Field of the invention

The invention relates to a library of replicating entities, each entity comprises a recombinant vector comprising a randomized nucleic acid sequence, having the reading frame structure [NXX] _n [CorAA] [NXX] _m [NZZ] ₀ , or [NZZ] ₀ [NXX] _m [CorAA] [NXX] _n . The invention further relates to a set of recombinant vectors, each vector comprises a randomized nucleic acid sequence, having said reading frame structure and to a set of randomized oligonucleotides, each oligonucleotide having said structure. Furthermore, the invention relates to a method for generating a library of replicating entities and to a method for identifying an amino acid polymer.

Background of the invention

Libraries of genes, small molecules, proteins or peptides are nowadays widely used for identifying novel compounds of particular pharmacological or chemical properties. One of the most successful strategies for identifying ligands from large biological libraries is the phage display method, which was developed more than 25 years ago. Following the first antibody libraries, random peptide libraries based on phage display were developed. Finally, screening approaches based on the concept of phage display libraries have also been introduced for eukaryotic cells, in particular yeast, but also for cells of higher organisms.

Despite several improvements of the techniques, screening results from such random peptide libraries are still not fully satisfying. In general, all approaches are based on randomly generated nucleic acid sequences, which are translated into a peptide within an organism, such that the library, at best, covers all possible variants of a peptide of a given length. However, as randomization is carried out on the level of the encoding nucleic acid sequence, already the number of nucleic acid sequences covering all possible variants of a peptide of only eight amino acids exceeds the size of a library that can be technically handled.

Moreover, the binding affinities of a peptide distinctly depend on its three-dimensional structure. As a consequence, many targets are bound by circular but not by linear peptides of corresponding sequences. Other targets, in contrast, are exclusively bound by linear peptides. Although ordinary peptide gene libraries theoretically include appropriate loops, their complexity is simply too small to cover the theoretically required amount of sequences. To address this problem, most screenings are performed using at least two different libraries, one including peptides that have been cyclised by including defined positions with codons for cysteines at or near the N- and C-terminus of the peptide. These cysteine residues can form a disulfide bond generating a loop structure. The use of several libraries is, however, time consuming and extremely elaborate, especially as most targets are either bound by linear peptides or by circular peptides.

Therefore, novel gene libraries are needed, which cover circular and linear versions of peptides alike with a reasonably high probability.

Summary of the Invention

In a first aspect, the invention relates to a library of replicating entities, each entity comprises a recombinant vector comprising a randomized nucleic acid sequence, having the reading frame structure [NXX] _n [CorAA] [NXX] _m [NZZ] ₀ , or [NZZ] ₀ [NXX] _m [CorAA] [NXX] _n wherein each NXX is independently a codon encoding for any amino acid except cysteine, CorAA is a codon encoding for cysteine or at least one other amino acid, each NZZ is independently a codon encoding for any amino acid, and n is an integer from 0 to 40, m is an integer from 1 to 20, o is an integer from 1 to 40, and at least 20 percent of CorAA encode for cysteine.

In a further aspect, the invention relates to a set of recombinant vectors, each vector comprises a randomized nucleic acid sequence, having the reading frame structure [NXX] _n [CorAA] [NXX] _m [NZZ] ₀ , or [NZZ] ₀ [NXX] _m [CorAA] [NXX] _n wherein each NXX is independently a codon encoding for any amino acid except cysteine, CorAA is a codon encoding for cysteine or at least one other amino acid, each NZZ is independently a codon encoding for any amino acid, and n is an integer from 0 to 40, m is an integer from 1 to 20, o is an integer from 1 to 40, and at least 20 percent of CorAA encode for cysteine. In a further aspect, the invention relates to a set of randomized oligonucleotides, each oligonucleotide having the structure [NXX] _n [CorAA] [NXX] _m [NZZ] ₀ , or [NZZ] ₀ [NXX] _m [CorAA] [NXX] _n wherein each NXX is independently a codon encoding for any amino acid except cysteine, CorAA is a codon encoding for cysteine or at least one other amino acid, each NZZ is independently a codon encoding for any amino acid, and n is an integer from 0 to 40, m is an integer from 1 to 20, o is an integer from 1 to 40, and at least 20 percent of CorAA encode for cysteine.

In a further aspect, the invention relates to a method for generating a library of replicating entities comprising the steps providing a set of randomized oligonucleotides of the invention, introducing each oligonucleotide into a replicating entity, and propagating the replicating entities as individual clones.

In a further aspect, the invention relates to a method for identifying an amino acid polymer able to interact with a target, comprising the steps providing a library of replicating entities of the invention, bringing the library into contact with the target, and enriching the replicating entities interacting with the target.

In a further aspect, the invention relates to the use of a set of randomized oligonucleotides of the invention for generating a library of replicating entities. Brief Description of the Drawings

Figure 1 shows a schematic drawing of the vectors pPepPr3A-stuffer (A) and pPepPr7B-stuffer (B) used to generate the phage display libraries ENTE-1 and ENTE-2, respectively.

Figure 2 shows the reading frame structure of the phage display library ENTE-1 including restriction sites and codon restrictions.

Figure 3 shows the distribution of amino acids at each position of the library ENTE-1.

Figure 4 shows a binding pattern analysis of the CD227 antibody and the BC2 antibody using the ENTE-1 library.

Detailed Description of the Invention

In a first aspect, the invention relates to a library of replicating entities, each entity comprises a recombinant vector comprising a randomized nucleic acid sequence, having the reading frame structure [NXX] _n [CorAA] [NXX] _m [NZZ] ₀ , or [NZZ] ₀ [NXX] _m [CorAA] [NXX] _n wherein each NXX is independently a codon encoding for any amino acid except cysteine, CorAA is a codon encoding for cysteine or at least one other amino acid, each NZZ is independently a codon encoding for any amino acid, and n is an integer from 0 to 40, m is an integer from 1 to 20, o is an integer from 1 to 40, and at least 20 percent of CorAA encode for cysteine.

The library of the invention is characterized in that the replicating entities encode for a large variety of peptides (also referred to as amino acid polymers), wherein the peptides are represented in a linear version and in a circular version. This is achieved by the specific structure of the reading frame of the nucleic acid sequence encoding for the peptide. In general, the nucleic acid sequence comprises two different parts, namely [NXX] _n [CorAA] [NXX] _m and [NZZ] ₀ . The part [NXX] _n [CorAA] [NXX] _m comprises at least two amino acids (n = 0; m = 1 ), of which no more than one amino acid (namely the one encoded by CorAA) may be a cysteine. The [NZZ] ₀ part consist of up to 40 amino acids and may comprise any combination of amino acids including cysteine. Besides the one position CorAA, which is occupied by a cysteine (C) in at least 20 percent of the nucleic acid sequences, all other codons are randomized. Thus, each NXX independently may be any codon encoding for an amino acid except cysteine and each NZZ independently may be any codon encoding for an amino acid. In other words, each NXX and each NZZ may be independently selected from codons listed in Table 1. Preferably, the codons encode for natural amino acids. However, one or more codons may encode for non-natural amino acids, when the library is generated using replicating entities or hosts of replicating entities, which are provided with tRNA molecules transferring non-natural amino acids.

Table 1 : Codons and encoded amino acids

12 GAA, GAG Glutamic acid

13 ACT, ACC, ACA, ACG Threonine

14 GGT, GGC, GGA, GGG Glycine

15 TGG Tryptophane

16 CAT, CAC Histidine

17 TAT, TAC Tyrosine

18 ATT, ATC, ATA Isoleucine

19 GTT, GTC, GTA, GTG Valine

20 TGT, TGC Cysteine*

* only for NZZ; In case of ribonucleic acid, thymine may be replaced by uracile.

In case the position CorAA is occupied by a codon encoding for cysteine and the part [NZZ]o of the nucleic acid sequence also comprises a codon encoding for cysteine, the resulting peptide will comprise two cysteines, which form a disulfide bond. This results in a circular version of the peptide. Moreover, due to the exclusion of cysteine from NXX, the occurrence of two closely adjacent cysteine residues in the [NXX] _n [CorAA] [NXX] _m part is avoided. This is advantageous, because two adjacent cysteines would both be able to form a disulfide-bond with a cysteine of the [NZZ] ₀ part, resulting in a nucleic acid sequence that encodes for two peptides of identical sequence, however, different conformation. Thus, restricting the number of cysteines in the [NXX] _n [CorAA] [NXX] _m part significantly improves the reliability of the library. Moreover, in contrast to usual cysteine constrained libraries, which carry a cysteine at each side of the randomized peptide, the library of the invention encodes for peptides having loops of different sizes, as the cysteine may be localized at any position.

Furthermore, due to its reading frame structure, the library of the invention provides the randomized peptides as a circular and a linear version. Namely, corresponding to a nucleic acid sequence encoding for a peptide having a cysteine at the CorAA position, the library also comprises a nucleic acid sequence encoding for a peptide, in which the CorAA position is occupied by a specific amino acid, other than cysteine. Additionally, at least 20 percent of the CorAA codons encode for cysteine. This is much more than in a standard randomized library. As a result, at least 20 percent of the replicating entities comprise a recombinant vector comprising a randomized nucleic acid sequence in which CorAA is a codon encoding for cysteine. Importantly, the peptides encoded by a randomized nucleic acid sequence with CorAA encoding for cysteine will form a disulfide bond, if a further cysteine is contained in the [NZZ] ₀ part. Corresponding peptides of identical sequence but with CorAA encoding for a different amino acid will remain linear. Thus, by the reading frame structure of the library, corresponding linear and circular peptides are covered by a single library. Additionally, by defining the position and probability of a cysteine in one part of the encoded peptide (namely the [NXX]n [CorAA] [NXXj _m part), the probability of a disulfide-bond within the peptide is specified. This results in a high statistical reliability of the library. Accordingly, a library of replicating entities is generated, in which each entity comprises a peptide, preferably presented on its surface. The peptides have the common structure of X _n C/AA X _m Z ₀ , with X being any amino acid except cysteine, C/AA being cysteine or at least one other amino acid and Z being any amino acid. Additionally, n is an integer from 0 to 40, m is an integer from 1 to 20, o is an integer from 1 to 40, and in at least 20 percent of the peptides C/AA is cysteine. An exemplary library may comprise a first set of replicating entities, each entity comprising a recombinant vector with a randomized nucleic acid sequence, having the reading frame structure [NXX] _n [TGY] [NXX] _m [NZZ] ₀ , and a second set of replicating entities, each entity comprising a recombinant vector with a randomized nucleic acid sequence, having the reading frame structure [NXX] _n [TCN] [NXX] _m [NZZ] ₀ , wherein the first set of replicating entities constitute at least 20 percent of the library. In this case, each NXX individually represents a codon encoding for any amino acid except cysteine, TGY represents a codon encoding for cysteine, TCN represents a codon encoding for serine, and each NZZ individually represents a codon encoding for any amino acid. Additionally, m is an integer from 0 to 40, n is an integer from 1 to 20, and o is an integer from 1 to 40.

In a preferred embodiment, the replicating entity is a cell or a virus, preferably the cell is a prokaryotic cell or a eukaryotic cell and/or the virus is a bacteriophage. The term "library" as used herein refers to a compilation of a large number of specimens, i.e. replicating entities, of the same kind, however, differing from each other such that the library covers a large variety of the respective specimen. Preferably, the library is formed by cells or viruses carrying DNA or RNA such that the replicating entity can translate the encoded peptide either itself (in case of a cell) or by means of a host (in case of a virus). Virus-based libraries of randomized peptides such as phage display libraries are well established and widely used. Accordingly, in a preferred embodiment, the library is a phage display library, preferably the phage is a filamentous phage. In a preferred embodiment, the phage is selected from the group consisting of M13, fd, fl, T and λ-phage. In addition to the well established technology of phage display libraries, cell-based libraries become increasingly important, in particular libraries based on bacteria or yeast but also based on fungi. The reading frame structure is likewise suitable for such libraries.

In a preferred embodiment, CorAA is a codon encoding for cysteine or 1 to 5 other amino acids, preferably 1 to 3 other amino acids, more preferred one other amino acid. To cover corresponding linear and circular peptides, the position CorAA within the [NXXJn [CorAA] [NXX] _m part of the nucleic acid sequence is occupied by either cysteine (allowing a loop formation within the resulting peptide, if a further cysteine is contained in the [NZZ] ₀ part) or one other amino acid that is not suitable for forming a loop (giving rise to a linear peptide). The other amino acid may be the same in all cases, i.e. may be one other amino acid, or may be different, e.g. one of 3 amino acids other than cysteine. For example, when generating the randomized nucleic acid sequences, a mixture of different nucleotide triplets including one nucleotide triplets representing a codon encoding for cysteine and three nucleotide triplets each representing a codon encoding for another amino acid is used. Randomly and depending on the abundance of each nucleotide triplet in the mixture, the codon encoding cysteine or a codon encoding one of the other amino acids is integrated into the randomized nucleic acid sequence. Preferably, CorAA encodes for cysteine or one other amino acid, as this results in an even and statistically predictable coverage of corresponding linear and circular peptides in the library.

In a preferred embodiment, CorAA is a codon encoding for cysteine (C) or at least one amino acid (AA) selected from the group consisting of tyrosine, phenylalanine, asparagine, aspartic acid, glutamine, glutamic acid, histidine, lysine, isoleucine, glycine, alanine, valine, threonine, proline, leucine, serine and arginine, preferably CorAA is a codon encoding for cysteine or serine. To cover corresponding linear and circular peptides, the position CorAA within the [NXX] _n [CorAA] [NXX] _m part of the nucleic acid sequence is occupied by one of only two possible amino acids, namely either cysteine (allowing a loop formation within the resulting peptide) or one other amino acid that is not suitable for forming a loop (giving rise to a linear peptide). The other amino acid may be any amino acid that is not suitable for forming a loop structure. However, serine is particularly preferred as it shows similar steric properties compared to cysteine except that it cannot undergo disulfide bonding. Thus, corresponding peptides carrying a cysteine or a serine at the CorAA position are particularly similar despite that one is provided in a circular conformation.

In a preferred embodiment, each amino acid is encoded by a single codon. For most randomized libraries, the encoding nucleic acid sequence is provided by randomly adding nucleotides (adenine, thymine, cytosine, guanine) to each other to form an oligonucleotide of random sequence. This oligonucleotide is then introduced into a vector as a reading frame, such that a randomized peptide is translated from the nucleic acid sequence. Due to the redundancy of the genetic code, however, this leads to the formation of 64 different codons (each comprising three nucleotides) that represent 20 amino acids and four functional codons (start and stop codons). However, in case a start or stop codon is present within the nucleic acid sequence, it most likely encodes for a non-functional peptide. Therefore, various approaches have been developed to reduce the number of stop codons. For example, libraries were designed, in which the third position of each codon must not contain adenosine, i.e. the third nucleotide is either guanine, cytosine or thymine (so-called NNB-library), or guanine or thymine (so- called NNK-library). This eliminates the stop codons TAA and TGA and reduces the number of codons to 32 (Dennis et al., 2002). The presence of the third stop codon TAG is usually overcome by using host organisms that provide a tRNA to translate this codon into an amino acid. Nevertheless, even in libraries of the NNK or NNB type, most of the amino acids are encoded by more than one codon, such that these libraries necessarily comprise redundant peptides. Thus, for covering all possible varieties of a peptide of a given length, such a library has to provide many more encoding sequences compared to the actually expressed peptide versions. For example, to cover a randomized peptide of seven amino acids, a fully randomized library would have to cover between 32 ⁷ to 64 ⁷ , i.e. 3.5 x 10 ¹⁰ to 4.4 x 10 ¹² sequences. This is not taking into deviations from the ideal composition, which can be more than ten-fold as commonly reported for phage libraries ft Hoen et al., 2012). Establishment and maintenance of such large libraries, however, is difficult if not impossible. For example, the number of DNA molecules that can be generated in vitro is limited and, most importantly, the efficiency of introducing the DNA into the replicating entities declines with the size of the library. Thus, using the techniques of the state of art, reliable libraries for random peptides of more than seven amino acids can hardly be generated.

In contrast, if each amino acid is encoded by a single codon, meaning all other codons encoding for the same amino acid are not used, the number of codons can be reduced to 20 exactly corresponding to the number of amino acids. This can be achieved by generating the nucleic acid sequence representing the reading frame by using oligonucleotides of three nucleotides of distinct sequence. In this case, each oligonucleotide represents one codon encoding for a distinct amino acid. The oligonucleotides are randomly combined to provide all versions of a peptide of a given length. If a single codon is used for each amino acid, the redundancy of peptides in the library is significantly reduced such that within the technical limits of the library a larger variety of peptides can be covered by a given number of sequence variants. This allows the generation of statistically more reliable libraries encoding for peptides of seven amino acids and more. For example, the diversity of a library required to cover nucleic acid sequences encoding for peptides of seven amino acids is significantly reduced compared to a usual NNK- or NNB-library, i.e. 26 fold or 460 fold, respectively. Moreover, using defined nucleotide triplets corresponding to specific codons, functional codons such as start and stop codons can be entirely excluded. This allows overcoming present limitations with respect to specific organisms that have to be used to avoid the introduction of stop codons. For example, for producing phage display libraries E. coli strains carrying suppressor tRNAs, e.g. supE or supF, are the most used organism as it translates the codon TAG into glutamine instead of recognizing the same as a stop codon (Bossi 1983).

In a preferred embodiment, at least 25 percent, preferably at least 50 percent, more preferred about 30 to 50 percent, most preferred about 50 percent of CorAA encode for cysteine. The higher the probability of a cysteine contained in the [NXX] _n [CorAA] [NXX] _m part, the higher is the number of peptides encoded by the library, which contain a disulfide-bond and are, thus, present in a circular conformation. For providing a library covering a linear and a circular version for each encoded peptide, about 50 percent of CorAA should encode for cysteine. The amount of CorAA encoding for cysteine can be determined, e.g. by adjusting the proportion of nucleotide triplets representing codons encoding for cysteine in the mixture of triplets used for CorAA when generating the randomized nucleic acid sequence. In a preferred embodiment, each NXX individually represents a codon encoding for any amino acid except cysteine and methionine. Methionine contains a thioether group which is subjected to gradual oxidation leading to the formation of methionine sulfoxid and methionine sulfone. Therefore, the amount of methionine present in a library, decreases over time. Experiments with phage display libraries revealed that already after a second round of selection hardly any methionine is present in the peptides of the library. This is particularly important as the oxidation products show different binding properties compared to methionine. Therefore, a peptide of the library containing an oxidation product of methionine may show binding affinities to a given target, whereas the same peptide containing methionine would not. Thus, reducing the amount of methionine in the peptides encoded by the library is advantageous. Accordingly, in a preferred embodiment, each NZZ individually represents a codon encoding for any amino acid except methionine. More preferred, neither NXX, NZZ nor CorAA encode for methionine such that the library exclusively encodes for peptides devoid of methionine. Instead of excluding methionine, a non-natural amino acid having a similar structure and showing similar interaction properties as methionine, but lacking the thioether group, may be incorporated.

In a preferred embodiment, each NXX individually represents a codon encoding for any amino acid except cysteine and tryptophane. Including tryptophane in the peptides encoded by the library promotes unspecific binding of the peptide to a target and, thus, increases the occurrence of false positive results. Consequently, the reliability of the library decreases with the amount of tryptophane occurring in the peptides encoded by the library. Moreover, experiments showed that an increased level of tryptophane provides bacteriophages with a survival advantage as they propagate more successfully. As a result, phages expressing peptides containing few or no tryptophane, which are expected to show more reliable binding properties, are outnumbered. After several rounds of selection this leads to an enrichment of false positive results due to high levels of tryptophane, whereas the truly binding peptides are lost. Finally, tryptophane is particularly expensive to produce and chemically unstable such that for large scale production peptides devoid of tryptophane are preferred. In summary, the reduction of tryptophane in the peptides encoded by the library is advantageous for several reasons. Accordingly, in a preferred embodiment, each NZZ individually represents a codon encoding for any amino acid except tryptophane. Preferably, neither NZZ, NXX nor CorAA encode for tryptophane such that the peptides encoded by the library are devoid of tryptophane.

In a particularly preferred embodiment, each NXX individually represents a codon encoding for any amino acid except cysteine, methionine and tryptophane and each NZZ individually represent a codon encoding for any amino acid except methionine and tryptophane.

Besides methionine and tryptophane, any other amino acid may be excluded from the peptides of the library by excluding the codons encoding for the respective amino acid from the randomized nucleic acid sequence. In a preferred embodiment, n is an integer from 2 to 20, preferably from 5 to 15.

In a preferred embodiment, m is an integer from 2 to 15, preferably from 5 to 10.

In a preferred embodiment, o is an integer from 2 to 20, preferably from 5 to 15.

In a preferred embodiment the randomized nucleic acid sequence comprises at least 21 nucleotides, preferably 21 to 20 nucleotides, more preferred 21 to 90 nucleotides, most preferred 24 to 60 nucleotides. The library of the invention is suitable for covering larger peptides than conventional libraries, because it avoids certain limitations as described above. For providing a randomized peptide of at least 7 amino acids, the nucleic acid sequences comprises 21 nucleotides in the reading frame. Using the library of the invention, however, peptides of more than 7 amino acids can be encoded. This is advantageous as larger peptides show more complex tertiary structures increasing their binding specificity to possible target molecules.

In a preferred embodiment, the library comprises at least 10 ⁵ preferably at least 10 ⁷ , most preferred at least 10 ⁹ replicating entities. Due to the reduced diversity of the library of the invention, already the size of 10 ⁵ replicating entities is suitable to cover at least 50 % of the number of replicating entities, which are needed to cover the variability required for a full coverage of all possible tetrameric peptides formed by the amino acids represented in all positions of the library.

In a further aspect, the invention is directed to a set of recombinant vectors, each vector comprises a randomized nucleic acid sequence, having the reading frame structure [NXXJn [CorAA] [NXX] _m [NZZ] ₀ , or [NZZ] ₀ [NXX] _m [CorAA] [NXX] _n wherein each NXX is independently a codon encoding for any amino acid except cysteine, CorAA is a codon encoding for cysteine or at least one other amino acid, each NZZ is independently a codon encoding for any amino acid, and n is an integer from 0 to 40, m is an integer from 1 to 20, o is an integer from 1 to 40, and at least 20 percent of CorAA encode for cysteine. The set of vectors is suitable to produce a library of replicating entities by introducing each vector into a replicating entity, e.g. a cell or a virus. The peptide encoded by the nucleic acid sequence is then produced either by the replicating entity itself (in case of a cell) or by use of a host (in case of a virus). The term "vector" as used herein refers to a DNA or RNA molecule used as a vehicle to introduce foreign genetic material in a replicating entity. The vector is replicated either by the entity itself or by use of a host. The vector may exist in the replicating entity as an individual molecule or incorporated into the entity's genome. Vectors include plasmids, viral vectors, phagemids, cosmids and artificial chromosomes.

In a preferred embodiment, the vector further comprises a regulated promoter controlling the expression of the randomized nucleic acid sequence, preferably the promoter is repressable. The activity of the promoter determines the amount of peptide produced and incorporated into the replicating entity. For example, in case of display libraries the peptide becomes localized at the surface of the replicating unit. Thus, the stronger the promoter the more peptides are presented at the surface of the replicating entity. A high amount of peptides, however, can cause an unspecific binding affinity of the replicating entity to a target molecule. As a result, the replicating entity would be enriched in selection processes due to its strong binding to the target, the peptide itself, however, would lack any specific binding properties. Thus, using a strong promoter increases the occurrence of false positive results. Therefore, it is preferred to use a regulated promoter such that the amount of peptide produced can be controlled. The regulation can for example occur via repression by compounds that can be added to the cell culture when propagating the replicating entity. Suitable promoters are e.g. those controlled by catabolics or metabolics of the replicating entity such as the Lac promoter or the PL promoter in case of bacteriophage systems. For example, using a Lac promoter, the production of peptides can be reduced by adding glucose to the bacterial culture when generating the bacteriophages. The P|_ promoter is even more preferred as it shows a rather low basic activity. The resulting bacteriophages carry only few peptides at their surface and thus lead to more precise and reliable screening results. A second effect of the promoter strength is an adverse counterselection. Since different peptide sequences will exhibit more or less toxic or growth limiting effects on the host, high expression usually depletes libraries of such genes coding for these peptides and generates a bias towards genes that are less harmful for the host. These clones have an advantage in replication over other sequences and significantly impair the selection.

In a preferred embodiment, the vector further comprises an endogenous gene of a replicating entity and the randomized nucleic acid sequence is located adjacent to the endogenous gene. This allows the production of a fusion protein, such that the peptide is translated as part of the endogenous protein of the replicating entity. Preferably, the nucleic acid sequence is positioned such that the peptide is fused to the N- or C- terminus of the endogenous protein. As a result, the peptide is processed and located within the replicating entity together with the endogenous protein. In a particular preferred embodiment, the endogenous gene encodes for a surface protein of the replicating entity, preferably for a phage coat protein, more preferred for gene III of M13. Expressing the peptide as a fusion protein with a surface protein of the replicating entity results in the presentation of the peptide on the surface of the replicating entity. For example, in case the entity is a cell, the protein may be a surface receptor or a membrane protein and the peptide may be fused thereto such that it is presented at the outside surface of the cell. Displayed on the surface of the cell, the peptide can well interact with any target of interest. Similarly, in a phage display library, the randomized nucleic acid sequence is positioned next to a gene encoding for a coat protein of the phage, preferably the gene III of M13. The peptide is then generated as a fusion protein of the coat protein and localized to the head of the phage, where it is free to interact with target molecules.

In a preferred embodiment, the vector further comprises a nucleic acid linker placed between the randomized nucleic acid sequence and the endogenous gene, the linker and the endogenous gene each comprises a restriction site, and the restriction site of the endogenous gene is located in the terminal portion of the gene that is located adjacent to the linker, such that upon cleavage the randomized nucleic acid sequence is fused to an inner portion of the endogenous gene. Upon cleavage the linker and the terminal portion of the endogenous gene are deleted such that the randomized nucleic acid sequence is fused to an interior region of the gene. Accordingly, the peptide is generated as a fusion protein of the endogenous protein, however, lacking the deleted terminus. In display libraries, unspecific binding often occurs due to an interaction of the peptide with the adjacent terminus of the endogenous protein. In other words, the binding affinity observed in the screening is due to the influence of the terminus of the endogenous protein to which the peptide was fused and can not be reproduced by the peptide only. By deleting the terminal portion and fusing the randomized nucleic acid sequence to an internal portion of the endogenous gene, the specificity of the peptide- target interaction can be validated in a second round of selection. In case the replicating entity no longer interacts with the target after deletion of the terminal part of the endogenous gene, the peptide is unlikely to show specific binding properties itself and should be excluded from further examinations. In case, however, the binding affinity still persists, the peptide is likely to possess specific binding properties.

In a preferred embodiment, the vector comprises a type Ms restriction enzyme cleavage site within the randomized nucleic acid sequence and a type lis restriction enzyme recognition site adjacent to the randomized nucleic acid sequence, wherein the cleavage site comprises a non-palindromic sequence. By cleavage and subsequent random re-ligation, it is possible to recombine randomized nucleic acid sequences, thereby increasing the variability of the set of recombinant vectors. Preferably, the restriction site is located between the part of the nucleic acid sequence comprising [NXXj _n [CorAA] [NXX] _m and the part of the nucleic acid sequence comprising [NZZ] ₀ . Cleavage and religation may be carried out after a first selection, increasing the variability of peptides already known to show certain binding affinities. Preferably, the restriction site is cleaved by a type lis enzyme, which cleaves in about 15 base pairs distance from the recognition site. Thus, the recognition site can be located outside the nucleic acid sequence while the enzyme still cleaves within the reading frame. As the restriction site is non-palindromic redirected ligation is ensured. By cleaving and religating the vectors, it is possible to increase the variability of vectors from 10 ⁴ to 10 ⁸ . In contrast to previously described technologies (e.g. WO9833901 (A2), the usage of libraries from codon based synthesis does not require any generation of special subsets.

In a preferred embodiment the type lis restriction enzyme recognition site is located within a sequence insert located between a first and a second part of the randomized nucleic acid sequence. This allows cleavage at the edge of the degenerated sequence and the generation of non-palindromic overhanging sequences at the cleavage site. For example, the vector may be used to generate a first library in the display vector, which is used for a first screening. Subsequently, the vectors are isolated from those replicating entities, which were enriched in the first screening, cleaved and recombined as described herein after removal of the linker. The particular advantage is that a primary library size of only the order of magnitude of the larger part of the randomized nucleic acid sequence allows generating a library of the maximum complexity through recombination. Since recombined DNA is more efficient in transforming the host (Collins et al., 2001), the overall work and the amount of oligonucleotides required for the cloning are significantly reduced.

In a further aspect, the invention relates to a set of randomized oligonucleotides, each oligonucleotide having the structure [NXX] _n [CorAA] [NXX] _m [NZZj ₀ , or [NZZ] ₀ [NXX] _m [CorAA] [NXX]n wherein each NXX is independently a codon encoding for any amino acid except cysteine, CorAA is a codon encoding for cysteine or at least one other amino acid, each NZZ is independently a codon encoding for any amino acid, and n is an integer from 0 to 40, m is an integer from 1 to 20, o is an integer from 1 to 40, and at least 20 percent of CorAA encode for cysteine. This set of oligonucleotides is suitable for producing the vectors and the library of the invention. The oligonucleotides encode for randomized peptides when introduced into the reading frame of a vector.

In a preferred embodiment, each codon encodes for a different amino acid. Thereby, a reduction to about 20 codons, corresponding to the number of amino acids, is possible. The number of codons can be reduced by using triplets of nucleotides corresponding to distinct codons, for generating the oligonucleotides. Each triplet of nucleotides corresponds to one codon representing one amino acid. For generating the oligonucleotides the triplets are randomly assembled. For example, when synthesizing the oligonucleotide, a mixture of nucleotide triplets each representing a codon of an amino acid, wherein each amino acid is represented once, is used for each NZZ position. Alternatively, the randomized oligonucleotides may be generated using Slonomics technology (Van den Brulle 2008). Additionally, for NXX positions a similar mixture of nucleotide triplets is used, however, lacking the nucleotide triplet representing the codon for cysteine. For the position CorAA a mixture of at least two different nucleotide triplets is used of which one encodes for cysteine and the other encodes for one other amino acid, preferably serine. As a result, at the position CorAA either cysteine or one other amino acid is located, whereas all other positions may encode for any amino acid, except of cysteine in the case of NXX. Moreover, the frequency with which the individual amino acids occur in the peptides, may be varied by adjusting the proportion of each nucleotide triplet contained in the mixture used for generating the oligonucleotide. In particular, the proportion of triplets representing codons for cysteine in the mixture for CorAA influence the abundance of circular peptides versus linear peptides. Furthermore, the proportion of triplets corresponding to codons of methionine, tryptophane and/or any other amino acid may be reduced in comparison to triplets encoding for other amino acids. Consequently, only few of the peptides encoded by the library will comprise a methionine and/or tryptophane.

In a preferred embodiment, each NXX and NZZ individually is one codon selected from each group, namely group 1 consisting of GCT, GCC, GCA, and GCG, group 2 consisting of TTA, TTG, CTT, CTC, CTA, and CTG, group 3 consisting of CGT, CGC, CGA, CGG, AGA, and AGG, group 4 consisting of AAA and AAG, group 5 consisting of AAT and AAC, group 6 consisting of ATG, group 7 consisting of GAT and GAC, group 8 consisting of TTT and TTC, group 9 consisting of CCT, CCC, CCA and CCG, group 10 consisting of CAA and CAG, group 1 1 consisting of TCT, TCC, TCA, TCG, AGT and AGC, group 12 consisting of GAA and GAG, group 13 consisting of ACT, ACC, ACA and ACG, group 14 consisting of GGT, GGC, GGA and GGG, group 15 consisting of TGG, group 16 consisting of CAT and CAC, group 17 consisting of TAT and TAC, group 18 consisting of ATT, ATC and ATA, group 19 consisting of GTT, GTC, GTA and GTG and group 20 consisting of TGT and TGC. As NXX may not encode for cysteine, group 20 only applies to NZZ. Furthermore, the codons used to encode for a specific amino acid may differ for NXX and NZZ. To increase the variability of the library, the number of codons is reduced, preferably to the number of amino acids represented in the peptides encoded by the library, in generally 20. To do so, each amino acid is represented by a single codon, such that during oligonucleotide synthesis no more than 20 different nucleotide triplets each representing a codon of a different amino acid, are used. Accordingly, from each group of codons encoding for the same amino acid, for example group 1 encoding for alanine, only one is selected and used. Thus, from each of the groups of codons listed in table 1 , one codon is selected and a corresponding nucleotide triplet is used for generating the randomized oligonucleotides of the invention. In a preferred embodiment, each NXX and NZZ individually is one codon selected from each group, namely group 1 consisting of GCT, GCC, GCA, and GCG, group 2 consisting of TTA, TTG, CTT, CTC, CTA, and CTG, group 3 consisting of CGT, CGC, CGA, CGG, AGA, and AGG, group 4 consisting of AAA and AAG, group 5 consisting of AAT and AAC, group 7 consisting of GAT and GAC, group 8 consisting of TTT and TTC, group 9 consisting of CCT, CCC, CCA and CCG, group 10 consisting of CAA and CAG, group 11 consisting of TCT, TCC, TCA, TCG, AGT and AGC, group 12 consisting of GAA and GAG, group 13 consisting of ACT, ACC, ACA and ACG, group 14 consisting of GGT, GGC, GGA and GGG, group 16 consisting of CAT and CAC, group 17 consisting of TAT and TAC, group 18 consisting of ATT, ATC and ATA, group 19 consisting of GTT, GTC, GTA and GTG and group 20 consisting of TGT and TGC. in this case the peptides encoded by the library are devoid of methionine and tryptophane.

In a preferred embodiment, each NXX individually is selected from the group consisting of AAA, AAT, ACT, ATA, CAG, CAT, CCA, CGT, CTG, GAA, GAC, GCC, GGT, GTT, TAC, TCT, TGG, TTT, and each NZZ individually is selected from the group consisting of AAA, AAC, ACT, ATC, CAG, CAT, CCA, CGT, CTG, GAA, GAC, GCT, GGT, GTT, TAC, TCT, TGC, TTC. Within one cell, not all codons are incorporated during translation with the same frequency. This is because tRNAs for some codons are less abundant than others. This can cause peptides having amino acids encoded by less frequent codons to be underrepresented in the library as fewer proteins are translated from such nucleic acid sequences. In the unfortunate event that the corresponding tRNA is entirely absent, translation of the peptide would be terminated prematurely, such that no peptide or a truncated peptide would be produced. Both reduces the reliability of a library. To avoid these problems, codons are selected, which are equally well processed and translated leading to a reliable distribution of amino acids within the peptides encoded by the oligonucleotides.

In a preferred embodiment, the oligonucleotide comprises at least 21 nucleotides, preferably 21 to 120 nucleotides, more preferred 21 to 90 nucleotides, most preferred 24 to 60 nucleotides.

In a preferred embodiment, the set of randomized oligonucleotides comprises at least 10 ⁵ , preferably at least 10 ⁷ , most preferred at least 10 ⁹ different oligonucleotides.

In a further aspect, the invention relates to a method for generating a library of replicating entities comprising the steps providing a set of randomized oligonucleotides of the invention, introducing each oligonucleotide into a replicating entity, and propagating the replicating entities as individual clones. The introduction of the oligonucleotides into the replicating entities may be, for example, achieved by incorporating each oligonucleotide into a vector such that the peptide is translated from the vector when introduced into a cell. For introducing the vector into the replicating entity various techniques depending on the kind of entity used are available. In case of a cell, e.g. a yeast cell, chemical or electrical transformation may be employed for introducing the vector into the cell. Alternatively, the vector may be transferred into a cell by transfection, e.g. using a virus particle as a carrier. After introduction, the vector may be maintained in the cell as an individual nucleic acid molecule, e.g. a plasmid, or integrated into the endogenous DNA or RNA. In case the replicating entity is a virus, e.g. a bacteriophage, the vector is introduced into a host cell, e.g. E. coli, which then produces the virus, or bacteriophage. In case of eukaryotic cells, sophisticated virus packaging cell lines are available. Finally, the replicating entities are propagated as clones, in case of bacteriophages by use of a host cell, such that each clone expresses one of the randomized peptides encoded by the library.

In a preferred embodiment, introducing the oligonucleotide into the replicating entity comprises incorporating the oligonucleotide into a recombinant vector comprising an endogenous gene of the replicating entity such that the oligonucleotide is located adjacent to the endogenous gene. The oligonucleotide is incorporated into the vector directly following a gene encoding for an endogenous protein of the replicating entity. This allows the expression of the randomized peptide as a fusion protein together with the endogenous protein of the replicating entity. Moreover, if the randomized peptide encoded by the oligonucleotide is expressed as a fusion protein together with a surface molecule, it is integrated into the viral envelope or phage head, respectively. Preferably, the vector is a phagemid.

In a preferred embodiment, the method further comprises the steps introducing each oligonucleotide into a recombinant vector, cleaving the recombinant vectors within the oligonucleotide, randomly ligating the vectors to form a concatamere, cleaving the recombinant vectors outside the oligonucleotide, and religating the vectors to generate novel circular recombinant vectors. By cleaving and randomly religating the vector, it is possible to significantly increase the variety of the library. Cleavage may be done after a first round of selection thereby specifically increasing the variability of peptides already showing certain binding affinity to the target. In detail, the oligonucleotides are designed as to include a cleavage site, preferably a cleavage site for a type lis restriction enzyme, e.g. as described herein. To ensure correct ligation, the cleavage site is designed such that upon cleavage non-palindromic termini are generated. The oligonucleotide is then integrated into a vector, e.g. a plasmid or phagemid, which is introduced into a cell, e.g. to produce bacteriophages. The replicating entities carrying the oligonucleotide are then subjected to a first round of selection to identify potential binding partners of the target molecule. In case of a phage display, the bacteriophages are brought into contact with the target and selected for those binding to the target. Subsequently, the vectors are extracted from the replicating entities that were found to interact with the target and digested with a restriction enzyme cleaving the recombinant vectors within the oligonucleotide. Ligation of the linearized vectors leads to the formation of a concatamere of multiple vectors. These concatameres are then cleaved with a restriction enzyme recognizing a restriction site located outside the oligonucleotide leaving linearized vectors, which are religated to form novel circular recombinant vectors. These vectors are then reintroduced in cells, e.g. to produce new bacteriophages. Interestingly, cleavage and religation of the vectors does not only increase the variety of the randomized peptides but the religated vectors were found to transform with much higher efficiency compared to the original vectors.

In a preferred embodiment, the method further comprises the steps including a nucleic acid insert into each oligonucleotide, introducing each oligonucleotide into a recombinant vector, cleaving the recombinant vectors to excise the insert, randomly ligating the vectors to form a concatamere, cleaving the recombinant vectors outside the oligonucleotide, and religating the vectors to generate novel circular recombinant vectors. Instead of designing the oligonucleotide such as to include a restriction site, the oligonucleotide may include an insert located between a first and a second part of the randomized nucleic acid sequence, which comprises the restriction site. Preferably the restriction site is a type lis restriction site.

In a further aspect, the invention relates to a method for identifying an amino acid polymer able to interact with a target, comprising the steps providing a library of replicating entities of the invention, bringing the library into contact with the target, and enriching the replicating entities interacting with the target. The library of the invention is particularly suitable for screening for interaction partners of a given target. The term "target" as used herein refers to any kind of molecule, preferably to a biomolecule, e.g. a peptide, protein or chemical compound. For example, using the present method, amino acid polymers acting as agonists or antagonists to a target receptor can be identified. Likewise, diagnostic tools for detecting a given target compound can be established based on the interaction of the amino acid polymer and the target compound. The library provides a large variety of randomized amino acid polymers (also referred to as peptides), as linear and circular versions. The peptides may be presented on the surface of the replicating entities, each entity presenting an individual amino acid polymer. Further, each peptide may be presented as such or as part of a larger molecule, e.g. a protein. The peptides are brought into contact with the target molecule by means of the replicating entities in a single screening. For example, the target molecule may be immobilized on a surface, to which the library e.g. the bacteriophages in case of a phage display library, are added. Subsequently, those replicating entities carrying a peptide, which is able to interact with the target are enriched, e.g. by a washing step removing non- and weak-binding replicating entities. The remaining bound entities may then be eluted and collected.

In a preferred embodiment, the method further comprises the steps sequencing the genetic material of the enriched replicating entities and determining an interaction between the amino acid polymer encoded by the replicating entities and the target. By sequencing the genetic material of the replicating entity, the randomized nucleic acid sequence is identified, which encodes for the amino acid polymer. Preferably, sequencing is performed using Next-Generation-Sequencing techniques (Metzker, 2005). After determining the nucleic acid sequence, the amino acid polymer is produced, e.g. by solid phase synthesis and its ability to interact with the target is verified. For example, the interaction between the amino acid polymer and the target may be determined by using a chemical reaction or a physical signal depending on the binding of the amino acid polymer to the target.

Examples

Material

Vent Proof-reading polymerase was provided from New-England Biolabs and dNTPs from Life Technologies. PCR-Oligos were provided from Eurofins Genomics. Point- mutations were performed with the QuickChange site-Directed Mutagenesis kit (Agilent). Restriction enzymes were provided either from Thermo Scientific or New England Biolabs. PCR-products were purified with the QIAquick PCR purification kit (Qiagen). Vector-DNA was purified using the Gel extraction kit (Omega, bio-tek). PCR- fragments were purified with the Nucleotide removal kit (Qiagen). DNA-ligations were carried out with T4 ligase (Thermo Scientific). TOP10F' cells were supplied from Life Technologies and the phage lambda lysogen TG1(A) was generated from E.coli K12 TG1 (originally obtained from D. Legendre, Universite Catholique de Louvain, Belgium) after infection with lambda phage.

Methods

Plasmid DNA containing the gene 3 driven by the pL-promoter was transformed in TG1(A) whereas plasmid DNA containing the gene 3 controlled by the Lacl-promoter was transformed in TOP10F'. KS(+) and derived KS(+)K DNA was transformed in TOP10F'. Both vectors used for the ENTE1 and ENTE2 libraries were constructed in several steps. PGR conditions and primers are described in the tables.

Construction of pPepPr3A-stoffer (ENTE1 )

All vectors were derived from the p AMPF vector (GenBank: 33637.1 ).

pPepPr1-mut1 : creation of a Bglll restriction site with the QuickChange site-Directed Mutagenesis kit (see table 2).

pPepPr1-mut2: creation of a second Bglll restriction site with the QuickChange site- Directed Mutagenesis kit (see table 2).

pPepPr1-mut3: elimination of the Chloramphenicol acetyltransferase gene. pPepPrl- mut2 was digested with the restriction enzyme Bglll. Digested pPepPr1-mut2 was gel purified and self-ligated.

pPepPr1-mut2: elimination of one Bpml restriction site with the QuickChange site- Directed Mutagenesis kit (see table 2).

pPepPr2: elimination of a second Bpml restriction site with the QuickChange site- Directed Mutagenesis kit (see table 2).

pPepPr2-mut1 : elimination of the BsmBI restriction site with the QuickChange site- Directed Mutagenesis kit (see table 2).

pPepPr2-mut2: elimination of the Bsal restriction site with the QuickChange site- Directed Mutagenesis kit (see table 2).

pPepPr3: elimination of the Earl restriction site with the QuickChange site-Directed Mutagenesis kit (see table 2). pPepPr3-stuffer: cloning of β-galactosidase gene between the Nael and BstXI restriction sites of pPepPr3. The gene was amplified through PGR (see table 1 ). The vector and β-galactosidase gene were digested with Nael and BstXI, purified and ligated. pPepPr3-stuffer-mut2: creation of a Bsgl restriction site in the leader sequence with the QuickChange site-Directed Mutagenesis kit (see table 2). pPepPr3A-stuffer (Figure 1A, SEQ ID NO.: 44): cloning again of the β-galactosidase gene between the Nael and BstXI restriction sites of pPepPr3. In addition, the primers were designed to create two BsmBI sites. The β-galactosidase gene was amplified through PGR. The vector and gene were digested with Nael and BstXI, purified and ligated.

Construction of pPEPPR7B-stuffer f ENTE2)

The pL-promoter of pPepPr3A-stuffer was replaced by the Lacl-promoter. 3 BstXI, one BsmBI and one Kpnl restriction sites were mutated. Two vectors, pFAB74 and KS(+) were used for subcloning. Lacl comes from pFAB74. Three BstXI resctriction sites of Lacl were mutated in pFAB74. Then, Lacl was cloned in KS(+) and 1 BsmBI and one Kpnl restriction sites were mutated. pFAB74mut5: elimination of one BstXI restriction site with the QuickChange site- Directed Mutagenesis kit (see table 2). pFAB74 contains the Lacl promoter. pFAB74mut54: elimination of one BstXI restriction site with the QuickChange site- Directed Mutagenesis kit (see table 2). pFAB74mut543: elimination of one BstXI restriction site with the QuickChange site- Directed Mutagenesis kit (see table 2).

KS(+)K: the Kpnl restriction site was removed after digestion with Kpnl and digestion with Klenow.

KS(+)KLacl: the Lacl-promoter was amplified through PCR (see table 2). KS(+)K and Lacl were digested with Xbal and EcoRI, purified and ligated.

KS(+)KLacl189: KS(+)KLacl was modified to remove one BsmBI restriction site in the Lacl-promoter. A 205-bp fragment was amplified by PCR (see table 2). One primer containing the Kasl restriction site was designed to introduce a mutation in the BsmBI restriction site close to Kasl. KS(+)K and the purified PCR-fragment were digested with Hpal and Kasl, purified and ligated.

KS(+)KLacl 189255: KS(+)KLacl189 was modified to remove one Kpnl restriction site in the Lacl-promoter. A 261 -bp fragment was amplified by PGR (see table 2). Primers were designed to contain a Kasl restriction site and a BstXI restriction site. Kpnl is compatible with BstXI therefore Kpnl and BstXI disappear after ligation. KS(+)KLacl189 was digested with Kasl and Kpnl, and the purified PCR-fragment was digested with Kasl and BstXI. Then, the vector and the insert were purified and ligated.

pPepPrSA-stuffer: the Bsgl restriction site was removed. The promoter / leader sequence was recovered from pPepPr2 through PGR, digestion with Bglll and Nael and cloned between the same restriction sites of pPepPr3A-stuffer. pPepPr2 does not have the restriction site Bsgl upstream of Nael.

pPepPr6A-stuffer: insertion of one Bglll restriction site in the 5'UTR of gene3 with the QuickChange site-Directed Mutagenesis (see table 2).

pPepPr7Ac2-stuffer: insertion of the Lacl between the two Bglll restriction sites. Lacl was amplified through PGR (see table 2). pPepPr6a-stuffer and Lacl were digested with Bglll, purified and ligated.

pPepPr7A-stuffer: truncation of a region upstream of Lacl. A 1000-bp fragment was amplified through PGR (see table 2). pPepPr7Ac2-stuffer and PGR product were digested with BstBI and EcoRV, purified and ligated. It leads also to the elimination of one Bglll restriction site upstream of Lacl.

pPepPr7B-stuffer (Figure 1 B, SEQ ID NO.: 45): reparation of the OmpA region. A 499- bp fragment was amplified through PCR (see table 2). pPepPr7A-stuffer and PCR product were digested with EcoRV and Nrul, purified and ligated. It results also of the elimination of the second Bglll restriction site downstream of Lacl.

Table 2: PCR conditions for construction of pPepPr3A-stuffer and pPEPPR7B-stuffer

PGR purpose / template Primer sets comments construction mut-R2 Bglll restriction site pPepPr2 pPepPr1-mut3 Bpml-mut-F and Elimination of one Bpml

Bpml-mut-R restriction site in the beta-lactamase gene pPepPr2-mut1 pPepPr2 BsmBI-mut-F1 and Elimination of one

BsmBI-mut-R1 BsmBI restriction site.

Creation of one Xhol site

pPepPr2-mut2 pPepPr2-mut1 Bsal-mut-F1 and Bsal- Elimination of one Bsal mut-R1 restriction site pPepPr3 pPepPr2-mut2 Earl-mut-F1 and Earl - Elimination of one Earl mut-R1 restriction site β-galactosidase pud 9 stuffer-BstXI-F and Amplification of a beta- stuffer-Nael-R galactosidase gene for cloning in pPepPr3 pPepPr3-stuffer- pPepPr3-stuffer Bsgl-mut-F1 and Bsgl- Creation of one Bsgl mut2 mut-R1 restriction site in the leader sequence β-galactosidase puc19 BsmBI-mut-F2 and Amplification of a beta- BsmBI-mut-R3 galactosidase gene gene for cloning in pPepPr3-stuffer-mut2 pFAB74mut5 pFAB74 Lac-BstXlm5-F and Elimination of one BstXI

Lac-BstXlm5-R restriction site pFAB74mut54 pFAB74mut5 Lac-BstXlm4-F and Elimination of one BstXI

Lac-BstXlm4-R restriction site pFAB74mut543 pFAB74mut54 Lac-BstXlm3-F and Elimination of one BstXI

Lac-BstXlm3-R restriction site

Lacl pFAB74mut543 Lac-F1 and Lac-R1 Amplification of Lacl for cloning in KS(+)K PCR purpose / template Primer sets comments construction

205-bp pFAB74mut543 Lacl-EcoRV-F and Amplification of a

Lacl-Kasl-R fragment of Lacl for cloning in KS(+)K

261 -bp pFAB74mut543 Lacl-F2 and Lacl- Amplification of a

BstXI-R fragment of Lacl for cloning in KS(+)K pPepPr6A-stuffer pPepPr5A-stuffer Bglll-mut-F3 and Bglll- Creation of one Bglll mut-R3 restriction site pPepPr7Ac2- KS(+)KLacl 189255 Lac-F1 and Lac-R1 Lacl amplification for stuffer cloning in pPepPr6A- stuffer

pPepPr7A-stuffer pFAB74mut543 Lac-F3 and Lac-R3

pPepPr7B-siuffer pPepPr7A-stuffer Lacl-EcoRV-F and

SD_OmpA-R2

ENTE1 ENTE1 oligo fwd1 and rev2 Entel oligo amplification for cloning in pPepPr3A-stuffer

ENTE2 ENTE2 oligo Ente2-F and Ente2-R Ente2 oligo amplification for cloning in pPepPr7B-stuffer

Table 3: Primer sequences

Lacl-R1 AATTTGAATTCAGCTGTTTCCTGTGTGAAATTG 24

Lac-EcoRV-F TGCGGATATCTCGGTAGTGG 25

Lac-Kasl-R ATTGGGCGCCAGGGTGGTTTTTCTTTTCACCAGTGACACGGGCAACAG 26

CTG

Lac-BstXI-R AATTCCAGGTACATGGAGCTCACTGCCCGCTTTCC 27

Bglll-mut-F3 CCAAGGAGGTCTAGATAACGAGGGAGATCTATGAAAAAGACAGCTATC 28

GC

Bglll-mut-R3 GCGATAGCTGTCTTTTTCATAGATCTCCCTCGTTATCTAGACCTCCTT 29

GG

Lacl-F3 TTAATTCGAAACCGTATTACCGCCTTTGAG 30

Lacl-R3 TTAAGATATCCGCACCAACGCGC 31

SD_OmpA-R2 TTTAATCGCGATAGCTGTCTTTTTCATTTTTTGCCCTCGTGTGAAATT 32

GTTATCCGC fwd1 GTGTCGACGTCTCCCGGC 33 rev2 ACGTCTCCCTCCGCTGGAG 34

Ente2-F GTGTCGACGTCTCCCGGC 35

Ente2-R CTGTCGACGTCTCCCTCC 36

Construction ENTE-1 library

For generating the ENTE-1 library (Figure 2) dsDNA oligo sequences having the structure GTGTCGACGTCTCCCGGCN##N##N++TSCN##N##N##N##NZZNZZNYYNZZNZZNZ ZNZZNNCTCCAGCGGAGGGAGACGTCGACAG (SEQ ID NO.: 37) and

CACAGCTGCAGAGGGCCGN##N##N++ASGN##N##N##N##NZZNZZNYYNZZNZZ N ZZNZZNNGTGGTCGCCTCCCTCTGCAGCTGTC (SEQ ID NO.: 38) were generated using the mixtures of trinucleotides as listed in table 4 Table 4:

^* Palindrome

Sequences were amplified through PGR using the fwd1 and rev2 primers (see table 3). PGR conditions were 5 min 95°C, 5 cycles (1 min 95 °C; 1 min 58 °C, 20 sec 72 °C). For 50μΙ, 10 mM dNTPs, 50 pmol each primer, 1 μΙ Vent polymerase (2 units), 100 ng (1 .71 pmol) dsDNA oligo, 5 μΙ 10x vent buffer. Several PGR products were purified using the QIAquick PGR purification kit (Qiagen) according the manufacturer's instructions. The purified product was digested with the BsmBI restriction enzyme in a concentration of 20 units / pg dsDNA for 4 hours at 37 °C. Esp3l was heat inactivated at 65 °C for 20 min. DNA was ethanol precipitated overnight at - 80 °C. DNA was centrifuged at 22,000 g for one hour at 4 °C, washed with 70 % ethanol and centrifuged again at 22,000 g for 20 min at 4°C. The DNA pellet was air dried and resuspended in H ₂ 0. pPepPr3A-stuffer plasmid dsDNA was digested for 4 hours at 37 °C with BsmBI restriction enzyme and Sacl in a concentration of 2.5 units and 5 units / pg DNA, respectively. Vector DNA was two times gel purified using the Gel extraction kit (Omega, bio-tek). The Esp3l-digested PGR product was ligated to Esp3l-digested pPEPPR3A vector. 285 fmol vector and 1440 fmol insert were incubated with 30 units T4 ligase in 30 μΙ, overnight at 16 °C. Ligation product was heat inactivated 10 min at 65 °C and ethanol precipitated overnight at - 80 °C. DNA was centrifuged at 22,000 g for one hour at 4 °C, washed with 70 % ethanol and centrifuged again at 22,000 g for 20 min at 4°C. The DNA pellet was air dried and resuspended in H ₂ 0. 70-80 μΙ electrocompetent TG1(A) in a concentration of 1.5-2.5 10 ¹¹ cells / ml were incubated with ca 60-100 ng ligation product in a 1-mm sample cuvette and pulsed by a Biorad electroporator apparatus set at 25pF, 200 Ω and 1800 V. 10-15 transformations were pooled and added to 25-30 ml pre-warmed SOC medium in a flask. Cells were shaken at 220 rpm at 37 °C for one hour and transformants were selected on dYT-plates supplemented with 200 pg / ml ampicillin. The plates were incubated at 30°C overnight. Transformants were counted and the complexity of the library was calculated. Each plate was washed with 25-30 ml dYT to remove the bacteria. The bacteria were collected in a flask and the optical density was estimated.

To recover the phage, bacterial cells were diluted to OD 0.3 in dYT supplemented with 400 pg /ml ampicillin and cultivated one hour at 37 °C with gentle agitation. Then, M13K07 helper phage was added at MOI of 20, cells were incubated for one hour at 37°C with gentle agitation. Then, the cells were incubated overnight at 30°C. Cells were discarded through centrifugation and phage particles present in the supernatant were precipitated with 20% PEG / 2.5 M NaCI. To prepare plasmid DNA from the cells, bacterial cells were diluted to OD 0.3 in dYT supplemented with 200 pg /ml ampicillin and cultivated overnight at 30 °C with gentle agitation. Cells were harvested and plasmid DNA was extracted and purified with the Nucleobond midiprep kit from Macherey-Nagel according the manufacturer's instructions.

Cosmix plexing was performed as follows: 50 pg DNA was digested in 50 μΙ with 0.35 unit Bpml / μg DNA for 4 hours at 37 °C. Bpml was heat inactivated at 65 °C for 20 min. Then ca 40 ig digested DNA was ligated with 40 units T4 ligase in 50 μΙ for one hour at room temperature. T4 ligase was heat inactivated at 65 °C for 10 min. Then, DNA was digested in 100 μΙ with 0.5 units Bgll / μg DNA for two hours at 37 °C. Bgll was heat inactivated at 65 °C for 20 min. DNA was self-ligated at a concentration of 20 ng / μΙ with 2.5 units / g T4 DNA ligase, overnight at 16 °C. T4 ligase was heat inactivated at 65 °C for 10 min. DNA was again digested with 1 unit BsmBI / μg DNA for 1 hour at 37 °C. BsmBI was heat inactivated at 65 °C for 20 min. DNA was centrifuged at 22,000 g for one hour at 4 °C, washed with 70 % ethanol and centrifuged again at 22,000 g for 20 min at 4°C. The DNA pellet was air dried and resuspended inHaO in a concentration of ca. 20 ng / μΙ.

Several transformations were performed with 20 ng DNA and 70-80 μΙ TG1(A) as described above. Transformants were selected on dYT-plates containing 200 pg / ml ampicillin. Transformants were counted and the complexity of the library was calculated.

Construction of the E TE-2 library

The dsDNA oligo sequences

GTGTCGACGTCTCCCGGCN##N##N##N##N##TSTGTTGTTGCAGGCACTGCACGC

CGTGCAGGCACCGTCGGTGTCTSTN##N##N##N##NZZNZZNZZNZZNZZGGAGGG AGACGTCGACAG (SEQ ID NO.: 39)

CACAGCTGCAGAGGGCCGN##N##N##N##N##ASACAACAACGTCCGTGACGTGC GGCACGTCCGTGGCAGCCACAGASAN##N##N##N##NZZNZZNZZNZZNZZCCTCC CTCTGCAGCTGTC (SEQ ID NO:: 40) were amplified through PGR using the fwdl and rev2 primers (Table 3). PGR conditions were 5 min 95°C, 10 cycles (1 min 95 °C; 1 min 60 °C, 1 min 72 °C).

For 50μΙ, 10 mM dNTPs, 50 pmol each primer, 1 μΙ Vent polymerase (2 units), 50 ng (1.71 pmol) dsDNA oligo, 5 μΙ 10x vent buffer. Several PGR products were purified using the QIAquick PGR purification kit (Qiagen) according the manufacturer's instructions. The purified product was digested with BsmBI restriction enzyme in a concentration of 40 units / pg DNA for 4 hours at 37 °C. DNA was purified with the Nucleotide removal kit according the manufacturer's instructions (Qiagen). pPepPr7B- stuffer DNA was digested overnight at 37 °C with BsmBI restriction enzyme in a concentration of 1.5 unit / pg DNA. DNA was gel purified using the Gel extraction kit according the manufacturer's instructions (Omega, bio-tek). Vector DNA was again digested two hours at 37 °C with 0.8 unit BsmBI, 0.4 unit EcoRI and 0.4 unit Sphl / pg DNA. Vector DNA was gel purified using the Gel extraction kit according the manufacturer's instructions (Omega, bio-tek). Linearized pPepPr7B was diluted in a concentration of 240 ng / pi. The BsmBI-digested PGR product was ligated to BsmBI- digested pPEPPR3A. 670 fmol vector and 3270 fmol insert were incubated with 30 units T4 ligase in 30 pi, overnight at 16 °C. The ligation product was purified with the Nucleotide removal kit (Qiagen). 70-80 pi electrocompetent TOP10F' cells in a concentration of 1.5-2.0 10 ¹¹ cells / ml were incubated with ca 50 ng ligation product in a 1-mm sample cuvette and pulsed by a Biorad electroporator apparatus set at 25pF, 200 Ω and 2200 V. 10-15 transformations were pooled and added to 50 ml pre-warmed SOC medium in a flask. Cells were shaken at 220 rpm at 37 °C for one hour and transformants were selected on LB-plates supplemented with 200 pg / ml ampicillin and 1 % glucose. The plates were incubated at 30°C overnight. Transformants were counted and the complexity of the library was calculated.

Cosmix plexing was performed as follows: 80 pg DNA was digested in 200 pi with 0.5 unit Bsgl / pg DNA, overnight at 37 °C. The ligation product was purified with the Nucleotide removal kit (Qiagen). DNA was precipitated with 1 volume of 20 % PEG / 2.5 M NaCI for 25 min at 37 °C. DNA was centrifuged at 22,000 g for 30 min at RT, washed with 70 % ethanol and centrifuged again at 22,000 g for 20 min at 4°C. The DNA pellet was air dried and resuspended in H ₂ O. Then, ca 35 pg digested DNA was ligated in a concentration of 500 ng / pi with 1.4 units T4 ligase / pg DNA for 36 hours at 16°C. T4 ligase was heat inactivated at 65 °C for 10 min. Then, total DNA was digested in 100 pi with 1 unit Bgll / pg DNA for three hours at 37 °C. Bgll was heat inactivated at 65 °C for 20 min. DNA was self-ligated in a concentration of 30 ng / pi with 2.8 units T4 DNA ligase / pg DNA, overnight at 16 °C. T4 ligase was heat inactivated at 65 °C for 10 min. DNA was precipitated with 1 volume of 20 % PEG / 2.5 M NaCI for 15 min at 37 °C. DNA was centrifuged at 22,000 g for 30 min at RT, washed with 70 % ethanol and centrifuged again at 22,000 g for 20 min at 4°C. The DNA pellet was air dried and resuspended in H ₂ 0. 70-80 μΙ TOP10F' electrocompetent cells were transformed with 130 ng DNA as described above.

Transformants were selected on LB-plates containing 200 pg / ml ampicillin and 1 % glucose, counted and the complexity of the library was calculated.

Results

Quality control of the ENTE-1 library

Phage DNA after the final transformation of the library was sequenced by Next Generation Sequencing. The amino acid distribution deviates only in position 2 significantly from the expected similar level of all amino acids, because the leaderpepiidase ompA preferentially processes cleavage sites with certain amino acids at this 2 ^nd position. In all other positions the number per amino acids does not differ by more than a factor of 2 between the maximal and minimal counts. In standard libraries this value is a factor of 10 or higher (based on data from Dias-Neto et al. 2009) Amino acid distribution in 871 ,069 sequences is shown in Figure 3.

The sequence redundancy from 871069 sequences obtained in a typical run is shown in table 5. The number of sequences found more than once is particularly small. In fact, taking into account PGR artefacts, it is almost negligible.

Table 5:

Effective selection of multiple sequences

Selections were carried out on FLAG M1 and FLAG M2 monoclonal antibodies, wherein the antibodies were generated with the peptide DYKDDDDK (SEQ ID NO.: 142). The core sequence required for binding the antibody is YK plus a negative charge in the proximity or a preceding small amino acid in the case of FLAG M1.

Already after the second round, more than 95% of those sequences reliably identified by standard sequencing show a binding motif comparable with the peptide used for immunization. In contrast, comparable efforts with common phage display libraries required three rounds of selection and result in 50% or less binding clones (Srila, W. & Yamabhai, M., 2013). FLAG M1 should, according to all published literature, only recognize an N-terminal epitope and FLAG M2 any comparable epitope. However, the obtained data (table 6 and table 7) suggests that recognition of the epitope may also be possible in other positions.

Table 6: M1 -Clones ENTE-1 Library (FLAG-tag: DYKDDDDK (SEQ ID NO.: 142))

SEQ ID NO.:

GAHLSQRVDYKEYKVSI 46

GAHLSQRVDYKEYKVSI 47

GVLHCDYKEKIYTQSSAS 48

GNQQCRQQLVDYKYSIYS 49

GPPPCIFYADYKY EGFS 50

GYRQSIQVDYKIRSERF 51

GYSWVSEWGFAYQVDYKIS 52

GHEHSWVQIDYKTAVRDS 53

GFT SLEVPYKQKQQL F 54

GIEMSILELVDYKANLYS 55

GEAPSYQYVDYKNIVDNS 56

GEV SYVDYKSPKKE AS 57

GVLHCDYKF LEYPKPNS 58

GDYRS FVYLDYKHKLEAS 59

GYKWSEFQQSQQGALFIS 60

GYKWSEFQQ SQQGALFIS 61

GYKWSEFQHFGQQGKYAS 62

GYK SEFYKDVKQQEGAS 63

GYKWSEFVQEEKKVNKDS 64

GYKWSEFHNQFPGVQDFS 65

GYK SETWRQVSNFQHAS 66

GYKWSETTHSVQVEAHAS 67

GYKWSEIHTVFEAAQVYS 68

GYKWSELYQVERDQY FS 69

GYKWSEYLIGKPHFEHDS 70

GYKWSQYHREDKLVQEIS 71

GYKWSQYHREDKLVQE IS 72

GYKWSQWHDPSKEAAYDS 73

GYKWSLFHKSEEQVDEYS 74

GYKWSLWLSELKQQNEAS 75

GYKWSAINPKIQQNQDFS 76

GYKWNSFNS SQYVPEPIS 77

GYKWNSFNSWSQYVPEPIS 78

GYKWCLVNQVQCNEQRAS 79

GYKWQSFQHNAEQHRKPYS 80

GYKFSEILRLDYHDLVNS 81

GYKFSEILRL DYHDLVN S 82

GYKFSEIFSVYGYEPHAS 83

GYKFSE FQ I SQADQPDS 84

GYKFSELQTRAYQPAVDS 85 GYKQKLYFAS 86

G KQKLYFAS 87

GQHVSVVQVGYKQKELNS 88

GYKLSELQSKTYFFPHFS 89

GAPASRGYKHKEYVRKCS 90

GVQSSKGYKAKEQFNKAS 91

GHFHSEVS KLKELIIYS 92

GVWVS NWGPHQSQQTNS 93

Binding motif: DYKxx(E)xx or GYKws(E)xx

Table 7: M2-Clones ENTE-1 Library (FLAG-tag: DYKDDDDK (SEQ ID NO.: 142))

SEQ ID NO.:

GQFFSTNDSHDYKDEDAS 94

GPEVSDYKDEDPFPYFS 95

GLESRSDNFIDYKDLDEDS 96

GNQGSWWQDYKQDDEFS 97

GPDPSNRDYKDWDVFSAS 98

GHQVCNYDFDYKDADKNS 99

GNPRSAEVYNDYKEQDIS 100

GEENCEHNDYKECDNSYYS 101

GEENCEHNDYKECDNSYYS 102

GEENCEHNDYKECDNSYYS 103

GVFPSVI FEDYKESDGDS 104

GYEQSKQPDYKWEDDHFS 105

GGTVCWLRDYKWEDEHFS 106

GGQHSEKDDYKWEDVRCS 107

GFNQSGFDYKIWDEQRIS 108

GVSGCYFDYKNCDETPDS 109

GHS SEAIDYKWQDIRDS 110

GPFWSTWVAVHDYKYEDS- 11 1

GNRQCYLDYKYEDHNAAS 112

GDDWSNYLDDYKLEDRYS 1 13

GGSQSHHEADYKLEDTYS 1 14

GGNTSWYEHDYKFEDQAS 1 15

GHQNSQWAWDYKHEDTFS 116

GFVVSPYDYKSEDTACFS 117

GANQSTDAYVDYKLLDYS 118

GANQSTDAYVDYKLLDYS 119

GHWQSAFDPDYKLTDTAS 120

GTVWSDGWSVDYKLADYS 121

GNIHSDYKLYDGTHATDS 122

GSLHSIWHQEDYKLQDFS 123

GDGWSKYFEDYKNCDTYS 124

GEVSSIQHWDYKNYDPNS 125

GPSTS NSDDYKFGDVDS 126

GLPVCGEELGIDYKFYDS- 127 GQSTCDDPWDYKCCDGNS 128

GFLASKWGHFEKDYKCYDS- 129

GHVLSDDFVDYKQ DLYS 130

GSLACDYKQYDPEVVRNS 131

GELHCFGENHDYKSADIS 132

GGRVCSYQDDYKSCEYS 133

GITLCAFHDYRWDDIQAS 134

GQFSSDYQISDYKELDYS 35

GKPHSDYVYNDCKQEDIS 136

GSPPCGWEAIQEYKLCDS- 137

GNQGCYKL PECYSVYNS 138

GAGYGCYLFYEIWYFGCCS 139

GVPPCNSEDKYCIDQFAS 140

GQYTCAWQ LLYQLCIFS 141

Binding motif: DYKxxDxx

Fingerprinting Antibody Epttoees

Monoclonal Antibodies CD227 (Becton, Dickinson and Company, BD Pharmingen™ Cat.No. 550486) and BC2 (PrimaBiomed USA Inc.) recognize the same site in the MUC-1 antigen PDXRP, with X being not preferred based on internal investigations applying peptide phage display and identifying less than 20 binding peptides per antibody. ENTE-1 library was used for two panning rounds on both antibodies. >200,000 sequences were determined from the first and second selection round and all patterns matching PDXRP were analyzed. The results (Figure 4) show a clearly differentiated binding pattern with respect to negatively charged glutamic acid (E) and hydrophobic amino acids like leucine (L), suggesting that the antibodies are not identical.

References

WO9833901 (A2)

Bossi, L. Context effects: translation of UAG codon by suppressor tRNA is affected by the sequence following UAG in the message. J. Mol. Biol. 1983; 164, 73-87.

Collins J, Horn N, Wadenback J, Szardenings M. Cosmix-plexing: a novel recombinatorial approach for evolutionary selection from combinatorial libraries. J Biotechnol. 2001 Jun;74(4):317-38.

Dennis MS, et al. Albumin binding as a general strategy for improving the pharmacokinetics of proteins. J Biol Chem 2002; 277:35035-35043.

Dias-Neto E, Nunes DN, Giordano RJ, Sun J, Botz GH, Yang K, Setubal JC, Pasqualini R, Arap W. Next-generation phage display: integrating and comparing available molecular tools to enable cost-effective high-throughput analysis. PLoS One. 2009 Dec 17;4(12):e8338.

Metzker ML. Emerging technologies in DNA sequencing. Genome Res. 2005 Dec;15(12):1767-76.

't Hoen PA, Jirka SM, Ten Broeke BR, Schultes EA, Aguilera B, Pang KH, Heemskerk H, Aartsma-Rus A, van Ommen GJ, den Dunnen JT. Phage display screening without repetitious selection rounds. Anal Biochem. 2012 Feb 15;421(2):622-31.

Srila W, Yamabhai M. Identification of amino acid residues responsible for the binding to anti-FLAG™ M2 antibody using a phage display combinatorial peptide library. Appl Biochem Biotechnol. 2013 Oct;171(3):583-9.

Van den Brulle et al. A novel solid phase technology for high-throughput gene synthesis. Biotechniques 2008; 45(3):340-3.