WO2025078637A1

WO2025078637A1 - Methods for making and screening a library of bacteriophages

Info

Publication number: WO2025078637A1
Application number: PCT/EP2024/078746
Authority: WO
Inventors: Ariel B. LINDNER; Antoine LEVRIER; Vincent Noireaux
Original assignee: Institut National de la Sante et de la Recherche Medicale INSERM; Universite Paris Cite; University of Minnesota Twin Cities; University of Minnesota System
Current assignee: Institut National de la Sante et de la Recherche Medicale INSERM; Universite Paris Cite; University of Minnesota Twin Cities; University of Minnesota System
Priority date: 2023-10-13
Filing date: 2024-10-11
Publication date: 2025-04-17
Anticipated expiration: 2026-04-13

Abstract

The invention relates to a cell-free method for making a library of phage variants comprising the steps of: (a) providing a library of phage genomes in mixture, and (b) subjecting said library of phage genomes to a step of cell-free transcription and translation. The invention also relates to the applications of such method, for instance in phage display or continuous evolution processes.

Description

METHODS FOR MAKING AND SCREENING A LIBRARY OF BACTERIOPHAGES

FIELD OF THE INVENTION

[0001] This disclosure pertains to the field of cell-free bacteriophage synthesis.

BACKGROUND

[0002] Bacteriophages, or phages, are bacterial viruses that are abundant in nature but harmless to humans. Phages are ubiquitous in nature, highly specific to bacteria, and, in addition, are extremely robust and can withstand even harsh conditions. Their role as the main regulators of the microbial balance among the diverse variety of bacteria existing in the ecosystem makes them naturally eminent, but, owing to many interesting characteristics, phages have also become an exceptional tool for many biotechnological applications. The ability of phages to couple their genome to their encapsulating proteins has given rise to a plethora of applications such as phage display (Smith et al, Rosenberg et al)⁴. Amongst other applications, this technique has been recognized as a powerful tool to screen and select binders on the basis of molecular recognition from phage-displayed libraries, assemblies of about 10 billion of phage clones each harboring a different variant of the displayed entity.

[0003] Under natural conditions, the coupling between the genome and the encapsulating proteins is due to the limited number of phages (typically one) that infect and propagate in a single bacterial host cell, assuring that phage-encoded proteins are assembled with their own coding genome. In laboratory settings, genetic phage variant libraries are either used directly for transformation or packaged in vitro used for infection of host cells to propagate and reveal their linked phenotypes.

[0004] Common applications of phages, such as phage display, have so far relied on the production and propagation of phage libraries in bacterial cell cultures. While this environment has been successfully used for decades, it also suffers from limitations, in particular due to the use of bacterial hosts. As a consequence, there remains a need for a robust, safer and more cost-effective way to synthesize and use phage libraries, e.g. in phage display applications for the identification of new peptide and antibody drugs.

SUMMARY

[0005] In one aspect, the invention relates to a cell-free method for making a library of phage variants comprising the steps of:

(a) providing a library of phage genomes in mixture, and

(b) subjecting said library of phage genomes to a step of cell-free transcription and translation.

[0006] In some embodiments, step (a) comprises a step of in vitro assembling the phage genomes from DNA fragments. [0007] In some embodiments, step (a) comprises:

(a1 ) digesting the overlapping DNA fragments by an exonuclease; and

(a2) annealing the overlapping DNA fragments to assemble a phage genome.

[0008] In some embodiments, step (a) comprises synthetizing the library of phage genomes in a same reaction mixture.

[0009] In some embodiments, step (a) comprises synthesizing phage genomes separately and mixing the phage genomes to provide a library of phage genomes in mixture.

[0010] In some embodiments, step (b) comprises contacting the library of phage genomes with a E coli lysate.

[0011] In some embodiments, said E. co// lysate comprises core RNA polymerase and sigma factor 70.

[0012] In some embodiments, said library of phage variants encodes a library of polypeptides, in particular library of heterologous polypeptides.

[0013] In some embodiments, each polypeptide of said library of polypeptides is encoded as a fusion protein with a phage protein, preferably a phage tail protein or a phage coat protein.

[0014] In some embodiments, said library of phage variants comprises a library of mutant phage genomes.

[0015] In another aspect, the invention relates to a method for screening a library of phage variants, comprising the steps of:

(i) making a library of phage variants according to the method of the present disclosure;

(ii) assaying a phenotype of the library of phage variants;

(iii) collecting a subset of phages based on their phenotype; and

(iv) optionally amplifying the collected phage genomes.

[0016] In some embodiments, the method further comprises a step of identifying the sequence of the collected phages.

[0017] In some embodiments, the method comprises repeating one or more cycles of steps (i) to (iv).

[0018] In another aspect, the invention relates to a method for screening a library of polypeptides, comprising the steps of:

I. making a library of phage variants according to the method of the present disclosure, wherein the genome of each phage variant of the library of phage variants comprises a polynucleotide encoding a polypeptide of a library of polypeptides;

II. contacting the library of phage variants with a test compound;

III. collecting a subset of phages which bind to the test compound; IV. optionally amplifying the collected phage genomes;

V. optionally repeating one or more cycles of steps I to IV or II to IV; and

VI. identifying the polypeptides of the library of polypeptides which are expressed by the collected phages, or the polynucleotides encoding such polypeptides.

[0019] In some embodiments, the polypeptides are encoded as fusion proteins with a phage protein, preferably a phage coat protein.

[0020] In some embodiments, the polypeptides are proteins, in particular are antibodies or antibody fragments or are scaffold proteins comprising one or more variable amino acid residues.

DETAILED DESCRIPTION OF THE INVENTION

[0021] The inventors have shown that cell-free transcription and translation carried out on a mixture of phage variants provides a coupling between the phage genotype and its phenotype, in the absence of compartmentalization. This phenotype-genotype coupling in a cell-free environment, was unexpected since, traditionally, phage-display experiments are carried out in compartmentalized environments, typically a bacterial cell, which isolates the different phage genomes from one another. The cell-free coupling evidenced by the inventors enables the cell-free synthesis of a mixture of phages variants, wherein, in the mixture, the phages express a phenotype which correspond to their genotype. This genotype-phenotype coupling has wide application such as, in phage display, therapy or drug screening.

Definitions

[0022] In order that the present disclosure be more readily understood, certain terms are defined. Additional definitions are set forth throughout the detailed description.

[0023] As used herein, the term “bacteriophage” or “phage” refers to a duplodnaviria virus that infects and replicates within bacteria and archaea. Bacteriophages are composed of proteins that encapsulate a DNA or RNA genome. Their genomes may encode as few as four genes (e.g. MS2) and as many as hundreds of genes. Phages replicate within the bacterium following the injection of their genome into its cytoplasm.

[0024] As used herein, the term “exonuclease” refers to an enzyme that cleaves nucleotides one at a time from the end of a polynucleotide chain via a hydrolyzing reaction that breaks phosphodiester bonds at either the 3’ or 5’ end. The “exonuclease” can be a 3’ to 5’ exonuclease or a 5’ to 3' exonuclease. E. coli exonuclease I and exonuclease III are two commonly used 3’-exonucleases that have 3’-exonucleolytic single-strand degradation activity. E. coli exonuclease VII and T7- exonuclease Gene 6 are two commonly used 5’ to 3’ exonucleases that have 5’-exonucleolytic single-strand degradation activity. [0025] As used herein the term, “ligase” refers to an enzyme that establishes a phosphodiester bond between nucleotides in a nucleic acid.

[0026] As used herein, the term “polymerase” refers to an enzyme which catalyzes the polymerization of ribonucleoside triphosphates (including deoxyribonucleoside triphosphates) to make nucleic acid chains.

[0027] As used herein, the term “cell-free”, e.g. in “cell-free method” or “cell-free reaction”, refers to a method or reaction, such as a biosynthetic reaction (e.g., transcription reaction, translation reaction, or both) which is carried out in vitro in the absence of living cells. A cell-free reaction is supported by a set of reagents capable of providing for or supporting such a biosynthetic reaction. For example, to provide for a transcription reaction, a cell-free reaction may comprise promoter-containing DNA, RNA polymerase, ribonucleotides, and a buffer system, to provide for a translation reaction, a cell- free reaction may comprise translation machinery, amino acids and an energy mixture. Cell-free systems can be prepared using enzymes, coenzymes, and other subcellular components either isolated or purified from eukaryotic or prokaryotic cells, including recombinant cells, or prepared as extracts or fractions of such cells. A cell-free system can be derived from a variety of sources, including, but not limited to, eukaryotic and prokaryotic cells, such as bacteria including, but not limited to, E. coli, thermophilic bacteria and the like.

[0028] As used herein, the term "in vitro" refers to an artificial environment and to processes or reactions that occur within an artificial acellular environment. In vitro environments can consist of, but are not limited to, test tubes. The term "in vivo" refers to the natural environment (e.g., an animal or a cell) and to processes or reaction that occur within a natural environment.

[0029] As used herein, the terms “polypeptide” “peptide” and “protein” are used interchangeably herein to refer to polymers of amino acids (natural or unnatural, e.g. synthetic) of any length. The terms also encompass an amino acid polymer that has been modified; for example, disulfide bond formation, glycosylation, lipidation, phosphorylation, or conjugation with a labeling component.

[0030] As used herein, the term “anneal” refers to specific interactions between strands of nucleotides wherein the strands bind to one another substantially based on complementarity between the strands as determined by Watson-Crick base pairing or Hoogstein-type hydrogen bonding. Base-stacking and hydrophobic interactions may also contribute to duplex stability. Conditions for hybridizing probes and primers to complementary and substantially complementary target sequences are well known. In general, whether such annealing takes place is influenced by, among other things, the length of the polynucleotides and the complementarity between the bases, the pH, the temperature, the presence of mono- and divalent cations, the proportion of G and C nucleotides in the hybridizing region, the viscosity of the medium, and the presence of denaturants. Such variables influence the time required for hybridization. Thus, the preferred annealing conditions will depend upon the particular application. Such conditions, however, may be routinely determined by the person of ordinary skill in the art without undue experimentation. It will be appreciated that complementarity need not be perfect; there can be a small number of base pair mismatches that will minimally interfere with hybridization between the target sequence and the single-stranded nucleic acids of the present teachings. However, if the number of base pair mismatches is so great that no hybridization can occur under minimally stringent conditions, then the sequence is generally not a complementary target sequence. Thus, “complementarity” herein is meant that the probes or primers are sufficiently complementary to the target sequence to hybridize under the selected reaction conditions to achieve the ends of the present teachings.

[0031] Percent complementarity of a nucleic acid sequence with a region of a target nucleic acid can be determined by alignment. For example, a compound in which 18 out of 20 nucleotides are complementary to a target sequence, i.e. with two mismatches between its sequence and the target sequence, would represent 90% complementarity. Percent complementarity of an sequence with a region of a target nucleic acid can also be determined using, for example, the GCG (Genetics Computer Group, Program Manual for the GCG Package, Version 7, Madison, Wisconsin) pileup program, or any of sequence comparison algorithms such as BLAST, FASTA or CLUSTALW.

[0032] As used herein, the term “identity” refers to the sequence similarity between two polypeptide molecules or between two nucleic acid molecules. When a position in both compared sequences is occupied by the same base or same amino acid residue, then the respective molecules are identical at that position. The percentage of identity between two sequences corresponds to the number of matching positions shared by the two sequences divided by the number of positions compared and multiplied by 100. Generally, a comparison is made when two sequences are aligned to give maximum identity (optimal alignment). The identity may be calculated by alignment using, for example, the GCG (Genetics Computer Group, Program Manual for the GCG Package, Version 7, Madison, Wisconsin) pileup program, or any of sequence comparison algorithms such as BLAST, FASTA or CLUSTALW.

[0033] As used herein, the term "heterologous " refers to a nucleic acid or polypeptide sequence that is not in its natural environment. For example, a heterologous sequence includes a sequence from one species introduced into another species, e.g. introduced from a foreign species in a bacteriophage genome. A heterologous sequence also includes a sequence native to an organism that has been altered in some way (e.g., mutated, added in multiple copies, linked to nonnative regulatory sequences, etc). Heterologous nucleic acids are distinguished from endogenous nucleic acids in that the heterologous nucleic acid sequences are typically joined to DNA sequences that are not found naturally associated with the nucleic acid sequence in the genome or are associated with portions of the genome not found in nature (e.g., genes expressed in loci where the gene is not normally expressed).

[0034] As used herein, the term "recombinant" refers to a non-natural DNA, protein, cell, seed, or organism that is the result of genetic engineering and was created by human intervention. A "recombinant DNA molecule" is a DNA molecule that does not naturally occur and as such is the result of human intervention, such as a DNA molecule comprised of a combination of at least two DNA sequences heterologous to each other.

[0035] As used herein, the term “transcription” refers to the formation or synthesis of an RNA molecule by an RNA polymerase using a DNA molecule as a template. [0036] As used herein, the term “translation” refers to the formation of a polypeptide molecule by a ribosome based upon an RNA template.

[0037] In one aspect, the invention relates to a cell-free method for making a library of phage variants comprising the steps of:

(a) providing a library of phage genomes in mixture, and

[0038] As used herein, the term “library” refers to a population of at least two entities, in particular at least 3, at least 5 or at least 10 entities, more particularly at least 100 entities, most particularly at least 1000 entities. Accordingly, the term “library of phage genomes” or “library of phage variants” relates to a population of at least two different phage genomes or phage variants. In some embodiments, the library of phage genomes comprises at least 3 different phage genomes, in particular at least 5 different phage genomes, more particularly at least 10 different phage genomes, even more particularly at least 100 different phage genomes, most particularly at least 1000 different phage genomes. For instance, the library of phage genomes comprises from 10 to 10,000 different phage genomes, in particular from 100 to 10,000 different phage genomes or from 100 to 100,000 different phage genomes.

[0039] Phage genomes are said different when they differ by at least one nucleotide mutation, in particular by a substitution, insertion or deletion of one or more nucleotides. In some embodiments, the genomes in the library of phage genomes differ by the mutation of 1 , 2, 3, 4, 5, 6, 7, 8, 9 or 10 nucleotides. In some embodiments, the genomes in the library of phage genomes differ by the substitution, insertion or deletion of one or more polynucleotide sequences. In some embodiments, the genomes in the library of phage genomes differ by the insertion of polynucleotide sequences, wherein a different polynucleotide sequence is inserted in each phage genome. In one embodiment, said polynucleotide sequence is the sequence of a gene or a gene regulatory sequence, e.g. a promoter or an enhancer.

In some embodiments, the gene is selected from an early gene, a DNA replication gene and a phage structure and assembly gene. In some embodiments, the gene is selected from a capsid gene and a tail gene, in particular a tail fiber gene. In some embodiments, the gene is a T7 gene selected from the genes shown in Table 1 or a gene having at least 50%, in particular at least 60%, more particularly at least 70%, 80% or 90%, most particularly at least 95 or 99% sequence identity with a gene shown in Table 1 . In one embodiment, the gene is an ortholog of a T7 gene shown in Table 1 in another phage species. In some embodiments, the gene is selected from genes 1 1 , 12 and 17 of the T7 phage.

Table 1 : List of the T7 phage genes and positions, with their functions when known and the size of the proteins (from GenBank V01 146.1 ).

[0040] At step (a), the library of phages genomes is provided in mixture. As used herein, the term "mixture" refers to any combination of at least two components and includes, for example, blends, dispersions, solutions, emulsions, suspensions, and combinations of any thereof. Entities are said “in mixture” or “mixed” when they are not compartmentalized, either biologically or artificially. Two entities are said to be compartmentalized when they are separated by physical boundaries that prevent them from mixing. Examples of compartments include biological compartments such as a cell, a cellular organelle or a liposome, or artificial compartments such as vials, tubes, containers etc... In some embodiments, the phage genomes are provided in a reaction mixture. A reaction mixture, as used herein, is a mixture comprising at least some reagents for cell-free transcription and/or translation, preferably all required reagents for cell-free transcription and/or translation. In some embodiments, the method comprises a prior step of mixing the phage genomes. In other embodiments, the method comprises a prior step of synthesizing the phage genomes in mixture.

[0041] In some embodiments, the phage genomes of the library of phage genomes are synthesized in vitro. In vitro synthesis of phage genomes is made in a cell-free manner.

[0042] Cell-free synthesis of phages is generally performed by assembly of DNA fragments, typically double-stranded DNA polynucleotides.

[0043] In some embodiments, the adjacent DNA fragments are overlapping DNA fragments. By “overlapping DNA fragments” it is meant that two adjacent DNA fragments comprise complementary sequences at their end such that, after digestion by exonuclease, two adjacent DNA fragments can anneal with each other. In other terms, the sense strand of a DNA fragment comprises a sequence complementary to the antisense strand of an adjacent DNA fragment. Two fragments are said adjacent is they are intended to be joined next to each other in the assembled phage genome.

[0044] Such fragments may be obtained by PCR amplification, e.g. from a template phage genome. Other means to obtain DNA fragments are also encompassed by the invention, such as synthetic DNA.

[0045] Once assembled, the DNA fragments form a phage genome. In addition to the phage sequence, the DNA fragments can also contain heterologous sequences, e.g. heterologous genes or chimeric genes.

[0046] The number of DNA fragments and their size are tailored depending on the size of the phage genome to assemble as well as of the size of recombinant fragments to be introduced. The inventors have shown that long fragments can be used to reconstruct phage genomes. In some embodiments, the DNA fragments are from 100 bp to 25 kbp, in particular from 100 bp to 20 kbp, more particularly from 100 bp to 15 kbp. Small fragments can be used, e.g. fragments of less than 5 kbp, in particular less than 2 kbp. For example, small recombinant fragments of less than 5kbp can be used to introduce a heterologous gene to be introduced within a phage genome. In some embodiments, 2 or more fragments are assembled into a phage genome. In some embodiments, 3 or more, 4 or more, 5 or more, 6 or more, 7 or more, 8 or more, 9 or more or 10 or more fragments are assembled into a phage genome. In some embodiments, the DNA fragments comprise one or more recombinant fragments.

[0047] In one embodiment, overlapping sequences have a sufficient length to allow ligase-free annealing of the fragments. Overlapping sequences of at least 5 bp are preferred, in particular at least 10 bp, more preferably at least 15 bp, still more preferably at least 20 bp. Typically, overlapping sequences of from 10 bp, 15 bp, 20 bp, 25 bp, 30 bp, 35 bp, 40 bp, 45 bp, 50 bp, 55 bp, 60 bp, 65 bp, 70 bp, 75 bp, 80 bp, 85 bp, 90 bp, 95 bp or 100 bp can be used. In particular, overlapping sequences of from 5 to 100 bp can be used, in particular from 5 to 75 bp, from 5 to 50 bp, from 10 to 100 bp, from 10 to 75 bp, from 10 to 50 bp, from 20 to 100 bp, from 20 to 75 bp or from 20 to 50 bp.

[0048] In some aspects of the invention, the overlapping sequences are orthogonal oligonucleotide sequences. The orthogonal oligonucleotide sequences are generally heterologous sequences added at the end of DNA fragments. The orthogonal sequences are unique for each pair of adjacent DNA fragments, such that one end of a DNA fragment only anneals to one end of an adjacent DNA fragment and not to a non-adjacent DNA fragment. Using a library of orthogonal oligonucleotides as overlapping sequences allow to assemble the phage genome in the desired order. This embodiment is particularly useful to insert recombinant fragments of DNA at a desired position within the genome. It prevents re-assembly of the DNA fragments in the form of a non-edited phage, i.e. phages which reassemble without the recombinant DNA fragment. In some embodiments, the method can comprise a prior step of synthesizing DNA fragments with orthogonal oligonucleotide sequences, wherein two orthogonal oligonucleotide sequences of two adjacent DNA fragments are overlapping. The orthogonal oligonucleotide sequences are added, e.g., by PCR amplification. The orthogonal overlapping sequences can have at least 5 bp, in particular at least 10 bp, preferably at least 20 pb.

[0049] The assembly of all DNA fragments typically forms a linear or circular DNA molecule which reconstructs a phage genome. Said genome can be genome-engineered, or can be unengineered, i.e. identical or substantially identical to a template phage genome, e.g. a naturally-occurring phage genome.

[0050] Different seamless cell-free assembly techniques of DNA fragments can be used to synthesize the phage genomes, in particular SLIC (Sequence and Ligation-Independent Cloning), In-Fusion, Gibson assembly, and Golden Gate assembly. SLIC uses 3'-5' exonuclease activity of T4 DNA polymerase to generate 5' overhangs at the ends of insert(s)and a linearized vector. By annealing the overlapping sequences of about 25 bp, insert(s) and the vector are assembled in vitro. In the same way, In-Fusion can assemble DNA fragments using vaccinia virus DNA polymerase with shorter DNA overlapping of about 15 bp. Gibson assembly uses thermostable DNA polymerase and DNA ligase in addition to thermolabile T5 5'-3' DNA exonuclease. After the insert(s) and a linearized vector, DNA is mixed with these enzymes at 50 °C, the exonuclease resects the ends of the DNA fragments to generate 3' overhangs and is inactivated by the heat. Then, the overlapping ends anneal, the DNA polymerase fills the gaps, and the DNA ligase repairs the nicks. Slightly different from these methods, Golden Gate assembly uses type IIS endonucleases. Since the type IIS restriction enzymes cleave DNA sequences distant from recognition sequences, it is possible to leave short single- stranded DNA overhangs of any sequence at the terminal after the cleavage. The DNA fragments with short overhangs are then ligated by T4 DNA ligase. Accordingly, in one embodiment, the assembly of the DNA fragments is assisted by one or more enzymatic activity selected from an exonuclease, an endonuclease, a ligase and a polymerase activity. In some embodiments, a combination of an exonuclease, a ligase and a polymerase is used. In other embodiments, T4 DNA polymerase is used. In other embodiments, IIS endonucleases are used.

[0051] The inventors have shown that exonuclease-digested fragments could anneal in a seamless manner, without a requirement for polymerase activity to fill the gaps and without a requirement for ligase activity for the ligation. The outer ends of the DNA fragments are digested by the exonuclease, containing the overlapping sequences, thereby generating protruding overhangs which are able to anneal.

[0052] Preferably, the exonuclease is the sole enzyme used in the in vitro assembly of the DNA fragments. In some embodiments, the assembly is carried out in the absence of polymerase in the reaction mix. In some embodiments, the assembly is carried out in the absence of ligase in the reaction mix. Preferably, the assembly is carried out in the absence of polymerase and ligase in the reaction mix. Alternatively or in addition, the assembly is carried out in the absence of IIS endonucleases in the reaction mix. In some embodiments, the exonuclease has no polymerase activity.

[0053] The exonuclease is preferably a 3’->5’ exonuclease. Particularly preferred exonucleases include exonuclease III, exonucleases which are homologous to exonuclease III and exonucleases having a comparable activity to exonuclease III. Exonuclease III is an enzyme that catalyzes the stepwise removal of mononucleotides from 3'-hydroxyl termini of duplex DNA. A limited number of nucleotides are removed during each binding event, resulting in coordinated progressive deletions within the population of DNA molecules. The preferred substrates are blunt or recessed 3'-termini, although the enzyme also acts at nicks in duplex DNA to produce single-strand gaps. 3'-protruding termini are resistant to cleavage. Exonuclease III has also been reported to have RNase H, 3'- phosphatase and AP-endonuclease activities. As used herein, exonuclease III includes E. coli exonuclease III (reference sequence Uniprot P09030 as set forth in SEQ ID NO:12). Homologous enzymes are also encompassed by the present invention. The invention encompasses wild-type exonuclease III as well as variant exonucleases III that maintain 3’->5’ exonuclease activity. Preferably, a homologous exonuclease has at least 40%, preferably at least 50%, still preferably at least 60% or 70%, more preferably at least 80 or 90%, most preferably at least 95% or 99% amino acid sequence identity with the amino acid sequence set forth in SEQ ID NO:12. Preferably, the exonuclease has at least 10% of the 3’->5’ activity of exonuclease III more preferably at least 20%, even more preferably at least 30%, 40% or 50%, still preferably at least 60%, 70% or 80%, most preferably at least 90%, 95% or 99% of the 3’->5’ activity of exonuclease III. Testing exonuclease activity may be carried out by measuring the fluorescence level of fluorescence probes, e.g. using the 3’ to 5’ Exonuclease Activity Assay (Sigma-Aldrich, #MAK416) Exonuclease, e.g. with a reference exonuclease III from New England Biolabs (#M0206S). Exonuclease III is commercially-available from various sources, for instance from New England Biolabs (#M0206S). In some embodiments, the exonuclease has 3'-phosphatase activity. In some embodiments, the exonuclease has AP- endonuclease activity. In some embodiments, the exonuclease has RNase H activity. Contacting the exonuclease with the DNA fragments is preferably carried out at a temperature of less than 20°C, preferably less than 15°C. Preferably, the exonuclease is contacted with the DNA fragments on ice. Low temperatures of contacting allow to control exonuclease activity.

[0054] The method advantageously comprises a step of inactivating the exonuclease after contacting the DNA fragments with the exonuclease and before annealing the DNA fragments. Inactivation enables to stop the exonuclease activity after digestion of the ends of the DNA fragments, such as to make annealing of the fragments possible. Preferably, a heat treatment is applied to inactivate the exonuclease. The reaction medium may be heated at a temperature of 60°C or more, preferably 70 °C or more, for instance a temperature of about 75°C.The heat treatment is applied for at least 30s, e.g. from 30s to 10mn. Heat treatment may be applied immediately after contacting the DNA fragments with the exonuclease. Indeed, exonuclease activity during the period of temperature increase to the inactivation temperature is generally sufficient to produce digested fragments which are capable of annealing and assemble a phage genome. Digestion time may also depend on the size of the annealing fragments.

[0055] Annealing of the digested ends is preferably carried out at a temperature of less than 45°C, preferably less than 40°C, more preferably less than 35°C, in particular from 20 to 45°C, preferably from 20 to 40°C or from 20 to 35°C. Annealing is advantageously carried out at room temperature. The digested DNA fragments may be annealed for at least 1 minutes, preferably at least 2 minutes. In particular, the digested DNA fragments may be annealed from 1 to 20 minutes, preferably 1 to 10 minutes, still preferably 2 to 10 minutes, most preferably about 3 to 10 minutes. Annealing is carried out without addition of a ligase and/or polymerase, contrary to the Gibson assembly method. Moreover, annealing can be effected in the same reaction mix as the exonuclease digestion, contrary to other existing methods which require an annealing buffer. Hence, no purification step is required between steps (a1 ) and (a2), i.e. exonuclease digestion, and annealing.

[0056] Preferably, the reaction mix containing the assembled genomes is not purified before the step of cell-free transcribing and translating. Indeed, the reactions in step (a) are compatible with the performance of subsequent steps of cell-free transcription and translation, without purification.

[0057] In one embodiment, the phage genomes of the library of phage genomes are synthesized separately and are subsequently mixed together. Accordingly, the method of the invention may comprise, before step (a), a step of separately synthesizing the phage genomes of the library of phage genomes, and a step of mixing the phage genomes. By “separately synthesizing and/or assembling”, it is meant that, following their synthesis and before being mixed, the phage genomes are compartmentalized, i.e. isolated from each other.

[0058] As used herein, the term “synthesizing” e.g. in “synthesizing a phage genome” encompasses any technique that results in the production of a phage genome. In particular, it encompasses assembling, preferably in vitro assembling, a phage genome from nucleic acid fragments, e.g. DNA fragments.

[0059] In another embodiment, the phage genomes of the library of phage genomes have been synthesized in mixture. Accordingly, the method of the invention may comprise, before step (a), a step of synthesizing the phage genomes of the library of phage genomes in mixture. In particular, said synthesis in mixture is a cell-free synthesis. Cell-free assembly of phage variants in mixture may be carried out by providing DNA fragments designed such as to assemble in a desired phage genome, in a desired order. Cell-free assembly of phage variants in mixture can use orthogonal overlapping sequences designed to assemble adjacent fragments of a given phage genome.

[0060] According to the invention, the library of phage variants is submitted to a cell-free transcription and translation. Various cell-free transcription and translation systems can be used, for instance the hybrid T7 cell-free platform that couples the T7 bacteriophage transcription to the translation of an organism such as E. coli. Another cell-free transcription and translation system which can be used is the PURE system, a fully purified cell-free TXTL, also based on the T7 transcription machinery (Shimizu et al., 2001 ).

[0061] Preferably, cell-free transcription and translation can use an E. co// cell-free system that is independent on the bacteriophage transcription machinery and uses the endogenous E. coil RNA polymerase and housekeeping sigma factor 70. Typically, such TXTL reaction medium is composed of an E. coli lysate, salts, and buffers that provide both the 20 amino acids and an ATP regeneration system. The lysate contains the translation machinery (e.g., E. coli core RNA polymerase, RNAP, and sigma factor 70) and the translation components (e.g., ribosomes and tRNA). The translation speed in a cell-free reaction is typically of between 5 and 10 nucleotides per second. To initiate transcription, RNAP and sigma factor 70 (o70) form a holoenzyme to bind to the promoter, in this case P70a, specifically recognized by o70. Because transcription is based on RNAP-o70 from E. coli, the TX repertoire is composed of hundreds of regulatory elements present in E. coli and other bacteria. The other six E. coli sigma factors are not present in the lysate, and therefore must be expressed and used with their specific promoters. Translation is carried out at a rate of roughly two amino acids per second by the same ribosomal machinery present in E. coli.

[0062] In some embodiments, cell-free transcription and translation comprises adding an E. coli lysate to the assembled phage genome. Preferably, the E. co// lysate comprises the E. coli core RNA polymerase and sigma factor 70. In some aspects, the method comprises further adding an energy mixture for ATP regeneration, wherein said energy mixture comprise a phosphate donor and a carbon source. An example of reaction mix for a TXTL reaction comprises comprised E. coli cell lysate, an energy and amino acid mixtures, maltodextrin and ribose, magnesium and potassium, PEG8000, water and the DNA to be expressed. The reaction can be incubated at 25-35°C, e.g. around 29 °C.

[0063] In some embodiments, the reaction mix at step (b) comprises at least 2% PEG8000, in particular at least 2.5% PEG8000, more particularly at least 3% PEG8000. In some embodiments the reaction mix comprises from 2 to 10% PEG8000, in particular from 2 to 5% PEG8000. All % are given w/v. In some embodiments, the viscosity of the solution is equivalent to the viscosity of a solution of from 2% to 10% PEG8000 solution, in particular from 2% to 5% PEG8000 solution, preferably as measured at 25°C. Dynamic viscosity may be measured in a thermostatized bath with a rotating concentric cylinder viscometer such as Contraves Rheomat 108 E/R. Such viscosity can be brought about by addition of a dispersant, preferably selected from polyethylene glycols. Preferably, the polyethylene glycols have a molecular weight from 200 to 20,000, in particular from 1000 to 20,000, more particularly from 1000 to 10,000. In some embodiments, the dispersant is selected from PEG4000, PEG6000 or PEG8000.

[0064] Exemplary reaction conditions are: 8.9-9.9 mg/mL protein (from crude extract), 3-6 mM Mg- glutamate, 40-100 mM K-glutamate, 2-4% PEG 8000, 3-4 mM of each amino acid, and an energy mix solution, described previously, composed of 0.33-3.33 mM DTT, 50 mM HEPES, 1 .5 mM ATP and GTP, 0.9 mM CTP and UTP, 0.2 mg/mL tRNA, 0.26 mM CoA, 0.33 mM NAD, 0.75 mM cAMP, 0.068 mM folinic acid, 1 mM spermidine, and 30 mM 3-PGA. For DNA-type phages, genome concentrations can be of 0.05-10 nM and for RNA-type phages the genome concentrations can be of 50-200 nM.

[0065] During cell-free transcription, mRNA is produced from the phage genome. During cell-free translation, the phage proteins are synthesized from the mRNA. These phage proteins assemble to form a phage (self-assembly). The self-assembled phage incorporates a phage genome which is packaged within the phage. The inventors have found that cell-free transcription and translation carried out on a mixture of phage variants provides a coupling between the phage genotype and its phenotype, in the absence of compartmentalization of the cell-free transcription and translation, in particular in the absence of cellular or subcellular compartmentalization. The genotype/phenotype coupling is reflected by the fact that at least a part of the phage constitutive proteins preferentially assemble with the phage genome that encodes said protein or proteins. By “preferentially”, it is meant that, in a mixture of phage genomes, phage constitutive proteins assemble more with a genome that encodes them, than with a genome that does not encode them. In other terms, a phage packages its genome in a non-random way: a phage has higher chances to package its own genome, i.e. the genome that encodes its constitutive proteins, than a different genome, i.e a genome that encodes variant constitutive proteins. In some embodiments, the tail proteins, in particular tail fiber proteins, preferentially assemble with the phage genome that encodes said proteins. In some embodiments the capsid proteins preferentially assemble with the phage genome that encodes said proteins. In some embodiments, the tail and capsid proteins preferentially assemble with the phage genome that encodes said proteins.

[0066] This genotype/phenotype coupling is reflected by the proportion of phages that are assembled with their own genomes. This proportion is typically higher than the proportion which would be found if the phage assembled in a random way.

[0067] The method of the invention can be used to make various type of phage variant libraries. In some embodiments, the phage is a filamentous phage. In some embodiments, the phage is an obligate lytic phage. In some embodiments, the phage is a non-lytic phage. In some embodiments, the phage is selected from the T7, T6, T5, lambda, VPAE1 , and FelixOl phages. In some embodiments, the phage is a phage that contains nonessential regions that allow exogenous gene insertions.

[0068] In some embodiments, the synthesized phage comprises at least 10kbp, in particular at least 20 kbp or 30 kbp, more particularly at least 50 kbp, still more particularly at least 75 kbp, even more particularly at least 100 kbp in its genome, most particularly at least 200kbp in its genome.

[0069] The genotype/phenotype coupling enables to use the method of the invention for a myriad of applications. Indeed, the cell-free genotype/phenotype coupling of the method can replace the genotype/phenotype coupling that traditionally occurs during phage propagation in bacterial host cells. As a consequence, the method can be used for applications such as phage display or continuous evolution.

[0070] In some aspects, the invention relates to the use of the method of the invention in phage display. "Phage display" refers to a technique by which variant polypeptides are displayed as fusion proteins to at least a portion of a protein on the surface of phage, in particular a coat protein. A utility of phage display lies in the fact that large libraries of randomized protein variants can be rapidly and efficiently selected for those sequences that bind to a target molecule with high affinity. Display of peptide and protein libraries on phage has been used for screening millions of polypeptides for ones with specific binding properties. Polyvalent phage display methods have been used for displaying small random peptides and small proteins through fusions to either gene III or gene VIII of filamentous phage. Wells and Lowman, Curr. Opin. Struct. Biol, 3:355-362 (1992), and references cited therein. In monovalent phage display, a protein or peptide library is fused to a gene III or a portion thereof, and expressed at low levels in the presence of, wild type gene III protein so that phage particles display one copy or none of the fusion proteins. Avidity effects are reduced relative to polyvalent phage so that selection is on the basis of intrinsic ligand affinity, and phagemid vectors are used, which simplify DNA manipulations. Lowman and Wells, Methods: A companion to Methods in Enzymology, 3:205-0216 (1991 ).

[0071] The present invention thus provides a cell-free phage display method.

[0072] In some embodiments, the invention relates to a method for screening a library of phage variants, comprising the steps of:

(i) making a library of phage variants by cell-free transcription and translation of a library of phage genomes in mixture, in particular as described in the present disclosure;

(ii) assaying a phenotype of the library of phage variants; and

(iii) collecting a subset of phages based on their phenotype.

[0073] In some embodiments, the method comprises a further step of amplifying the collected phage genomes. Said amplification may be effected by propagation in bacterial host cells. Said amplification may also be carried out by direct amplification of the genomes, e.g. by PCR. PCR amplified fragments can then be reassembled according to the cell-free assembly methods described in the present specification.

[0074] In some embodiments, the method comprises a further step of analyzing the sequence of the collected phage genomes. Said analysis step can be effected, e.g. by sequencing such as NGS sequencing or by the use of PCR amplification, or by the use of a DNA probe.

[0075] In some embodiments, the phenotype is the binding of a polypeptide expressed by the phage variant to a ligand. In other embodiments, the phenotype is infection and/or lysis of a host bacterial cell.

[0076] In some embodiments, phage variants of the library of phage variants differ by the presence of one or more mutations in their genomes. Said one or more mutations may be one or more nucleotide substitutions, insertions or deletions.

[0077] In some embodiments, steps (i) to (iii) or (ii) to (iii) are repeated for one or more cycles. Accordingly, the phages collected at step (iii) and optionally amplified can be used as a new library of phage variants to be assayed in a further occurrence of step (ii). Alternatively, the genome of the phages collected at step (iii) and optionally amplified can be isolated and used in a further occurrence of step (i) to prepare a library of phage variants by cell-free transcription and translation. Cycles of steps (i) to (iv) or (ii) to (iv) can be repeated for at least 2 rounds, at least 3 rounds, at least 4 rounds or at least 5 rounds. In particular, 2, 3, 4, 5, 6, 7, 8, 9 or 10 rounds of cycles of steps (i) to (iv) or (ii) to (iv) are carried out. Preferably, 2 to 15 rounds, in particular 2 to 10 rounds, more particularly 2 to 8 rounds of cycles are carried out. In a phage display application, repeating the cycles will narrow the phage variant library onto the phage variants displaying the desired phenotype, e.g. binding to a defined ligand.

[0078] In some embodiments, the phage genomes are mutated between each cycle of steps. Mutations can be introduced randomly or in a directed manner. Through such mutagenesis, continuous evolution can lead to the identification of novel phage variants with a desired property.

[0079] In some embodiments, the library of phage variants encodes a library of polypeptide variants. In some embodiments, the polypeptides are proteins. In some embodiments, the library according to the invention comprises polypeptides of the immunoglobulin superfamily, such as antibody polypeptides or T-cell receptor polypeptides. Advantageously, the library may comprise individual immunoglobulin domains, such as the VH or VL domains of antibodies, or the Vp or Va domains of T-cell receptors. In a preferred embodiment, therefore, repertoires of, for example, VH and VL polypeptides may be individually prescreened using a generic ligand and then combined to produce a functional repertoire comprising both VH and VL polypeptides. Such a repertoire can then be screened with a target ligand in order to isolate polypeptides comprising both VH and VL domains and having the desired binding specificity. In some embodiments, the polypeptide library is a small peptide library. Other examples of libraries which can be used include libraries either based on nanobodies (VHH) or libraries based on scaffold proteins not related to antibodies, e.g. scaffold proteins with one or more variable amino acid residues. The variable amino acid residues may be located in variable loops that have the potential to be diversified. Example of scaffold protein is the CheY protein used in Gomes et al., 2023.

[0080] In some embodiments, the invention therefore relates to a method for screening a library of polypeptides, comprising the steps of:

I. making a library of phage variants by cell-free transcription and translation of a library of phage genomes in mixture, in particular as described in the present disclosure, wherein the genome of each phage variant of the library of phage variants comprises a polynucleotide encoding a polypeptide of a library of polypeptides;

II. contacting the library of phage variants with a test compound;

III. collecting a subset of phages which bind to the test compound;

IV. optionally amplifying the collected phage genomes;

V. optionally repeating one or more cycles of steps I to IV or II to IV; and

VI. identifying the polypeptides of the library of polypeptides which are expressed by the collected phages or the polynucleotides encoding such polypeptides.

[0081] In some embodiments, the test compound is a ligand. The library of phage variants expresses potential binders for this ligand, i.e. peptide or protein binders.

[0082] In some embodiments, the test compound is a substrate for an enzyme. In such embodiments, the library of phage variants may be an enzyme library expressing enzymes which can potentially catalytically process the test compound. Step III may thus comprise collecting a subset of phages which bind to and process the test compound, in particular enzymatically process the test compound.

[0083] In some embodiments, the library of phage variants is prepared by cloning the library of polypeptides variants into a phage. In some embodiments, the library of phage variants is prepared by in vitro assembly of the polynucleotides encoding the library of polypeptide variants, with DNA fragments encoding the phage genome.

[0084] In one embodiment, the polypeptides are encoded as fusion proteins with a phage protein, preferably a phage coat or tail protein. The coat protein can be a full-length coat protein or any portion thereof capable of effecting display of the polypeptide on the surface of the genetic package. Exemplary of M13 coat proteins are phage coat proteins, such as, but not limited to, (i) minor coat proteins of filamentous phage, such as gene III protein (g 11 Ip, cp3), and (ii) major coat proteins (which are present in the viral coat at 10 copies or more, for example, tens, hundreds or thousands of copies) of filamentous phage such as gene VIII protein (gVI lip, cp8); fusions to other phage coat proteins such as gene VI protein, gene VII protein, or gene IX protein (see, e.g., WO 00/71694); and portions (e.g., domains or fragments) of these proteins, such as, but not limited to domains that are stably incorporated into the phage particle, e.g. such as the anchor domain of g II Ip, or g VI I Ip. Additionally, mutants of g VI I Ip can be used which are optimized for expression of larger peptides, such as mutants having improved surface display properties, such as mutant gVIlp (see, for example, Sidhu et al. (2000) J. Mol. Biol. 296:487-495). Examples of T7 phage proteins that can be used in screening applications are proteins encoded by the gp10a gene or gp10b gene. Accordingly, in some embodiments, the polypeptides are encoded as fusion proteins with a phage protein encoded by the gp10a gene or gp1 Ob gene of the T7 phage.

[0085] In some embodiments, steps I to IV or II to IV of the method for screening a library of polypeptide variants are repeated for one or more cycles. Accordingly, the phages collected at step III and optionally amplified at step IV can be used as a new library of phage variants to be assayed in a further occurrence of step II. Alternatively, the genomes of the phages collected at step III and optionally amplified at step IV can be isolated and used in a further occurrence of step I to prepare a library of phage variants by cell-free transcription and translation. Cycles of steps I to IV or II to IV can be repeated for at least 1 round, at least 2 rounds, at least 3 rounds, at least 4 rounds or at least 5 rounds. In particular, 2, 3, 4, 5, 6, 7, 8, 9 or 10 rounds of cycles of steps I to IV or II to IV are carried out. Preferably, 2 to 15 rounds, in particular 2 to 10 rounds, more particularly 2 to 8 rounds of cycles are carried out. In a phage display application, repeating the cycles will narrow the phage variant library onto the phage variants displaying the desired phenotype, e.g. binding to a defined ligand.

[0086] The test compound is preferably immobilized on a support. Any type of test compound can be used depending on the objective of the phage display procedure. The test compound may be selected from an antigen, a full protein, a part of a protein such as a protein domain, a receptor, a chemical compound, e.g. a small molecule and a drug, drug candidate or pro-drug.

[0087] Step III may comprise eluting the phages that do not bind to the ligand, and collecting the phages that bind to the ligand.

[0088] In some embodiments, the method comprises a step of producing the polypeptide(s) identified by the screening method. Methods for producing proteins and peptides are well known from the skilled person.

[0089] In addition to the identification of peptide/polypeptides which bind a ligand or process a substrate, phage display exhibit various other applications (Yue et al., 2022). For instance, phage display has applications in the field of diagnosis to detect and identify new biomarkers, such as serological biomarkers. Through the phage display technique, biomarkers specific of a given condition can be identified. Phage display can also be used to detect pathogens such as bacterial pathogens. Phage display has for instance been shown to be efficacious in the detection of E. coli in drinking water. There are also applications of phage display in cancer therapy and in targeted therapies, such as targeted bacterial therapy or targeted gene therapy.

[0090] The method of the invention can also be used for identifying phage variants that are infective to a host bacterial cell.

[0091] In some embodiments, the invention relates to a method for screening a library of phage variants, comprising the steps of:

(i) making a library of phage variants by cell-free transcription and translation of a library of phage genomes in mixture, in particular as described in the present disclosure; (ii) infecting a culture of bacterial cells;

(Hi) collecting the phages that infect and/or lyse the bacterial cells.

[0092] In one embodiment, the method further comprises a step of analyzing the sequence of the collected phage genomes.

[0093] The invention also relates to a library of phage variants prepared according to the method described herein.

[0094] The invention also relates to the phages identified by the screening method.

[0095] The invention also relates to the use of cell-free transcription and translation to prepare a library of phage variants. The embodiments relating to the above methods are applicable to these aspects of the invention.

[0096] The invention further relates to a kit for the preparation of a library of phage variants, wherein said kit comprises:

- a library of phage genomes; and

- at least one reagent for cell-free transcription and translation.

[0097] The invention further relates to a kit for the preparation of a library of phage variants, wherein said kit comprises:

- overlapping DNA fragments designed to assemble into a plurality of phage genomes;

- at least one exonuclease, preferably exonuclease III; and

- at least one reagent for cell-free transcription and translation.

[0098] Said reagent for cell-free transcription and translation may be selected from an energy mixture, an amino acid mixture, a cell or a cell lysate, in particular a bacterial cell or cell lysate, in particular an E. co// cell or E. co// lysate.

[0099] Any aspect of the method as described in the present disclosure is applicable to the above aspects.

[0100] The invention further relates to a phage, preferably a T7 phage, comprising a mutation in gene 17, gene 1 1 and/or gene 12 of its genome. The mutation may be the substitution, deletion of one or more nucleotides, preferably resulting in the mutation of an amino acid in the sequence of the protein encoded by the gene. The mutation may be as described in the Examples. In some embodiments, the mutated amino acid is selected from G521 , S541 , A500, N501 , G521 , S541 , G480, A539, V544 of the protein encoded by gene 17. In some embodiments, the mutated amino acid is selected from E166 and L3 of the protein encoded by gene 1 1 . In some embodiments, the mutated amino acid is selected from E784, T773, Q6, A193, Q446, G780, S559, E73 and L682 of the protein encoded by gene 12. In some embodiments, the mutation is selected from the mutations E166A and L3F in the protein encoded by gene 1 1 . In some embodiments, the mutation is selected from the mutations E784G, T773A, Q6R, A193V, Q446R, G780R, S559L, E73G and L682P in the protein encoded by gene 12. Corresponding mutations by conservative replacement of the mutant amino acids are also encompassed herein.

[0101] Various embodiments as described in the present detailed description can be combined according to the present invention unless clearly specified.

[0102] The invention will be further illustrated by the following figures and examples. However, these examples and figures should not be interpreted in any way as limiting the scope of the present invention.

[0103] Sequence listing

[0104] SEQ ID NO: 1 shows the sequence of the T7A-s primer.

[0105] SEQ ID NO: 2 shows the sequence of the T7A-as primer.

[0106] SEQ ID NO: 3 shows the sequence of the T7B-s primer.

[0107] SEQ ID NO: 4 shows the sequence of the T7B-as primer.

[0108] SEQ ID NO: 5 shows the sequence of the T7C-S primer.

[0109] SEQ ID NO: 6 shows the sequence of the T7C-as primer.

[0110] SEQ ID NO: 7 shows the sequence of the T7D-s primer.

[0111] SEQ ID NO: 8 shows the sequence of the T7D2-as primer.

[0112] SEQ ID NO: 9 shows the sequence of the T7E-s primer.

[0113] SEQ ID NO: 10 shows the sequence of the T7E-as primer.

[0114] SEQ ID NO: 1 1 shows the sequence of the T7F-s primer

[0115] SEQ ID NO: 12 shows the amino acid sequence of exonuclease III.

LEGENDS OF THE FIGURES

[0116] Other features, details and advantages will be shown in the following detailed description and on the figures, on which:

[0117] Fig. 1 shows an exemplary workflow of phage assembly and synthesis. The phage genome is amplified by PCR into fragments of 12 kbp or less with overlapping DNA sequences. Gene addition, mutation, and deletion are introduced at any permissive locus. The PCR products are cleaned up and annealed in vitro. The DNA assembly reaction is directly added to a cell-free extension (CFE) reaction to produce phages. Engineered phages are propagated and titrated on the desired host and directly used for NGS sequencing and downstream applications.

[0118] Fig. 2 shows evidence of a genotype/phenotype (g/P) linkage in batch TXTL reactions, a. One pot assembly of T7 phages from five parts leads to two possible phage genomes carrying a tail fiber mutation or not. Batch TXTL expression of this equimolar co-assembly leads to six types of phages, phages encapsulating a mutant tail fiber genome or not and displaying combinations of mutant and wild-type tail fibers, b. Titer of the co-expression on rfaC versus E. coli B determines the mut/MUT + mut/MIX proportion of phages in the co-expression. c. The presence of purified ReLPS leads to the selective inhibition of phages with mutant tail fibers. E. coli B titer of the phage/LPS mixture determines the wt/WT + mut/WT initial proportion of phage in the co-expression. Subsequent amplification on E. coli B transforms mut/WT phages into mut/MUT phages. Titers on rfaC and E. coli B determine the mut/WT over wt/WT proportion of phages in the ejection mixtures, d. Table compiling the proportion of phage types measured in the co-expression and the predicted proportion for a random tail fiber mix with the hypothesis that the six trimer tail fibers of phage result from the random combinations of three monomers, e. Co-expression of mutant and WT tail-fiber phages with various decreasing proportions of mutant tail fiber. Other hypotheses regarding the random assembly of tail fibers are also considered.

[0119] Fig. 3 shows co-expression experiment evidencing a link between genotype and phenotype, a. Spotting assay of the four phage solutions on strains E. coli B and rfaC expressed at final 0.1 nM of DNA. C-E: co-expression. E-V: equal volume. Titers on E. coli B: WT/WT 4.2 ± 0.8 x 10⁹ PFU/ml, M/M 2.9 ± 0.1 x 10⁹ PFU/ml, C-E 4.9 ± 0.4 x 10⁹ PFU/ml, E-V 3.9 ± 0.4 x 10⁹ PFU/ml. Titers on rfaC: XNT/XN no plaque detected <10⁵ PFU/ml, M/M 6.3 ± 0.9 x 10⁸ PFU/ml, C-E 4.4 ± 0.4 x 10⁸ PFU/ml, E-V 4.0 ± 0.7 x 10⁸ PFU/ml The first spot corresponds to a dilution of 100. b. Ejection assay with 0.4 mg/ml of ReLPS (KLA) 37 °C, overnight, applied to the four phage solutions. Spotting assay of the four phage solutions on strains E. coli B and rfaC after ejection. Titers on E. coli B: WT/WT 2.3 ± 0.4 x 10⁷ PFU/ml, M/M no plaque detected for all solution spotted, 0 PFU/ml, C-E 1 .7 ± 0.5 x 10⁷ PFU/ml, E-V 1 .6 ± 0.2 x 10⁸ PFU/ml. The first spot corresponds to a dilution of 1000 of the original TXTL reaction. Inset: 25 pl of each of the four phage solutions were spotted without any dilution. No plaques were detected after ejection on the rfaC strain, c. Amplification in E. coli B and spotting on E. coli B and rfaC. Titers on E. coli B: WT/WT 1 .2 ± 1 x 10¹¹ PFU/ml, C-E 1 .7 ± 0.4 x 10¹° PFU/ml, E-V 2.1 ± 0.5 x 10¹¹ PFU/ml. Titers on rfaC: WT/WT 6.7 ± 3 x 10⁴ PFU/ml, C-E 2.5 ± 0.1 x 10⁶ PFU/ml, E-V 1 .0 ± 0.3 x 10⁵ PFU/ml. The first spot corresponds to no dilution. No lysis occurred for the ejection for M/M only. Each condition is three reactions spotted once.

[0120] Fig. 4 shows an evaluation of randomness of tail fiber assembly in TXTL. a. Phage fraction in several hypotheses and in the experimental results, b. Ratio ReLPS infectious phage vs mutant tail fiber DNA dilution factor.

[0121] Fig. 5 shows the use of the phage synthesis workflow to engineer T7 phages that infect E. co// strains with any type of rough LPS. a. The T7 genome is assembled using five T7 WT gene parts (T7A, T7B, T7C, T7D1, T7D2) and fragment 1 146-1662 of the tail fiber gene 17 obtained by mutagenic PCR. b. cell-free assembly workflow for T7 LPS mutant libraries. T7 genomes with E0, E1 , E2, and E3 were separately assembled and expressed in TXTL. TXTL reaction containing the phage libraries was directly spotted on the E. co// strain harboring ReLPS. Twelve T7 variants were selected, and their tail fiber was sequenced for each of the three E. coli mutant strains IpcA, rfaC, and rfaE (all ReLPS), 36 phages total. E. coli B strain has a type RbLPS. c. The mutation frequency of the four mutated tail fiber DNA fragments was determined by NGS. The graphs are indexed from 1 to 516 which corresponds to the bp 1 146-1662 of gene 17. d. The tail fibers of the selected phages were sequenced to establish the mutation landscape, especially the T7-ReLPS. The blue zones show the external loops of the tail fiber tip. Grey bars are silent mutations, red bars are non-silent mutations, e. This heatmap compiles all the mutations of the 3 x 12 phages, the native amino acid is indicated below every mutation and the color scale indicates how many times a mutation is found in the 36 variants, f. This table reports the most frequent amino acid mutations among the 36 T7 variants that lead to ReLPS infection as a single mutation.

[0122] Fig. 6 shows the generation of T7 phage variants with mutated tail fibers at four different rates of mutations. The genome was assembled from six DNA fragments. Four different mutated tail fiber tips DNA fragments (gene 17 1 146-1662) E0, E1 , E2, E3 were obtained by PCR. The probability of nucleotide substitution for each fragment library are indicated in the tables. For instance, for E1 , when a G is mutated, it is mutated to an A in 60.7% of the cases, to a C in 1 1 .7% of the cases, and to a T in 27.6% of the cases (and the sum is 100%). Four T7 variants batches were spotted on an E. coli B lawn (T7-E0, T7-E1 , T7-E2, T7-E3). The experiment was repeated three times, and each time spotted three times. The TXTL reaction was diluted in LB by factor of tens and spotted from left to right. The first spot on the left corresponds to a dilution of 100 of the TXTL reaction. Negative controls: (i) a TXTL reaction with added water (no DNA) was spotted three times on an E. coli B lawn, (ii) a TXTL reaction from the cell-free workflow without the insert (mutated tail fiber) but with all the other DNA parts was spotted three times on an E. coli B lawn without dilution. No plaques were formed. A slight inhibition of E. coli B growth is observed.

[0123] Fig. 7 shows a spotting of T7-E1 , T7-E2 and T7-E3 on rfaC, IpcA, rfaE, rfaD, rfaG, ClearColi and Seattle 1946. No variant infected ClearColi and Seattle 1946. After the cell-free workflow, 10 pL of TXTL reaction were diluted hundred times. 25 pL of each dilution was spotted eight times for T7- E1 and four times for T7-E2 and T7-E3. A few to a dozen of plaques were detected on each ReLPS strain. Variants were picked for sequencing.

[0124] Fig. 8 shows a spotting assay to test T7 phage variants generated by the cell-free workflow on all the ReLPS E. co// strains (E. coli B used as a control, rfaC, rfaE, IpcA and rfaD mutant strains are all ReLPS). Each column 1 -5 has twelve variants, column 6 is T7 WT. T7 WT phages were isolated from single plaques from T7-E0 assembly spotted on E. coli B. All the isolated variants amplified on their host could infect all the ReLPS strains except for some rfaG variants. No WT T7 infected ReLPS strains except rfaD. 3.5 pL of each undiluted clarified phage lysate were spotted.

[0125] Fig. 9 shows a spotting assay to test T7 phage variants generated by the cell-free workflow on strains with different LPS: rfaF (Rd2LPS), rfaG (Rd1 LPS), rfal (ReLPS), rfaJ (RbLPS), rfaB (side chain modification LPS). Each column 1 -5 has twelve variants (the same as in Fig. 8), column 6 is T7 WT. T7 WT phages were isolated from single plaques from T7-E0 assembly spotted on E. coli B. Contraction of host range is visible on rfaE. some rfaC, rfaE and IpcA T7 variants were not able to infect the rfaF strain while T7 WT can. 3.5 pL of each undiluted clarified lysate were spotted. [0126] Fig. 10 shows a mutation landscape for twelve T7 phage variants generated by the cell-free workflow that infect E. coli mutant strains rfaD (ReLPS) and for ten T7 variants that infect rfaG (ReLPS). These variants poorly infected by T7 WT. These mutation patterns are retrieved by ORACLE (Huss et al). Top row: diagram showing the mutation landscape after treating the sequencing data. Silent mutations are grey and non-silent mutations are red. The blue zones show the external loops of the tail fiber tip. Bottom row: tables that show the annotated mutations. The bottom row in each table shows the positions of the native amino acids in the 172 amino acid segment (0 to 172, corresponding to amino acids 382 to 554 in gp17) that was mutated. The brackets show the amino acid substitution.

[0127] Fig. 11 shows the use of the workflow to synthesize T7 phages with mutations only found in the tail genes 11 and 12, by assembling genomes from four parts: T7A, T7B, T7D from T7 WT and T7C from variants as indicated in the figure. Each phage was assembled twice. Serial dilution was spotted once on E. coli B and rfaC mutant strain. No plaques on rfaC suggesting that the tail mutations are not responsible for the gain of function. The first spot on the left corresponds to a dilution of 10 of the TXTL reactions. A single plaque was picked and amplified in E. coli B for tail only IpcA variant 4 and rfaC variant 1 . These phages were fully sequenced.

[0128] Fig. 12 shows the use of the workflow to synthesize T7 phages with mutations only found in the tail fiber gene 17, by assembling genomes from four parts: T7A, T7B, T7C from T7 WT and T7D from variants as indicated in the figure. Each phage was assembled twice. Serial dilution was spotted once on E. coli B and rfaC mutant strain. Low EOP are observed for the tail-fiber only assemblies on rfaC compared to E. coli B suggesting a poor fitness of the phage in absence of tail mutation in gene 11 and gene 12. The first spot on the left corresponds to a dilution of 10 of the TXTL reactions. A single plaque was picked and amplifies in E. coli B for tail fiber only IpcA variant 4 and rfaC variant 1 . These phages were fully sequenced.

Examples

Example 1 : materials and methods

[0129] Reagents. The genomic DNA of phage T7 was purchased from Boca Scientific (# 310025). The DNA ladder for DNA electrophoresis was purchased from Invitrogen (10-787-018). The exonuclease III was purchased from NEB (#M0206S). DNA oligonucleotides were ordered from Integrated DNA Technologies (IDT), standard desalting. The bacterial strains were obtained from various sources as described in Table 3. Plasmid DNAs were obtained as follows: pTXTL-P70a- deGFP (Arbor Bioscience, #502056), pTXTL-T7p14-deGFP (Arbor Bioscience, #5021 1 1 ), pTXTL- P70a-T7rnap (Arbor Bioscience, #502082), pTXTL-T7p14-mmCherry (Arbor Bioscience, #502141 ). The LPS EH100 Ra (L9641 ) mutant and smooth LPS from E. coli 01 1 1 :B4 (L5293) were purchased from Sigma-Aldrich. The ReLPS was purchased from Avanti Polar Lipids (Kdo2-Lipid A (KLA), 699500). The phage lysates were sterilized with 0.22 pm centrifuge filter tubes (Costar #8160). The phage serial dilutions were performed with filter tips (Dutscher, #014210, #014220). The phage kinetics were performed in flat bottom transparent 96 bacterial culture well-plates with lids (Thermo Scientific Optical-Bottom Plates, #265301 ) on a Synergy H1 multi-mode microplate reader (Agilent). Phage spotting was performed on square Petri dishes (Greiner Bio-One #688102). Microscopy was done with an Olympus 1X81 inverted epi-fluorescence microscope mounted with a thermoplate (Tokai Hit). An imaging spacer (Grace Bio-Labs #654006) and microscopic slides (Fisher, #12-550-A3), and cover slides (Fisher #12542C) were used for bacterial lysis microscopic experiments. The purchased T7 genome (39.9 kbp, GenBank V01 146.1 ) was verified by NGS observing two mutations compared to GenBank V01 146.1 : (i) insertion of an A base after position 1896 in the gene 0.7; (ii) A to G mutation at position 22629 (N227S) in gene 9.

[0130] Cell-free transcription-translation. Cell-free gene expression was carried out using an E. coli TXTL system described previously(Sun et al, 2013, Garenne et al, 2021 ) with one modification. The strain BL21 -ArecBCD Rosetta2 was used, in which the recBCD gene set is knocked out to prevent the degradation of linear DNA(Batista et al). The preparation and usage of the TXTL system were the same as reported before(Sun et al, 2013, Garenne et al, 2021 ). Briefly, E. co// cells were grown in a 2xYT medium supplemented with phosphates. Cells were pelleted, washed, and lysed with a cell press. After centrifugation, the supernatant was recovered and preincubated at 37 °C for 80 min. After a second centrifugation step, the supernatant was dialyzed for 3 h at 4 °C. After a final spin-down, the supernatant was aliquoted and stored at -80 °C. The TXTL reactions comprised the cell lysate, the energy and amino acid mixtures, maltodextrin (30 mM) and ribose (30 mM), magnesium (2-5 mM) and potassium (50-100 mM), PEG8000 (3-4%), water and the DNA to be expressed. The reactions were incubated at 29 °C, in either 1 .5 ml tubes or on 96 well plates. For phage titration, the TXTL reactions were diluted with Luria broth (LB).

[0131] DNA amplification. The Q5 high-fidelity PCR polymerase (NEB #M0491 ) was used to amplify the fragment tail fiber fragment E0 (gene 771 146-1662, gp17 382-554). PCR mutagenesis was carried out with the Agilent Genemorph II Random Mutagenesis Kit (#200550) according to manufacturer instructions. Low (E1 ), medium (E2), and high (E3) mutation rates were obtained by adding respectively 500 ng, 50 ng, and 0.5 ng of E0 in the initial PCR mix (50 pl) and performing 30 PCR cycles. The PCR fragments for the assembly were otherwise amplified with KOD OneTM PCR Master Mix (No. KMM-101 ) according to manufacturer instructions. Either 1 pL of 1 ng/pL of T7 genome DNA or 1 pL of 10⁵ PFU/ml of clarified phage lysate obtained from a single plaque was used as template DNA. All PCR reactions were purified with a PCR clean-up kit (Invitrogen™ K310001 ) and DNA concentration was normalized to 20 nM in deionized water.

[0132] DNA assembly and TXTL reactions. Purified PCR fragments were mixed at an equimolar concentration of 0.5-5 nM. An equal volume of a 2x assembly mix (20 mM Tris-HCI pH 7.9, 100 mM NaCI, 20 mM MgCI2, 10% (w/v) PEG8000, 2 mM Dithiothreitol, 1 U/pL Exonuclease III) and 2.5 pL DNA mix were mixed on ice. Exonuclease inactivation was done by incubation of the DNA assembly reaction at 75 °C for 1 min. Single-stranded ends annealing was done by incubation at room temperature for 5 min. The mix was directly added to a TXTL reaction. Typically, 2.5 pL of the assembly mix was added to 7.5 pL of TXTL reaction. Negative controls were done by mixing all the PCR fragments except one replaced by an equal volume of water. The fragments were individually amplified by PCR and purified using standard procedures. For the PCR amplification, a few picograms of genomes or a few microliters of a clarified phage lysate are added directly as a template to the PCR mixture. The size of each fragment was verified by standard DNA gel electrophoresis and their concentration was measured by spectrophotometry. The fragments were mixed at an equimolar concentration in the nanomolar range, treated with the E. coll exonuclease III and annealed. The exonuclease, rapidly deactivated during the procedure, produces single-stranded gaps at the 3’ termini of the overlapping fragments enabling the annealing of complementary fragments. 2.5 pl of T7 DNA assembly reaction were added to a 7.5 pl TXTL reaction to express the annealed genomes. T7 phage genomes assembled from four parts were confirmed by NGS.

[0133] NGS. Sequencing of the mutagenic PCR fragments was done by NGS (Illumina). DNA samples were converted to Illumina sequencing libraries using Illumina’s Truseq NanoDNA Sample Preparation Kit (Cat. # 20015964). During library creation, amplicon DNA was end-repaired with the adaptors, and indexes were ligated to each sample. The libraries did not undergo any PCR cycling. The final library size distribution was validated using capillary electrophoresis and quantified using fluorimetry (PicoGreen) and Kapa q-PCR. Pooled libraries were denatured and diluted to the appropriate clustering concentration. The libraries were then loaded onto the MiSeq paired-end flow cell and clustering occurred onboard the instrument. Once clustering was complete, the sequencing reaction immediately began using Illumina’s 4-color SBS chemistry. Upon completion of read 1 , 2 separate 8 or 10 base pair index reads were performed. Finally, the clustered library fragments were re-synthesized in the reverse direction thus producing the template for paired end read 2. Base call (.bcl) files for each cycle of sequencing were generated by Illumina Real Time Analysis (RTA) software. Primary analysis and de-multiplexing were performed using Illumina’s bcl2fastq software 2.20. The result of the bcl2fastq workflow was de-multiplexed in FASTQ files. Reference mapping of the obtained reads and variant calling were done with open-source Galaxy software (BWA-MEM and iVar, and quality Q > 30 were used). Nucleotide mutations were analyzed with a Python program.

[0134] Phage spotting assay. 1 .5% agar-LB plates were pre-incubated at 37 °C for 1 h. 10 mL of 0.7% soft agar was kept at 55 °C in a water bath. 100 pL of overnight bacterial culture were mixed with the soft agar and vortexed gently. The soft agar was slowly dispensed onto the agar LB plates plate to cover uniformly the entire surface of the agar plate. The soft-agar plates were left at room temperature for 15 min on a flat surface to solidify. Serial ten-fold dilutions in LB of either cell-free phage reaction or clarified phage lysates were prepared in 200 pL. Spotting: for each phage dilution, 3.5 pL were dropped onto the soft agar. For negative control TXTL reactions, the whole reaction was diluted in LB at a final volume of 25 pL and spotted in one droplet onto the soft agar layer. After spotting, the plate was left for 15 min on the bench to let droplets absorb onto the soft agar. The plates were incubated at 37 °C, facing down, for 4 h. Plaques were counted at the dilution where 1 <#plaques<20 per spot. Titers were typically calculated from three serial dilution spotting, plate uncertainties are estimated at the time of counting (duplication or plate belonging to the spot). For calculations on PFUs, Uncertainties were propagated using Python module Uncertainties: a Python package for calculations with uncertainties, Eric O. Lebigot. [0135] Phage infection kinetic assay. Infection kinetics were carried out in 96 well plates in a Synergy Hi m microplate reader (Agilent). In each well, 180 pL of the host culture in LB (initial OD600 0.01 -0.04) were mixed with 20 pL of different serial phage dilutions. Each condition was replicated in four different wells. Positive controls consisted of 180 pL hosts in LB + 20 pL LB. Optical density at 600 nm and fluorescence intensity (excitation 580 nm, emission 610 nm for mCherry) was blanked against wells containing 200 pL of LB at each timestep. A lid was added to the 96 well plates to reduce evaporation during acquisition. The microplate reader was set to 37 °C with continuous double orbital shaking at 200 rpm. Optical density and fluorescence intensity were measured in each well every 3 min during 5-10 h. The mean and standard deviation of each condition were calculated at each timestep.

[0136] Microscopy. 100 pL of 2% agar were poured onto a slide with a spacer, covered with a cover slip, and let to solidify at 4 °C 1 h. 2 pL of host cells (OD600 ~ 0.2) and 2 pL phage lysate ~10⁵ PFU/ml were added on the agar pad. The agar slide was incubated at 37 °C for 15 min to allow cells to decant on the agar layer. A slide was added on top of the agar layer and cell growth and lysis were recorded on an epifluorescence microscope (40x objective). Time Lapse movies were assembled by recording one image every 2 min at 37 °C. Image composites and sequences were obtained with Imaged.

[0137] LPS-phage in vitro assay. LPS stock solutions were prepared in deionized water at 1 mg/ml and sonicated at 60 °C 30 min. T7 clarified lysates (10⁷-10¹⁰ PFU/ml) were mixed with LPS (final 200 pg/ml) in a final 50 pL volume and incubated at 37 °C 3 h. The phage LPS mixtures were serially diluted in LB and spotted on E. coli B lawn. For the p/G experiment, ReLPS was used at 400 pg/mL and incubated at 37 °C overnight to inhibit the potential remaining ReLPS+ phage phenotypes.

Example 2: phage synthesis workflow.

[0138] The T7 genome was re-assembled from long PCR fragments (<12 kbp) with overlapping sequences using a cheap assembly mix containing only an exonuclease, followed by heat inactivation of the enzyme. Annealed fragments are directly expressed in TXTL without additional steps enabling the synthesis and selection of T7 phage variants that integrate gene addition, deletion, and mutation. The workflow, achieved in under one day, delivers phages at titers comparable to titers obtained from bacterial lysate (10¹⁰'¹¹ PFU/ml) (Fig. 1 ).

Example 3: Cell-free transcription-translation.

[0139] The myTXTL system (Sun et al, 2013, Garenne et al, 2021 ) uses the endogenous E. coli core RNA polymerase and sigma factor 70 present in the lysate as the sole primary transcription proteins. This system does not contain any remaining live E. coli cells. Genes are expressed either from plasmids or linear dsDNA. In this work, all the TXTL reactions were carried out in batch mode at the scale of 1 -10 pl, either in 1 .5 mL tubes or in well plates. In batch mode, 80-100 pM of deGFP protein are produced after 12 h of incubation from an E. coli promoter (P70a (Garenne et al, 2019)). With the bacteriophage T7 promoter, 100-120 pM deGFP proteins are produced in 3-6 h. The wild-type phage T7 (T7 WT) is synthesized from its genome at a concentration of 10¹⁰-10¹¹ PFU/ml (plaqueforming units per milliliter) after 3 h of incubation.

Example 4: Evidence of a phenotype/genotype linkage in batch TXTL.

[0140] The workflow used in the present invention achieves phage synthesis without compartmentalization. The inventors reckoned that the fast kinetics of the T7 phage coat and tail proteins’ cooperative assembly to encompass the phage genome, following their coupled transcription and translation, may limit their diffusion and cross-binding to the non-self-phage genome in the viscous TXTL mix, leading to genotype to phenotype linkage without the need for encapsulation or in vivo infection and propagation.

[0141] To determine whether phage genotype/phenotype (g/P) coupling prevails in the cell-free synthesis in the absence of physical compartmentalization, a simple two-phage experiment was devised, consisting of co-expression of two phages whose genomes are annealed in the same DNA assembly reaction. Both phages, T7-vWT (T7 variant WT phage infecting only E. coli B, WT host) and T7-rfaD-1 phage, carry gene t2 G784E. T7-rfaD-1 carries an additional mutation (gene 17 S541 R) rendering it infectious on both WT and ReLPS rfaC hosts. Co-assembly was done with an equimolar mix of T7D and T7D-S541R fragments (Fig. 2a). Expression of this one-pot dual genome assembly mix in TXTL reactions is expected to result in six different g/P types of phages in the absence of linkage, where “MIX” represents a mixture of incorporated WT and S541 R mutant tail fibers (Fig 2a).

[0142] The theoretical fully randomized relative abundance of the six different g/P hybrids was calculated under the most permissive assumption, namely that the integration of at least one S541 R mutant monomer into one of the six trimeric tail fibers of a given phage is necessary and sufficient to infect rfaC ReLPS (Fig. 2d). The following set of experiments allowed us to quantify the abundance of the six phages g/P hybrids in the co-synthesis of equimolar assemblies of T7-vWT and T7-rfaD-1 (Fig. 2d, Fig. 3).

[0143] The fraction of mut/MUT and mut/MIX phages (Fig. 2b), 50±18%, was derived by titrating the equimolar co-expression on WT and rfaC strains (4 ± 1 .0 x10⁸ and 8 ± 1 .0 x 10⁸ PFU/mL, respectively), as wt/WT or mut/WT cannot infect the rfaC strain while wt/MIX or wt/MUT cannot propagate on strain rfaC after infection. As controls, co-synthesis of separately assembled T7-vWT and T7-rfaD-1 phages resulted in the expected 50% of rfaC vs WT infection (2.8 ± 0.9 x 10⁸ and 6 ± 1 .4 x 10⁸ PFU/mL, respectively). While the EOP ratio of separately synthesized T7-rfaD-1 between rfaC and WT hosts (5.8 ± 0.7 x 10⁷ and 3.2 ± 0.6 x 10⁸ PFU/mL, respectively) was 0.18 ± 0.04, no plaques from separately synthesized T7-vWT were detected on the rfaC strain indicating that <10⁴ PFU/mL of ReLPS emerging variant phages were present in the T7-vWT assembly. As expected, we found comparable phage titers on the WT host for T7-rfaD-1 , T7-vWT or their equimolar co- synthesis or their equal volume of their separate synthesis (~6 x 10⁸, 3 x 10⁸, 8 x 10⁸, 6 x 10⁸ PFU/mL, respectively).

[0144] The fraction of wt/WT and mut/WT (Fig. 2c), 6±2%, was determined by amplifying and titrating the I/ T50 ± 18%, infective phages, following incubation of the equimolar co-expression phage mix with purified ReLPS (0.4 mg/mL, 37 °C overnight), in comparison to an equal volume mix of the separately assembled genes (1 .9 ± 0.4 x 10⁶, 1 .7 ± 0.4 x 10⁷ PFU/mL, respectively). Under such conditions, all other variants are phenotypically ReLPS+ and eject their genome. Indeed, no plaques on rfaC were detected in the ReLPS overnight incubation of the T7-rfaD1 mutant tail fiber expression, the co-expression, and the equal volume mix. Separately assembled T7-vWT phages were poorly susceptible to ReLPS when incubated with 0.4 mg/mL ReLPS for an extended period at 37 °C (10- fold decrease of titer from ~6 x 10⁸ PFU/mL to ~5 x 10⁷ PFU/mL).

[0145] Finally, the fractions of wt/WT (Fig. 2c), 6 ± 2%, and mut/WT, 0.03 ± 0.02%, were determined by amplifying on l/VT host and titrating the wt/WT and mut/WT fraction (above) on WT and rfaC hosts (6 ± 2 x 10¹° and 2 ± 1 x 10⁷ PFU/ml, respectively). A few undefined plaques were also detected on rfaC from the equal volume mix of pre-assembled T7-vWT and -rfaD-1 phages after the ReLPS ejection and amplification on the l/VT host. We attribute these plaques to new mutant phages emerging from T7-vWT.

[0146] The above results together with the following two hypotheses based on the symmetry due to equimolarity and considering that point mutation does not entail tail fiber assembly bias:

(i) (wt/WT)/(wt/WT+wt/MIX+wt/MUT) = (mut/MUT)/(mut/MUT+mut/MIX+mut/WT)

(ii) (wt/MUT)/(wt/WT+wt/MIX+wt/MUT) = (mut/WT)/(mut/MUT+mut/MIX+mut/WT) allowed us to determine the relative fractions of the six phage variants expected from the initial coassembly mix (Fig. 2d). A clear disparity is evident between the most permissive randomized hypothesis and the experimental results, indicating significant g/P coupling. The fraction of pure g/P (wt/WT and mut/MUT) coupling amounts to 12% of what is expected from full linkage and four orders of magnitude greater than the randomized hypothesis. g/P coupling is also manifested by the stark (x200) asymmetry between pure and opposing p/Gs.

[0147] It was next tested whether T7-rfaD-1 phages from the cell-free assembly could be selectively enriched as above but with different ratios (1 :2 - 1 :10,000) of T7D-S541R to T7D-WT at constant DNA concentration (50 pM). The ratio of plaque counts of the resulting phages between rfaC (T7- rfaD-1 phages) and WT (both phages) hosts closely follows the dilution regime, as expected from g/P linkage, and not the hypothetical random assembly regime (Fig. 2f). Taken together, these experiments support the existence of local interaction between the tail fibers, the procapsid, and their encoding genome, indicative of localized encapsulation underlying the observed g/P coupling.

[0148] The randomness of tail fiber assembly in TXTL was assessed according to the following calculations. T7 contains six tail fibers. Each tail fiber is a trimer of gp17. In the case of the coexpression of the two-phage system in batch TXTL, three hypotheses can be made, ordered by permissiveness: HO, H1 , H2 on how T7 tail fibers are assembled in the absence of g/P coupling and pure mixing in TXTL. Here, pure (100%) g/P linked phage is defined as a phage that displays six tail fibers composed of only Mutant tail fibers if it encapsulates a mutant genotype or only WT tail fibers if it encapsulates a wt genotype. We further make the assumptions that Mutant and WT tail fibers are equally expressed and assemble at the same rates. Conversely, an inversely linked phage is defined as a phage that displays six tail fibers composed of only Mutant tail fibers if encapsulates a wt genotype or only WT tail fibers if it encapsulates a mutant genotype. We define D as the ratio of mutant genomes.

In the equimolar co-assembly D is

In the dilution experiment D is - ; — ; — ;

[0149] HO: Hypothesis presented in the manuscript. 18 monomers assemble randomly in 6 tail fibers. If the phage has a mutant genotype, only 1 mutant monomer confers a ReLPS+ phenotype. In this condition:

(i) the fraction of pure g/P link and pure inverse g/P link in an equimolar mix is: D¹⁸ =

(ii) the fraction of phages that are ReLPS infectious phages is: (1 — (1 — £>)¹⁸) x D

[0150] H1 : Only identical monomers can assemble in tail fiber trimer. 6 tail fiber trimers assemble randomly onto a prophage. At least 1 mutant tail fiber confers ReLPS+ phenotype. In this condition:

(i) the fraction of pure g/P link and pure inverse g/P link in an equimolar mix is: D⁶ =

(ii) the fraction of phages that are ReLPS infectious phages is: (1 — (1 — D)⁶) x D

[0151] H2: 18 monomers assemble in 6 tail fiber trimers randomly. At least 1 tail fiber that is MMM confers ReLPS+ phenotype (MMM is a mutant tail fiber from 3 mutant monomers). In this condition:

(i) the fraction of pure g/P link and pure inverse g/P link in an equimolar mix is: (£)³)⁶ =

(ii) the fraction of phages that are ReLPS infectious phages is: (1 - (1 - D X £)³)⁶) x D

[0152] We compared the three hypotheses for the equimolar assembly experiment and the dilution experiment (Figure 4). Interestingly while H1 results in greatest fraction of pure link we still find experimentally four times more pure linkage and the asymmetry between pure link and inversely link fractions. H1 however predicts less well than HO for the dilution experiment when compared to the experimental results due to the condition that only identical monomers assemble in a tail fiber. Finally, H2 leads to the furthest predictions for both experiments. Indeed, Mutant monomers become quickly diluted in WT monomers reducing the chances that three identical monomers assemble in a tail fiber. Overall, these three simple hypotheses cover how tail fiber assemble in a perfectly mixed reaction in the absence of g/P coupling and do not predict our experimental results. While the equimolar coexpression experiment show that bulk assembly results in only a small fraction of pure linkage, the g/P link is however not lost with dilution. This suggests that a coupling exist. This coupling could stem from an interaction between phage proteins and its genome or the specific conditions of a TXTL reaction. [0153] Example 5: Selection for T7 phage host range expansion via g/P linkage. The tip of the tail fiber gene 17 (amino acids 472-554), determinant of the phage’s host range (Garcia-Doval et al, Molineux et al, Qimron et al), can be exchanged between phages or mutated to adapt to new hosts (Fraser et al, 2006, Fraser et al, 2007; Lin et al, Heineman et al). Recently (Huss et al), a library of 1660 exhaustive single mutations of the T7 tail fiber tip residues was probed using yeast cloning, site-specific recombination with a helper plasmid, Cas9-gRNA-based progressive variant phages, and final amplification of the phage variant library in the absence of helper plasmid. Tail fiber mutations rendering T7 variants infectious to truncated LPS E. coli strains rfaD (ReLPS) and rfaG (Rdl LPS) were identified (Tables 4 to 6). In another study, a continuous evolution setup generated T7 variants with tail and tail fiber gene mutations specific to different rough LPS types (Holtzman et al) (Tables 4 to 6). Here, similar mutation patterns was retrieved with the simplified, faster, and cheaper cell-free protocol (Fig. 5a, 4b). The T7 genome was assembled using five T7 WT gene parts (T7A, T7B, T7C, T7D1, T7D2) and fragment 1 146-1662 of the tail fiber gene 17 obtained by mutagenic PCR.

Table 2: primers used for the fragment amplification

[0154] A fragment consisting of the 172 C-terminal residues (516 bp) of the T7 WT tail fiber was amplified by high-fidelity PCR (fragment E0). We used fragment E0 to create 3 libraries of randomly mutated fragments with increasing PCR mutational load (E1 , E2, and E3; see Methods). NGS analysis (2 x 10⁵ reads/fragment) of the raw sequences revealed a uniform distribution of mutations averaging 1 , 2.5, and 5 mutations per fragment for E1 , E2, and E3, respectively (Fig 5c). The cell- free workflow was used to assemble these libraries with the rest of the genome, amplified by PCR in five overlapping parts, resulting in four T7 variant phage pools: unmutated T7-E0, and T7-E1 , T7-E2, and T7-E3 libraries, with diminishing respective titers on l/ Thost of 1 x 10¹¹ , 3 x 10¹⁰, 1 x 1 O¹⁰ and 7 x 10⁹ PFU/ml (as compared to 1 x 10¹¹ of rebooted T7 WT), suggesting an overall detrimental mutational load (Fig. 6). No phages were detected in a control in the absence of the tail fiber fragment in the assembly mix. Given the established g/P linkage, the three libraries were directly spotted on our panel of strains. As expected, none of the phage libraries infected the LPS-deficient smooth or ClearColi E. co// strains. Hundreds of plaques were obtained from T7-E0, -E1 , and -E2 libraries infecting previously described LPS phenotypes of rfaG (Rd1 LPS) and rfaD (Huss et al) (Table 3). 0155]

Table 3: Spotting assay of the four batches of T7 phages (T7-E0, T7-E1 , T7-E2, T7-E3) on different E. coli strains

[0156] Importantly, -100 clear and circular plaques per 1 ng of T7-E1 and T7-E2 TXTL-assembled libraries were successfully and systematically obtained by direct plating on ReLPS rfaC, IpcA, and rfaE E. coli B hosts that were not infected by T7 WT (or the T7-E0 control). This is in stark contrast to our estimation of the probability of selecting a single infective library genetic mutant that might have randomly packaged with an infecting phenotype in a well-mixed milieu and assuming hundred possible infective mutations of an estimated 2 x 10⁶ variants per 1 ng assembled library (<6 x 10³ PFU). This strongly supports a g/P linkage. Interestingly, titers were somewhat lower on the rfaC as compared to the rfaE or IpcA host strains although they are believed to share the same ReLPS serotype. This may suggest different membrane properties between the knockouts.

[0157] Twelve clonal phages from each of the three hosts (Fig. 5d-f) were purified, propagated in their respective host, phenotyped on all hosts (Fig. 7, 8, 9) , and sequenced (Fig. 10). Unlike T7 WT, all mutants exhibited a broad infection spectrum. Interestingly, some T7 variants isolated from ReLPS strains could not infect the rfaF host. This observation is reminiscent of a contraction of the host range discussed previously (Holtzman et al).

[0158] Sequencing of the 36 clones (500 bp upstream and downstream of the inserted tail fiber fragment) revealed more amino acid mutations in T7-E2 and T7-E3 as compared to T7-E1 . The mutations were located mainly in the exterior loops of the tip of the tail fiber (Fig. 5 d, e, f), matching the top ten residues G521 , S541 , A500, N501 , G480, D540) discovered previously by systematic mutagenesis for rfaD infection (Huss et al). Moreover, four common substitutions: G521 R, N501 K, S541 R , and G480R were identified. Comparable results are found for rfaG (tables 4 to 6). This suggests that the workflow can quickly generate and identify the best gain of function variants. The presence of silent mutations within the tail fiber gene assures that these clones resulted from the error-prone libraries and did not appear de novo when propagating on their host. Whole genome sequencing of six of the phage clones (two for each of rfaC, IpcA, and rfaE mutant E. coli strains) identified further mutations in tail genes 11 and 12(Qimron et al, Holtzman et al). The method allowed us to reconstruct separately gene 11 and gene Invariants and gene 1 /variants, to demonstrate that ReLPS strain infection was solely dependent on the selected gene 17 mutations (Fig. 11 , 12). Subsequent mutations in the tail genes 11 and 12 provide an increase in the fitness of the phages within the new host.

[0159]

Table 4: comparison of the mutations identified on gene 17 with the results of Huss et al. (Mapping the functional landscape of the receptor binding domain of T7 bacteriophage by deep mutational scanning eLife) 10:e63775.

Table 5: comparison of the mutations identified on gene 11 with previously identified mutations

Table 6: comparison of the mutations identified on gene 12 with previously identified mutations

Bibliography

1 . Salmond, G. P. C. & Fineran, P. C. A century of the phage: past, present and future. Nat Rev Microbiol 3, 777-786 (2015).

2. Stanley, S. Y. & Maxwell, K. L. Phage-Encoded Anti-CRISPR Defenses. Anna Rev Genet 52, 445-464 (2018).

3. Dedrick, R. M. et al. Engineered bacteriophages for treatment of a patient with a disseminated drug-resistant Mycobacterium abscessus. Nat Med 25, 730-733 (2019).

4. Veterinary use of bacteriophage therapy in intensively-reared livestock - PubMed. https://pubmed.ncbi.nlm.nih.gov/31831017/.

5. Harada, L. K. et al. Biotechnological applications of bacteriophages: State of the art. Microbiol Res 212-213, 38-58 (2018).

6. Lemire, S., Yehl, K. M. & Lu, T. K. Phage-Based Applications in Synthetic Biology. Anna Rev Virol 5, 453-476 (2018).

7. Endersen, L. et al. Phage Therapy in the Food Industry. Anna. Rev. Food Sci. Technol. 5, 327-349 (2014). 8. Matsuzaki, S. & Uchiyama, J. Phage Pharmacokinetics: Relationship with Administration Route, in Phage Therapy: A Practical Approach (eds. Gorski, A., Mi^dzybrodzki, R. & Borysowski, J.) 43-57 (Springer International Publishing, 2019). doi:10.1007/978-3-030-26736-0_2.

9. Daube, S. S., Arad, T. & Bar-Ziv, R. Cell-free co-synthesis of protein nanoassemblies: tubes, rings, and doughnuts. Nano Lett 7, 638-641 (2007).

10. Vonshak, O. et al. Programming multi-protein assembly by gene-brush patterns and two- dimensional compartment geometry. Nat Nanotechnol 15, 783-791 (2020).

11. Bundy, B. C., Franciszkowicz, M. J. & Swartz, J. R. Escherichia coli-based cell-free synthesis of virus-like particles. Biotechnol Bioeng 100, 28-37 (2008).

12. Lu, Y., Chan, W., Ko, B. Y., VanLang, C. C. & Swartz, J. R. Assessing sequence plasticity of a virus-like nanoparticle by evolution toward a versatile scaffold for vaccines and drug delivery. Proc Natl Acad Sci U S A 112, 12360-12365 (2015).

13. Principi, N., Silvestri, E. & Esposito, S. Advantages and Limitations of Bacteriophages for the Treatment of Bacterial Infections. Front Pharmacol 10, 513 (2019).

14. Pryor, J. M., Potapov, V., Pokhrel, N. & Lohman, G. J. S. Rapid 40 kb genome construction from 52 parts. http://biorxiv.org/lookup/doi/10.1101/2020.12.22.424019 (2020) doi :10.1101/2020.12.22.424019.

15. Kiro, R., Shitrit, D. & Qimron, U. Efficient engineering of a bacteriophage genome using the type l-E CRISPR-Cas system. RNA Biology 11 , 42-44 (2014).

16. Liu, Y., Huang, H., Wang, H. & Zhang, Y. A novel approach for T7 bacteriophage genome integration of exogenous DNA. J Biol Eng 14, 2 (2020).

17. Grigonyte, A. M. et al. Comparison of CRISPR and Marker-Based Methods for the Engineering of Phage T7. Viruses 12, 193 (2020).

18. Rustad, M., Eastlund, A., Marshall, R., Jardine, P. & Noireaux, V. Synthesis of Infectious Bacteriophages in an E. coli-based Cell-free Expression System. J Vis Exp 56144 (2017) doi:10.3791/56144.

19. Garenne, D., Thompson, S., Brisson, A., Khakimzhan, A. & Noireaux, V. The all-E. coliTXTL toolbox 3.0: new capabilities of a cell-free synthetic biology platform. Synthetic Biology 6, ysab017 (2021 ).

20. Garenne, D. & Noireaux, V. Cell-free transcription-translation: engineering biology from the nanometer to the millimeter scale. Curr Opin Biotechnol 58, 19-27 (2019).

21. Noireaux, V., Bar-Ziv, R. & Libchaber, A. Principles of cell-free genetic circuit assembly. Proc Natl Acad Sci U S A 100, 12672-12677 (2003).

22. Takahashi, M. K. et al. Characterizing and prototyping genetic networks with cell-free transcription-translation reactions. Methods 86, 60-72 (2015).

23. Sun, Z. Z., Yeung, E., Hayes, C. A., Noireaux, V. & Murray, R. M. Linear DNA for rapid prototyping of synthetic biological circuits in an Escherichia coli based TX-TL cell-free system. ACS Synth Biol 3, 387-397 (2014).

24. Tayar, A. M., Karzbrun, E., Noireaux, V. & Bar-Ziv, R. H. Synchrony and pattern formation of coupled genetic oscillators on a chip of artificial cells. Proc Natl Acad Sci USA 114, 11609-11614 (2017).

25. Pardee, K. etal. Portable, On-Demand Biomolecular Manufacturing. Ce//167, 248-259. e12 (2016).

26. Noireaux, V. & Libchaber, A. A vesicle bioreactor as a step toward an artificial cell assembly. Proc Natl Acad Sci U SA 101 , 17669-17674 (2004).

27. Godino, E. et al. Cell-free biogenesis of bacterial division proto-rings that can constrict liposomes. http://biorxiv.Org/lookup/doi/10.1101/2020.03.29.009639 (2020) doi: 10.1101/2020.03.29.009639.

28. Gaut, N. J. etal. Programmable Fusion and Differentiation of Synthetic Minimal Cells. ACS Synth Biol 11 , 855-866 (2022).

29. Shin, J. & Noireaux, V. An E. coli cell-free expression toolbox: application to synthetic gene circuits and artificial cells. ACS Synth BioI l , 29-41 (2012).

30. Garamella, J., Marshall, R., Rustad, M. & Noireaux, V. The All E. coli TX-TL Toolbox 2.0: A Platform for Cell-Free Synthetic Biology. ACS Synth Biol 5, 344-355 (2016).

31 . Shin, J., Jardine, P. & Noireaux, V. Genome Replication, Synthesis, and Assembly of the Bacteriophage T7 in a Single Cell-Free Reaction. ACS Synthetic Biology 1 , 408-413 (2012).

32. Rustad, M., Eastlund, A., Jardine, P. & Noireaux, V. Cell-free TXTL synthesis of infectious bacteriophage T4 in a single test tube reaction. Synthetic Biology 3, (2018).

33. Nozaki, S. Rapid and Accurate Assembly of Large DNA Assisted by In Vitro Packaging of Bacteriophage. ACS Synth Biol 11 , 4113-4122 (2022).

34. Sun, Z. Z. et al. Protocols for Implementing an Escherichia coli Based TX-TL Cell-Free Expression System for Synthetic Biology. JoVE 50762 (2013) doi:10.3791/50762.

35. Cheng, L. et al. Harnessing stepping-stone hosts to engineer, select, and reboot synthetic bacteriophages in one pot. Cell Rep Methods 2, 100217 (2022).

36. Nethery, M. A., Hidalgo-Cantabrana, C., Roberts, A. & Barrangou, R. CRISPR-based engineering of phages for in situ bacterial base editing. Proceedings of the National Academy of Sciences 119, e2206744119 (2022).

37. Guan, J. etal. Bacteriophage genome engineering with CRISPR-Cas13a. Nat Microbiol 7, 1956-1966 (2022).

38. Chan, L. Y., Kosuri, S. & Endy, D. Refactoring bacteriophage T7. Mol Syst Biol 1 , (2005).

39. Gibb, B., Hyman, P. & Schneider, C. L. The Many Applications of Engineered Bacteriophages-An Overview. Pharmaceuticals (Basel) 14, 634 (2021 ).

40. Mitsunaka, S. et al. Synthetic engineering and biological containment of bacteriophages. Proc. Natl. Acad. Sci. U.S.A. 119, e2206739119 (2022).

41. Zhang, Y., Werling, U. & Edelmann, W. SLiCE: a novel bacterial cell extract-based DNA cloning method. Nucleic Acids Res 40, e55 (2012).

42. Subramanian, S. K., Russ, W. P. & Ranganathan, R. A set of experimentally validated, mutually orthogonal primers for combinatorially specifying genetic components. Synth Biol (Oxf) 3, ysx008 (2018).

43. Hausmann, R. The T7 Group, in The Bacteriophages (ed. Calendar, R.) 259-289 (Springer US, 1988). doi:10.1007/978-1 -4684-5424-6_8.

44. Rosenberg, A. et al. T7Select® Phage Display System: A powerful new protein display system based on bacteriophage T7.

45. Raetz, C. R. Biochemistry of endotoxins. Annu Rev Biochem 59, 129-170 (1990).

46. Qimron, U., Marintcheva, B., Tabor, S. & Richardson, C. C. Genomewide screens for Escherichia coli genes affecting growth of T7 bacteriophage. Proc Natl Acad Sci U SA 103, 19039- 19044 (2006).

47. Frontiers | The Role of Outer Membrane Proteins and Lipopolysaccharides for the Sensitivity of Escherichia coli to Antimicrobial Peptides. https://www.frontiersin.Org/articles/10.3389/fmicb.2018.02153/full.

48. Mamat, U. et al. Endotoxin-free protein production — ClearColi™ technology. Nat Methods 10, 916-916 (2013).

49. Huss, P., Meger, A., Leander, M., Nishikawa, K. & Raman, S. Mapping the functional landscape of the receptor binding domain of T7 bacteriophage by deep mutational scanning. Elife 10, e63775 (2021 ).

50. Holtzman, T. et al. A continuous evolution system for contracting the host range of bacteriophage T7. Sci Rep 10, 307 (2020).

51. Gonzalez-Garcia, V. A. et al. Conformational Changes Leading to T7 DNA Delivery upon Interaction with the Bacterial Receptor. Journal of Biological Chemistry 290, 10038-10044 (2015).

52. Smith, G. P. Filamentous fusion phage: novel expression vectors that display cloned antigens on the virion surface. Science 228, 1315-1317 (1985).

53. Garcia-Doval, C. & van Raaij, M. J. Structure of the receptor-binding carboxy-terminal domain of bacteriophage T7 tail fibers. Proc Natl Acad Sci U SA 109, 9390-9395 (2012).

54. Molineux, I. J. No syringes please, ejection of phage T7 DNA from the virion is enzyme driven. Mol Microbiol 0, 1-8 (2001 ).

55. Fraser, J. S., Yu, Z., Maxwell, K. L. & Davidson, A. R. Ig-like domains on bacteriophages: a tale of promiscuity and deceit. J Mol Biol 359, 496-507 (2006).

56. Fraser, J. S., Maxwell, K. L. & Davidson, A. R. Immunoglobulin-like domains on bacteriophage: weapons of modest damage? Curr Opin Microbiol 10, 382-387 (2007).

57. Lin, T.-Y. etal. A T3 and T7 recombinant phage acquires efficient adsorption and a broader host range. PLoS One 7, e30954 (2012).

58. Heineman, R. H., Springman, R. & Bull, J. J. Optimal foraging by bacteriophages through host avoidance. Am Nat 171 , E149-157 (2008).

59. Liyanagedera, S. B. W. et al. SpyPhage: A Cell-Free TXTL Platform for Rapid Engineering of Targeted Phage Therapies. ACS Synth Biol 11 , 3330-3342 (2022).

60. Simoliunas, E. et al. Incomplete LPS Core-Specific Felix01 -Like Virus vB_EcoM_VpaE1 . Viruses 7, 6163-6181 (2015).

61 . Whichard, J. M. et al. Complete genomic sequence of bacteriophage felix o1 . Viruses 2, 710-730 (2010).

62. Batista, A. C. et al. Differentially Optimized Cell-Free Buffer Enables Robust Expression from Unprotected Linear DNA in Exonuclease-Deficient Extracts. ACS Synth Biol 11 , 732-746 (2022).

63. Huss, P., Meger, A., Leander, M., Nishikawa, K. & Raman, S. Mapping the functional landscape of the receptor binding domain of T7 bacteriophage by deep mutational scanning. Elife 10, e63775 (2021 ).

64. Shimizu, Y., Inoue, A., Tomari, Y. et al. Cell-free translation reconstituted with purified components. Nat Biotechnol 19, 751-755 (2001 ).

65. Yue, H., Li, Y., Yang, M„ & Mao, C. (2022). T7 phage as an emerging nanobiomaterial with genetically tunable target specificity. Advanced Science, 9(4), 2103645.

66. Gomes, M., Fleck, A., Degaugue, A., Gourmelon, F., Leger, C., Aumont-Nicaise, M., ... & Valerio- Lepiniec, M. (2023). Design of an artificial phage-display library based on a new scaffold improved for average stability of the randomized proteins. Scientific Reports, 13(1 ), 1339.

Claims

[Claim 1] A cell-free method for making a library of phage variants comprising the steps of:

(a) providing a library of phage genomes in mixture, and

[Claim 2] The method of claim 1 , wherein step (a) comprises a step of in vitro assembling the phage genomes from DNA fragments.

[Claim 3] The method of claim 1 or 2, wherein step (a) comprises:

(a1 ) digesting the overlapping DNA fragments by an exonuclease; and

(a2) annealing the overlapping DNA fragments to assemble a phage genome.

[Claim 4] The method of any one of claims 1 to 3, wherein step (a) comprises synthetizing the library of phage genomes in a same reaction mixture.

[Claim 5] The method of any one of claims 1 to 3, wherein step (a) comprises synthesizing phage genomes separately and mixing the phage genomes to provide a library of phage genomes in mixture.

[Claim 6] The method of any one of claims 1 to 5, wherein step (b) comprises contacting the library of phage genomes with a E coli lysate.

[Claim 7] The method of claim 6, wherein said E. coli lysate comprises core RNA polymerase and sigma factor 70.

[Claim 8] The method of any one of the preceding claims, wherein said library of phage variants encodes a library of polypeptides, in particular a library of heterologous polypeptides.

[Claim 9] The method of claim 8, wherein each polypeptide of said library of polypeptides is encoded as a fusion protein with a phage protein, preferably a phage tail protein or a phage coat protein.

[Claim 10] The method of any one of claims 1 to 7, wherein said library of phage variants comprises a library of mutant phage genomes.

[Claim 11] A method for screening a library of phage variants, comprising the steps of:

(i) making a library of phage variants according to the method of any of claims 1 to 10;

(ii) assaying a phenotype of the library of phage variants;

(iii) collecting a subset of phages based on their phenotype; and

(iv) optionally amplifying the collected phage genomes.

[Claim 12] The method of claim 1 1 , further comprising a step of identifying the sequence of the collected phages.

[Claim 13] The method of claim 1 1 or 12, comprising repeating one or more cycles of steps (i) to (iv).

[Claim 14] A method for screening a library of polypeptides, comprising the steps of:

I. making a library of phage variants according to the method of any one of claims 1 to 10, wherein the genome of each phage variant of the library of phage variants comprises a polynucleotide encoding a polypeptide of a library of polypeptides; II. contacting the library of phage variants with a test compound;

III. collecting a subset of phages which bind to the test compound ;

IV. optionally amplifying the collected phage genomes;

V. optionally repeating one or more cycles of steps I to IV or II to IV; and

[Claim 15] The method of claim 14, wherein the polypeptides are antibody, antibody fragments or are scaffold proteins comprising one or more variable amino acid residues.