol (DODT) (2 mL) for 2h. Finally, the peptidyl-resin was washed with DCM and DMF, neutralized with 10% DIEA/DMF, washed with DMF and DCM and dried. 1.0 mg of peptide- bound resin was cleaved with NaOH 0.1 M 5mM DTT (50 pL) for 30 mins, quenched with HCI 0.1 M (50 pL) and a 50 pL aliquot was injected in the LC-MS. The dry peptide-bound resin was used for the enzymatic reaction.

2.7 Synthesis of S4 by chemical macrocyclization of S3

To a stirred solution of HATU (74 mg, 0.19 mmol), DIEA (53 pi, 0.42 mmol) in dry DMF (14 ml) at RT, a solution of AGIGFPVG (37 mg, 0.051 mmol) in dry DMF (3 ml) was added by syringe pump in 2 hrs. After 24 hrs the resulting mixture was concentrated in vacuo, diluted with DCM (20 ml) and washed with 1 M HCI (7x2 ml). The mixture was extracted with DCM (10x2 ml) and the combined organic phases were washed three times with water (10 ml), dried MgS04 and concentrated in vacuo. The crude residue was purified by HPLC (9.0 mg, yield = 25.0%).

2.8 PatGm _ac macrocyclization reaction off bead of S1, S4-S6, S11

7 mg of peptide-bound PEGA1900 resin was swelled in water for 30 mins. The reactions were conducted in 20 mM bicine, 500 mM NaCI, and 5% DMSO solution, pH = 8.1 and incubated at 37 °C on a shaking platform for 1 week. The reaction set-ups were prepared in the following order; final concentrations and weight of resin:

7 mg of peptide-bound PEGA1900 resin (S1 , S4-S6, S11)

DMSO; 5%

20 mM bicine, 150 mM NaCI, pH = 8.1 buffer

5mM NaCI; 500 mM PatGmac enzyme; 60 mM

Trypsin enzyme; 12 pg

Total volume = 1 ml

A 50 pi aliquot of the reaction mixture was then injected in the LC-MS for monitoring the enzymatic reaction.

2.9 Trypsin cleavage of S4-S6

7 mg of peptide-bound PEGA1900 resin was swelled in water for 30 mins. The reactions were conducted in 20 mM bicine, 500 mM NaCI, and 5% DMSO solution, pH = 8.1 and incubated at 37 °C on a shaking platform overnight. The reaction set-ups were prepared in the following order; final concentrations and weight of resin:

7 mg of peptide-bound PEGA ₁₉₀₀ resin (S4-S6)

DMSO; 5%

20 mM bicine, 150 mM NaCI, pH = 8.1 buffer

5M NaCI; 500 mM

Trypsin enzyme; 12 pg

Total volume = 1 ml

A 50 mI aliquot of the reaction mixture was then injected in the LC-MS for monitoring the enzymatic reaction.

2.10 PatGmac macrocyclization of S13

40 mg of peptide-bound tentagel resin (0.21 mmol/g) were cleaved for 30 mins with NaOH 0.1 M (3 ml) and after quenched with HCI 0.1 M (3 ml). Then, the PatGmac buffer (20 mM bicine, 150 mM NaCI, pH = 8.1), NaCI (350 mM), DMSO (5%) and PatG mac (50 mM) were gradually added. The incubation was carried out at 37 °C until full conversion of the cleaved linear peptide (MALDI monitoring). The reaction set-ups were prepared in the following order; final concentrations and weight of resin:

40 mg of peptide-bound tentagel resin (S13)

DMSO; 5%

20 mM bicine, 150 mM NaCI, pH = 8.1 buffer

5M NaCI; 500 mM

PatGmac enzyme; 50 pM

Total volume = 30 ml

The reaction mixture was finally extracted three times with n-butanol (BuOH). BuOH (1/1 , v/v) was added to the aqueous reaction, vigorously mixed, and then centrifuged for 10 mins at high speed to help separate the two phases. The combined BuOH fractions were evaporated under reduced pressure to dryness. The crude was solubilized in a minimum volume of H ₂ 0/MeOH for HPLC purification.

2. 11 One-pot heterocyclization, macrocyclization of 1

40 mg of peptide-bound chemmatrix resin (0.53 mmol/g) were cleaved for 30 mins with NaOH 0.1 M 5mM DTT (2 ml) and after quenched with HCI 0.1 M (2 ml). Then, the one pot buffer (100 mM bicine, 150 mM NaCI, pH = 9.0) and LynDfusion (10 pM) were gradually added. The incubation was carried out at 29 °C until completion of the heterocyclization step (MALDI monitoring). The reaction set-ups were prepared in the following order; final concentrations and weight of resin:

40 mg of peptide-bound chemmatrix resin (1)

100 mM bicine, 150 mM NaCI, pH = 9.0 buffer

LynDfusion enzyme; 10 mM

Total volume = 30 ml

Then, the salt concentration was increased to 500 mM, DMSO was added (to help with the solubilization of products), and PafGmac enzyme was added to the reaction mixture. Below are the final concentrations for the macrocyclization reaction:

DMSO; 5%

5M NaCI; 500 mM

PafGmac enzyme; 50 mM

The reaction mixture was finally extracted three times with BuOH. BuOH (1/1 , v/v) was added to the aqueous reaction, vigorously mixed, and then centrifuged for 10 mins at high speed to help separate the two phases. The combined BuOH fractions were evaporated under reduced pressure to dryness. The crude was solubilized in a minimum volume of HSO/MQOH for HPLC purification.

2.12 ArtGox oxidation of 4A

The purified cycle 4A (1 .3 mg, 0.0018 mol) was solubilized in 25 ml of 100 mM bicine, 150 mM NaCI, pH = 9.0 buffer (5% DMSO). FMN cofactor (200 mM) and ArtGox (5 pM) were gradually added. The incubation was carried out at 29 °C for roughly 3hrs until quantitative oxidation (MALDI monitoring). The reaction mixture was finally extracted three times with n-buthanol (BuOH). BuOH (1/1 , v/v) was added to the aqueous reaction, vigorously mixed, and then centrifuged for 10 mins at high speed to help separate the two phases. The combined BuOH fractions were evaporated under reduced pressure to dryness. The crude was solubilized in a minimum volume of H ₂ 0/Me0H for HPLC purification.

2.13 One-pot heterocyclization, macrocyiization, oxidation of S16

0.2 mg of peptide-bound chemmatrix resin (0.53 mmol/g) were cleaved for 30 mins with NaOH 0.1 M 5mM DTT (50 pi) and after quenched with HCI 0.1 M (50 pi). Then, the one pot buffer (100 mM bicine, 150 mM NaCI, pH = 9.0) and /cDfusion (30 pM) were gradually added. The incubation was carried out at 29 °C for roughly 12hrs until completion of the heterocyclization step (MALDI monitoring). The reaction set-ups were prepared in the following order; final concentrations and weight of resin:

0.2 mg of peptide-bound chemmatrix resin (S16)

100 mM bicine, 150 mM NaCI, pH = 9.0 buffer

/W/cDfusion enzyme; 30 pM

Total volume = 1 ml

Then, the salt concentration was increased to 500 mM, DMSO was added (to help with the solubilization of products), and PafGmac enzyme was added to the reaction mixture. Below are the final concentrations for the macrocyclization reaction:

DMSO; 5%

5M NaCI; 500 mM

PafGmac enzyme; 50 pM The incubation was carried out at 29 °C for roughly 24hrs until completion of the macrocyclization step (MALDI monitoring). The reaction mixture was finally extracted with BuOH. BuOH (1/1 , v/v) was added to the aqueous reaction, vigorously mixed, and then centrifuged for 10 mins at high speed to help separate the two phases. The BuOH fraction was evaporated with a SpeedVac. The crude was solubilized in 60 pi of H ₂ 0/ e0H 1/1 and 50 pi were injected in the LC-MS.

2.14 One-pot heterocyclization, macrocylization, oxidation of 6, S16, S26, S30, S35, S39, S43

40 mg of peptidyl-chemmatrix resin (0.53 mmol/g) cleaved for 30 mins with NaOH 0.1 M 5mM DTT (2 ml) and after quenched with HCI 0.1 M (2 ml). Then, the one pot buffer (100 mM bicine, 150 mM NaCI, pH = 9.0) and M/cDfusion (30 mM) were gradually added. The incubation was carried out at 29 °C until completion of the heterocyclization step (MALDI monitoring). The reaction set-ups were prepared in the following order; final concentrations and weight of resin:

40 mg of peptide-bound chemmatrix resin (6, S16, S26, S30, S35, S39, S43)

100 mM bicine, 150 mM NaCI, pH = 9.0 buffer

LynDfusion enzyme; 10 mM

Total volume = 30 ml

The incubation was carried out at 29 °C for 48hrs until completion of the macrocyclization step (MALDI monitoring). Afterwards, FMN cofactor (200 mM for 23,25,26 and 1 mM for 24) and ArtGox enzyme (5 mM for 23,25,26 and 30 mM for 24) were gradually added. The incubation was carried out at 29 °C for 3hrs (23,25,26) or 2 days (24) until quantitative oxidation (MALDI monitoring). The reaction mixture was finally extracted three times with BuOH. BuOH (1/1 , v/v) was added to the aqueous reaction, vigorously mixed, and then centrifuged for 10 mins at high speed to help separate the two phases. The combined BuOH fractions were evaporated under reduced pressure to dryness. The crude was solubilized in a minimum volume of H ₂ 0/MeOH for HPLC purification.

2.15 Production of Combinatorial Library

2.15.1 Solid support for solid phase peptide synthesis

SynPhase™ Lanterns (Mimotopes Pty Ltd, AU) were selected for synthesis of the library as they eliminate the limitations of weighing and tracking of compounds which are attributed to traditionally used resins for solid phase peptide synthesis (Duvall et al, 2010 Curr Protoc Chem Biol. 2010;2(3):135-151). Synphase Lanterns are also easier to handle and wash; the latter is important in the Fmoc-based coupling process of SPPS as the lantern’s cylindrical shape enables free flow of solvents. Inefficient washings can lead to incomplete coupling, that is often due to residual volumes of cleavage agents from deprotection steps and coupling by-products being retained within resins.

2.15.2 Combinatorial library design

Previous in-house studies have shown that the best positions for incorporating diversity are at the 5th and 7th position, from the N-terminal, of the peptide sequence of the patellamide analogue‘ITACITAC’.

Hence, diversification was achieved by employing 2-point diversity at these amino acid positions (Table 3). Other constant sites in the sequence allowed for inclusion of heterocyclizable residues (Cys, Ser and Thr). To create diversity, ten amino acids including natural and unnatural amino acids were selected. These included those reported to promote potential for obtaining cell permeable compounds, such as asparagine and fluorinated amino acids (Shah et al J Enzyme Inhib. Med. Chem 2007, 22, 5, 527-540; Buckton et al Org. Lett. 2018, 20, 3, 506-509). Compound set 1 (below) shows all amino acids applied in the design of the library.

The patellamide analogue‘ITACITAC’ was used as the parent compound and incorporated the macrocyclase enzyme recognition sequence‘AYD’ at its C-terminal with a spacer 8-amino-3, 6- dioxaoctanoic acid (Doc) between the linker and aspartic acid (D). Other positions in the compound were kept constant and a grid approach was used to design a 100-membered library (Table 3).

2.15.3 Synthesis

A mix-split-parallel’ approach was applied in the synthesis of the library. For this, all lanterns were initially mixed to couple the linker and the 1 st monomer of the sequence (Fmoc-Doc). The HMBA linker (80 mM) was coupled to non-functionalized SynPhase PA D-Series Lanterns (8 pmol loading) in a double coupling reaction overnight, initially using 96 mM HOBLH2O and 80 mM diisopropylcarbodiimide (DIC), followed by 80 mM OxymaPure with 96 mM DIC. Linear compounds were synthesized using Fmoc chemistry. The Fmoc-Doc spacer (80 mM) was coupled to the linker via esterification reaction with 80 mM 1 -(2- mesitylenesulfonyl)-3-nitro-1 ,2,4-triazole and 60 mM 1 - methylimidazole. This was followed by manually coding the lanterns using colored cogs and spindles (Mimotopes, Pty Ltd, AU) for easy identification of each synthesized compound. After which lanterns were manually split into the 5 mL syringe reactors (3 lanterns per reactor) of a Biotage Syrowave parallel synthesizer, to synthesize each compound. Coupling of amino acids to lanterns was performed with final concentrations of 0.2 M HBTU, 0.2 M amino acids and 0.4 M DIPEA/NMP solution. Stocks of 0.5M, 0.5M and 2M of HBTU, amino acid and DIPEA/NMP, respectively, were used in the synthesizer. HBTU and amino acids were dissolved in DMF except for 4- fluorophenylalanine and N-methyl alanine which were dissolved in a solution of DMSO and DMF (1 :1 v/v). For the latter, amino acids were initially dissolved in half volume of DMSO and volume made up with DMF. Double coupling reactions, one hour each at room temperature, were performed for all amino acids in a final 2.5 mL volume. The library was synthesized in batches of 20 compounds at a time. All reactions were performed at room temperature, the overall workflow is represented in Figure 8.

2.15.4 Tagging compounds in the library

To keep track of compounds within the library, a color-based tagging system was employed by using commercially available cogs and spindles which easily fit into the lantern and are chemically compatible with the solvents used during SPPS. The spindles were used to identify the amino acids in the 7th position, each amino acid in the 7th position had a different colored spindle. Attached to each spindle were cogs which denoted the amino acids in the 5th position of each peptide sequence.

2.15.5 Multi-well Cleavage of Linear compounds from Support

Selected batches of lanterns were mixed and N-terminal Fmoc-groups of the final compounds deprotected at room temperature for 30 minutes with 20% v/v piperidine/DMF in a 250 mL SCHOTT Duran bottle, with holes drilled into the cap to drain waste solvents. Lanterns were washed sequentially with DMF, DCM, methanol and diethylether and dried under a gentle stream of nitrogen gas. Washes were done at least 5 times with each solvent. After which, the side-protecting groups were deprotected with a mix of trifluoroacetic acid/triisopropylsilane/dithiotreitol and water (94:2.5:2.5:2.5 ratio) for 2 hours at room temperature in the same bottle. TFA deprotection can be reduced to 1 hour using the CEM Razor Rapid Peptide Cleavage System. After cleavage, lanterns were washed with copious amounts of DCM, then 5% v/v DIPEA in DCM and finally with DCM and dried before cleavage.

To cleave the linear compounds from the lanterns, cogs and spindles on each lantern were removed and 1 -2 lanterns per compound placed in one-well of a deep 2 mL 96-well plate. The well positions were coded with the standard alphanumeric coding present on commercially available multi-well plates. Each alphanumeric code was linked to the color-code of the lantern and stored digitally for downstream decoding of compounds.

Linear compounds were cleaved using 0.1 M NaOH and 5 mM DTT for 30 minutes at room temperature, as previously described and neutralized with a standard solution of HCI (0.1 N). 100 mM of the cleaved linear product may then be retrieved, for example into a daughter plate, for‘one-pot’ enzymatic transformation using heterocyclase, macrocyclase and oxidase enzymes.

Selected linear compounds may be biosynthetically transformed to cyclic products using heterocyclase (MicDfusion), macrocyclase (PatGmac), oxidase (ArtGox) and/or other biosynthetically transforming enzymes in a multi-well plate format. MicDfusion creates oxazoline- and thiazoline-containing heterocycles from serinesAhreonines and cysteines via cyclodehydration, while PatGmac creates macrocycles via C-N terminal cyclization of the heterocycles. The resulting heterocyclic and/or macrocyclic compounds may then be extracted/purified using a manual multi-well plate liquid-liquid extraction process with 1 -butanol. To obtain the final compounds, the solvents may be dried with a heated stream of nitrogen gas on a Biotage® SPE Dry 96 equipment.

2.15.6 Improvement of MicDfusion reaction conditions

For the MicDfusion enzyme, under standard conditions, the desired 4 heterocycle product is produced with a stochiometric 3 heterocycle impurity. To improve yield of the desired product, an experimental analysis was undertaken based on a Plackett-Burman screening design. Using ITACITACAYD-Doc as a test substrate, 6 factors were investigated of which, increasing the concentration of Substrate (100 mM to 300 mM), MicDFusion (from 30 mM to 45 mM) and dithiothreitol (from 5 mM to 12 mM) caused a 72% (average of 2) decrease in the impurity.

2.15.7 Optimized procedure (0.5 mL scale)

An Eppendorf tube (1 .5 mL) containing Bicine/NaCI (100/150 mM), dithiothreitol (12 mM), ATP (5 mM) and MgCL (5 mM) may be added to a heating block (40 °C), after 10 minutes, ITACITACAYD-Doc (300 mM) and MicDfusion (45 mM) may be added and the reaction allowed to stand at 40 °C for 16 hr. See below for purification.

2.15.8 Additional purification protocols Additional extraction/purification protocols for macrocyclic compounds have been developed using acetonitrile. An equal volume of cold acetonitrile (stored in fridge) was mixed with the reaction mixture and the reaction vessel (e g. 15 mL Falcon tube) was cooled in an ice bath for 10 minutes. After the formation of a biphasic system, the top layer was removed via liquid/liquid extraction (using a separating funnel). The aqueous layer was re-extracted with an equal volume of cold acetonitrile, concentrated under reduced pressure and freeze-dried from water. Greater than 90% of macrocycle product was extracted in the first portion of acetonitrile, with no product detected in the aqueous layer (LCMS analysis) after the second extraction.

2.15.9 Analysis of compounds

Crude linear compounds were analyzed using Direct Detect™ infrared spectroscopy for measurement of yields and LCMS for mass-based verification of synthesized compounds. An ammonium carbonate (5 mM)/acetonitrile method was used for the LCMS run with a gradient from 5 - 95% acetonitrile for 15 mins, with Macherey Nagel EC 250/4.6 NUCLEODUR 300-5 C18 ec column, 1 mL/min flow with MS scan from 100 - 3000 m/z. At intervals, after a transformation, such as heterocyclization, macrocyclization or oxidation, compounds may be analyzed by LCMS to confirm the transformation before moving to the next enzymatic reaction.

3. _ Results

3.1 One-pot processing of basic cleavable resin-bound peptides

Studies began with an exploration of the use of the macrocyclase PatGmac in the production of cyclic peptides by cyclization of resin-bound peptides. This strategy was appealing because product be released from the bead only when a successful cyclization had occurred. Whilst it did prove possible to make cyclic peptides this way using PatGmac, the amount of peptide produced was in general too low (Table 1) and so an alternative strategy was selected.

In our second approach, we used automated solid phase synthesis to prepare a linear peptide precursor in which the side chains of the amino acids had been deprotected (Figures 1 C and 2). This required the use of an acid-stable linker system. The system may also require a spacer unit between the linker and the developing peptide chain with the AYD-motif that is essential for PatGmac-rnediated processing. Subsequent release of the peptide into a buffer solution may be achieved under basic conditions. After adjustment of the pH, the use of different enzyme combinations would lead to the controlled production of a range of complex cyclic peptide derivatives based on the naturally occurring patellamide system (Figures 1 C and 2).

This second approach was successfully achieved using a HMBA linker system on amino-ChemMatrix synthesis resin (100-200 mesh, 0.53 mmol/g) with a Doc spacer unit (Doc = 8-amino-3, 6-dioxaoctanoic acid) (Figures 3 and 4). The HMBA linker was chosen as a suitable linker for the following reasons: (i) resin-bound peptides can be deprotected with TFA, while remaining attached to the resin; (ii) the cleavage of the HMBA linker to give C-terminal acid peptides does not require harsh conditions (e.g. sodium borohydride, hydrazine, ammonia are not required as with other linkers). The cleavage conditions required for the HMBA linker are compatible with the subsequent addition of enzymes; (iii) commercially available HMBA resins (tentagel/chemmatrix resin) are significantly cheaper than PEGA1900.

The peptide sequence VGAGIGWPAYD-Doc (Doc = 8-amino-3,6-dioxaoctanoic acid) was prepared (Figure 4A). The incorporation of tryptophan in the sequence allowed monitoring of the reaction at 280 nm and the Doc spacer was inserted to prevent diketopiperazine formation at the dipeptide stage. After experimentation of alternative basic buffers for the HMBA cleavage, the most efficient procedure was to employ the following sequence: (i) 30 mins 0.1 M NaOH treatment; (ii) quench with an equal volume of 0.1 M HCI and (iii) dilute to the final volume with the PatGmac enzyme and buffer (Table 2). The macrocycle was then extracted with n-butanol, concentrated by rotary evaporation and purified by RP- HPLC to give macrocycle S15, which was characterized by LC-MS, MS-MS and NMR (see below for analytical data, yield = 45%; 1 .1 mg).

The peptide sequence VGAGIGWCAYD-Doc (1) was prepared and incubated with LynDfusion in the presence of ATP and MgCh overnight. The engineered fusion enzyme LynDfusion ²⁶ allows removal of the leader sequence and to process a short linear peptide (2, 12 amino acids; figures 3A and 4B). The linear thiazoline-containing peptide 3 was detected by MALDI TOF analysis but not isolated (Figures 3A and 4B). Subsequent incubation with PatGmac yielded macrocycle 4 (Figure 4B), which was produced as a mixture of two separable diastereomers in 40% yield. Cycle 4 shows two separable peaks in a 4:1

(4A/4B) ratio in the analytical HPLC trace. Both compounds had identical HRMS and MS-MS fragmentation data. However, the NMR spectra of these two compounds were different, but confirmed that 4A and 4B had the same peptide sequence [cyclic VGAGIGWThz]. Interestingly when the separated diastereomers were treated with the FMN dependent oxidase ArtGox ⁴² , able to oxidize thiazoline to thiazole and oxazoline to oxazole (Figure 3A and 4C), only one of the two (4a) was successfully converted to the corresponding thiazole 5.

After synthesis of the resin bound peptide S16 (VTACITWCAYD-Doc), cleavage of the HMBA linker led to the release of peptide S17 which was incubated with LynDfusion in the presence of ATP and MgCL overnight and the product S18 was detected by MALDI-TOF analysis but not isolated. Subsequent incubation with PatGmac yielded macrocycle S19. ArtGox oxidation of S19 led to S20 (figure 4D). The cycles S19 and S20 were extracted in n-butanol and characterized as crude samples by LC-MS, MS-MS.

In addition, the resin bound peptide S16 (VTACITWCAYD-Doc) was cleaved from the resin to give peptide S17 which was incubated with MicDfusion in the presence of ATP and MgCh overnight and the product S21 was detected by MALDI-TOF but not isolated. Subsequent incubation with PatGmac yielded macrocycle S22 (figure 4E). Very different final products (S19 and S20 above and S22 in this case) were therefore delivered from the same resin bound peptide using the approach described herein.

The peptide VTACITFCAYD-Doc (6) was synthesized and cleaved from the resin as before to give peptide S23 (VTACITFCAYD-Doc) which was then incubated with MicDfusion in the presence of ATP and MgCh overnight to generate the linear peptide with four heterocycles 7, which was detected by MALDI-TOF but not isolated (Figures 3B and 4F). Subsequent incubation with PatGmac yielded macrocycle S24. The cycle S24 was purified by prep-HPLC (system P4) and characterized by LC-MS, MS-MS, NMR.

MicDfusion is an engineered ²⁶ version of the enzyme MicD from the microcyclamide pathway (M.

aeruginosa ) ⁴⁴ . While LynDfusion will only introduce thiazolines, MicDfusion introduces thiazoline, oxazoline and methyl oxazoline heterocycles into a peptide backbone (from cysteine, serine and threonine residues, respectively).

The peptide VTACITWCAYD-Doc (S16) was synthesized and cleaved as described above to give S17 (VTACITWCAYD-Doc) in solution. S17 was incubated with MicDfusion in the presence of ATP and MgCh overnight and the product S21 was detected by MALDI but not isolated (Figure 4G). Subsequent incubation with PatGmac and ArtGox oxidation yielded macrocycle S25 (Figure 4G). The cycle S25 was purified by prep-HPLC (system P4) and characterized by LC-MS, MS-MS, CD and NMR.

The peptide VTACITACAYD-Doc (S26) was synthesized and cleaved from the resin to give S27

(VTACITWCAYD-Doc). S27 was incubated with MicDfusion in the presence of ATP and MgC overnight and the product S28 was not isolated (figure 4H). Subsequent incubation with PatGmac and ArtGox oxidation yielded macrocycle S29 (figure 4H). The cycle S29 was purified by prep-HPLC and characterized by LC-MS, MS-MS and NMR.

The peptide VTACITFCAYD-Doc (6) was synthesized and cleaved from the resin as described above to give peptide S23. S23 was then incubated with MicDfusion in the presence of ATP and MgC overnight and the product 7 was detected by MALDI but not isolated (Figure 4J). Subsequent incubation with PatGmac and ArtGox oxidation yielded macrocycle 8 (epi-epi-patellamide C, Figure 4J). The cycle 8 was purified by prep-HPLC and characterized by LC-MS, CD, MS-MS and NMR. Macrocycle 8 (epi epi patellamide C) was isolated as a single diastereomer (Figure 3B). Surprisingly, comparison of the NMR spectrum of 8 with the literature reported NMR for patellamide C ⁴⁴ showed that the two cycles were not identical (Figure 5A). Since we were unable to source any patellamide C for a doping NMR experiment, a one-pot synthesis of patellamide D was carried out in the same way as for patellamide C. The NMR spectrum of patellamide D produced in house (S33, epi epi patellamide D) again differs from the authentic cycle and interestingly resembles the NMR of epi epi patellamide C (8). Indeed, a doping experiment was accomplished employing epi epi patellamide D and native patellamide D, which finally demonstrates that an unassigned epimer of authentic patellamide is obtained (Figure 5B).

Resin bound peptide S30 (ITACITFCAYD-Doc) was cleaved to give S31 as described above. S31 was then incubated with MicDfusion in the presence of ATP and MgCh overnight and the product S32 was detected by MALDI-TOF analysis but not isolated (Figure 4K). Subsequent incubation with PatGmac and ArtGox oxidation yielded macrocycle S33 (epi-epi-patellamide D, Figure 4K).

Resin bound peptide S16 was cleaved to give S17 (VTACITWCAYD-Doc). S17 was then incubated with MicDfusion in the presence of ATP and MgCh overnight and the product S21 was detected by MALDI- TOF but not isolated (figure 4L). Subsequent incubation with PatGmac and prolonged ArtGox oxidation yielded macrocycle S34 (figure 4L). The cycle S34 was purified by prep-HPLC (system P4) and characterized by LC-MS and NMR.

Resin bound peptide S35 was cleaved to give S36 (VSACISFCAYD-Doc). S36 was then incubated with MicDfusion in the presence of ATP and MgCh overnight and the product S37 was not isolated (figure 4M). Subsequent incubation with PatGmac and prolonged ArtGox oxidation yielded macrocycle S38 (figure 4M). The cycle S38 was purified by prep-HPLC (system P4) and characterized by LC-MS and NMR.

Resin bound peptide S39 was cleaved to give S40 (VCACACACAYD-Doc). S40 was then incubated with MicDfusion in the presence of ATP and MgCh overnight and the product S41 was detected by MALDI- TOF analysis but not isolated (figure 4N). Subsequent incubation with PatGmac and prolonged ArtGox oxidation yielded macrocycle S42 (figure 4N). The cyclic peptide S42 was purified by prep-HPLC (system P4) and characterized by LC-MS and NMR

The peptide ITACITACAYD-Doc (11) was synthesized and cleaved from the resin to give S43

(ITACITACAYD-Doc). S43 was incubated with MicDfusion in the presence of ATP and MgC overnight and the product 12 was not isolated (figure 40). Subsequent incubation with PatGmac and prolonged ArtGox oxidation yielded macrocycle 13 _e (figure 40). Cycle 13 _e was purified by prep-HPLC (system P5) and characterized by LC-MS, MS-MS and NMR.

OxZ model dipeptides

With the purpose of identifying the structural differences between our analogue and the natural species we primarily focused on the differences between the two NMR spectra as regards patellamide C (figure 3A). For example, a significant difference is related to the J (a-CH, b-CH) coupling values in the OxZ ring. In the case of epi epi patellamide C (8) J 6.7, 7.5 Hz for the two OxZ rings, which are higher than the corresponding ones for patellamide C (J 2.5 Hz). Patellamide C, D (figures 1A, 1 B) contain OxZ units in anti arrangement (aS.pR) for the two substituents. According to the proposed enzymatic mechanism of MicD (figure 6A) ⁴⁶ , L-threonine gets converted to an anti (aS.pR) OxZ. A possible hypothesis for the difference in the J couplings of the OxZ units for the two structures under discussion may be that the MicDfusion heterocyclization mechanism is different from the expected one and instead equal to the chemical formation of OxZ rings from L-threonine and a//o-threonine (SN2 like), which proceeds through the activation of the threonine b-OH as a better leaving group ⁴⁷ . Such an alternative mechanism would lead to a syn (aS.pS) OxZ as a result of a b-inversion.

In order to get insights into the actual OxZ conformation for our analogues, the synthesis of Fmoc-L-lle- OxZ-O-allyl model compounds (9,10) (figure 6B) with either an anti (a S,pR) arrangement or a syn (aS,pS) arrangement of the two substituents was planned. The experimental and DFT computed ./ couplings for the OxZ units in the case of 8, 9, 10, patellamide C are listed in figure 6C. The experimental J coupling of the anti model dipeptide (9, J = 7.0 Hz) is close to the experimental J values measured for epi epi patellamide C (8, J = 7.5, 6.7 Hz). This evidence supports the argument that our patellamide analogue contains an anti (aS, R) OxZ as in authentic patellamide. In addition, the DFT predicted J values are overall in good agreement with the experimental data. Stereochemical assignment of the patellamide analogues

After demonstrating that our patellamide analogues contain OxZ rings with the expected arrangement of the substituents (e.g. anti) and on the way to the final goal, we considered to simplify the patellamide chemical structure in terms of total number of stereocenters (potentially prone to epimerization) by enzymatically converting OxZ to OxH. The stereochemical issue had to originate from a different part of the molecule and as a result of this argument, the absolute configuration was assumed not to change after the introduction of OxH.

A chiral analysis performed on cycle S24 (V-OxZ-A-ThZ-l-OxZ-F-ThZ) ⁴⁸ , which is the ThZ precursor of epi epi patellamide C (8), showed that the stereocentres next to OxZ preserve their stereochemical integrity during the enzymatic incubation and therefore epi epi patellamide C (8) can adopt one out of four different absolute configurations since only the two stereocenters next to ThH are prone to exchange (figure 7A). Beforehand, we produced a few analogues containing ThH and/or OxH in order to expand the scope of our chemoenzymatic route and establish a robust technology, such as VTA-ThH-ITW-ThH (S20), V-OxH- A-Th H- 1-OxH-W-Th H (S34), V-OxH-A-ThH-l-OxH-F-ThH (S38), V-ThH-A-ThH-A- ThH-A-ThH (S42). OxH was formed from either L-threonine (S34) or L-serine (S38).

At a later stage it was necessary to identify a patellamide structure with the following requirements: i) Content of OxH and ThH (see before); ii) simple chemical structure (preferably symmetric); iii) synthesizable employing either enzymes or pure synthetic chemistry approaches in order to perform an NMR comparison of the enzymatic and synthetic cycle.

A good candidate was the structure 1-OxH-A-ThH-l-OxH-A-ThH (figure 5B), which becomes symmetric only if the oalanine stereocentres next to ThH are either LL or DD. In first instance, we excluded the possibility of an asymmetric cycle. Indeed, we decided to prepare the candidate structure (13 _e ) following our reported approach (figure 5C), and the synthetic LL cycle (13 _s ) and DD cycle (14) employing a previously reported chemical synthesis. To our delight, the NMR spectrum of 13 _s was equal to 13 _e ⁴⁹ and the peaks pattern was consistent with a symmetrical species.

Then, a doping NMR experiment was made by progressively introducing in a single NMR tube 13 _e , a mixture of 13 _e and 14 and a mixture of 13 _e , 13 _s and 14. To note, after the introduction of 14 two set of peaks were visible considering the fact that two different species were mixed (figure 7D). The final addition of 13 _s to the previous mixture leads to an increment in the intensity of the peaks of 13 _e (figure 7D) as a definitive proof that 13 _s and 13 _e are the same compound.

Macrocycles hold out considerable promise. Previous work has shown that the enzymes involved in the patellamide pathway are highly promiscuous. In vivo and in vitro studies have shown that non-natural amino acids can be incorporated into macrocycles. More recently, it has been shown that non proteinaceous moieties (triazoles) can be introduced during the enzymatic transformations. Thus the cyanobacterial enzymes, such as the enzymes in the patellamide pathway, may constitute a

biotechnological tool-kit for chemical synthesis. So far, there is the lack of a general method that combines solid-phase synthesis with enzymatic transformations to generate structurally complex cyclic peptides. Such a method would be paramount in the peptide therapeutics field since it can be directly applied in library production and ultimately the screening of cyclic peptides, and the identification of new biologically active macrocycles. Existing studies have been based around one substrate (in vivo one gene) one product; even where the substrates are made on solid phase they are purified before use. These approaches are incompatible with large-scale library generation. Our initial approach of on bead enzymatic modification proved unsuccessful with very low yields and the main limitation has been to source a resin with pores large enough to accommodate enzymes.

We have demonstrated that a combination of resin-bound peptides, a cleavable linker and enzymatic transformations can be carried out progressively in“one-pot” strategy avoiding the need for intermediate purification, improving scalability of macrocycle production (mg scale). We have been able to introduce heterocycles (thiazoline, oxazoline), macrocyclizing substrates and tuning the oxidation to yield thiazole(s) and oxazoles. This synthetic strategy combines the advantages of solid-phase synthesis to prepare a library of highly diverse substrates with enzymes that introduce modifications difficult to achieve by simple chemical means. Employing the novel route, we managed to make 13 patellamide analogues, which have been fully characterized by NMR and MS-MS.

In house production of patellamide C, D allowed us to discover that the current in vitro technology leads to an epimer of authentic patellamide. The stereochemical assignment of epi epi patellamide C, D has been achieved starting with the chemical synthesis of OxZ model compounds and DFT calculations.

After, the chemical and enzymatic synthesis of a suitable OxH, ThH containing analogue provides the final assignment by doping NMR. Above all, it has been shown for the first time that the stereocenters next to ThH do not undergo a spontaneous epimerization under the in vitro enzymatic conditions.

Combinatorial libraries of patellamide analogues using a split/mix/parallel strategy

A‘split-mix-parallel’ solid phase peptide synthesis (SPPS) approach was used in a semi-automated workflow to synthesise a 100-member combinatorial library of linear patellamide analogues on

Synphase™ Lanterns, as described above.

Table 1 : Concentrations of macrocycle S7 and cleaved linear peptides S8-S10 determined by the quantification experiments with the EDANS tagged substrates S4-S6.

Table 2: HMBA cleavage of the resin bound peptide S13 using different basic buffers compatible with PatGmac. 1 .0 mg of S13 was cleaved at RT with 60 pL basic buffer for 1 day, quenched with HCI 0.1 M (50 pL) and a 50 pL aliquot was injected in the LC-MS. Peak area (mAU*s) of S14 was measured for each condition tested. The cleavage yield was determined by comparison with the peak area of S14 obtained with the standard quantitative 30 mins NaOH 0.1 M treatment. Shorter treatments proved to be ineffective. No improvement on the standard conditions was observed.

Table 3: Grid design of a representative 100-member library of linear octapeptides using ten different amino acids at each of two positions, with the remaining 6 positions fixed. Each cell represents a different compound with amino acid (aa) combinations from the 7th and 5th positions of the peptide sequence. These linear compounds serve as precursors for the creation of cyclic compounds.

References

1. Falanga, A. et al. Molecules 22, (2017).

2. Nielsen, D. S. et a!. Chem. Rev. 117, 8094-8128 (2017).

3. Ozores, H. L, Amorin, M. & Granja, J. R. J. Am. Chem. Soc. 139, 776-784 (2017).

4. Albericio, F. & Kruger, H. G. Future Med Chem 4, 1527-1531 (2012).

5. Katsara, M. et al. Curr. Med. Chem. 13, 2221-2232 (2006).

6. Thell, K. et al. Proc. Natl. Acad. Sci. U.S.A. (2016).

7. Kawamoto, S. A. et al.. J. Med. Chem. 55, 1 137-1 146 (2012).

8. Moellering, R. E. et al. DNature 462, 182-188 (2009).

9. Santos, G. B., Ganesan, A., Emery, F. S. Chem. Med. Chem. 11 , 2245-2251 (2016).

10. Nevola, L. et al. Angew. Chem. Int. Ed. 52, 7704-7708 (2013).

11 . Davies, J. S. J. Pept. Sci. 9, 471-501 (2003).

12. White, C. J. & Yudin, A. K. Nat Chem 3, 509-524 (2011).

13. Bandyopadhyay, A. & Gao, J. J. Am. Chem. Soc. 138, 2098-2101 (2016).

14. Wu, X., Bu, X., Wong, K. M., Yan, W. & Guo, Z.. Org. Lett. 5, 1749-1752 (2003).

15. Kohli, R. M., Burke, M. D., Tao, J. & Walsh, C. T.. J. Am. Chem. Soc. 125, 7160-7161 (2003).

16. Sivonen, K. et al Appl. Microbiol. Biotechnol. 86, 1213-1225 (2010).

17. Ireland, C. & Scheuer, P. J.. J. Am. Chem. Soc. 102, 5688-5691 (1980).

18. Ireland, C. M., Durso, A. R., Newman, R. A. & Hacker, M. P. J. Org. Chem. 47, 1807-181 1 (1982).

19. Schmidt, E. W. et al. Proc. Natl. Acad. Sci. U.S.A. 102, 7315-7320 (2005).

20. Schmitz, F. J. et al. J. Org. Chem. 54, 3463-3472 (1989).

21. Milne, B. F et al. J. Chem. Soc., Perkin Trans. 2, 1076-1080 (2002).

22. Schmidt, E. W. et al. Proc. Natl. Acad. Sci. U.S.A 102, 7315-7320 (2005).

23. Koehnke, J. et al. Nat. Struct. Mol. Biol. 19, 767-772 (2012).

24. McIntosh, J. A. et al. J. Am. Chem. Soc. 132, 15499-15501 (2010).

25. Tianero, M. D. B et al. J. Am. Chem. Soc. 134, 418-425 (2012).

26. Koehnke, J. et al. Nat. Chem. Biol. 11 , 558-563 (2015).

27. Oueis, E. et al Angew. Chem. Int. Ed. Engl. 55, 5842-5845 (2016).

28. Oueis, E. et al ChemistryOpen 6, 1 1-14 (2017).

29. Ruffner, D. E., Schmidt, E. W. & Heemstra, J. R. ACS Synth Biol 4, 482-492 (2015).

30. Scott, C. P. et al Proc. Natl. Acad. Sci. U.S.A. 96, 13638-13643 (1999).

31. Tavassoli, A. & Benkovic, S. J. Nat. Pmtocols 2, 1126-1133 (2007).

32. Schmidt, U. & Gleich, P. Angew. Chem. Int. Ed. Engl. 24, 569-571 (1985).

33. Kato, S., Hamada, Y. & Shioiri, T.. Tetrahedron Letters 27, 2653-2656 (1986).

34. Schmidt, U. & Griesser, H. Tetrahedron Leters 27, 163-166 (1986).

35. Garcia-Reynaga, P. & VanNieuwenhze, M. S.. Org. Let. 10, 4621-4623 (2008).

36. Caba, J. M et al J. Org. Chem. 66, 7568-7574 (2001).

37. Houssen, W. E. et al. Angew. Chem. Int. Ed. Engl. 53, 14171-14174 (2014).

38. Sardar, D., Lin, Z., Schmidt, E. W Cell. Chem Biol. 22, 907-916 (2015). 39. Schoof, S., Arndt, H. D. Chem. Commun. 7113-71 15 (2009J.

40. Merritt, E. A. et al Synthesis 22, 3535-3541 (2007);

41 . Garcia-Reynaga, P., VanNieuwenhze, M. S. Tetrahedron Lett. 53, 4989-4993 (2012).

42. Koehnke, J. et al.. Opin. Struct. Biol. 29, 112-121 (2014).

43. Ishida, T., In, Y., Shinozaki, F., Doi, M.. J. Org. Chem. 60, 3944-3952 (1995).

44. Ishida, K„ Nakagawa, H. & Murakami, M.. J. Nat. Prod. 63, 1315-1317 (2000).

45. Authentic patellamide D was provided from Marcel Jaspars group at the University of Aberdeen.

46. Koehnke, J. et al. Angewandte Chemie International Edition 52, 13991-13996 (2013).

47. Wipf, P., Miller, C. P. & Grant, C. M. Tetrahedron 56, 9143-9150 (2000).

48. A commercially available chiral analysis carried out on cycle S24 (V-OxZ-A-ThZ-l-OxZ-F-ThZ) showed that there is no loss of stereochemical integrity for the stereocenters next to OxZ. GC-MS analysis of hydrolyzed S24 was performed by C.A.T. GmbH.

49. 13 _e is the enzymatic cycle and 13 _s the synthetic one.

50. Wipf, P., Fritch, P. C., Geib, S. J., Sefler, A. M . J. Am. Chem. Soc. 120, 4105-41 12 (1998).

Sequences

MSYYHHHHHHDYDDYDIPTTENLYFQGEDEIESAGVSASEVESSATKQKVALHPHDLDER IPGLADLHNQTLGDPQI TIVIIDGDPDYTLSCFEGAEVSKVFPYWHEPAEPITPEDYAAFQSIRDQGLKGKEKEEAL EAVIPDTKDRIVLNDHA CHVTSTIVGQEHSPVFGIAPNCRVINMPQDAVIRGNYDDVMSPLNLARAIDLALELGANI IHCAFCRPTQTSEGEEI LVQAIKKCQDNNVLIVSPTGNNSNESWCLPAVLPGTLAVGAAKVDGTPCHFSNWGGNNTK EGILAPGEEILGAQPCT EEPVRLTGTSMAAPVMTGI SALLMSLQVQQGKPVDAEAVRTALLKTAIPCDPEWEEPERCLRGFVNIPGAMKVLFG Q

SEQ ID NO: 1 PatGmac (with His tag underlined)

MSHHHHHHDYDENLYFQGSQLSSQLAELSEEALGDAGAGAGAGAGAGAMQSTPLLQI QPHFHVEVIEPKQVYLLGEQ ANHALTGQLYCQILPLLNGQYTLEQIVEKLDGEVPPEYIDYVLERLAEKGYLTEAAPELS SEVAAFWSELGIAPPVA AEALRQPVTLTPVGNI SEVTVAALTTALRDIGISVQTPTEAGSPTALNWLTDDYLQPELAKINKQALESQQTWLLV KPVGSVLWLGPVFVPGKTGCWDCLAHRLRGNREVEASVLRQKQAQQQRNGQSGSVIGCLP TARATLPSTLQTGLQFA ATEIAKWIVKYHVNATAPGTVFFPTLDGKIITLNHSILDLKSHILIKRSQCPTCGDPKIL QHRGFEPLKLESRPKQF TSDGGHRGTTPEQTVQKYQHLI SPVTGWTELVRITDPANPLVHTYRAGHSFGSATSLRGLRNTLKHKSSGKGKTDS QSKASGLCEAVERYSGIFQGDEPRKRATLAELGDLAIHPEQCLCFSDGQYANRETLNEQA TVAHDWI PQRFDASQAI EWTPVWSLTEQTHKYLPTALCYYHYPLPPEHRFARGDSNGNAAGNTLEEAILQGFMELVE RDGVALWWYNRLRRPAV DLGSFNEPYFVQLQQFYRENDRDLWVLDLTADLGI PAFAGVSNRKTGSSERLILGFGAHLDPTIAILRAVTEVNQIG LELDKVPDENLKSDATDWLITEKLADHPYLLPDTTQPLKTAQDYPKRWSDDIYTDVMTCV NIAQQAGLETLVIDQTR PDIGLNWK VTVPGMRHFWSRFGEGRLYDVPVKLGWLDEPLTEAQMNPTPMPF

SEQ ID NO: 2 LynD fusion (with His tag underlined)

MSHHHHHHDYDENLYFQGSRLTAGQLSSQLAELSEEALGDAGLEASGAGAGAGAGAK LMQSTPLLQIQPHFHVEVIE PKQVYLLGEQAMYALTGQLYCQILPLLDGQHSREQIVEKLDGEVPSEYIDYVLDRLAEKG YLTEAAPELSSEVAAFW SELGIAPPVAAEALRQSVTLTPVGNI SEVTVAALTTALRDIGI SVQTPTEAGSPTALNWLTDDYLQPELAKINKQA LESQQTWLLVKPVGSVLWLGPVFVPGKTGCWDCLAHRLRGNREVEASVLQQKQAQQQRNG QSGSVIGCLPTARATLP STLQTGLQFAATEIAKWIVKHHVKATAPGTVFFPTLDGKI ITFNHTVIDLKSHVLVRRSQCPSCGDRQILHRQGFEP VKLVSRRKHFTHDGGHRAFTPEQTVQKYQHLVSPITGWTELVRLTDPANPLVHTYKAGHA FGSATTLRGLRNTLKY KSSGKGKTDIQSRASGLCEAIERYSGI FQGDEPRKRATLAELGDLALHPESLLYFSNTQYANREELNAQGSAAAYRW IPNRFDVSQAIDWTPWSLTEQKHKYVPTAFCYYGYPLPEEQRFCKADSNGNAAGNTLEEA ILQGFLELVERDSIAM WWYNRIRRPAVDLSTFDEPYFVDLQQFYQGQNRELWVLDVTADLGI PAFAGFSRRTVGTSERI SIGFGAHLDPTIAI LRALTEVSQVGLELDKIPDDKLDGESKDWMLNVTVENHPWLAPDPSVPMKTASDYPKRWS DDIHTDVMNCVKTAQTA GLEVMVLDQTRPDIGLNVVKVI I PGMRTFWTRFGQGRLYDIPVKLGWLDAPLAEEELNQTNI PF

SEQ ID NO: 3 MicD fusion (with His tag underlined)

MGSSHHHHHHGSDSEVNQEAKPEVKPEVKPETHINLKVSDGSSEIFFKIKKTTP:LR RIMEAFAKRQGKEMDSIRFIY P^JiU AP-O-P^LDMEDNDIIEAHREQI^GDIPTTENLYFQGRTYPFAVSLNSTIQVSTTADGYA ISPANTDPGQS IAPSMVTLPAITGMGDALAHLQAGTATLQQLTQTLSAREGVEAGEQLAATLQQMGDRGWL QYAVLPLAIAEPMVESA ELDLNSPHWTQAKVSLSRFAYQRSHAGGMVLESPLSKFRVKLLDWRSSAILAQLAQPQPL GWVTPPPQIGAETAYQF LNLLWATGFLTVETEAPELKLWEFHNLLFHSRCRQGRHDYPTGDIAASLDIWDEFPWKPP MSGHIVPLPQLSIDAI RQRDKTLTTAIEKRASIREYDENHPITIEQLGELLYRTARIKEIYTHDAEQAELLKAQFG EDFDWGELSRRPYPCGG AMYELEIYLAVRRCAGVKPGLYHYDPLNHQLAQIDAADADIQALLKDAHQSSGEQGMPQV LLMITARFGRLFRKYRS LAYALVLKHVGVLYQNLYLVATNMGLAPCALGAGDSDRFAQATGLDYWESSVGEFMLGSL

SEQ ID NO: 4 ArtGox (with His tag full underlined, sumo dashed underline)

MTIMAIANRVPYNYLREQRIQFMHAHQDAFDVSTVFPLPLFEKLVTELEGSNVIELS CKIEADKLLAGRFLIFD

QENNWHQSLAQALQFLDS IESRVGVEINRESLDKFLAAHINSGKIMGISTGLDLRPELENSSVKIHIMLGENSE

ELVRTAIAIDGSHYPVELAQVLLKDTMMIGFDFFLNGHSEVELYISCSRKKDSLPNN RGESTRYYIRQKFSPKV

SSLLDASDFFVGGFSKANVEPVLYYAFENIKDIPKYFVFNDLGNRVYDFCRSQDSIT MTWIGINERDLDRERLN

NFRLYYRRSFG

SEQ ID NO: 5 LynF Chemical Structures

Compound set 1 - amino acids used to design 100-member library