WO2004056995A1

WO2004056995A1 - Process for producing an in vitro peptide expression library

Info

Publication number: WO2004056995A1
Application number: PCT/GB2002/005909
Authority: WO
Inventors: Ole Jørgen MARVIK; Herald Reiersen; Ole Henrik Brekke
Original assignee: Affitech As; Owen, Deborah
Priority date: 2002-12-23
Filing date: 2002-12-23
Publication date: 2004-07-08
Also published as: AU2002368488A1

Abstract

The present invention provides a method of producing a peptide expression library which displays a plurality of different peptides, wherein displayed peptides coupled to the genetic material encoding them, said method comprising the following steps: (a) preparing a library of nucleic acid molecules wherein each member of the library comprises a nucleotide sequence encoding a nucleic acid binding amino acid sequence, an oligonucleotide tag and a nucleotide sequence encoding the peptide which is to be displayed said nucleic acid binding sequence being capable of coupling to the nucleic acid molecule which encodes it; (b) attaching said library or a subpopulation thereof to one or more solid supports, wherein said attachment occurs by specific hybridisation of the oligonucleotide tag of a library member to a complementary tag which is attached to a solid support or a discrete area of a solid support so that each solid support or discrete area of a solid support has attached a distinct type of individual library members; and (c) translating the library members. Nucleic acid and peptide expression libraries and the members thereof form yet further aspects of the invention, as do kits comprising these libraries and members thereof. Further aspects involve the use of said libraries to screen and identify peptides or proteins or other entities exhibiting desired properties or the use of said library to screen against another library.

Description

PROCESS FOR PRODUCING AN IN VITRO PEPTIDE EXPRESSION LIBRARY

The present invention relates to methods of producing an in vi tro peptide or protein expression library, the expression library so produced and the members thereof . The invention also relates to the use of the library to screen and identify peptides or proteins or other entities exhibiting desired properties or the use of the library to screen against another library. The invention further relates to kits comprising said libraries.

Molecular libraries are important tools in many areas of molecular and cellular biology and in the identification and development of new drugs and therapeutic and diagnostic agents. Such libraries frequently contain genetic material, for example fragments of genes or nucleic acid sequences associated with a suitable vehicle which allows the fragments to be amplified and manipulated, e.g. a plasmid. Such libraries may be screened at the nucleic acid level, i.e. by screening for the desired sequence or sequences using a nucleotide probe. Alternatively such libraries can be designed so that the peptides encoded by the nucleic acid fragments or genes are expressed and screening can be carried out using an appropriate "target" molecule to which it is desired that the selected protein should bind, or a peptide with a desired catalytic or enzymatic activity can be selected for. The latter type of libraries where the encoded peptides are expressed are known as "expression libraries" .

A major use of molecular libraries today is to identify nucleic acid molecules which encode peptides which are specific binding partners for proteins, nucleic acid molecules or other chemical entities, i.e. ligands which bind to particular receptors, antigens, etc .

As the screening of such libraries usually involves identifying the peptide of interest by identifying the gene or nucleic acid fragment that encodes it, the expression libraries often have the expressed peptide associated in some way with the genetic material which encodes it. In this way, once the desired peptide is selected the nucleic acid encoding it is automatically identified and no deconvolution of the library is required. This is generally achieved by displaying the peptides encoded by the library members on the surfaces of organisms such as viruses, particularly bacteriophage, or bacterial cells or yeast cells as a fusion with a surface protein of the organism in question. In this way the genetic material encoding the displayed peptide can be retained inside the genetic package and can easily be isolated and analysed once appropriate organisms have been selected. The technique known as phage display is a well known example of such a techniques. These techniques are also referred to as " in vivo" expression techniques as the peptides are expressed on the surface of viable organisms .

As many of the disadvantages associated with the conventional in vivo expression techniques discussed above arise from the in vivo nature of the expression libraries there has been some suggestion for the use of in vi tro, cell free expression libraries. For example, ribosome display libraries have been described in which a correctly folded complete protein carrying different display peptides in different members of the library and its encoding mRNA both remain attached to the ribosomes . This is achieved by ensuring that the protein chain and the mRNA does not leave the ribosome (ie. there is no stop codon and the ribosomes are stabilized) . Such expression libraries are the subject of several patent applications, for example as published in O92/02536 (The Regents of the University of Colorado) , O93/03172 (University Research Corporation) and O91/05058 (Kawasaki) .

Ribosome libraries suffer from certain limitations. RNA is very sensitive to RNAses and is thus difficult to work with. In addition, to maintain attachment to the ribosomes requires special conditions, e.g. the continued presence of magnesium ions which can create problems for screening and other steps . Most importantly, all the steps after translation, especially during the screening and selection procedure, may not be performed with harsh reagents as the ribosome: RNA link must be retained.

A further selection technique, called PROfusion, has been developed by Phylos, Inc. (www.phylos.com) . This involves covalently attaching puromycin to the 3 ' terminal of mRNA using a synthetic linker. In in vi tro translation a ribosome travels along the mRNA generating a protein the C-terminus of which covalently binds to the puromycin leaving a protein bound to its own mRNA. While this covalent attachment is advantageous, the

PROfusion technique still requires complicated chemical procedures to achieve the two covalent attachments.

A more recent in vi tro expression system is described in W098/37186 (Actinova Limited) . In this system a peptide expression library may be generated in which the specific translation products of the genetic material in the library are directly and covalently attached to the encoding DNA sequence. This then obviates the need for cellular genetic packages with their inherent limitations during the construction and screening of the expression library. The covalent attachment of the expressed peptide to its encoding DNA sequence is made possible by the inclusion in the DNA of a sequence which encodes a so-called cis-acting protein or a pseudo-cis-acting protein which binds covalently to its own encoding DNA. The DNA of the library members also includes the coding sequence for the peptide for display. When the genetic material of the library members is translated, the DNA-binding protein and the display peptide form as a single polypeptide, which becomes covalently bound to the encoding DNA thus displaying the display peptide in association with the

DNA encoding it. Screening of such a library is carried out by conventional techniques wherein the target ligand is bound to a solid support, the library members are passed over the solid support, the solid support is washed and the library members which remain bound are identified.

Although such a system has overcome a number of the disadvantages associated with previous in vivo and in vi tro systems it still suffers from a number of drawbacks. For example, as the library members are all present in solution, manipulation and handling of the individual library members is difficult. Further disadvantages in such prior art techniques include (i) the considerable uncertainties that exist in estimating the actual diversity in the library, (ii) it is not possible to accurately estimate the efficiency of the translation, (iii) it is not possible to directly estimate the affinity between the expressed peptide bait and the target ligand, and (iv) a high background of unspecific binding is observed since the "bait" molecule (i.e. the expressed peptide) is in solution. In addition this system is not appropriate for library versus library screening.

In this regard, there is currently great interest in techniques and systems which would also allow the isolation of interaction partners among members of two different display libraries, e.g. a cDNA expression library versus an antibody library. The ability to do this would be a very valuable tool for elucidating the function of novel gene products. Such "library versus library" screening is not generally possible with current in vivo systems like phage display or known in vitro systems where the target ligand used to screen the library is conveniently present in an immobilized form whilst the candidate binding peptides are present in mobile form, e.g. on the phage particles in a liquid medium or otherwise present in a liquid medium. Surprisingly, a new and improved method of producing in vi tro peptide expression libraries has now been found which overcomes these drawbacks .

Thus, viewed from one aspect the present invention provides a method of producing a peptide expression library which displays a plurality of different peptides, wherein displayed peptides are coupled to the genetic material encoding them, said method comprising the following steps: (a) preparing a library of nucleic acid molecules wherein each member of the library comprises a nucleotide sequence encoding a nucleic acid binding amino acid sequence, an oligonucleotide tag and a nucleotide sequence encoding the peptide which is to be displayed, said nucleic acid binding amino acid sequence being capable of coupling to the nucleic acid molecule which encodes it;

(b) attaching said library or a subpopulation thereof to one or more solid supports, wherein said attachment occurs by specific hybridisation of the oligonucleotide tag of a library member to a complementary tag which is attached to a solid support or a discrete area of a solid support so that each solid support or discrete area of a solid support has attached a distinct type of individual library members; and

(c) translating the library members.

Any in vi tro library technique by which peptides can be generated and remain associated with the nucleic acid encoding them can be used in the methods of the present invention. Thus, any library of nucleic acid molecules can be used in step (a) of the method providing that the members of the library comprise a sequence encoding the peptide which is to be displayed and an oligonucleotide tag which is capable of attaching said library or a subpopulation thereof to one or more solid supports in accordance with step (b) of the method and further providing that the library members comprise a means of coupling the displayed peptides to the genetic material encoding them.

Thus, viewed from an alternative aspect the present invention provides a method of producing a peptide expression library which displays a plurality of different peptides, wherein displayed peptides are coupled to the genetic material encoding them, said method comprising the following steps :

(a) preparing a library of nucleic acid molecules wherein each member of the library comprises an oligonucleotide tag, a nucleotide sequence encoding the peptide which is to be displayed, and a means of coupling the displayed peptide to the nucleic acid molecule which encodes it; (b) attaching said library or a subpopulation thereof to one or more solid supports, wherein said attachment occurs by specific hybridisation of the oligonucleotide tag of a library member to a complementary tag which is attached to a solid support or a discrete area of a solid support so that each solid support or discrete area of a solid support has attached a distinct type of individual library members; and (c) translating the library members. Thus, in such alternative aspects rather than the members of the library comprising as a "means of coupling" a nucleotide sequence encoding a nucleic acid binding amino acid sequence which is capable of coupling to the nucleic acid molecule which encodes it, the members of the library might comprise as a "means of coupling" an alternative non-nucleic acid based molecule, moiety or entity, for example a non-nucleic acid based chemical molecule, moiety or entity (e.g. puromycin) which can facilitate the coupling of the displayed peptides to the genetic material encoding them.

Examples of such alternative libraries include those based on the technology of Phylos, Inc. (discussed above) . The Phylos technology relies on a puromycin molecule being covalently attached to the 3 ' end of mRNA using a synthetic linker. This puromycin molecule can covalently bind to the C-terminus of the peptide expressed from the mRNA molecule thereby giving rise to an expressed peptide linked to the mRNA encoding it. Thus the library of nucleic acid molecules used in step (a) may contain members which comprise mRNA molecules (which contain a sequence encoding the peptide which is to be displayed) , an oligonucleotide tag and a chemical moiety (for example a puromycin moiety) which is capable of coupling the expressed displayed peptide to the mRNA molecule encoding it .

Alternatively the library of nucleic acid molecules used in step (a) may comprise DNA molecules which comprise an oligonucleotide tag and a sequence encoding the peptide which is to be displayed and are capable of being transcribed to give an RNA molecule (which also contains the sequence encoding the peptide which is to be displayed) to which a chemical moiety (for example a puromycin moiety) which is capable of coupling the displayed peptide (once expressed) to the mRNA molecule encoding it can be attached.

Advantageously, the attachment of library members to the solid supports by specific hybridisation as described in step (b) of the method may optionally be followed by a subsequent step in which the attachment is stabilised by covalent attachment by for example enzymatic or chemical means. The method of the invention allows the display of large numbers of distinct peptide molecules in association with the nucleic acid molecules which encode them on discrete solid supports or areas of solid supports. This method overcomes the disadvantages associated with previous in vivo and in vi tro expression systems. The libraries generated by the methods of the invention can be generated in vi tro, thus overcoming all the above discussed disadvantages with in vivo systems . In addition the use of solid supports facilitates easy manipulation and handling and enables an estimation of library quality by allowing direct sequencing to determine the complexity of the library and the quantitation of protein production (i.e. the efficiency of translation) by for example using fluorescent control ligands and FACS analysis as discussed further below.

Moreover the screening process can be automated and has the properties of a homogeneous assay (i.e. an assay where all the steps can be performed and all the components added without washing or a need to remove components from the assay) . This is a major advantage of the invention over the prior art in vi tro and in vivo techniques as it allows the possibility of obtaining the ratio of bound target ligand to total expressed library member per individual solid support/area of solid support directly (by for example FACS analysis) , thereby giving an immediate indication as to the affinity between the particular library member and the target ligand.

Furthermore, compared to other display libraries the invention offers several advantages related to screening efficiency. First, there is no need to break the interaction between the selected library member (s) and the target ligand. Hence, this system ensures that even very high-affinity candidates are recovered. In addition, in contrast to, for instance, phage display where the target ligand is immobilised while the library members (phage particles) are mobile in a liquid phase, the present invention "immobilises" the library members (on beads or other solid supports) while the target ligand is in solution. In consequence, one would obtain a reduced background of low-affinity or non-specific binders that are selected in phage display (and other display techniques where the target ligands are immobilised) due to avidity effects. Further, as the system is completely cell-free, there is no counter selection against toxic proteins (which may kill or disrupt the function of an in vivo host) or proteins that are expressed poorly in vivo . Moreover, and importantly since the expression library members are obtained on a solid phase it is possible to screen such libraries against another library in solution. It would be extremely difficult, if not impossible, to effectively screen two solution based libraries against each other. Thus, the libraries produced by the methods of the present invention are advantageously compatible with library versus library screening.

Although the step of preparing the library of nucleic acid molecules may be carried out using in vivo systems, this need not be the case and the library itself may be generated in vi tro. In any event, the remaining steps of the method, including the steps of attaching the library members or a subpopulation thereof to the solid supports and the expression (translation) of the library members are carried out in vi tro . The screening of the library is also carried out in vi tro . This has many advantages as discussed herein.

The term "library" as used herein refers to a composition comprising a collection of molecular entities. The molecular entities making up said libraries can be nucleic acid molecules or peptide molecules or combinations of the two and thus the term "library" as used herein includes a composition comprising a collection of more than one, i.e. a plurality of different nucleic acid molecules, or a plurality of different peptides and includes libraries where said peptides are present associated with their encoding nucleic acids. The libraries of the present invention thus contain combinations of nucleic acid molecules and peptide molecules. Each different library member is present in one or more copies. Preferably multiple copies of each library member are present in the library as a whole.

The term "peptide" is used herein to relate to oligo- and polypeptides including proteins, protein fragments, etc. The term "display peptide" is used to refer to the portion of the variable peptide component of the library members which is utilized in library screening, e.g. it is screened for its ability to bind to a target ligand or the presence of a desired catalytic or enzymatic activity. The display peptide may if desired form part of a multipeptide complex in which the other peptide components are for example display peptides or nucleic acid binding peptides; however it is preferred that the display peptide is part of a single peptide which also includes a nucleic acid binding moiety.

The term "coupled to said encoding sequence" or "capable of coupling to the nucleic acid molecule which encodes it" as used herein in the context of an amino acid sequence encoded by the library members is intended to indicate that the amino acid sequence (once produced from its encoding nucleic acid by transcription and translation) will bind to/associate with a site (a binding site) in its own encoding sequence. This binding/association is thus caused by a protein: nucleic acid interaction and the amino acid sequence responsible for this is sometimes referred to herein as the nucleic acid binding moiety. Thus, an appropriate nucleic acid binding moiety for use in the present invention includes any nucleic acid binding protein or functionally equivalent fragment, derivative or variant thereof which associates with its encoding genetic material to form an operative link. As used herein, the "encoding" nucleic acid or nucleotide sequence is intended to refer to a nucleic acid molecule which, when translated, produces a translation product which contains the peptide to be displayed in the expression library, and the nucleic acid binding moiety. The binding site with which the binding moiety interacts is also contained within the encoding nucleic acid sequence. The location of the binding site with which the binding moiety interacts in the encoding nucleic acid can vary in location, e.g. it can be within the region encoding the peptide to be displayed and the nucleic acid binding moiety, or can be positioned at either side of this region either directly adjacent or at a distance therefrom. The main requirement is that the binding site which binds the nucleic acid binding moiety and the nucleic acid encoding the display moiety and the nucleic acid binding moiety are located on the same nucleic acid molecule.

The terms "coupled to", "associates", "associated", "interacts", etc. as used herein in connection with the interaction between molecular entities, for example a protein: nucleic acid interaction or a protein: chemical moiety interaction, refers to any association or interaction which is of sufficient strength to be maintained, i.e. remain functionally intact under the conditions of the various treatments to which the particular library member in which the association is found is subjected to during the preparation and screening of the library. Such interactions include non-covalent and, more preferably, covalent chemical bonding. A covalent linkage between for example the nucleic acid binding moiety and its binding site on the encoding nucleic acid has certain advantages in that the displayed peptide will not be released from the nucleic acid by ionic conditions and solvents that would disrupt bacteriophages or ribosomes. Furthermore, covalent attachment allows selection to be carried out at a wider range of temperatures, over longer periods of time and with intermediate freezing steps. Thus selection is much more convenient as well as potentially much more rigorous . The library of nucleic acid molecules to which the display peptides become coupled may comprise any type of nucleic acid molecules, e.g. DNA (e.g. genomic DNA or cDNA) or RNA (e.g. mRNA) or a mixture thereof. If said library comprises DNA molecules then these are generally present in a double stranded form. In this case the display peptide may become coupled to either strand of the double stranded molecule .

Proteins which interact with the nucleic acid sequence that encodes them are known in nature and are referred to herein as "cis-acting proteins". Generally and preferably the interaction of the cis-acting protein with its encoding nucleic acid is by way of covalent linkage. If the encoding nucleic acid is double stranded (e.g. DNA) then the protein can become coupled to either strand.

The same amino acid sequence which couples to its encoding nucleic acid is generally used for each member of a particular expression library. Thus preferred cis- acting proteins for use in the invention are those which only interact with the binding site on their own encoding sequence and not with the equivalent binding site in other library members. This is important as it facilitates manipulation of the library to be carried out in vi tro since the encoded cis-acting proteins can find their encoding nucleic acid despite the presence of other nucleic acids exhibiting equivalent binding sites. Such "true" cis-acting proteins are known in nature and examples include those which are involved in initiating replication. It should of course be borne in mind that no biological system is 100% accurate. Thus, cis-acting molecules that are not 100% cis-acting, but show a preferred cis-activity or a strongly preferred cis- activity or are essentially "true" cis-acting proteins will also work in the methods claimed herein and are included within the scope of the term "cis-acting proteins" .

Rolling circle type of replication is commonly used among circular replicons of different origins, for example single-stranded (ss) and double-stranded (ds) DNA phages (Van Mansfield et al . (1984), Adv. Exp. Med. Biol., 179, p221-230; Baas & Jansz (1988), Cur. Topics

Microbiol . Immunol., 136, p31-70) , ssDNA plasmids (Gruss & Ehrlich (1989), Microbiol. Rev., 53, p231-241; Novick (1989), Ann. Rev. Microbiol., 43, p537-565) , ssDNA plant viruses (Stenger et al . (1991), PNAS, 88, p8029-8033; Saunders et al . (1991), Nucl . Acids Res., 19, p2325- 2330) , ss and ds DNA animal viruses (Berns (1990) , Microbiol. Rev., 54, p316-329; Dasgupta et al . (1992), J. Mol. Biol., 228, pl-6) and ds DNA bacterial plasmids (Kha , 1997, Microbiol. Molec. Biol. Rev., 61(4), p445- 455) . In the systems studied, the initiation proteins possess a nicking-closing and topoisomerase-like activity. One of the best studied systems is that of the ssDNA phage φX174, where the A protein nicks the ori site in the viral strand of the replicative form and forms a covalent link to the 5¹ end of the cleaved strand. The 3' end is thereafter extended by the host polymerase displacing the 5 ' viral strand and after one round of replication the parental viral strand is religated and the A protein is transferred to the progeny strand to initiate a new round of replication

(Baas & Jansz, 1988, supra) . The phage P2 protein A has also been found to cleave the ori site in the coding region of the A gene at a site which is devoid of secondary structure and bind to the 5 ' end of the cleaved strand (Liu & Haggard-Ljungquist (1994), Nucl. Acids Res., 22, p5204-5210) . The description of constructs containing such cis-acting sequences is described in W098/37186, the teaching of which is incorporated herein by reference.

Thus in a preferred aspect the present invention provides a method as defined above wherein the amino acid sequence which is capable of coupling to the nucleic acid which encodes it (i.e. the nucleic acid binding moiety) includes or is derived from a cis-acting protein or a functionally equivalent fragment, derivative or variant thereof. Appropriate cis-acting proteins for use in the present invention include the family of replication proteins including the phage P2 A protein, which are related by sequence (preferably exhibiting at least 60% sequence identity, more preferably at least 70, 80 or 90% identity) , organisation and mode of replication; such as equivalent proteins from phage 186 (Sivaprasad et al., 1990, J. Mol. Biol., 213, p449-463), HP1 (Esposito et al . , 1996, Nucl. Acids Res., 24, p2360- 2368), PSP3 (Bullas et al . , 1991, Virology, 185, p918- 921) or the A protein of φX174 as mentioned above and functionally-equivalent fragments, derivatives and variants thereof.

As used herein, "functionally-equivalent" fragments, derivatives and variants include peptides related to or derived from a native protein as defined herein (e.g. a cis-acting protein), wherein the amino acid sequence has been modified by single or multiple amino acid substitutions, additions and/or deletions, which may alternatively or additionally include the substitution with or addition of amino acids which have been chemically modified, e.g. by deglycosylation or glycosylation, but which fragments, derivatives or variants nevertheless retain the desired functionality, e.g. exhibit cis nucleic acid-binding properties. Conveniently, such derivatives or variants may have at least 60, 70, 80 or 90% sequence identity to the native protein from which they are derived. Functionally-equivalent variants also include natural biological variations (e.g. allelic variants or geographical variations within a species) , or functionally equivalent variants isolated from a different species of organism and prepared using known techniques (for example, by screening nucleic acid libraries from different organisms with a probe based on the sequence of the native protein and isolating the nucleic acids encoding the variants) . In general, functionally-equivalent peptides or proteins may be prepared either by chemical synthesis or in recombinant form using the known techniques of site-directed mutagenesis, random mutagenesis, enzymatic cleavage and/or ligation of nucleic acids. Preferably, the cis-acting proteins (i.e. the nucleic acid binding moieties) for use in methods of the invention are derived from the phage P2 DNA replication initiation system. The P2 A protein recognizes a defined initiator sequence located within the P2 A gene on the very same DNA molecule which codes for it ( cis- action) and specifically nicks the coding strand while forming a covalent bond with one of the free end bases at the nick site (Liu & Haggard-Ljungquist , 1994, supra) . The sequence of the P2 A gene has been reported and thus library constructs comprising all or part of the phage P2 A gene which will allow the association of display moieties with the genetic material encoding them as described above can be designed in a straightforward manner (Liu et al . (1993), J. Mol. Biol., 231, p361- 374) .

The library of nucleic acid molecules can be prepared by methods which are standard and well known in the art. Alternatively a known library of nucleic acid sequences which has already been developed can be used as the basis for the library of nucleic acid molecules used in the methods of the present invention. Generally the techniques used to prepare the library constructs will be based on known genetic engineering techniques. Each library construct and hence each member of the library of nucleic acid molecules has a number of components. The first of these is a nucleic acid, preferably DNA, encoding the actual peptide which is to be displayed in the expression library (i.e. encoding the display peptide) . These nucleic acid sequences and thus the peptides which they encode generally vary between different library members and provide the library diversity.

The display peptides encoded may be relatively short peptides, e.g. up to 100 residues, preferably 5 to 50 residues, more preferably 7 to 20 residues or may be peptides encoded for by larger nucleic acid fragments or whole genes, e.g. the nucleic acid molecules encoding the display peptides may be a cDNA library or an mRNA library generated from a particular cell type or types or may be a fractionated genomic DNA library. The nucleic acid sequences encoding the appropriate display peptides molecules can be generated by conventional methods and any known library or newly developed library of nucleic acid molecules may be used in this regard. Such diverse nucleic acid sequences encoding the display peptides can be generated by conventional methods including the cloning of natural or artificial protein repertoires e.g. antibodies derived from B cells, cDNA libraries or mRNA libraries from various cell types or peptide repertoires based on synthetic oligonucleotides . Preferred libraries of nucleic acid molecules for use in the present invention encode antibody molecules or antibody fragments in any appropriate format, e.g. may encode whole antibody molecules or antibody fragments such as single chain antibodies (e.g. scFv antibodies), Fv antibodies, Fab antibodies, Fab '2 fragments, diabodies, etc.

Once generated the nucleic acid molecules encoding for different library members can also be further diversified using standard techniques, for example by mutation involving the addition, deletion and/or substitution of one or more nucleotides in a controlled (e.g. site directed mutagenesis) or random manner, or by domain swapping, cassette mutagenesis, chain shuffling etc. Synthetic nucleotides may be used in the generation of the diverse nucleic acid sequences. Thus, all or part of the nucleic acids encoding the display peptides can be synthesised chemically or be derived from various organisms or cell types.

Each expression library member also contains a means of coupling the displayed peptide to the nucleic acid molecule which encodes it. As described above such "means of coupling" may be provided by any suitable non- nucleic acid based molecule, moiety or entity, for example a chemical molecule, moiety or entity (e.g. puromycin) . Preferably said "means of coupling" is provided by a nucleic acid sequence or region which encodes for the nucleic acid binding moiety (e.g. a cis- acting protein) as described above which enables the association of the display peptide with its encoding nucleic acid via the interaction of the amino acid sequence (e.g. the cis-acting protein) with its binding site on the encoding nucleic acid. An appropriate fragment encoding at least the region of the nucleic acid-binding moiety which is necessary to achieve association (and preferably covalent binding) with its encoding nucleic acid and the binding site for said nucleic acid binding moiety must be present in such library members. Thus, the nucleic acid encoding the nucleic acid binding moiety may be varied by for example the addition, deletion and/or substitution of one or more bases or regions providing that the resulting expressed peptide retains its functional activity, i.e. can still associate with a site on the nucleic acid encoding it .

The particular nucleic acid sequences which encode or provide a binding site for the nucleic acid binding moiety must also be provided at a suitable site in each such expression library member, although these sequences may be moved from their naturally occurring position or additional such sequences may be introduced. The requirement is simply that the location of the region which binds the nucleic acid binding moiety must be such that the nucleic acid binding moiety (and the display peptide) remains associated with/attached to the nucleic acid which encodes it once it is expressed. Again the nucleic acid sequence of the binding site for the nucleic acid binding moiety may be varied by for example the addition, deletion and/or substitution of one or more residues or regions providing that the resulting nucleotide sequence can still associate with the nucleic acid binding moiety.

The sequence encoding the display peptide can be located anywhere within the nucleic acid molecule encoding the library member providing that when both the nucleic acid binding moiety (e.g. a cis-acting protein) and the display peptide moiety are expressed, the display peptide moiety does not interfere with the binding of the nucleic acid binding moiety to its binding site and the display peptide moiety is itself expressed in a form available for library screening.

Thus, the sequence encoding the display peptide moiety may be inserted within, lie adjacent to or fall outside the region encoding the nucleic acid binding moiety provided that the nucleic acid binding moiety and display peptide moiety once expressed form part of the same expressed peptide or multipeptide complex, preferably the same expressed peptide. In addition, the expressed display peptide moiety (and nucleic acid binding moiety) must remain attached or associated with the nucleic acid which encodes them.

In the preferred embodiment of the invention where the nucleic acid binding moiety is the cis-acting protein P2A, at least the sequence TCGGA, for example in the sequence GCGCCTCGGAGTCCTGTCAA, should be included in the DNA encoding the library members. The sequence TCGGA is recognized by (i.e. is a binding site for) the DNA-binding moiety (i.e. the P2A protein) and forms a covalent bond therewith. Alternatively, a functionally- equivalent fragment, derivative or variant of this sequence which is recognized by the P2A protein can be included in the library members. In these embodiments conveniently the sequence encoding the display peptide moiety is located such that it is expressed in the C- terminal portion of the expressed library member.

As described above the natural reaction mechanism of a number of the cis-acting proteins (nucleic acid binding moieties) which can be used in accordance with the methods of the invention involves binding of the cis-acting protein to a region on the encoding nucleic acid and nicking the coding strand, after which the nucleic acid binding moiety (and the linked display peptide if this reaction takes place with a library construct of the present invention) becomes covalently attached to the 5 ' end of the nucleic acid molecule created during the nicking process.

Under these circumstances, depending on the construction of the nucleic acid constructs of the library and the placement of the display peptide sequences within them, the nucleic acid binding moiety may be covalently attached to a nucleic acid fragment which does not contain the library sequences (i.e. the display peptide sequences) due to cleavage of this fragment from the remainder of the nucleic acid molecule. Thus it should be ensured that the genetic material encoding the display moiety is carried on the part of the coding strand which becomes covalently attached to the nucleic acid binding moiety or at least remains associated with it. If nucleic acid is retained in a double-stranded form following translation and during selection (e.g. by maintaining the libraries under conditions which induce hybridisation) , then the template (or non-coding) strand of the double stranded nucleic acid molecule will ensure that both coding strands are associated with the nucleic acid-binding moiety. Alternatively the binding site for the nucleic acid binding moiety and the site encoding the display peptide sequences should be chosen such that the nucleic acid-binding moiety covalently attaches to the part of the coding strand encoding the display moiety. This may be achieved by insertion of the display peptide encoding region at or near the carboxyl encoding terminal side of the binding site for the nucleic acid binding moiety (wherein the latter may also be displaced from its natural position) .

In addition, when the invention is performed using certain cis-binding proteins such as P2A, or their functionally-equivalent fragments, derivatives or variants, whilst the nucleic acid binding moiety will associate covalently with the nucleic acid encoding it, this represents a kinetic intermediate and, if replication is occurring in the system, the nucleic acid binding moiety will religate the coding strand and detach from this strand transferring to a further coding sequence with an intact protein binding site.

Advantageously, replication may be avoided in in vi tro systems such as that of the present invention by performing the translation reaction in the absence of dNTPs or in the presence of other means of inhibiting the polymerase reaction. In the absence of replication conditions the nucleic acid binding moiety will remain associated with the nucleic acid encoding it allowing screening of the expression library to be performed. Methods where the translation reaction is performed in conditions not conducive to replication (e.g. in the absence of dNTPs or in the presence of other means of inhibiting the DNA polymerase reaction) form preferred embodiments of the invention. In cases using in vi tro systems where dNTPs are required to be present for some reason, a mutant/modified nucleic acid binding moiety can be used which remains covalently attached to its encoding nucleic acid. In the case of P2A for example, Y450F which comprises a substitution of the tyrosine at amino acid position 450 of the A protein with phenylalanine may be used.

Each expression library member also contains an oligonucleotide tag which is specific for a particular library member. Methods for generating sequences of sets of diverse oligonucleotide tags which can be incorporated into the expression libraries of the invention are described in the art. See for example WO96/41011 (Spectragen, Inc.) which describes the design and generation of a set of minimally cross-hybridising oligonucleotide tags wherein each oligonucleotide in the set varies from every other member of the same set by at least two nucleotides. The number of tags in any particular set is clearly dependent on the length of the nucleotide sequences making up the tags and the chosen limit set for the permissible degree of similarly (cross-hybridization) between the tags of the same set. Further diversity is however achieved by joining a number of tags of a set together, i.e. if each individual tag of a set is referred to as a subunit additional minimally cross-hybridizing tags can be generated by joining a number of subunits together and also by varying the order in which the subunits are joined. In this way a very large number of different tags can be produced and moreover, due to the selection of the tag sequences so that they are minimally cross hybridising the different tags should not show significant interactions with each other. As mentioned above, possible methods for generating appropriate oligonucleotide tags for use in the present invention are described for example in W096/41011 and the teaching of this document is incorporated herein by reference. Generally however the tags are synthesised either individually or preferably combinatorially from subunits of a set of oligonucleotide tags which is preferably a minimally cross hybridizing set. When synthesised combinatorially, an oligonucleotide tag preferably consists of a plurality of subunits, each subunit consisting of an oligonucleotide of 3 to 9 nucleotides in length wherein each subunit is selected from the same minimally cross-hybridizing set. In such embodiments, the number of oligonucleotide tags available depends on the number of subunits per tag, on the length of the individual subunits and on the degree of cross-hybridization allowed between the subunits making up the tags. Very large repertoires of tags can be produced, for example repertoires up to approximately 1 x 10⁸ tags have been described in the literature. Once designed oligonucleotide tags can be incorporated into the library members at an appropriate location using standard molecular biology techniques such that all or substantially all different library members have different tags attached.

Hence generally to prepare the members of the library of nucleic acid molecules described herein the first step is to obtain a repertoire of nucleic acid sequences which encode the display peptides of interest . This library pool is then associated with a means of coupling the displayed peptide to the nucleic acid molecule which encodes it. For example the library pool may be inserted using standard molecular biology techniques in frame into a genetic construct encoding an amino acid sequence which associates specifically with its encoding sequence (i.e. in frame with a nucleic acid sequence encoding a nucleic acid binding moiety) , said insertion being carried out in a way which does not impair the normal function of the nucleic acid binding moiety to recognise and associate with its encoding sequence. In preferred embodiments of the invention the library gene pool is inserted into a construct encoding the cis-acting P2 protein A gene in a way which does not impair protein A's normal function of recognising and attaching itself covalently to the coding strand.

This repertoire of nucleic acid molecules can then be "tagged" by incorporating the oligonucleotide tags as described above into the repertoire of library members such that all or substantially all different library members have different tags attached. As mentioned above the tags can be incorporated at an appropriate location using standard molecular biology techniques. For example a repertoire of oligonucleotide tags can be incorporated into a diverse population of library members by direct enzymatic ligation or by amplification, e.g. by PCR using primers containing the tag sequences. Alternatively tags can be conjugated to different library members by excising library members, e.g. in this case the constructs comprising the nucleic acid molecules encoding the nucleic acid binding moiety which associates specifically with its encoding sequence and the display peptides, from their existing vectors, and religating them into vectors containing a repertoire of tags (each vector molecule containing one tag) . In any event, regardless of the methods used and the order of method steps used to construct the final library of nucleic acid molecules, each member of the final library of nucleic acid molecules will comprise a nucleotide sequence encoding an amino acid sequence which is capable of coupling to the nucleic acid molecule which encodes it (or some other means of coupling the display peptide to the nucleic acid molecule which encodes it) , an oligonucleotide tag and a sequence encoding the peptide which is to be displayed. The nucleic acid sequences encoding the library members, i.e. nucleic acid molecules comprising a nucleic acid sequence encoding a nucleic acid binding moiety sequence capable of binding to a binding site for said nucleic acid binding moiety, a nucleic acid sequence encoding a display peptide sequence, and an oligonucleotide tag nucleic acid sequence, and their complexes with the peptides they express form further aspects of the invention.

In addition, a library comprising a plurality of different nucleic acid sequences as defined above and expression libraries encoded by said nucleic acid sequences, for example the expression libraries formed from said different nucleic acid sequences by the methods of the invention, provide yet further aspects of the invention.

The library constructs may optionally additionally contain other appropriate components, for example origins of replication, inducible or non-inducible promoters for initiating transcription, antibiotic resistance genes and markers, general tags or reporter molecules, primer binding sites to enable amplification of the constructs by e.g. PCR, or other desirable sequence elements. Appropriate sources and positioning of such additional components within the library constructs so that they perform their desired function would be well within the normal practice of a skilled person in the art.

The inclusion of markers or reporter molecules can be particularly useful in the expression libraries of the present invention. Such markers or reporter molecules may be directly or indirectly detectable and include for example radiolabels, fluorescent labels or labels which can be detected enzymatically. Other markers which can be used include one partner of a binding pair such as streptavidin:biotin. Such markers are typically general markers which are present in all the library constructs and can be used to detect the presence of nucleic acid molecules of the libraries. In addition, the strength of signal detected on a particular solid support or area of a solid support can also be used to quantitate the amount of a particular library member present on a particular solid support or area of a solid support. Such quantitation of the amount of a library member present on a particular solid support can prove extremely useful in determining the affinity of a target ligand for a library member as discussed below. Alternatively or additionally, if for example one or more of the nucleic acids making up the library molecules are labelled with different labels, information about the content, e.g. the sequence of the nucleic acid molecules can be obtained.

As discussed above each library member has an oligonucleotide tag such that different library members (i.e. library members encoding different display moieties) have different tags. In order to ensure that all or at least substantially all of the different library members have different tags, ideally a repertoire of tags substantially greater than the total number of library molecules is used. The term

"substantially all" as used herein in reference to the attachment of tags to library members is meant to reflect the statistical possibility that more than one different library member may become attached to the same tag, i.e. the situation where not all the different library members have different tags. Preferably substantially all means that greater than 80%, more preferably greater than 90% and most preferably greater than 95% of the different library members have unique tags attached.

"Substantially greater" as used herein includes the scenario where the number of different tags is greater than two times the number of library molecules. More preferably the complexity of the tag repertoire is at least 10 times and even more preferably at least 100 times that of the library molecules.

If the total number of molecules making up the library exceeds the complexity of the tag repertoire or the tag repertoire is not substantially greater than the total number of molecules making up the library, the potential problem of having more than one different library member having the same tag may be overcome by taking a sufficiently small sample (referred to above as a "sub-population") of tagged library members from the full library of tagged library members and taking this subpopulation through the methods of the invention as described above. Further details as to methods and strategies for dealing with these problems are again described in WO96/41011.

It is an important aspect of the invention that as well as designing and generating the oligonucleotide tags which are to be incorporated in the expression library members, oligonucleotides which are complementary to each of the tags are also generated. These complementary tags are sometimes referred to herein as "anti-tags" and can be generated in the same way as the oligonucleotide tags. Such anti-tags can be either single stranded or double stranded.

The term "oligonucleotide" as used herein includes linear oligomers of natural or modified monomers or linkages, including deoxyribonucleosides, ribonucleosides, peptide nucleic acids (PNAs) , and the like, capable of specifically binding to a target nucleotide sequence by way of a regular pattern of monomer-to-monomer interactions, such as Watson-Crick type of base pairing, base stacking, Hoogsteen or reverse Hoogsteen types of base pairing, or the like.

Usually the monomers making up the oligonucleotides are linked by phosphodiester bonds, however analogs of phosphodiester linkages including phosphorothioate, phosphorodithioate, phosphoranilidate, phosphoramidate, and the like can also be used to link all or some of the monomers in an oligonucleotide. It will be clear to a person skilled in the art if the use of an analog linkage is appropriate. For example some analogue linkages, e.g. phosphoroamidate linkages can result in oligonucleotides which are less susceptible to degradation and confer greater stability on duplexes in which they are incorporated. Usually oligonucleotides of the invention comprise the four natural nucleotides. However, they may also comprise non-natural nucleotide analogs. It is clear to those skilled in the art when oligonucleotides having natural or non-natural nucleotides may be employed, for example where processing by enzymes is involved usually oligonucleotides consisting of natural nucleotides are necessary and for oligonucleotides where increased stability is important some non-natural nucleotides may be used. Thus, due to the general increase in stability of oligonucleotides containing modified linkages and/or modified nucleotides sometimes the tags in the library members will comprise naturally occurring nucleotides that permit processing by enzymes and hence manipulation, whereas the corresponding anti-tags attached to the solid support may comprise non-natural nucleotide analogs, e.g. may be PNA molecules, that are more stable in either or both a single stranded and double stranded form. Oligonucleotide tags of the invention may range in length from 12 to 60 nucleotides or base pairs. Preferably, oligonucleotide tags range in length from 18 to 40 nucleotides or base pairs. More preferably, oligonucleotide tags range in length from 25 to 40 nucleotides or base pairs.

A "complementary tag" as used herein refers to an oligonucleotide sequence which can specifically hybridise with an oligonucleotide tag to form a perfectly matched duplex or triplex. In embodiments where specific hybridization results in a triplex, the oligonucleotide tag may be selected to be either double stranded or single stranded. Thus, where triplexes are formed, the term "complement" is meant to encompass either a double stranded complement of a single stranded oligonucleotide tag or a single stranded complement of a double stranded oligonucleotide tag. In the methods of the present invention the "anti- tags" are attached to one or more solid supports. The material of the solid support can be any that is suitable for the attachment of nucleic acid molecules. Such supports are well known in the art and include plastic, glass, cellulose, synthetic polymers (e.g. polystyrene and, especially, polystyrene-divinyl benzene copolymers) , silicon, silica gels and the like.

The solid supports may take any appropriate physical form for example the supports may be microparticles, beads, plates, slides, chips, membranes etc., but preferred supports for use in the present invention are particulate or plate like in nature. Particularly preferred supports are beads which may optionally be magnetic or at least magnetisable, for example polymer beads carrying superparamagnetic particles may be used. Such supports are well known and documented in the art and it is within the normal practice of a skilled person to select the most appropriate support for use in the methods of the invention. Particulate supports, e.g. beads are particularly preferred because they are easy to manipulate in vi tro and facilitate automation. For example if the beads are magnetic they can be separated from a sample by a magnetic field and then washed. Alternatively if the beads are labelled, for example with a fluorescent tag the beads can be separated by flow cytometry. Suitable labelled, unlabelled, and magnetic beads are available commercially from Dyno Specialty Polymers AS of Lillestrøm, Norway and Dynal Biotech ASA. Particular examples of magnetic beads which can be used in the methods described herein are the M-270 or M-280 beads from Dynal Biotech ASA, Norway. Particular examples of non-magnetic beads which can be used in the methods described herein are 5 μm glycidyl methacrylate microbeads (Bangs Laboratories, Carmel, IN) . The important point in the present methods is that each individual species of anti-tag is attached either to its own discrete solid support (for example if the solid support is a particulate form - e.g. beads - each bead has only one species of anti-tag attached) or its own discrete area of a solid support (for example if the solid support is in a planar form, e.g. a plate, chip or membrane or is in a two or three dimensional comb-like or brush-like form where each "tine" or "bristle" can have a different anti-tag attached to it) . Generally, a plurality of copies of the same species of anti-tag is attached to the same solid support or area thereof. The number of anti-tags attached to an individual solid support can be decided depending on the method concerned and the form of the support. However, in the case of beads, up to 100,000 copies of the same anti-tag can be attached to one bead. An appropriate size of bead (or indeed other type of solid support) will depend on the number of tags it is desired to attach to said support and could be readily determined by a skilled person. Particles/beads as small as 1 μm in diameter can be used, as can larger beads, e.g. up to approximately 500 μm in diameter. Particular preferred beads for use in this regard are 5 μm glycidyl methacrylate microbeads (Bangs Laboratories, Carmel, IN) or M-270 or M-280 magnetic beads (Dynal Biotech ASA, Oslo, Norway) , which are approximately 270 μm and 280 μm in diameter, respectively.

Anti-tags may be synthesised in si tu on the surface of the solid support such that populations of identical tags are produced in specific regions. That is, the surface of each support (in the case of a bead) , or of each region (in the case of an array) , is derivatised by only one type of anti-tag which has a particular sequence. The total population of such beads or regions contains a repertoire of anti-tags with distinct sequences, wherein said repertoire refers to the complementary sequences to a set of oligonucleotide tags discussed above. Methods of synthesising nucleic acid sequences on solid supports are well known and documented in the art. Such nucleic acid on solid supports may also be amplified enzymatically on the spot in order to achieve a desired level of molecules.

Alternatively but less preferably, the anti-tags may be synthesised separately and then attached to the solid supports. Methods of attaching nucleic acid molecules to solid supports are well known and described in the art. Generally, the 3' end of the anti-tag is attached to the solid support and the 5' end remains free.

Once the anti-tags have been attached to the solid phase (s) either by synthesis in si tu ox by attachment after synthesis, the library of nucleic acid molecules as described above can be "sorted" onto the solid supports. This is enabled by the hybridization of a particular oligonucleotide tag in a library member to its specific anti-tag located on a discrete solid support or area thereof. The library of nucleic acid molecules will generally have been amplified by a suitable method before this step occurs and each member of the library may thus be present in multiple copies, but each copy of the same library member will have the same oligonucleotide tag. Hence if the amplified library is brought into contact with the solid support (s) which have anti-tags attached as described above, each member of the library can be sorted to its own solid support or area of a solid support resulting in solid supports (or areas thereof) with multiple copies of the same library member attached. Appropriate methods for doing this are described in WO96/41011. Suitable methods for amplifying the libraries before attachment would be well known to a person skilled in the art and include in vivo amplification by inserting the library members into appropriate host organisms or cells e.g bacteria such as E. coli, viruses, bacteriophages, yeast, prokaryotic cells, eukaryotic cells, insect cells or fungal cells, and growing them. Alternatively the library members could be amplified in vi tro using methods such as PCR with appropriate primers . As mentioned above tags interact with their appropriate anti-tags by hybridization thereto. Oligonucleotide tags may be single stranded and be designed for specific hybridization to single stranded anti-tags by duplex formation involving Watson-Crick base pairing or for specific hybridization to double stranded tag complements by triplex formation involving Hoogsteen interactions. Oligonucleotide tags may also be double stranded and be designed for specific hybridization to single stranded tag complements by triplex formation.

In a preferred embodiment of the invention both the tags and anti-tags are in a single stranded form for hybridisation. In such embodiments the anti-tag can be attached to a bead in a single stranded form. However, due to the techniques used for the generation and amplification of the libraries it is likely that the oligonucleotide tags present within the library members will be in a double stranded form. If so the tags have to be made into a single stranded form before hybridisation is allowed to take place. Any appropriate method can be used to obtain tags in a single stranded form. However, a general method for exposing an oligonucleotide tag in a single stranded form after amplification involves digesting a library member (s) with the 3*→5' exonuclease activity of T4 DNA polymerase, or a like enzyme. When used in the presence of a single deoxynucleoside triphosphate in the reaction mixture, such a polymerase will cleave nucleotides from 3 ' recessed ends present on both strands of a double stranded library member until a complement of the single deoxynucleoside triphosphate is reached on the template strand. When such a nucleotide is reached the 3'→5' digestion effectively ceases, as the polymerase ' s extension activity adds nucleotides at a higher rate than the excision activity removes nucleotides. The fact that this "stripping" reaction can be controlled on a specific strand by the presence of a single deoxynucleoside triphosphate in the reaction mixture, in that the stripping of the tag terminates once a . complement of the single deoxynucleotide triphosphate is reached on the template strand, means that preferred tags for use in the invention are constructed with only three of the four possible nucleotides A, T, C or G. In these embodiments the presence of the fourth remaining nucleotide in the constructs designates the end of the tag and provides an appropriate substrate wherein the tag sequence can be "stripped" as described above thereby meaning that single stranded tags can be readily prepared for loading library members onto solid phase supports by hybridisation to anti-tags.

In such embodiments where the library of nucleic acid molecules is in a double stranded form, i.e. is double stranded DNA and a single stranded oligonucleotide tag thereon is exposed by the above- described stripping reaction, the coding strand of the DNA can either be the strand that hybridises to the anti-tag or can be the strand complementary to the strand that hybridises to the anti-tag. In other words either the coding strand or the non-coding (template) strand can hybridise to the anti-tag. Which strand is used to hybridise with the anti-tag has implications on which direction the RNA polymerase travels relative to the solid support during transcription as RNA polymerase travels along the non-coding strand in the 3 ' to 5 ' direction (or in other words follows the coding strand in the 5 ' to 3' direction) . Optionally, after hybridisation, any gap between the anti-tag and the hybridised template can be filled in with T4 DNA polymerase in the presence of all dNTPs, followed by a ligase reaction according to standard procedures well known in the art (see PNAS Vol. 97(4), 1665-1670, 2000). This procedure will result in covalent attachment and hence stabilisation of the attachment of the anti-tag to the expression library member.

Although the technology described herein in general is referred to as covalent display technology, it is actually in most instances semi-covalent , since it will rely on DNA hybridisation. The only time when the complex attached to the solid support is truly covalent is when the gap between the anti-tag and the hybridised conjugate is filled in with T4 polymerase, and then ligated with T4 ligase (as described above) . In addition, in order for the attachment to be truly covalent the same strand that is ligated must be used as the coding strand (to which the nucleic acid binding moiety covalently attaches) and the inserted display peptide should be located on the C-terminal portion of the protein. In all other circumstances, the interaction between the bead and the display peptide would be dependent on DNA hybridisations.

Of course where hybridisation of double stranded tags to single stranded anti-tags involves triplex rather than duplex formation a stripping reaction such as that described above to expose the tag in a single stranded form is not necessary.

As the tags are used to attach the library constructs to the solid supports, the tag sequences are generally located towards the 5 ' or 3 ' ends of the library constructs so that attachment to the solid support can occur without interfering with the expression of the nucleic acid binding moiety (and its interaction with its binding site on the encoding sequence) or the expression of the display protein.

In order to facilitate attachment of the tagged library members (or a subpopulation thereof) to the anti-tags on the solid supports the supports and the library members are brought into contact under conditions that permit specific hybridization of the tags with their respective complementary tags (anti- tags) . Such conditions which permit hybridization are well known and documented in the art.

If the solid supports are particulate, such as beads, the supports and library members may simply be mixed together under the appropriate conditions. If the solid support is a planar structure, e.g. is a plate, column or a membrane, the library members can be passed over the surface of the support in an appropriate way and under appropriate conditions to promote hybridisation. Whatever the method by which the supports and library molecules (or subpopulation of library molecules) are brought into contact the library members become sorted onto particular beads or regions of the support so that each support or region of the support has one kind of library member attached.

The populations of library members attached to the supports can then be manipulated on the solid support. This manipulation can take a number of forms and steps. For example once the library members have been allowed to attach, the supports can be subjected to one or more washing steps to remove unbound library members. In addition, the fact that the libraries are attached to a solid support facilitates an estimation of library quality using direct in si tu sequencing to determine the complexity of the nucleic acid molecules. Any method of direct sequencing in si tu on solid supports can be used in this regard (see e.g. that described in WO96/12014 and in Brenner et al . , PNAS 17: 1665-1670 (2000)) and preferably the sequencing of a number of members of the library is carried out in parallel.

The essential manipulation step however for the methods of the current invention is the translation of the library members attached to the solid supports. Such a translation step converts the solid support based nucleic acid library into a solid support based protein expression library and, moreover, due to the design of the library constructs as described above, i.e. the use of a nucleic acid sequence encoding a nucleic acid binding moiety or the use of another means of coupling the displayed peptide to the nucleic acid which encodes it in conjunction with the tag - anti-tag technology, the proteins displayed are associated on a particular solid support (e.g. a bead) or a region of a solid support with the nucleic acid sequence encoding them. The translation step is carried out in vi tro by inducing transcription and translation of the nucleic acid library members. Usually the transcription and translation steps will be coupled. Systems for inducing coupled transcription-translation are well known and described in the art and some are commercially available. The fact that the nucleic acid libraries are located on solid supports makes the in vi tro manipulations required for inducing expression (i.e transcription and translation) very easy in that the required reagents can simply be added to the solid supports at the appropriate times and under the appropriate conditions. Of course if the members of the library of nucleic acid molecules comprise RNA (e.g. mRNA) then no transcription step will be necessary and only translation will be required to get expression of the display peptides.

The fact that translation is carried out in vi tro is particularly advantageous as it allows the incorporation of many co- and post-translational modifications (which may be made chemically or enzymatically, during or after the translation step) , some of which were not previously possible when translation was performed in vivo . For example phosphorylation or sulphation, formation of disulphide bonds, glycosylation or isomerization may be performed. These reactions may be accomplished in vi tro by supplementing the extract with the enzyme responsible for the modification. Non-natural amino acids may also be introduced, by for example chemically charging a t- RNA or by modifying the amino acid on a charged t-RNA. After the expression of the display proteins has been induced a number of further manipulations can optionally be carried out. Firstly it will generally be desirable to carry out one or several washing steps in order to remove all soluble reactants and any unbound molecules which might otherwise reduce the efficiency of the subsequent interactions between the displayed proteins and other molecules (e.g. target ligands) .

The peptide or protein expression libraries produced as a result of the expression step form a further aspect of the invention.

Often one would wish to screen the expression library produced for molecules which interact with a particular target ligand. Thus, a further aspect of the present invention provides a method of identifying and/or purifying a library member exhibiting desired properties from a peptide expression library as defined hereinbefore, said method comprising the step of screening a library of the invention for molecules which display certain properties. A preferred aspect of the invention provides a method of identifying and/or purifying from a peptide expression library as defined hereinbefore a display peptide which is a specific binding partner for a target ligand or exhibits desired properties, said method comprising the steps of a) screening a library of the invention for display peptides which bind to a particular target ligand and b) identifying and/or purifying the relevant library member .

Such screening could for example be carried out by conventional techniques such as attaching the target ligand to an affinity column or a similar solid support and passing a bead protein library generated as described herein over the solid phase so that the beads with proteins that interacted with the target ligand would bind and the non-binding particles could be washed away. In addition, the target ligand and the library members might be bound to small particles (for example beads of the order of 0.1-5 μm in diameter) after which complexes formed between beads bound to target ligand and beads bound to an appropriate display peptide or peptides can be separated from beads which have not formed complexes by any suitable method, e.g. agglomeration, precipitation or centrifugation, etc. of bead complexes. One appropriate method of separation would involve the incorporation in the library members and target ligands of a tag or marker which has a characteristic signal that can be detected either directly or indirectly (e.g. fluorescent labels, etc.). Complexes formed between target ligand and library members can then be separated from non-complexed beads using flow cytometry, for example on the basis of the intensity of fluorescent signal (if the tags of the library members and target ligand are the same) or based on the presence or interaction of two different fluorescent signals (if the tags of the library members and target ligand are different) .

Another suitable screening method could involve the incorporation into the library members attached to beads of a member of a specific binding pair (e.g. streptavidin or biotin) at a position such that the binding of a target ligand to the display peptide of a library member will result in the binding site of the specific binding pair member being masked or otherwise disrupted. Screening could then be carried out by the addition of beads (or other suitable solid supports) to which the other member of the specific binding pair was attached. Beads which have not bound to the target ligand become complexed with the solid support to which the other member of the specific binding pair is attached. These complexed beads can then be separated from beads which have bound to the target ligand and therefore cannot become complexed to the solid support, by any suitable method, e.g. agglomeration, precipitation or centrifugation etc. of bead complexes. Alternatively, if the library is immobilised on a planar solid support the target ligand could be labelled with a detectable label such as a radiolabel, a fluorescent label or an enzymatically detectable label and brought into contact with the library. When the non-bound substances have been washed away, the library members interacting with the target ligand could then be identified using appropriate detection techniques. A preferred method of screening however, which is particularly suitable when the protein expression library of the invention is bead-based, is to label the target ligand with a fluorescent label, bring the target ligand into contact with the bead expression library so that it can bind to any appropriate library members, and then sort the beads using for example a FACS (fluorescence activated cell sorting) machine on the basis of the fluorescent signal . In such embodiments the sorting can either be carried out based on the presence or absence of the fluorescent signal, or can be more selective and can be carried out on the basis of the strength of the fluorescent signal, i.e. based on the amount of target ligand bound to the individual bead - display peptide complex. Sorting by a FACS machine has the advantage that it allows automated sorting and subsequent analysis of both high affinity and low affinity binding partners and up to 20 x 10⁶ individual beads can be sorted in this way with current instrumentation. In addition, the intensity of fluorescence (which can be monitored by FACS analysis) can give an immediate indication of the affinity between the display peptide and the ligand providing that the total amount of display peptide per individual bead is known and the ratio of bound ligand to total display peptide per individual bead can be determined. Determining the total display peptide per individual bead can easily be carried out by for example providing the library constructs with a second general tag as discussed above which can either be detected directly (for example if it is a fluorescent tag) or indirectly via interaction with a differently labelled control ligand.

Once the particular beads/areas of solid support of interest have been identified the genes/fragments encoding the candidate binding molecules can be recovered from the beads/solid supports and analysed, optionally after further amplification either in vivo or in vi tro using systems for example as described above. If amplification is to be carried out in vivo the particular genes/fragments must be inserted into an appropriate organism as part of a vehicle or vector which will allow in vivo amplification. However, if in vi tro amplification is carried out, for example using PCR, amplification can be carried out while the candidate molecules are still bound to the solid supports. Amplification in vi tro is thus preferred. Especially preferably a desired bead pool can be added directly to a PCR mix for amplification of the genes/fragments encoding the corresponding candidate binders. Alternatively, the sequence of the anti-tag can be used to identify the display peptide sequence. Although, as discussed above, appropriate amino acid sequences which couple to the nucleic acid which encodes them for use in the libraries of the present invention are derived from cis-acting proteins found in nature, for example the family of replication proteins including the phage P2A protein, the libraries and methods of the present invention may be used to identify additional and alternative amino acid sequences which are capable of coupling to the nucleic acid sequence which encodes them. This could be carried out for example by selecting a relatively short nucleic acid sequence (for example between 4 and 30 base pairs in length, preferably between 4 and 10 base pairs in length) which occurs rarely in nature to act as a binding site for the "coupling" amino acid sequence which is to be selected and using this nucleic acid sequence (which is preferably labelled, e.g. with a fluorescent label) as a target ligand to screen an expression library of the present invention wherein the display peptides of the library are candidate amino acid sequences which are screened for their ability to bind to the target nucleic acid sequence. Once an appropriate amino acid which can couple/bind to a known nucleic acid sequence has been selected, new expression libraries of the invention can be designed and prepared wherein the library members comprise the nucleotide sequence encoding the selected nucleic acid binding amino acid sequence to provide the nucleic acid binding sequence capable of coupling to the nucleic acid molecule encoding it, and the nucleotide sequence of the target ligand used in the above described selection process to provide the binding site/coupling site for said nucleic acid binding amino acid sequence. The nucleic acid sequences of such newly designed libraries would also comprise oligonucleotide tags and nucleic acid sequences encoding the peptides to be displayed in the expression libraries as described above. Once appropriate genes/nucleic acid fragments encoding candidate display peptides with particular properties have been identified, the gene pool encoding candidate peptides can be subjected to affinity maturation. Such affinity maturation (i.e. the identification of candidate binders with further improved properties) can be performed by carrying out any conventional form of mutagenesis, including but not limited to error-prone PCR, cassette mutagenesis, and fragment shuffling, prior to repetition of the screening cycle .

A yet further aspect of the invention provides the use of a protein expression library as described herein to identify one or more peptides which interact/bind specifically to one or more target molecules or to identify one or more peptides with a desired activity, e.g. a desired catalytic or enzymatic activity. Library screening will preferably be carried out by fluorescence activated sorting, e.g. as described in WO 99/35293 the contents of which are hereby incorporated by reference .

Once one or more binding partners have been identified using the methods of the invention these peptides or proteins (or the nucleic acid encoding them) can be isolated and purified. Thus a further aspect of the invention provides a method of identifying and/or purifying a library member exhibiting desired properties from a peptide expression library as defined hereinbefore, said method comprising at least the steps of a) screening a library of the invention and b) selecting and isolating the relevant library member.

When one or more protein or peptide candidates have been selected, identified and/or purified using the methods of the invention, these candidates, or a component, fragment, variant, or derivative thereof may be manufactured and if desired formulated with at least one pharmaceutically acceptable carrier or excipient . Thus, a yet further aspect of the invention provides a method of manufacturing a specific binding agent comprising the steps of identifying a specific binding partner for a target ligand or identifying a peptide which exhibits desired properties according to the methods of the invention as described above, manufacturing said identified binding partner or peptide, or a component, fragment, variant, or derivative thereof and optionally formulating said manufactured binding partner or peptide with at least one pharmaceutically acceptable carrier or excipient. Peptides (or components, fragments, variants, or derivatives thereof) , identified, manufactured or formulated in this way form yet further aspects of the invention.

Said variants or derivatives of a peptide or specific binding partner include peptoid equivalents, molecules with a non-peptidic synthetic backbone and peptides related to or derived from the original identified peptide wherein the amino acid sequence has been modified by single or multiple amino acid substitutions, additions and/or deletions which may alternatively or additionally include the substitution with or addition of amino acids which have been chemically modified, e.g. by deglycosylation or glycosylation. Conveniently, such derivatives or variants may have at least 60, 70, 80 or 90% sequence identity to the original peptide from which they are derived.

The design of the expression library system discussed herein means that it can be used for the screening of libraries against other libraries. This is extremely advantageous and the use of the expression libraries of the invention as described herein in library against library screening is a yet further aspect of the invention. The major design feature which makes library versus library screening possible is the fact that in the expression libraries of the present invention it is the display peptides (i.e. the candidate molecules of the library) and not the target ligands that are immobilised/attached to a solid phase. This means that it is in turn possible to introduce particular target ligands or alternatively an entire library of further target ligands into the system in solution (see for example the above discussion on screening the expression library using one or more labelled target ligands in solution) . This is in contrast with for example phage display libraries where the target ligand is immobilised while the display moieties/ candidate molecules (i.e. the phage particles) are mobile in a liquid phase.

Screening two bead-based libraries against each other using this system is possible, although it is likely that avidity problems will be encountered, as every bead is multivalent (although this potential disadvantage could perhaps be used to precipitate large complexes of reaction pairs) . However, using the system described herein a bead-based library can be exposed to a monovalent or reasonably monovalent library which is in solution, e.g. a phage display library, a non-bead based P2-A library or a ribosomal display library thereby enabling the isolation of pairs of binding partners from the two libraries. For instance an antibody library displayed on phage may be allowed to react with a complete c-DNA library translated into protein molecules on beads (using the methods of the present invention) , or vice versa, the antibody library may be a solid phase library of the present invention and may be screened against a protein library displayed on phage. In this way, binding partners to almost every different display peptide molecule could be obtained in one single screening and sorting cycle. This is extremely advantageous as it can be regarded as approximately the equivalent of running up to 20 million individual screening reactions in parallel.

Yet further aspects of the present invention provide kits comprising the nucleic acid sequences as defined above encoding the library members, i.e. said kits comprise nucleic acid molecules comprising sequences encoding a display peptide and a nucleic acid binding moiety, at least one site of attachment for the binding moiety and an oligonucleotide tag. As an alternative to the nucleic acid binding moiety and site of attachment therefore any other means of coupling the display peptides to the nucleic acid molecule which encodes it (e.g. as described above) may be included in the nucleic acid molecules comprising the kits of the invention.

Such kits can be used for preparing an expression library of the invention according to the methods as defined herein. Optional further components of said kits include a set of anti-tag sequences that are complementary to the set of tags incorporated into the nucleic acid sequences of the kit. Said anti-tag sequences can be provided in solution or preferably attached to discrete solid supports or regions thereof. Kits of the invention may optionally further contain reagents for inducing expression of the nucleic acid molecules of the library, for example reagents necessary for inducing in vitro transcription and translation. Kits of the invention may also contain one or more of the following components: labels, detection reagents for said labels (including optionally reference standards for such detection) and solid phases, e.g. beads or particles .

Claims

1. A method of producing a peptide expression library which displays a plurality of different peptides, wherein displayed peptides coupled to the genetic material encoding them, said method comprising the following steps:

(a) preparing a library of nucleic acid molecules wherein each member of the library comprises a nucleotide sequence encoding a nucleic acid binding amino acid sequence, an oligonucleotide tag and a nucleotide sequence encoding the peptide which is to be displayed said nucleic acid binding sequence being capable of coupling to the nucleic acid molecule which encodes it ;

(c) translating the library members.

2. The method of claim 1 wherein the amino acid sequence which is capable of coupling to the nucleic acid which encodes it includes or is derived from a cis- acting protein or a functionally equivalent fragment, derivative or variant thereof.

3. The method of claim 2 wherein said cis-acting protein is a phage replication protein.

4. The method of claim 3 wherein said cis-acting protein is the phage P2A protein.

5. The method of any one of claims 1 to 4 wherein said display peptides are antibody molecules or antibody fragments.

6. The method of any one of claims 1 to 5 wherein said support is particulate in nature.

7. A polynucleic acid comprising a nucleic acid sequence providing a peptide binding site, a nucleic acid sequence encoding a nucleic acid binding amino acid sequence capable of binding to said peptide binding site, a nucleic acid sequence encoding a display peptide sequence, and an oligonucleotide tag nucleic acid sequence .

8. The polynucleic acid molecule of claim 7 wherein said nucleic acid binding amino acid sequence and said display peptides are as defined in any one of claims 2 to 5.

9. A complex comprising a peptide and a polynucleic acid according to claim 7 or claim 8 wherein said peptide is expressed from said polynucleic acid.

10. A nucleic acid library comprising a plurality of polynucleic acid molecules as defined in claim 7 or claim 8.

11. A peptide library comprising a plurality of complexes as defined in claim 9.

12. A solid support having attached thereto a library as defined in claim 10 or claim 11.

13. A method of identifying and/or purifying a library member exhibiting desired properties, said method comprising the step of screening a library as defined in claim 11 for molecules which display certain properties.

14. A method of identifying and/or purifying a display peptide which is a specific binding partner for a target ligand or exhibits desired properties, said method comprising the steps of a) screening a library as defined in claim 11 for display peptides which bind to a particular target ligand or exhibit desired properties and b) identifying and/or purifying the relevant library member.

15. A method of manufacturing a specific binding agent comprising the steps of identifying a specific binding partner for a target ligand or identifying a peptide which exhibits desired properties according to the method of claims 13 or 14, manufacturing said identified binding partner or peptide, or a component, fragment, variant, or derivative thereof and optionally formulating said manufactured binding partner or peptide with at least one pharmaceutically acceptable carrier or excipient .

16. Peptides or binding partners identified, manufactured or formulated according to the method of claim 15.

17. The use of a peptide library as defined in claim 11 to identify one or more peptides which bind specifically to one or more target molecules or which exhibit desired properties.

18. The use of a peptide library as defined in claim 11 in library against library screening.

19. A kit comprising the polynucleic acid molecules or libraries as defined in any one of claims 7 to 10.

20. The kit of claim 19, wherein said polynucleic acid molecules are provided on a solid support .

21. The kit of claim 19 or claim 20 wherein said kit further comprises reagents for inducing expression of said polynucleic acid molecules or libraries.