WO1998048008A1

WO1998048008A1 - Methods for identifying nucleic acid molecules encoding (poly)peptides that interact with target molecules

Info

Publication number: WO1998048008A1
Application number: PCT/EP1998/002420
Authority: WO
Inventors: Andreas Plückthun; Jozef Hanes; Lutz Jermutus
Original assignee: Plueckthun Andreas; Jozef Hanes
Priority date: 1997-04-23
Filing date: 1998-04-23
Publication date: 1998-10-29
Also published as: DE69834032D1; ATE321849T1; US6348315B1; CA2289691C; JP4086325B2; DE69834032T2; EP0975748A1; US6589741B2; JP2001521395A; CA2289691A1; US20020115083A1; EP0975748B1

Abstract

The present invention relates to methods for identifying nucleic acid molecules encoding (poly)peptides that interact with target molecules. The method of the present invention is particularly characterized by an in vitro translation step under conditions that allow formation of polysomes in the presence of antisense oligonucleotides complementary to the tag-coding sequence of ssrA-RNA. The present invention further relates to kits that are useful for carrying out the method of the invention.

Description

METHODS FOR IDENTIFYING NUCLEIC ACID MOLECULES ENCODING (POLY)PEPTIDES THAT INTERACT WITH TARGET MOLECULES

The present invention relates to methods for identifying nucleic acid molecules encoding (poly)peptιdes that interact with target molecules The method of the present invention is particularly cnaracteπzed by an in vitro translation step under conditions that allow formation of polysomes in the presence of antisense oiigonucleotides complementary to the tag-coding sequence of ssrA-RNA The present invention further relates to kits that are useful for carrying out the method of the invention

Evolutionary methods may bring the refinement to protein engineering which is beyond the powers and accuracy of rational design today Evolution can be defined as a succession of "generations", cycles of genetic diversification followed by Darwinian selection for a desired phenotypic property In classic experiments, nucleic acids have been evolved for physical properties (Saffhill, R , Schπeider-Bernloehr, H , Orgel, L E & Spiegelman, S (1970) J Mol Biol. 51 , 531 -539) in vitro, and in this case, the substance conferring the phenotype was identical to the genetic material Oligonucleotide ligands, usually single stranded RNA, have been identified for many targets by SELEX (Gold, L , Polisky, B Uhlenbeck, O & Yams, M (1995) Annu Rev. Biochem 64, 763-797, Irvine, D , Tuerk, C & Gold, L (1991 ) J Mol Biol 222, 739-761 ), in which a synthetic DNA library is transcribed, the RNA selected for binding, reverse transcribed and amplified over several rounds Early experiments with proteins as the carrier of the phenotype, clearly of much broader applicability, had relied on living cells for effecting the coupling between gene and protein, either directly or via the production of phages or viruses (Phizicky, E M & Fields, S (1995) Microbiol Rev 59, 94-123) Since in this type of experiment the DNA library as the information carrier for encoded protein diversity has to be transformed or transfected into bacterial or eukaryotic cells, the available diversity was severely limited by the low efficiency of DNA uptake (Dower, W J & Cwirla, S E (1992) in Guide to Electroporation and Electrofusion, eds Chang, D C , Chassy, B M Saunders, J A & Sowers, A E (Academic Press, San Diego), pp 291-301 ) Furthermore, in each generation, the DNA library had to be first gated into a rep cable genetic package by which diversity was again decreased In addition, many promising variants would have to be selected against in the host environment Only very few studies (Yang, W P , Green, K , Pmz-Sweeney, S Bπones, A T , Burton, D R & Barbas 3rd , C F (1995) J Mol Biol 254, 392- 403) have carried protein optimization through more than one generation using methods such as phage display, since this requires repeated switching between in vitro diversification and in vivo screening — a laborious process

With the goal of circumventing or improving this process, a number of laboratories have designed novel systems that are based on the immediate vicinity and physical connection of mRNA and corresponding (poly)peptιdes during translation Thus, a series of studies have shown that specific mRNAs can be enriched by immunoprecipitation of polysomes (Schechter, I (1973) Proc Natl Acad Sci U S A 70, 2256-2260, Payvar, F & Schimke, R T (1979) Eur J Biochem 101, 271-282, Kraus, J P & Rosenberg, L E (1982) Proc Natl Acad Sci U S A 79, 4015-4019) Recently, Mattheakis and coworkers reported an affinity selection of a short peptide from a library using polysomes, in order to connect genotype and phenotype in vitro (Mattheakis, L C , Bhatt, R R & Dower, W J (1994) Proc Natl Acad Sci U S A 91 , 9022-9026, WO95/11922)

This system employs an in vitro translation system that is preferentially coupled to an in vitro transcription system The translation system allows the simultaneous isolation of mRNA and (poly)peptιde in a polysome complex after a suitable screening step for the (poly)peptιde Preferably, the (poly)peptιdes in Mattheakis' system are comprised of two components, one of which is the peptide to be screened and the second is a tether segment that binds to the mRNA Co- isolatioπ of mRNA and (poly)peptιde in the polysome complex can possibly be improved with the help of a translation stalling sequence even though the existence of such sequences is still unclear for E coll This sequence possibly enhances the overall stability of the polysome complex by decreasing the translation rate and thus allows for suitable conditions for the concomitant screening and isolation of (poly)peptιde and corresponding mRNA

Similar work has earlier been reported by Gold and colleagues (WO 93/03172) and Kawasaki and co-workers (WO 91/05058) Although the above-described systems have established a means of characterizing a nucleic acid via the identification of a protein encoded by said nucleic acid, there are practical limitations with respect to the efficiency of the πbosome displays of the nascent (poly)peptιde The technical problem underlying the present invention was therefore to increase the efficiency of (i) synthesis of a collection of stable RNA molecules and (n) translation of said RNA molecules, and thereby to achieve an increased efficiency of the use of polysomes in screening The solution to said technical problem is achieved by providing the embodiments characterized in the claims

Accordingly, the present invention relates to a method for identifying a nucleic acid molecule encoding a (poly)peptιde that interacts with a target molecule comprising the following steps

(a) translating a population of mRNA molecules devoid of stop codons in the correct reading frame in an in vitro translation system, said translation system either comprising antisense oligonucleotides complementary to the tag-coding sequence of ssrA-RNA or being free of ssrA-RNA, under conditions that allow the formation of polysomes, (b) bringing the polysomes so formed into contact with said target molecules under conditions that allow the interaction of the (poly)peptιdes encoded by said mRNA molecules and displayed by said polysomes with said target molecules,

(c) separating polysomes displaying (poly)peptιdes that interact with said target molecules from polysomes displaying no such (poly)peptιdes, and

(d) identifying the nucleic acid molecule encoding a (poly)peptιde displayed in a polysome that interacts with said target molecules

The term "(poly)peptιde" as used in the present invention relates both to peptides as well as to polypeptides Said (poly)peptιdes may either comprise a natural or a recombinantly engineered ammo acid sequence The latter alternative also includes fusion proteins

According to the present invention, the term "polysome" refers to a complex formed by at least one, preferably several ribosomes and mRNA during translation

The population of mRNA molecules may be of varying origin For example, it may be derived from a cDNA library In an alternative embodiment, it may be directly derived from cells or tissue Particularly advantageous is also the use of the present invention in mutagenized (poly)peptιdes to find improved variants Alternatively, synthetic protein or peptide libraries or antibody libraries can be used

The term "tag-coding sequence of ssrA-RNA" relates to a nucleic acid sequence encoding the ammo acid sequence AANDENYALAA This sequence has been described in Keiler et al , Science 221 (1996), 1990-1993

The antisense o gonucleotides comprised in the translation system employed in the method of the invention are of a suitable length to hybridize to the tag-coding sequence of ssrA-RNA and block translation thereof under conditions that allow the formation of polysomes

A translation system being free of ssrA-RNA can, for example, be derived from E coli strains lacking a functional ssrA gene such as X90 ssrA1 cat (Keiler et al , Science 221 (1996), 1990-1993), N2211 or NM101 (Tu et al , J Biol Chem 270 (1995), 9322-9326), W3110 ΔssrA (Komine et al , Proc Natl Acad Sci U S A 91 (1994), 9223-9227), K7823 or K6676 (Retallack and Friedman, Cell 83 (1995), 227-235)

Conditions that allow the interaction of the (poly)peptιdes encoded by the translated mRNA molecules and displayed by said polysomes with corresponding target molecules can, without undue burden, be established by the person skilled in the art Such conditions are, for example, derivable from the teachings of WO 95/11922, WO 93/03172 and WO 91/05058 or from the appended examples As is well known in the art, said conditions also rely on the screening procedure that is employed for detecting said interactions

The separation of the polysomes that display (poly)peptιdes which interact with the target molecules from polysomes displaying no such (poly)peptιdes can be effected according to known procedures Again, the separation technique employed may well depend on the screening system that is used A convenient method of separating the aforementioned polysomes is, for example, based on affinity chromatography wherein the target molecules are bound to the column material

The identification of the nucleic acid molecule encoding the selected (poly)peptιde can be achieved by any suitable means and is most conveniently achieved by sequencing the nucleic acid molecule, for example, by sequencing the mRNA or the DNA, after cloning into a vector For identification of the mRNA, it may be removed from the πbosome by treatment with EDTA or by acid elution followed by standard RNA purification using a kit (see Example 3) or by competitive elution using a soluble target molecule, followed by standard RNA purification using a kit

Preferably and most advantageously, steps (a) to (c) are carried out two or more times prior to the identification step (d) This measure results in the less ambiguous identification of the desired nucleic acid with the concomitant minimization of false positive polysomes and thus nucleic acids This embodiment for identifying the desired nucleic acid is particularly preferred, if several rounds of selection are necessary to isolate the specifically interacting (poly)peptιde- target molecule pair

In accordance with the present invention, it has surprisingly been found that by including antisense o gonucleotides complementary to the tag-coding sequence of ssrA-RNA results in a manifold increase in the efficiency of polysome display This result is all the more unexpected since the prior art referred to above had already tried by various routes and means to establish such optimal conditions

The present invention therefore provides a system for the phenotypic selection of target molecules such a ligands with (poly)peptιdes that are preferably complete, native protein molecules

In a preferred embodiment of the present invention, the mRNA molecules comprise a stem loop at their 3' end

In this embodiment of the invention, degradation of the mRNA by exonucleases is precluded to a significant extent

Most preferably, a spacer is fused to the reading frame of the (poly)peptιde to tether the emerging, folded (poly)peptιde to the putative (poly)peptιde channel of the πbosome Said spacer preferentially encodes 57 to 1 16 ammo acids Tethering of emerging (poly)peptιde to the (poly)peptιde channel of the nbosome is an additional advantageous means to enable co-selection of (poly)peptιde and corresponding mRNA since it might significantly slow down the dissociation of the polysome

In a further most preferred embodiment said stem loop region at the 3' end of the mRNA molecules encodes said spacer Thereby, the length of the total 3' region can be kept to a minimum, if both spacer (required at the protein level) and stem loop (required at the RNA level) can be encoded by the same DNA

A further preferred embodiment of the present invention relates to the above- recited method wherein said mRNA molecules comprise a stem loop structure at their 5' end

Like the stem loop structure at the 3' end, the stem loop at the 5' end of the mRNA serves the purpose of avoiding a successful exonuclease attack towards the mRNA It is particularly preferred that the mRNA comprises both a 5' and a 3' stem loop structure In this embodiment, the mRNA molecule would structurally resemble natural mRNAs

In an additional preferred embodiment of the method of the invention, said in vitro translation system is supplemented with inhibitors of πbonuclease

Preferably, said πbonuclease inhibitors are transition state analogs and most preferably they are vanadyl πbonuclease complexes

In accordance with the present invention, it was found that in particular vanadyl ribonuclease complexes may advantageously be used to further increase the efficiency of nbosome display This result is particularly surprising since said complexes at the same time partially inhibit protein synthesis

A further preferred embodiment of the invention relates to a method wherein the polysomes in steps (a) to (c) are stabilized by (a) the addition of magnesium salts, preferably magnesium acetate, after the formation of polysomes, and/or

(b) a means that forms a bridge between the mRNA and the corresponding (poly)peptιde, and/or

(c) a low temperature after the translation and/or during the screening process

The above recited means have been proven to further enhance the stability of the polysomes Thus, a 50 mM magnesium acetate concentration in the reaction buffer significantly stabilized the nbosome complexes against dissociation The term "low temperature" in the above context is intended to mean a temperature that allows a successful screening to take place Preferably said low temperature is in the range of 0 to 5°C

Preferably, the translation is carried out in a prokaryotic translation system Particularly preferred is an E coli based translation system such as the S-30 E coli translation system

Alternatively, the translation system may be carried out in a eukaryotic translation system

In a further preferred embodiment of the method of the present invention step (d) comprises

(da) reverse transcribing said mRNA,

(db) optionally amplifying the resulting cDNA,

(dc) optionally cloning the optionally amplified cDNA, and

(dd) determining the sequence of said cDNA

This embodiment for identifying the nucleic acid of interest is preferred, if the population of mRNA molecules is too large to identify the desired species in a single round Furthermore, it allows repeated and detailed testing of identified molecules, since the population of mRNA molecules becomes "immortalized" by cloning

Reverse transcription allows sequencing using the most convenient DNA sequencing technology developed by Maxam and Gilbert as well as by Sanger and colleagues (see, e.g , Sambrook et al , "Molecular Cloning, A Laboratory Manual", second edition 1989, CSH Press, Cold Spring Harbor). The amplification of cDNA, preferably by PCR, with or without subsequent cloning into a suitable vector, further significantly facilitates the identification of the desired nucleic acid molecule. In various cases, amplification of the nucleic acid molecule will be a prerequisite for allowing the investigator to subsequently identify said nucleic acid molecule

In an additional preferred embodiment of the method of the invention, DNA is transcribed into mRNA in the presence of a reducing agent, such as β- mercaptoethaπol and/or DTT, prior to step (a).

The inclusion of a reducing agent such as β-mercaptoethanol and/or DTT into the reaction buffer is known to increase the stability of DNA-polymerase. Accordingly a reducing agent in the buffer contributes to an increase in the yield of mRNA which, in turn, results in an overall improvement of the nbosome display

It is particularly preferred to remove said reducing agent after transcription and prior to step (a). This method of the invention is most preferred in cases where the (poly)peptide to be screened may comprise species that assume their native conformation by forming disulfide bridges. An example of such (poly)peptides are members of the immunoglobulin superfamily. By introducing this preferred embodiment, the present invention directly contradicts the prior art referred to above that suggests using a combined transcription/translation system. Furthermore, the present invention also relates to a method for identifying a nucleic acid molecule encoding a (poly)peptιde that interacts with a target molecule comprising the following steps

(a) transcribing a population of DNA molecules devoid of stop codons in the correct reading frame into the corresponding population of mRNA molecules in the presence of a reducing agent,

(b) removing said reducing agent from said population of mRNA molecules,

(c) translating said population of mRNA molecules in an in vitro translation system under conditions that allow the formation of polysomes,

(d) bringing the polysomes so formed into contact with said target molecules under conditions that allow the interaction of the (poly)peptιdes encoded by said mRNA molecules and displayed by said polysomes with said target molecules,

(e) separating polysomes displaying (poly)peptιdes that interact with said target molecules from polysomes displaying no such (poly)peptιdes, and

(f) identifying the nucleic acid molecule encoding a (poly)peptιde displayed in a polysome that interacts with said target molecules

In a preferred embodiment the reducing agent used in step (a) is β- mercaptoethanol and/or DTT

In a further preferred embodiment of the method of the present invention, the (poly)peptιdes comprise domains of the immunoglobulin superfamiiy, and preferably of the immunoglobulin family

For example, said (poly)peptιdes may comprise complete T cell receptor or antibody chains or parts thereof such as domains of antibodies, for example the V_H or V_L regions

It is particularly preferred that the (poly)peptιdes are single chain antibodies or fusion proteins comprising such single chain antibodies In the latter alternative, the fusion partner of said antibody chains preferably is a tag that is employed for tethering the nascent (poly)peptιde to the corresponding mRNA

The present invention also preferably relates to a method wherein the translation system is supplemented with at least one comound selected from the group consisting of protein disulfide isomerase, oxidized or reduced glutathione, E coli protein DsbA and molecular chaperones such as DnaK, DnaJ, GrpE, GroEL or

GroES

The above compounds, alone or in combination, may enhance the stability, solubility and/or native folding capacities of the nascent (poly)peptιde

The protein disulfide isomerase may be of bacterial or eukaryotic origin The compound/enzyme that is included into the system would be selected by the person skilled in the art according to the type of protein that is screened For example, if he screens a library comprising antibody domains he would, in accordance with the teachings of the present invention, include a eukaryotic protein disulfide isomerase As could be shown by the present invention (see the appended examples) the polysome display system is significantly improved by incorporating said enzyme into the translation reaction system

In a further preferred embodiment of the method of the present invention, nonspecific interactions between the polysomes and/or the polysomes and the target molecules and/or, optionally, the polysomes and the matrix on which the target molecules are immobilized, formed during the step of bringing the polysomes into contact with said target molecules are inhibited or reduced by the addition of a blocking compound

In a most preferred embodiment, said blocking compound is a polyanionic compound like hepaπn Hepaπn has been suggested to be included as RNase inhibitor (WO 91/05058), but it has surprisingly been found in accordance with the present invention that it additionally decreases non-specific binding It can be assumed that hepann as polyanionic compound competes with the polyanionic mRNA as part of the polysome complexes for non-specific binding sites, rendering the addition of polyanionic compounds such as hepann in polysome display a generally applicable method for decreasing non-specific binding

In another most preferred embodiment, said blocking compound is sterilized milk The addition of nonfat milk has already been suggested for polysome display (WO 95/11922) However, according to the present invention it has been found that no RNA could be isolated when milk was used during affinity selection Surprisingly, when sterilized milk was used, RNA isolation was again possible, and the amount of non-specific binding was substantially decreased

Furthermore, the present invention relates to a kit comprising

(a) antisense oligonucleotides complementary to the tag-coding sequence of ssrA-RNA,

(b) optionally a vector suitable for cloning nucleic acids encoding (poly)peptιdes to be screened,

(c) optionally, πbonuclease inhibitors, preferably transition state analogs, and most preferably vanadyl πbonucleoside complexes,

(d) optionally, at least one compound selected from the group consisting of a protein disulfide isomerase, oxidized or reduced glutathione, E coli DsbA, and molecular chaperones, and

(e) optionally oligonucleotides encoding 5' or 3' stem loops, spacers or terminators without stop codons

In a preferred embodiment the kit according to the invention may furthermore comprise

(f) S-30 translation extract,

(g) PCR components, (h) reverse transcπptase, (i) an RNA sequencing kit, (j) a DNA sequencing kit, either alone or in combination

Finally, the present invention relates to a kit comprising

(a) an in vitro cell-free translation extract free of ssrA-RNA,

(c) optionally, nbonuclease inhibitors, preferably transition state analogs, and most preferably vanadyl πbonucleoside complexes,

(f) PCR components,

(g) reverse transcnptase, (h) an RNA sequencing kit,

(i) a DNA sequencing kit, either alone or in combination

The kit of the present invention can conveniently be used to carry out the method of the present invention

The figures show

Fig. 1 Principle of in vitro nbosome display for screening native protein (scFv) libraries for ligand (antigen) binding 1 , a DNA scFv library is first amplified by PCR, whereby a T7 promoter, nbosome binding site and stem loops are introduced, and then transcribed to RNA 2, after purification, mRNA is translated in vitro in an E coli S-30 system in the presence of different factors enhancing the stability of πbosomal complexes and improving the folding of the scFv antibodies on the ribosomes Translation is stopped by cooling on ice, and the nbosome complexes are stabilized by increasing the magnesium concentration 3, the desired nbosome complexes are affinity selected from the translation mixture by binding of the native scFv to the immobilized antigen Unspecific nbosome complexes are removed by intensive washing 4, the bound nbosome complexes can then be dissociated by EDTA (4b), or whole complexes can be specifically eluted with antigen (4a) 5, RNA is isolated from the complexes 6, Isolated mRNA is reverse transcribed to cDNA, and cDNA is then amplified by PCR This DNA is then used for the next cycle of enrichment, and a portion can be analyzed by cloning and sequencing and/or by ELISA or RIA

Fig. 2 (A) Schematic drawing of the scFv construct used for nbosome display 77 denotes the T7 promoter, SD the nbosome binding site, F the 3 ammo acid long FLAG (Knappik, A & Pluckthun, A (1994) BioTechniques 17, 754-761 ) with N- terminal methiomne, V_L and V_H the variable domains of the scFv fragment, L the linker, spacer the part of protein construct connecting the folded scFv to the nbosome, 5'sl and 3'sl the stem loops on the 5'- and 3'-ends of the mRNA The arrow indicates the transcπptioπal start The strategy for the PCR amplification is shown

(B) The amount of the scFvhag mRNA bound to the polysome complexes is influenced by the secondary structure of its ends and the length of the spacer connecting the folded scFv to the nbosome Different constructs of scFvhag mRNA were used for one cycle of nbosome display (constructs d - c9) None of them contained a stop codon Each species was tested separately After affinity selection of the nbosome complexes from 100 μl of translation mixtures, mRNAs were isolated and analyzed by Northern hybridization The presence of the 5' stem loop is labeled (+) or (-), that of the 3' stem loop is labeled as (-) when absent, (I) when derived from the Ipp terminator or (t) when derived from the T3Te terminator The spacer length indicates the number of ammo acids following the scFv and connecting it to the nbosome, including the translated stem loop The lengths of RNA in kb is derived from RNA molecular weight marker III (Boehnnger Mannheim) The bar graph shows the quantified amounts of mRNA from fiuoroimager analysis

(C) Effect of additives on the amount of mRNA bound to the nbosome complexes The mRNA of the scFvhag construct c7 was used for one cycle of nbosome display Samples in Fig 2B lane c7 and Fig 2C lane 6 are identical When indicated, 3 5 mM antι-sstv4 o gonucleotide ON10 (5'-

TTAAGCTGCTAAAGCGTAGTTTTCGTCGTTTGCGACTA-3'), 35 mg/ml protein disulfide isomerase (PDI) or 1 mg/ml vanadyl nbonucleoside complexes (VRC) was included in the translation Translations were stopped by addition of Mg(Ac)₂ to the final concentration indicated (in mM) and by cooling on ice After affinity selection of the nbosome complexes from 100 μl of translation mixtures imRNAs were isolated and analyzed by Northern hybridization The bar graph shows the quantified amounts of mRNA from fiuoroimager analysis

Fig. 3 Anti-ssn antisense ohgonucleotide ON10 (Fig 2C) decreases the molecular mass of the longest protein species (arrows) In vitro translation was performed using ³⁵S-methιonιne and scFvhag c7 mRNA Reactions were carried out in the absence (-) or presence (+) of 3 5 mM ohgonucleotide ON10 An SDS- PAGE of translation products is shown

Fig. 4 (A) Enrichment of the scFvhag nbosome complexes from mixtures with scFvAL2 by nbosome display mRNA of scFvhag c5 was diluted 10⁸ times with mRNA of scFvAL2, and the mixture was used for nbosome display After affinity selection of scFvhag nbosome complexes, mRNA was isolated, reverse transcribed to cDNA using ohgonucleotide ON5, cDNA was amplified by PCR using oligonucleotides ON3 and ON7 and analyzed by agarose eiectrophoresis Lanes 1 - 5 are PCR products amplified after the 1 st to 5th cycles of the nbosome display, respectively M is the 1 kb DNA ladder (Gibco BRL) used as MW marker PCR products corresponding to scFvhag and scFvAL2 are labeled (B) Enrichment of either scFvhag c5 or scFvAL2 nbosome complexes by nbosome display as a function of immobilized antigen mRNAs of scFvhag and scFvAL2 were mixed in a 1 1 ratio and used for one cycle of nbosome display After affinity selection, mRNA was isolated, reverse transcribed, amplified by PCR and analyzed by agarose electrophoresis as in Fig 4A The same 1 1 mixture was affinity selected on immobilized transferrin (tr) as a control ampicillin-traπsfernn conjugate (amp), or hag-peptide-transferπn-conjugate [hag) PCR products corresponding to scFvhag and scFvAL2 are labeled

Fig. 5 (A) Analysis of RNA pool after the 3^rd round of polysome display in vitro translation of RNA pool and RIA assay of labelled polysome complexes in the presence of 2% milk (a) binding to BSA, (b) binding to BSA-GCN4(7P14P), (c) binding to BSA-GCN4(7P14P) on inhibition with 1 μM GCN4(7P14P), (d) binding to BSA-GCN4(7P14P) on inhibition with 1 μM hag peptide, (e) binding to BSA- GCN4(7P14P) on inhibition with 1 μM fluorescein

(B) Analysis of RNA pool after the 3^rd round of polysome display, in vitro translation of RNA pool and RIA assay of labelled polysome complexes in the absence of 2% milk (a) binding to BSA, (b) binding to BSA-GCN4(7P14P), (c) binding to BSA-GCN4(7P14P) on inhibition with 1 μM GCN4(7P14P), (d) binding to BSA-GCN4(7P14P) on inhibition with 1 μM hag peptide, (e) binding to BSA- GCN4(7P14P) on inhibition with 1 μM fluorescein

(C) Analysis of RNA pool after the 3^rd round of polysome display by RIA RNA pool after 3 rounds of affinity selection performed in the absence of milk (A), or in the presence of mιlk(B) in vitro translation of RNA pool and binding assay of labelled polysome complexes in the presence of 2% milk (a) binding to BSA-GCN4(7P14P), (b) binding to BSA-GCN4(7P14P) on inhibition with 1 μM GCN4(7P14P)

The examples illustrate the invention In the appended examples, it is shown that an scFv fragment of an antibody which requires its correctly assembled three-dimensional structure to bind the antigen (a hydrophihc peptide), can be enriched 10⁸-fold by nbosome display and its sequence "evolves" during the process

Example 1 : Construction of scFv antibody fragments

The scFvhag construct c7 was amplified by PCR in two steps from an fd phage displaying the scFv 17/9 (Schulze-Gahmen, U , Rini, J M & Wilson, I A (1993) J Mol Biol 234, 1098-1 1 18, Krebber, C , Spada, S , Desplancq, D & Pluckthun A (1995) FEBS Lett 377, 227-231 ) containing a (Gly4Ser)₃ linker, using in the first step oligonucleotides ON1 (5'-

TTTC C C GAATTGTG AGC GG ATAAC AATAG AAATAATTTTGTTTAACTTTAAGAA GGAGATATATCCATGGACT ACAAAGA-3'), which introduces a nbosome binding site (RBS), and ON2 (5'-TTTAAGCAG CTCGATAGCAGCAC-3') In the second step ON3 (5'-ATATATGTCGACGAAATTAATACGACTCACTATAG GGAGACCACAACGGTTTC CCGAATTGTG-3'), which introduces the T7 promoter and the 5'-loop, and ON4 (5'-AGACCCGTTTAGAGGCCCC AAGGGGTTATGGAATTCACCTTTAAGCAGCT C-3'), which introduces a modified Ipp terminator loop with stop codon removed, were used The spacer, fused C-terminally to the scFv, is derived from ammo acids 211 to 299 of gene III of filamentous phage M13mp19 (Krebber, C , Spada, S , Desplancq, D & Pluckthun, A (1995) FEBS Lett 377, 227-231 ), and the translated Ipp terminator adds another 24 ammo acids

The other constructs were prepared by PCR from c7 using the following primer pairs d , ON5 (5'-TCAGTAGCGACAGAATCAAG-3') and ON6 (5'-GAAATTAA TACGACTCACTATAGGGTTAACTTTAGAAGGAGGTATATCCATGGACTACAAA GA-3'), which introduces the T7 promoter and RBS without the 5'-stemloop, the spacer is derived from ammo acids 21 1 - 294 of gene III, c2, ON4 and ON6, the same spacer as c7, c3, ON6 and ON7 (5'-GGCCCACCCGTGAA 1 ύ

GGTGAGCCTCAGTAGCGACAG-3'), the spacer is derived from ammo acids 21 1 - 294 of gene III following 7 ammo acids of the translated 7^"3Te terminator, c4, ON3 and ON5, the same spacer as d , c5, ON3 and ON7, the same spacer as c3, c6, ON2 and ON3, the spacer is derived from ammo acids 21 1 - 299 of gene III following the first 4 ammo acids of the Ipp terminator, c8, ON3 and ON8 (5'- TTTAAGCAGCTCATCAAA ATCACC-3'), the spacer is derived from ammo acids 21 1 - 264 of gene III following the first 4 ammo acids of the Ipp terminator, c9, ON3 and ON4, the spacer is derived from ammo acids 21 1 to 264 of gene III following 24 ammo acids of the Ipp terminator, using construct c8 as a template The scFvAL2 construct (V_L-(Gly4Ser)₄- V_H) was amplified by PCR in two steps from plasmid pAK202 (Krebber, A , Bornhauser, S , Burmester, J , Honegger, A , Willuda, J , Bosshard, H R & Pluckthun, A (1997) J Immunol Meth 201 , 35- 55), using first oligonucleotides ON1 and ON7 and in the second step ON3 and ON7 The spacer is derived from ammo acids 240 to 294 of gene III of filamentous phage M13mp19 (Krebber, A , Bornhauser, S , Burmester, J , Honegger, A , Willuda, J , Bosshard, H R & Pluckthun, A (1997) J Immunol Meth 201 , 35-55), and the translated T3Te terminator adds another 7 ammo acids

Example 2: RT-PCR and in vitro transcription

Reverse transcription was performed using Superscript reverse transcπptase (Gibco BRL) according to the manufacturer's recommendation PCR was performed using Taq polymerase (Gibco BRL) in the presence of 5% DMSO (4 mm at 94°C, followed by 3 cycles of 30 sec at 94°C, 30 sec at 37°C, 2 mm at 72°C, followed by 10 similar cycles at 60°C instead of 37°C, 20 similar cycles at 60°C instead of 37°C with elongation at 72°C prolonged by 15 sec per cycle and finished by 10 mm at 72°C) PCR products were analyzed by agarose gel electrophoresis and purified from the gel and reamplified, if the amount and quality was not sufficient, or directly used for transcription without additional purification In vitro transcription was performed as described (Pokrovskaya, I D & Gurevιch, V V (1994) Anal Biochem 220, 420-423)

Example 3: Model system and quantification of yields of affinity selection

As a model system, we used single-chain Fv (scFv) fragments of antibodies (Huston, J S , Levmson, D , Mudgett-Hunter, M , Tai, M S , Novotny, J , Margolies, M N , Ridge, R J , Bruccolen, R E , Haber, E , Crea, R & Oppermanπ, H (1988) Proc Natl Acad Sci U S A 85, 5879-5883), in which the variable domain of the light chain (V_L) is connected via a flexible linker to the variable domain of the heavy chain (V_H) To tether the folded protein to the nbosome and not interfere with folding we fused a spacer to the C-terminus of the scFv fragment Since the antibody domains form disulfide bonds, and the RNA polymerase requires β-mercaptoethanol for maximal stability, the effect of performing transcription in a separate reaction was investigated Further, the conditions for oxidative protein folding during translation (Ryabova, L , Desplancq, D , Spirin, A & Pluckthun, A (1997) Nature Biotechnology, 15, 79-84) were optimized (see below)

Polypeptide release is an active process requiring three polypeptide release factors in E coli (Grentzmann, G , Brechemier-Baey, D , Heurgue-Hamard, V & Buckingham, R H (1995) J Biol Chem 270, 10595-10600, Tuite, M F & Stansfield, I (1994) Mol Biol Rep 19, 171 -181 , Tate, W P & Brown, C M (1992) Biochemistry 31 , 2443-2450) and one nbosome recycling factor which releases the mRNA (Janosi, L , Shimizu, I & Kaji, A (1994) Proc Natl Acad Sci U S A 91 , 4249-4253) The consequence of release factor binding is normally the hydrolysis of the peptidyl tRNA between the ribose and the last ammo acid by the peptidyl transferase center of the nbosome (Tate, W P & Brown, C M (1992) Biochemistry 31 , 2443-2450) Our system is devoid of stop codons, and thus a fraction of the polypeptide may not be hydrolyzed off the tRNA and remain attached to the nbosome and thus be available for affinity selection In vitro translation in an E coli S-30 system was performed according to Chen and Zubay (Chen, H Z & Zubay, G (1983) Methods Enzymol 101 , 674-690) with several modifications Translation was carried out for 10 mm at 37°C in a 1 10 μl reaction which contained the following components 50 mM Tns-Ac, pH 7 5, 30 mM NH4Ac, 12 3 mM Mg(Ac)2, 0 35 mM of each ammo acid, 2 mM ATP 0 5 mM GTP, 1 mM cAMP, 0 5 mg/ml E coli tRNA, 20 μg/ml fohnic acid, 100 mM KAc, 30 mM acetylphosphate (Ryabova, L A , Vmokurov, L M , Shekhotsova, E A , Alakhov, Y B & Spirm, A S (1995) Anal Biochem 226, 184-186), 1 5% PEG 8000, 33 μg/ml rifampicm, 1 mg/ml vanadyl nbonucleoside complexes (VRC), 23 μl of E coli MRE600 extract (Chen, H Z & Zubay, G (1983) Methods Enzymol 101 , 674-690) and 90 μg/ml of mRNA

Translation is stopped by cooling on ice, and the nbosome complexes are further stabilized against dissociation by 50 mM magnesium acetate (Holschuh, K & Gasseπ, H G (1982) J Biol Chem 257, 1987-1992) Chloramphenicol at 50 μM concentration, assumed to also induce nbosome stalling (Mattheakis, L C , Bhatt R R & Dower, W J (1994) Proc Natl Acad Sci U S A 91 , 9022-9026 Moazed, D & Noller, H F (1987) Nature 327, 389-394), did not improve the selection properties (data not shown) The whole complex consisting of synthesized protein, nbosome and mRNA was then bound to the affinity matrix washed and eluted with competing ligand, or the ribosomes were dissociated with EDTA (Fig 1 ) In particular, the samples were diluted four-fold with ice-cold washing buffer (50 mM Tns-Ac pH 7 5, 150 mM NaCI, 50 mM Mg(Ac)₂ and 0 1 % Tween-20) and centnfuged for 5 mm at 4°C at 10 000 x g to remove insoluble components Microtiter plates coated with hag-transferrin conjugate were prewashed with ice-cold washing buffer, and the supernatant from the centnfuged translation mixture was applied (200 μl per microtiter well) and the plate was gently shaken for 1 hour in the cold room After 5 washes with ice-cold washing buffer, the retained nbosome complexes were dissociated with ice-cold elution buffer (100 μl per well, 50 mM Tns-Ac, pH 7 5, 150 mM NaCI, 10 mM EDTA, 50 μg/ml of E coli tRNA) for 10 mm in the cold room, and released mRNA was recovered by ethanol precipitation or by isolation using the Rneasy kit (Qiagen) We did not find it necessary to preparatively isolate polysomes at any stage The efficiency of nbosome display was found to be two orders of magnitude lower when using a coupled in vitro transcription-translation (data not shown) We developed the nbosome display system in two stages, first by engineering the gene structure for in vitro transcription, translation and folding, then by optimizing the translation reaction itself Each test started with 10 μg of input mRNA, this corresponds to ~ 1 5 x 10¹³ molecules The input mRNA was subjected to a single round of affinity selection translation in vitro, capture of nbosome complexes on immobilized target hgaπd and release of mRNA The released mRNA was then quantified by Northern analysis

For Northern analysis, RNA electrophoresis and transfer to a Nytran membrane (Schleicher & Schuell) were carried out as described (Goda, S K & Mmton, N P (1995) Nucleic Acids Res 23, 3357-3358) with a Turboblotter (Schleicher & Schuell) Hybridization was performed for 12 hours at 60°C (Church, G M & Gilbert, W (1984) Proc Natl Acad Sci U S A 81 , 1991 -1995) Hybridization was carried out with the ohgonucleotide ON9 (5'-

ACATGGTAACTTTTTCACCAGCGGTAACGG-3'), which anneals to the V_L region of scFvhag mRNA This ohgonucleotide was labeled by 3' tailing with digoxigen - 1 1 -dUTP/dATP using the DIG Ohgonucleotide Tailing Kit (Boehnnger Mannheim) Washing conditions were as follows 2 x SSC 0 5% SDS for 5 mm at room temperature, 2 x SSC, 0 5% SDS for 2 x 30 mm at 60°C, 0 1 x SSC for 10 mm at room temperature The hybridized ohgonucleotide probe was detected using the DIG DNA Labeling and Detection Kit with the chemilumiπescent substrate CSPD (Boehnnger Mannheim) and exposure to X-ray film, or with the fluoπmetπc substrate Attophos (Boehnnger Mannheim) and analysis using a fiuoroimager (Molecular Dynamics)

Example 4: Effect of mRNA structure on yield

In designing the nbosome display system, we first engineered the flanking regions of the scFv gene (Fig 2A) The gene should be transcribed very efficiently from the PCR product, and its mRNA should be stable against nucleases For the 5' end, we used the T7 promoter and the natural T7-gene10 upstream region, which encodes a stemloop structure directly at the beginning of the mRNA (Studier, F W , Rosenberg, A H , Dunn, J J & Dubendorff, J W (1990) Methods Enzymol 185, 60-89) At the 3' end, we fused a spacer region of 57 - 116 ammo acids to the reading frame of the scFv to tether the emerging folded polypeptide to the putative polypeptide channel of the nbosome and to give it enough distance to not interfere with folding This spacer encodes at the RNA level a 3' stemloop, either the terminator of the E coli lipoprotem (Studier, F W , Rosenberg, A H , Dunn, J J & Dubendorff, J W (1990) Methods Enzymol 185, 60-89) [Ipp term) or the early terminator of phage T3 (Reynolds, R Bermudez-Cruz, R M & Chamberhn, M J (1992) J Mol Biol 224, 31-51 ) (T3Te), to increase the stability of mRNA against exonucleases In a direct comparison of flanking regions (Fig 2B) after one round of translation and selection, we obtained the best result for constructs possessing a 5'-loop (from T7-g10), a 3'-loop (from the Ipp terminator) and the longest spacer of 116 ammo acids The yield of mRNA after one round of nbosome display improved from less than 0 001 % of input mRNA to 0 015 % (15 times)

Example 5: Effect of translation conditions on yield

We then tested the effect of various compounds present during translation Nucleases were found to be efficiently inhibited by vanadyl πbonucleoside complexes (VRC), which should act as transition state analogs (Berger, S L (1987) Methods Enzymol 152, 227-234) VRC at 1 mg/ml maximized the yield of isolated mRNA from nbosome complexes after one round of affinity selection when it was present during translation (Fig 2C), even though protein synthesis was partially inhibited (data not shown), and omitting VRC led to a several-fold decreased efficiency of the nbosome display (Fig 2C) In contrast, Rnasin (Mattheakis, L C , Bhatt, R R & Dower, W J (1994) Proc Natl Acad Sci U S A 91, 9022-9026) had no effect on the efficiency of the system We did not find evidence for significant proteolytic degradation of the scFv synthesized under these conditions, since prolonged incubation of released product (up to 300 m ) did not alter the electrophoretic pattern (data not shown)

From a systematic study on the in vitro translation of soluble scFv fragments in the presence of molecular chaperones and disulfide forming catalysts (Ryabova, L , Desplancq, D , Spiπn, A & Pluckthun, A (1997) Nature Biotechnology, 15, 79- 84), we found that binding activity is obtained if and only if disulfide formation and rearrangement is allowed to take place during translation and folding The strong beneficial effect of protein disulfide isomerase (PDI) was verified for the nbosome display system, in which the protein is not released It can be seen that PDI improves the performance of the nbosome display for scFv fragments three-fold (Fig 2C), and thus catalyzes the formation and isomeπzation of disulfide bonds on the πbosome-bound protein

Recently, a peptide tagging system was discovered in E coli, whereby proteins translated from mRNA, devoid of a stop codon, are modified and released from the ribosomes by the addition of a C-termmal tag, encoded by sstv4-RNA, and thereby marked for degradation (Keiler, K C , Waller, P R & Sauer, R T (1996) Science 271 , 990-993) It is shown in the present invention that this degredation could be inhibited by an antisense ohgonucleotide complementary to the tag- coding sequence of ssrA-RUA It was further surprising to observe that such a possible inhibition had an effect on the nbosome display Indeed, a four-fold higher efficiency of nbosome display is visible in the presence of anti-ssrA ohgonucleotide (Fig 2C), and the MW of the longest protein product is decreased, presumably by preventing the attachment of the degradation tag (Fig 3) Combining PDI and the antι-ssrΛ ohgonucleotide led to a twelve-fold increased efficiency of the nbosome display system (Fig 2C) By the combination of proper mRNA secondary structure and various compounds present during translation we could increase the yield of mRNA after one round of affinity selection 200 times — from less than 0 001 % to 0 2% of input mRNA This number expresses the combined efficiency of covalent attachment to the nbosome, protein folding, hgand binding, nbosome capture and RNA release and amplification

Example 6: Specific enrichment of target mRNA through multiple rounds of selection

In a test of the optimized system, we investigated how well a mixture of two proteins can be enriched for function Two scFv antibody mRNAs, constructed identically according to Fig 2A, both possessing the 5' and the 3' T3Te loop, one encoding the anti-hemagglutiπin scFv 17/9 (Schulze-Gahmen, U , Rini, J M & Wilson I A (1993) J Mol Biol 234, 1098-1 1 18) (scFvhag), the other the aπti- beta-lactam antibody AL2 (Krebber, A , Bornhauser, S , Burmester, J , Honegger, A , Willuda, J , Bosshard, H R & Pluckthun, A (1996) J Immunol Meth , in press) (scFvAL2), were mixed at a ratio of 1 10⁸ Their PCR products differ slightly in length, because of differences mainly in the spacer length, and can thus easily be distinguished (Fig 4A) After 5 cycles according to Fig 1 , undergoing selection on immobilized hag-peptide, 90 % of the nbosome complexes contained scFvhag We can thus conclude that the enrichment is about 2 orders of magnitude per cycle under these conditions To verify that enrichment really occurred through affinity selection, we tested enrichment of a 1 1 mixture of both mRNAs on an irrelevant surface and saw no change from the input ratio (Fig 4B) Furthermore, from the identical 1 1 mRNA mixture either antibody could be enriched, depending on which antigen was immobilized (Fig 4B)

Example 7: Analysis of scFv antibody fragments after affinity selection

After the 5th round of selection, the PCR products were hgated into a vector, transformed into E coli and single clones were analyzed The experimental protocoll was as follows After the 5th round of nbosome display PCR products were cloned into the vector pTFT74 (Ge, L , Knappik, A , Pack, P , Freund, C & Pluckthun, A (1995) in Antibody Engineering, eds Borrebaeck, C A K (Oxford University Press, New York), In vitro coupled transcription-translation in the S-30 E coli system was performed using 50 μg/ml of plasmid DNA under similar conditions as described above with the following modifications A coupled transcription-translation was carried out for 30 mm at 37°C, and the reaction mixture was supplemented with 2000 U/ml of T7 RNA polymerase and 0 5 mM UTP and TTP The mixture contained 50 μCi/ml of 35S-methιoπιne and 0 35 mM of each ammo acid except methionine After translation, the reaction mixture was diluted four-fold with PBS and bound to immobilized hag-peptide in a microtiter well After 60 mm incubation with gentle shaking, microtiter wells were washed 5 times with PBST and bound radioactive protein was eluted with 0 1 M tnethylamine Eluted protein was quantified in a scintillation counter Of 20 clones sequenced, 18 had the scFvhag sequence, and 2 had the scFvAL2 sequence, demonstrating that the 10⁸-fold enrichment was successful Of the 18 scFvhag clones 13 gave ELISA signals within 43 - 102 % of wild-type, mhibitable by soluble hag peptide, 2 were reduced to 14 and 18 %, and 3 were significantly reduced in binding (less than 10 %), probably a result of errors introduced during the last round of PCR amplification (Table 1 ) Thus, the selective pressure to maintain antigen binding, executed by binding and elution from immobilized antigen is clearly operating, albeit in the context of an ongoing genetic diversification through PCR errors

The sequence analysis showed that the clones contained between 3 and 7 base changes, with 90% transitions and 10% transversions, in good agreement with the known error properties of Taq polymerase (Keohavong, P & Thilly, W G (1989) Proc Natl Acad Sci U S A 86, 9253-9257) Each clone has gone through a total of 165 cycles of PCR, interrupted by 5 phenotypic selections for binding, and an error rate of 3 3 x 10^'5 can be calculated under these conditions, where only mutations which have survived the selection are counted At the protein level, the selected clones carry between 0 and 4 exchanged ammo acids, distributed over V_L, V_H and the linker All mutations are independent of each other (Table 1 ) and give therefore an indication of a large range of neutral mutations compatible with function, and perhaps even of improvements the system has selected for. While the exact properties of the selected molecules requires further analysis, we note the selection of a proline at the beginning of the linker, which may facilitate the required turn formation (Tang, Y., Jiang, N., Parakh, C. & Hilvert, D. (1996) J. Biol. Chem. 271 , 15682-15686) and the selection of acidic residues, which are known to increase solubility (Dale, G. E., Broger, O, Langen, H., D'Arcy, A. & Stϋber, D. (1994) Protein Eng. 7, 933-939; Knappik, A. & Pluckthun, A. (1995) Protein Eng. 8, 81-89).

Example 8: Construction of polysome display library

We constructed a library of scFvs in the orientation VL-VH with the two domains being connected by a (Gly4Ser)4 linker. To introduce a spacer, a protein tethering an scFv to the ribosome and allowing to fold an scFv on it, we first ligated the library to the vector pAK200 (Krebber, A., Bornhauser, S., Burmester, J., Honegger, A., Willuda, J., Bosshard, H. R. & Pluckthun, A. (1997) J. Immunol. Meth. 201 , 35-55) which resulted on the protein level in scFvs fused to the C- terminal part of gene III of filamentous phage M13mp19. In the second step we introduced stem loops, which stabilize mRNA against RNases, and the other important features by PCR. The final library construct contained a T7 promoter, 5'-stem loop and Shine-Dalgamo sequence upstream of the scFv coding sequence and a spacer consisting of 129 bases of gene III filamentous phage M13mp19 (amino acids 250-293) following by 21 bases of translated early terminator of phage T3 (T3Te), introducing as well a 3'-stemloop, downstream of scFv.

Cloning of the library followed the methods described in literature (Krebber, A., Bornhauser, S., Burmester, J., Honegger, A., Willuda, J., Bosshard, H. R. & Pluckthun, A. (1997) J. Immunol. Meth. 201 , 35-55). In short, mRNA was extracted from spleen cells of 6 mice immunized with either GCN4(7P14P) peptide (RMKQLEPKVEELLPKNYHLENEVARLKKLVGER) coupled to KLH or biotinylated GCN4(7P14P) coupled to avidin, and transcribed to cDNA using random hexamer primers After VL and VH PCR amplification followed by assembly PCR, PCR products were directly diluted 3x in Sfi\ reaction buffer, digested and separated using agarose gel electrophoresis Sfι\ digested DNAs were extracted from agarose gel by Amicon spin columns, concentrated by isopropanol precipitation and dissolved in sterile water

Purified PCR products (150 ng of each) were hgated to Sfi\ site of pAK200 vector overnight at 16°C (molar ratio insert vector = 1 2)

Ligation of library to pAK200 vector was very efficient by agarose gel electrophoresis and restriction analysis we found that 100% library scFv DNA was hgated at least to one side of the vector and more than 50% was connected to spacer part of the plasmid

In order to introduce features necessary for polysome display ligation mixtures were amplified in two steps by PCR using in the first step oligonucleotides SDA (5'-AGACCACAACGGTTTCC-

CTCTAGAAATAATTTTGTTTAACTTTAAGAAGGAGATATATCCATGGACTACAA AGA-3'), which introduces a nbosome binding site (RBS), and ON7 (see Example 1 ), which introduces a translated T3Te terminator, and in the second step oligonucleotides T7B (5'-ATAC-

GAAATTAATACGACTCACTATAGGGAGACCACAACGG-3'), which introduces the T7 promoter and the 5'-loop, and ohgonucleotide ON7 PCR products were directly used for in vitro transcription without additional purification, and RNA was purified by LiCI precipitation (Pokrovskaya, I D & Gurevich, V V (1994) Anal Biochem 220, 420-423) RNAs from both mini-libraries were pooled in equal ratio and used for polysome display

Example 9: In vitro translation of a library of scFv antibody fragments

In vitro translation in an E coli S-30 system was performed as described in Example 6 with small modifications In short in vitro translation was carried out for 8 mm at 37°C in a 220 μl reaction which contained the following components 50 mM Tns-Ac, pH 7 5, 30 mM NH4Ac, 12 3 mM Mg(Ac)2, 0 35 mM of each ammo acid, 2 mM ATP, 0 5 mM GTP, 1 mM cAMP, 0 5 mg/ml E coΛ tRNA, 20 μg/ml folinic acid, 100 mM KAc, 30 mM acetylphosphate, 1 5% PEG 8000, 33 μg/ml rifampicm, 1 mg/ml vanadyl nbonucleoside complexes (VRC), 3 5 mM anti-ssrA ohgonucleotide, 0 3 μM PDI, 51 4 μl of E coli MRE600 extract and 90 μg/ml of mRNA

Example 10: Screening of scFv library

Affinity selection of polysome complexes and RNA isolation

In a first approach, affinity selection was performed as described in Example 6 with the improvement of using hepann during selection We observed that hepann decreases unspecific binding of polysome complexes to specific (GCN4(7P14P)-BSA) as well as to unspecific surface (milk or BSA)

The translation (see Example 9) was stopped by adding Mg(Ac)2 to a final concentration of 50 mM, and the translation mixture was cooled on ice The samples were diluted four-fold with ice-cold washing buffer (50 mM Tns-Ac pH

7.5, 150 mM NaCI, 50 mM Mg(Ac)2, 2.5 mg/ml hepann and 0 1 % Tween-20) and centnfuged for 5 mm at 4°C at 10 000 x g to remove insoluble components Microtiter plates coated with GCN4(7P14P)-BSA conjugate were prewashed with ice-cold washing buffer, and the supernatant from the centnfuged translation mixture was applied (200 μl per microtiter well) and the plate was gently shaken for 1 h in the cold room After 5 washes with ice-cold washing buffer without hepann, the retained polysome complexes were dissociated with ice-cold elution buffer (100 μl per well, 50 mM Tns-Ac, pH 7 5, 150 mM NaCI, 10 mM EDTA, 50 μg/ml of E. coli tRNA) for 10 mm in the cold room, and released mRNA was recovered by isolation using the RNeasy kit (Qiagen) and used for RT-PCR After in vitro transcription of PCR products, RNA was purified by LiCI precipitation (Pokrovskaya, I. D & Gurevich, V V (1994) Anal. Biochem. 220, 420-423) and used either for RIA analysis (see below) or for the next round of polysome display To further improve affinity selection, the use of competing proteins during selection was evaluated Using milk during affinity selection of polysome complexes to GCN4(7P14P)-peptιde resulted in no RNA isolated after a polysome display cycle probably due to milk's RNase activity Pure proteins (e g BSA, casein or transfernn) did not decrease unspecific binding Surprisingly, we observed that milk could be used during affinity selection of polysome complexes if it was sterilized Thus, it had no influence on RNA stability and substantially decreased unspecific binding to the surface

In this second approach, the conditions described above were used with the modification of having 2% sterilized milk in the buffer during affinity selection

Radioimmunoassay (RIA)

After each round of polysome display, RNA of the whole recovered pool was in vitro translated in the S-30 E coli system using similar conditions as described for the library above with the following modifications translation was carried out for 30 mm at 37°C, the reaction mixture contained 50 μCi/ml of 35S-methιonιne and 0 35 mM of each ammo acid except methionme, and anti-ssrA ohgonucleotide as well as PDI were absent After translation, the reaction mixture was diluted four fold with PBST/milk to a final concentration of 2% milk and bound to immobilized GCN4(7P14P) peptide in a microtiter well After 30 mm incubation with gentle shaking, microtiter wells were washed 5 times with PBST and bound radioactive protein was eluted with 0 1 M tnethylamine Eluted protein was quantified in a scintillation counter Inhibition RIA was carried out by premcubation of the translation mixture diluted to PBST/2% milk with the GCN4(7P14P) peptide at different concentrations at least 1 hour at room temperature prior to binding to immobilized antigen

ELISA of single clones

After the 3rd round of polysome display PCR products were cloned into the vector pTFT74 (Ge, L , Knappik, A , Pack, P , Freuπd, C , & Pluckthun, A (1995) Antibody Engineering, ed Borrebaeck, C A K (Oxford University Press, New York)) Plasmids of single clones were isolated, transcribed in vitro (3), RNA was purified by LiCI precipitation (3) and used for in vitro translation in the S-30 E coli system using similar conditions as described for the library above with the following modifications translation was carried out for 30 mm at 37°C, and anti- ssrA ohgonucleotide, VRC as well as PDI were absent After translation, the reaction mixture was diluted four fold with PBST/milk to a final concentration of 2% milk, with the mixture containing GCN4(7P14P) peptide at different concentrations, and pre-mcubated for 1 h at room temperature Binding to immobilized GCN4(7P14P) peptide in a microtiter well was carried out for 30 mm with gentle shaking, and bound scFv protein was detected using the monoclonal anti myc-tag antibody 9E10 and a polyclonal anti-mouse/peroxidase conjugate (Pierce)

Results

After the 3rd round of polysome display without addition of milk, we analysed the pool of binders by RIA (using 2% milk) for binding to the GCN4(7P14P) peptide and found that the pool binds and can almost completely be inhibited by 1 μM peptide No inhibition with hag peptide as well as with fluorescein was observed (Fig 5A) However, analysis of single clones of the pool by ELISA (using 2% milk) revealed that among 24 clones only 3 bound the GCN4(7P14P) peptide and could be inhibited with it (positive clones) The other 21 clones did not bind the target peptide and were probably binding unspecifically to the surface Further 2 polysome display rounds did not increase the RIA signal of the pool indicating that there was no further enrichment for binders Because we did not use milk during affinity selection of polysome complexes and it is known from phage display experiments that milk decreases unspecific binding of phages to surfaces we repeated RIA of pools as well as ELISA of single clones in the absence of milk RIA of the pool in the absence of a milk revealed that the pool contains unspecific binders as well (Fig 5B) Analysis of single clones showed that binding properties of 3 positive clones were not influenced by the presence of 2% milk during ELISA, however, unspecific binders bound to the surface in the absence of milk but did not bind at all in the presence of 2% milk.

Affinity selection was repeated with addition of 2% milk. After the 3rd round we again analysed the pool by RIA for binding to the GCN4(7P14P) peptide. In that case the pool gave about 4 times more bound protein than the pool obtained without milk and was nearly completely inhibited by 1 μM peptide (Fig.5C). Analysis of single clones out of this pool by ELISA revealed that approximately 75% of clones were positive.

Table 1 : Mutations in scFvhag fragments selected by nbosome display in vitro.

Clone Mutation V|_ Mutation VH Mutation relative rel. RIA Inhibition

No. Linker RIA signal (%) signal (+hag)

12 K45R P41 R 102 6 94

10 S30P 101 5 95

^*6 100 4 96

*2 96 5 95

7 G16D 89 5 95

3 Y49H 86 5 95

9 E55G, E105G A23V S5P 86 5 94

1 T20A G12E 80 3 96

18 N(27d)D D10G, T50N 76 5 93

4 V13A, K28R V12A 72 6 91

13 V58I Y79C 65 8 87

16 L83R 151V, K75E, 63 5 92 A1 13V

14 K30E D61 G S5P 43 6 85

17 F71 S 18 6 66

8 E17G, K18E, 14 6 56 K30E

5 L1 1 P G13D 9 5 51

1 1 S(27e)F, N90S V48I 9 7 22

15 G64D, S77A G15S 6 6 2

Mutations in the scFv of the antibody 17/9 (Schulze-Gahmen, U., Rini, J. M. & Wilson, I. A. (1993) J. Mol. Biol. 234, 1098-1118) are numbered according to Kabat, E. A., Wu, T. T., Perry, H. M., Gottesmann, K. S. & Foeller, C. (1991 ) in Sequences of Proteins of Immunological Interest, vol. I. (US Department of Health and Human Services, 5th ed. ), pp. 151 and 464, in the variable domains, and sequentially in the linker. Clones are sorted according to a relative RIA signal. Wild-type clones 2 and 6 are marked by an asterisk. RIAs were normalized to the same amount of full length protein and inhibited by 0.1 mM hag peptide (denoted as +hag), to verify binding specificity.

Claims

1. A method for identifying a nucleic acid molecule encoding a (poly)peptide that interacts with a target molecule comprising the following steps:

(a) translating a population of mRNA molecules devoid of stop codons in the correct reading frame in an in vitro translation system, said translation system either comprising antisense oligonucleotides complementary to the tag-coding sequence of ssrA-RNA or being free of ssrA-RNA, under conditions that allow the formation of polysomes;

(b) bringing the polysomes so formed into contact with said target molecules under conditions that allow the interaction of the (poly)peptides encoded by said mRNA molecules and displayed by said polysomes with said target molecules;

(c) separating polysomes displaying (poly)peptides that interact with said target molecules from polysomes displaying no such (poly)peptides; and

(d) identifying the nucleic acid molecule encoding a (poly)peptide displayed in a polysome that interacts with said target molecules.

2. The method according to claim 1 wherein said mRNA molecules comprise a stem loop at their 3' end.

3. The method according to claim 2 wherein a spacer encoding preferentially 57 to 116 amino acids is fused to the reading frame of the (poly)peptide to tether the emerging, folded (poly)peptide to the putative (poly)peptide channel of the nbosome.

4. The method according to claim 2 or 3 wherein said stem loop at the 3' end of said mRNA molecule encodes said spacer.

5. The method according to any one of claims 1 to 4 wherein said mRNA molecules comprise a stem loop structure at their 5' end.

6. The method according to any one of claims 1 to 4 wherein said in vitro translation system is supplemented with inhibitors of ribonuclease.

7. The method according to claim 6 wherein said inhibitors of ribonuclease are transition state analogs.

8. The method according to claim 7 wherein said transition state analogs are vanadyl ribonucleoside complexes.

9. The method according to any one of claims 1 to 7 wherein the polysomes in steps (a) to (c) are stabilized by

(a) the addition of magnesium salt, preferably magnesium acetate, after the formation of polysomes; and/or

(b) a means that forms a bridge between the mRNA and the corresponding (poly)peptide; and/or

(c) a low temperature after the translation and/or during the screening process.

10. The method according to any one of claims 1 to 9 wherein the translation system is a procaryotic translation system.

11. The method according to any one of claims 1 to 9 wherein the translation system is a eucaryotic translation system.

12. The method according to any one of claims 1 to 10 wherein step (d) comprises

(da) reverse transcribing said mRNA;

(db) optionally amplifying the resulting cDNA; (dc) optionally cloning the optionally amplified cDNA, and

(dd) determining the sequence of said cDNA

The method according to any one of claims 1 to 12 wherein, prior to step (a), DNA is transcribed into mRNA in the presence of a reducing agent

The method according to claim 13 wherein the reducing agent is removed prior to step (a)

A method for identifying a nucleic acid molecule encoding a (poly)peptιde that interacts with a target molecule comprising the following steps

(b) removing said reducing agent from said population of mRNA molecules,

The method of any one of claims 13 to 15 wherein the reducing agent is β-mercaptoethanol and/or DTT O

The method according to any one of claims 1 to 16 wherein the (poly)peptιdes comprise domains of the immunoglobulin superfamily, and preferably of the immunoglobulin family

The method according to claim 17 wherein the (poly)peptιdes are single chain antibodies

The method according to any one of claims 1 to 17 wherein the translation system is supplemented with at least one compound selected from the group consisting of protein disulfide isomerase, oxidized or reduced glutathione, E coli DsbA and molecular chaperones

The method according to any one of claims 1 to 19 wherein non-specific interactions between the polysomes and/or the polysomes and the target molecules and/or, optionally, the polysomes and the matrix on which the target molecules are immobilized, formed during the step of bringing the polysomes into contact with said target molecules are inhibited or reduced by the addition of a blocking compound

The method according to claim 20 wherein said blocking compound is hepann or sterilized milk

Kit comprising

(c) optionally, ribonuclease inhibitors, preferably transition state analogs, and most preferably vanadyl nbonucleoside complexes, (d) optionally, at least one compound selected from the group consisting of a protein disulfide isomerase, oxidized or reduced glutathione, E. coli DsbA, and molecular chaperones; and

(e) optionally oligonucleotides encoding 5' or 3' stem loops, nbosome binding sites, spacers and/or terminators without stop codons.

23. The kit according to claim 22 further comprising

(f) S-30 translation extract;

(g) PCR components; (h) reverse transcriptase; (i) an RNA sequencing kit;

(j) a DNA sequencing kit either alone or in combination.

24. Kit comprising

(a) an in vitro cell-free translation extract free of ssrA-RNA;

(b) optionally a vector suitable for cloning nucleic acids encoding (poly)peptides to be screened;

(c) optionally, ribonuclease inhibitors, preferably transition state analogs, -and most preferably vanadyl nbonucleoside complexes;

(d) optionally, with at least one compound selected from the group consisting of a protein disulfide isomerase, oxidized or reduced glutathione, E. coli DsbA, and molecular chaperones; and

25. The kit according to claim 24 further comprising

(f) PCR components;

(g) reverse transcriptase; (h) an RNA sequencing kit;

(i) a DNA sequencing kit either alone or in combination.