The invention furthermore relates to fusion proteins comprising two or more chitin-binding domains, to uses thereof, and to affinity chips and their application.

Because of its high affinity and selectivity, immunoaffinity chromatography (IAC; Cutler, 1996; Jack, 1996; Jones, 1995; Sii & Sedana, 1991) could be in principle of great value in the rapid purification of proteins to high degrees of purity.

The antibody could be used as an affinity matrix in a one-step purification of natural proteins or be used in conjunction with methods based on peptide tags for the isolation of recombinant proteins (Ford et al., 1991; Nilsson et al., 1997). Further, the antibody may be directed against the peptide tag itself, thereby using the same tag twice in an orthogonal manner, as each column will remove other impurities.

However, because of the large investment of time and effort and, thus, cost in generating monoclonal antibodies (mAb) and producing them at the scale required, IAC has only rarely been used in the purification of proteins on mg scales. As the

mAb will normally be covalently coupled to expensive chemically activated column material, the use of a fresh column for each experiment, even though desirable, would be prohibitive in terms of labor and cost.

Recently, progress in display technologies such as phage display (Smith & Scott, 1993; Winter et al., 1994), ribosome display (Hanes et al., 2000; Plückthun et al., 2000), the protein fragment complementation assay (PCA; Mössner et al., 2001) as well as the availability of antibody libraries (Knappik et al., 2000; Vaughan et al., 1996) as a source for virtually all antibodies has solved the problem of obtaining recombinant antibodies with the desired specificity. Therefore, IAC with antibody fragments, e. g. scFv fragments, which take advantage of modern library technologies and which can be easily expressed and purified in E. coli in large amounts, would be particularly attractive for the parallel purification of proteins for proteomics projects. However, no convenient method has been described for immobilizing the recombinant antibody fragments. Random crosslinking (Arnold- Schield et al., 2000; Berry et al., 1991; Berry et al., 1993; Molloy et al., 1995) to the columns may obstruct the binding site, and the smaller the fragment, the larger the chance that the most reactive group may be positioned by chance such that it interferes with antigen binding. The directed immobilization strategies of mAbs (Turkova, 1999) rely on their Fc parts, e. g. coupling via their carbohydrate moieties or coupling by first binding to protein A or protein G columns, followed by crosslinking. Evidently, these methods cannot be applied to recombinant antibody fragments.

There are several approaches for immobilizing recombinant antibody fragments reported in the literature. Most of them use chemically activated matrices

(Arnold-Schield et al., 2000; Berry et al., 1991; Berry et al., 1993; Molloy et al., 1995), with all the problems described above for mAbs. Streptavidin matrices (Kleymann et al., 1995; Weiss et al., 1994), which are rather expensive, as streptavidin itself has to be purified and chemically coupled to an activated matrix, have been used either with a tag which is biotinylated in vivo, but only 15% of the Fab fragment was biotinylated, leading to a low yield of immobilized molecules. If instead a weak-binding purely peptidic tag (Strep tag) is used (Kleyman et al., 1995), more functional molecules can be immobilized, but the whole complex of Fv fusion protein and antigen is eluted by biotin, thus requiring a second purification step, dramatically adding to the cost of using such columns on a large scale.

Thus, the technical problem of the present invention is to provide a general approach for quickly purifying a protein from a complex mixture in a one- step procedure.

The solution to this technical problem is achieved by providing the embodiments characterized in the claims. The technical approach of the present invention, an affinity chromatography method wherein the ligand of interest can be eluted from the ligand-antibody complex in a one-step procedure, is neither provided nor suggested by the prior art.

Thus, the present invention relates to a method of separating a ligand comprised within a mixture comprising the steps of: (a) immobilizing a fusion protein on a matrix, said fusion protein comprising (i) an antibody fragment moiety having a specificity for said ligand and (ii) an affinity domain moiety that attaches to the matrix to form a matrix-fusion protein complex ; (b) contacting said mixture with said matrix-fusion protein complex; (c) removing components of said mixture not binding

to the antibody fragment moiety of said matrix-fusion protein complex ; and (d) eluting said ligand from said matrix-fusion protein complex.

Accordingly, the present invention relates to the use of the fusion protein in the purification of a ligand. Preferably the present invention relates to the use of said fusion protein in the method of separating a ligand from a mixture according to the present invention. In a preferred embodiment of the present invention, the fusion protein is expressed in E. coli.

The term"mixture"refers to a situation wherein the ligand of interest is not present in pure form, but associated with one or more other molecules, compounds or impurities, for example in a solution or in solid form.

The term"antibody"is used as a synonym for immunoglobulin. Antibody fragments according to the present invention may be Fv (Skerra & Pluckthun, 1988), scFv (Bird et al., 1988; Huston et al., 1988), disulfide-linked Fv (Glockshuber et al., 1992; Brinkmann et al., 1993), Fab, (Fab') 2 fragments, single VH domains or other fragments well-know to the practitioner skilled in the art, which comprise at least one variable domain of an immunoglobulin or immunoglobulin fragment and have the ability to bind to a target. Particularly preferred is the scFv fragment format.

In the context of the present invention, the term"binding specificity" refers to the ability of an antibody or antibody fragment to discriminate between a <BR> <BR> ligand of interest and one or more control molecules. Thus, "binding specificity"is not an absolute, but relative parameter. One way of determining"binding specificity"is to compare binding of an antibody fragment to the ligand of interest with binding to a set of unrelated controls, such as BSA, KLH, milk powder, albumin <BR> <BR> and lysozyme. However, "binding specificity"may also relate to the ability of an

antibody or antibody fragment to discriminate between a ligand of interest and a closely related homologous molecule as control. The assays used for determining "binding specificity"may have different formats well-know to one of ordinary skill in the art. For example, conventional ELISA techniques may be employed, wherein an optical density read-out for binding to the ligand of interest, which is greater by more than the standard deviation than the read-out for binding to the control, may be considered to represent"binding specificity". Preferably, the optical density is at least two times, or more preferably at least five times, greater than the mean optical density of binding to the control. Other assays known in the art may also be used to determine specific binding.

In the context of the present invention, the choice of an appropriate "binding specificity"of the antibody fragment will depend on the complexity of the mixture being used. For separation of a given ligand from a mixture containing a whole variety of unrelated molecules, an antibody fragment may be used which exhibits"binding specificity"when compared to a set of unrelated control molecules.

When a separation of closely related molecules is required, the antibody fragment being used preferably exhibits"binding specificity"for the ligand of interest when compared to these closely related molecules. The successful use of an antibody fragment with appropriate"binding specificity"can be determined by analyzing the purity of the ligand being obtained, which should preferably exceed 80%, more preferably 90%, and most preferably 95%.

In the context of the present invention, the term"antibody fragment moiety"refers to a moiety, or part, of the fusion protein, which comprises at least a functional fragment of an antibody. A"functional fragment"refers to a fragment of

an antibody comprising at least a variable domain having the ability to bind to a corresponding binding partner. Such a corresponding binding partner may be called antigen. In the context of the present invention, the corresponding binding partner for the antibody fragment moiety may be the ligand, or it may be a part of the ligand..

In the context of the present invention, the term"ligand"describes a molecule of interest, at least part of it being a corresponding binding partner for the antibody fragment moiety. A ligand according to the invention can be a protein, wherein the antibody fragment moiety has specificity for that protein. As a further example, a ligand according to the present invention can be a protein comprising an affinity tag, such as a his-tag, wherein the antibody fragment moiety has specificity for the affinity tag. A his-tag is a peptidic tag comprising at least five consecutive histidine residues (Hochuli et al., 1988; Lindner et al., 1992).

In a preferred embodiment, the antibody fragment is a single chain Fv (scFv) fragment.

In another preferred embodiment, the antibody fragment is a miniantibody. In the context of the present invention, the term"mini-antibody" refers to multimeric antibody fragments, eg. dimers, comprising two or more antibody fragments, each fused to a self-associating domain (Pack & Pluckthun, 1992; Pack et al., 1993; Pack, 1994).

In the context of the present invention, the term"affinity domain moiety"refers to another moiety, or part, of the fusion protein, comprising a (poly) peptide/protein that is able to tightly interact with a matrix.

Examples for such an affinity domain moiety include, but are not limited to, maltose

binding protein (MBP), protein A, cellulose binding domain, a polypeptide including a His-Tag, or a sufficiently long portion of a polypeptide such that affinity binding characteristics of the affinity domain can be used to immoblize the fusion protein to the matrix via said affinity domain. Examples of affinity domains and suitable matrices that can be used as a solid support for the immobilization of a fusion protein are known to one of ordinary skill in the art and can be found, e. g. in various reviews occurring in the Journal of Biochemical and Biophysical Methods (J Biochem Biophys Methods 2001 Oct 30; 49, (1-3). The fusion protein may contain an affinity domain as a first moiety and, for example, an antibody fragment as a second moiety.

In this context, "tightly interact"means, that the affinity of said affinity domain moiety to the matrix, as described by its Ko, is < 106, preferably < 10-7, more preferably < 10-8, and most preferably < 10-9. Ko is measured in units of moles per liter. The smaller the value of the dissociation constant Ko, the stronger is the binding between two molecules, e. g. the affinity domain to the matrix.

In yet a further preferred embodiment, the affinity domain comprises one or more carbohydrate-binding domains. Thereby, one is able to bind such fusion proteins to, eg., cellulose or chitin and to immobilize the fusion proteins to matrices comprising such carbohydrates. Up to now, these methods have been used only to purify the fusion with the protein of interest on a carbohydrate column with subsequent cleavage of the partners (Chong et al., 1997; Shpigel et al., 1998), or simply to immobilize proteins (Ramirez et al., 1993; Shpigel et al., 1999).

Preferably, the carbohydrate-binding domain is a chitin-binding domain and wherein said matrix consists of immobilized chitin. The chitin-binding domain can be a modified variant of a wild type chitin-binding domain, wherein a free

cysteine is exchanged by an alanine or a serine. In a yet further preferred embodiment, the affinity domain of the present invention is part of a chitinase from Bacillus circulans WL-12.

In yet another preferred embodiment, the affinity domain moiety of the fusion protein comprises two or more chitin-binding domains. Most preferably the present invention relates to a fusion protein comprising 2 to 4 chitin-binding domains.

Preferably, the fusion protein is immobilized to said matrix directly out of a crude extract. In the context of the present invention, the term"crude extract" refers to an extract obtainable by whole cell lysis. Further preferred, the fusion protein is purified prior to immobilization to said matrix. In the context of the present invention, the term"matrix"refers to a stationary phase, such as a solid phase material, which provides the structural support and the affinity domain interaction sites for the immobilization of the fusion proteins.

In a preferred embodiment of the present invention, the elution of said ligand from said matrix-fusion protein complex is achieved by a pH shift. The"pH"is a measure of acidity and alkalinity of a solution that is a number on a scale on which a value of 7 represents neutrality and lower numbers indicate increasing acidity and higher numbers increasing alkalinity and on which each unit of change represents a tenfold change in acidity or alkalinity and that is the negative logarithm of the effective hydrogen-ion concentration or hydrogen-ion activity in gram equivalents per liter of the solution. pH shift, accordingly, refers to an experimental step, wherein a change of the pH is achieved by either the addition of solutions with a higher or

lower pH as compared to the conditions employed in the"mixture-matrix contacting" and"non-bound component removing" (i. e., washing) steps, so that the specific interaction between the antibody fragment and the ligand is destroyed, and the interaction of the affinity domain with the matrix remains intact.

The specific elution of the ligand from the matrix-fusion protein complex can be achieved under alkaline or acidic conditions. For alkaline lutions, the pH is preferably around 10. Most preferably the pH is at 10. For acidic lutions, the pH is preferably around 3.2 or around 2, depending on interaction between the ligand and the antibody fragment.

In another aspect, the present invention relates to an affinity chip for the identification of one or more ligands from a mixture. A chip may comprise a substrate and a set of immobilized fusion proteins, as described herein. Each fusion protein may contain an antibody fragment moiety out of a set of different antibody fragments with specificities for a set of different ligands. The other moiety of the fusion protein can be an affinity domain moiety. In this regard, the fusion protein is attached to the substrate of said chip via said affinity domain moiety.

In the context of the present invention, the term"affinity chip"refers to a substrate upon which at least one, and often a plurality of, probe chemicals such as oligonucleotides or antibody fragments, are adherent. The terms"affinity chip"and "biochip"are used as synonyms. A biochip is useful for analysis of sample fluids using a method according to the invention. Target components of the sample fluid that react with complementary probes on the biochip can thereby be detected; biochips with an array of probe chemicals thereupon allow simultaneous screening of samples for a variety of target components.

The possibility of specific elution of the ligand bound to the matrix-fusion protein complex does not only enable the recovery of the ligand according to the method of the present invention, but additionally allows the recovery of the matrix- fusion protein complex for further use. In the case of an affinity chip according to the present invention, this leads to the possibility of being able to reuse the chip after specific elution of the ligands, thus reducing the cost of using such biochips on a large scale.

Particularly preferred is an affinity chip having a set of fusion proteins, wherein each member of said set of fusion proteins is coupled to the chip at a spatially addressable position. By knowing the position of every specific fusion protein and its binding specificity, a simple read-out can be performed after having determined those positions on the affinity chip, where specific binding to the fusion protein have occurred.

In another aspect, the invention relates to a method for the parallel detection of one or more ligands in a mixture by using an affinity chip according to the present invention, comprising the steps of: a) contacting said mixture with said affinity chip, wherein said affinity chip comprises a set of fusion proteins, and wherein said each of said fusion proteins comprise an antibody fragment moiety; b) removing the components of said mixture not binding to the antibody fragment moiety of said fusion proteins; and c) identifying one or more ligands bound to said affinity chip.

The present invention relates to a fusion protein comprising two or more chitin- binding domains.

The present invention also relates to a nucleic acid sequence encoding a fusion protein according to the present invention. The term"nucleic acid molecules" refers to RNA or DNA molecules, either single stranded or double stranded. In another aspect, the present invention relates to a vector comprising a nucleic acid sequence according to the present invention.

Furthermore, the present invention relates to host cells comprising a nucleic acid sequence according to the present invention, or a vector according to the present invention. A host cell may be any of a number systems commonly used in the production of proteins, including but not limited to bacteria, such as E. coli (see. e. g. , Ge et al, 1995) or Bacillus subtilis (Wu et al., 1993); fungi, such as yeasts (Horwitz et al., 1988; Ridder et al., 1995) or filamentous fungus (Nyyssonen et al., 1993); plant cells (Hiatt, 1990, Hiatt & Ma, 1993; Whitelam et al., 1994); insect cells (Potter et al., 1993; Ward et al., 1995), or mammalian cells (Trill et al., 1995).

Although the present invention has been described for the use of antibody fragments for the separation of a ligand from a mixture, the invention is not limited to antibodies, but additionally can be used with any molecule with binding specificities for a ligand, provided that the interaction between such molecule and its ligand allows the specific elution of said ligand as described in the present invention.

Examples for molecules include, but are not limited to, receptors or functional fragments thereof, ankyrin-type repeat molecules, leucine-rich repeat (LRR) molecules or lipocalins.

The present invention can be further understood with reference to the following examples. These examples are illustrative and, hence, are not intended to limit the scope of the invention.

Example In the following example, all molecular biology experiments are performed according to standard protocols (Ausubel et al., 1999).

Construction and expression of CBD-fusion proteins Plasmid construction The plasmids for the periplasmic expression of scFv-CBD fusion proteins (Fig. 5) are based on the pAK-series (Krebber et al., 1997). The dHLX-part of pAK500 was replaced by a linker-CBD-fragment via EcoRI and HindIII. The linker-CBD-fragment was amplified out of the plasmid pTYB1 using the primers CBD5'_AvrII and CBD3'_NheI.

CBD5'Av/II : 5 CATCCGGAATTCGGCGGTGGCTCCGAAGGCGGTGGCAGCGAAGGTGGCGGCCTAGGCACC AC AAATCCTGGTG-3'contained the linker sequence (SGAEFGGGSEGGGSEGGGLG) including an EcoRI-and an Av4I-site. CBD3'_NheI : 5'- GTACCCAAGCTTAGCTAGCTTGAAGCTGCCACAAG-3'introduced a NheI-site and a HindIII-site. The EcoRI-site was used to clone a mIgG3hinge-dHLX fragment (derived from pACKdHLX (Pack & Pluckthun, 1992) for dimerization (Fig. 5). The Noel site allowed cloning of a second CBD fragment also amplified out of pTYBl, including the linker sequence. The primers CBD5'EcoRI : 5'-CCGGAATTCGCTAGCGGTGGCCTGACC-3' and CDB3'_NheI were used to introduce NheI-sites at both ends. The resulting plasmids were called pKB100_wt (one CBD), pKBIOOdHLX-Wt (dimeric miniantibody-two CBDs), pKB100 wtilwt (tandem CBDs) and pKBlOOdHLX_wtilwt (dimeric miniantibody-two tandem CBDs). The ribosome binding site (RBS) was exchanged for the stronger T7G10

RBS of pAK400 (Krebber et al., 1997) and the gene for coexpression of Skp was introduced (Bothmann & Pluckthun, 1998) resulting in the pKB200-series (Fig. 5). The exchange of cysteine at position 30 (amino acid 30 of the isolated CBD, according to the numbering of NEB's pTYB-vector series; amino acid 677 in whole chitinase Al, see PDB code 1ED7 ; Ikegami et al., 2000) to serine and to alanine was performed by site directed mutagenesis of all CBD fragments. The gene for the anti-his tag scFv fragment (mutl2) was cloned as SFII cassette replacing the tet-resistance cassette. The scFv fragments directed against FkpA (6B1, 7B2, 9B3 and 11B4), as well as the scFvs K14G2 (anti-gpD scFv) and N7A9 (anti-SHP scFv) were selected out of the HuCAL (Knappik et al., 2000) by two rounds of automated phage display (Krebs et al., 2001). Non symmetrical Sty-sites (Krebber et al., 1997) were attached to the scFv cassettes of the HuCAL series by PCR and cloned into a pKB200 derivative containing gIIIpss (Krebber et al., 1997) as stuffer, necessary to supply the EcoRI site. All HuCAL scFvs described here were introduced as MfeI/EcoRI fragments, thus removing the second Sf/l-site. The stuffer fragment was then exchanged for the CBDiICBD-cassette by EcoRI/HindIII cloning (Fig. 5).

Expression of scFv-CBD fusions The plasmids encoding the 12 different anti-his tag-CBD fusion proteins (Fig. 1 and 2; pKB2Hmutl2-series, see Fig. 5) were transformed in the E coli K12 strain SB536 (Bass et al., 1996) (F-, WG1, AfhuA (tonA), AhhoAB (SacII), shh). Small scale expressions were performed at 25 °C using 50 ml of SB-medium (20 g/t tryptone, 10 g/ ! yeast extract, 5 gui NaCI, 50 mM K2HPO4) containing 30 ug/ml chloramphenicol.

Cultures were inoculated from a 5 mi preculture to OD550=0. 1. Expression was induced with 1 mM IPTG at an OD550 between 1.0 and 1.5. Cells were harvested 3

hours after induction by centrifugation. Cell pellets were resuspended in MBS buffer (20 mM Mes/NaOH pH 6.5, 500 mM NaCl, 0.1 mM EDTA), normalized to their end OD550 using 2.5 ml of MBS per 1 unit OD550. Whole cell extracts were prepared by French Press lysis at 10, 000 psi and 1 ml of crude extract was centrifuged in an Eppendorf tube for 60 min at maximum speed and 4 °C. The supernatants containing the soluble material and the pellets were analyzed in an anti-FLAG blot. Large scale expressions of the scFv-CBD fusion proteins were carried out in culture volumes from 750 ml to 1 I in 5 I baffled shake flasks as described for the small cultures.

Expression of antigens The expression of the antigens FkpA (Ramm & Pluckthun, 2000), GpHD (Forrer & Jaussi, 1998), GpHDL-cCrk (Forrer & Jaussi, 1998) and SHP (Yang et al., 2000) was carried out exactly following published protocols. His-tagged proteins (scFv 4D5-his, GroES-his, PhoA-his, CS-his and GFP-his) were expressed in the E coli strain JM83 (F-, aral, A (lac-proAB), rpsL (StrR), thil, 80, A (/aZ) M15). Cultures (1 I) were grown at 25 °C in 5 I baffled shake flasks using dYT-medium (16 g/l tryptone, 10 g/l yeast extract, 5 g/l NaCI) and were induced at OD550 = 1 with 1 mM IPTG (Lindner et al., 1997). Cells were harvested 5 h after induction.

The chitin binding domain (CBD; Chong et al., 1997; Ikegami et al., 2000; Watanabe et al., 1994) of Bacillus circulans WL-12 chitinase AI (Swiss Prot. Nr. P20533) was used. Preliminary tests of a fusion consisting of one CBD and a mutated version of the anti-his tag scFv 3D5 (Lindner et al., 1997) showed bleeding from the column, suggesting that the binding of one CBD to chitin must be very weak. This observation is in contrast to expectations raised in the literature (Chong et al., 1997; Shpigel et al., 1998; Ramirez et al., 1993; Shpigel et al., 1999), but consistent with

quantitative determinations of the affinity of a cellulose binding domain from Trichoderma reesei to cellulose (Linder et al., 1996; Linder et al., 1998).

Consequently, stable binding to the beads, which is observed under many conditions, relies on the high molar concentration of chitin on the beads, and is dynamic in nature (Linder et al., 1998; Carrard et al., 2000).

Therefore two CBDs in tandem were fused, dimerized the scFv to a miniantibody CBD fusion or combined both strategies to generate a protein with 4 CBDs (Fig. 1).

While all of these constructs showed better binding (Fig. 2B-wt-constructs), periplasmic expression was decreased compared to the original scFv-CBD construct (Fig. 2A-wt-constructs). As the CBD used contains a single unpaired cysteine, the periplasmic production of the fusion proteins is impaired. Therefore the cysteine residue was replaced by serine or alanine, and indeed both mutations lead to much better expression of all constructs possessing two or four CBDs. Two constructs (scFv-C30S-C30S and scFv-C30A-C30A) showed about the same expression level than the simple scFv-CBD fusion.

Batch binding experiment Cell pellets of the anti-his tag fusion proteins were resuspended in TBST buffer (50 mM Tris/HCI pH 8,1 M NaCl, 0.1 mM EDTA, 1 % Triton X-100) using 1 ml of buffer per 1 g of cells. DNase I was added and cell disruption was achieved by French-Press lysis. After centrifugation the supernatant was passed through a 0.22 um filter. To compensate for differences in expression levels of the constructs, estimated from Western-Blot analysis, crude extracts were diluted with TBST buffer to different extents. For the binding experiment those diluted crude extracts, now containing the

same amount of fusion protein, were shaken with equilibrated chitin beads.

Equilibrated chitin beads were prepared by using 100 NI of beads in an ethanol suspension and washing them 3 times with 1 ml of TBST buffer in a 2 ml tube. To these beads 500 ul of diluted crude E. coli extract containing the scFv-CBD fusion proteins was added. The mixture was shaken at 4 °C for 1 h. The beads were washed 3 times with 1 ml of TBST buffer. An aliquot was taken and analyzed by SDS-PAGE to ensure that the same amount of each fusion protein had bound to the beads. The dissociation from the beads was determined by shaking the beads in 200 u elution buffer (50 mM Caps/NaOH pH 10,500 mM NaCl, 0.1 mM EDTA) for 1 h at 4 °C. The beads were sedimented and the supernatant containing dissociated molecules was analyzed by an anti-FLAG blot.

SDS-PAGE analyses were carried out under reducing conditions according to standard protocols using 12 % and 15 % polyacrylamide gels. Western-Blots with the monoclonal anti-FLAG antibody M1 were carried out as described (Ge et al., 1995; Knappik & Pluckthun, 1994).

For their use as affinity ligands stable binding of the fusion proteins to the column material is decisive. Therefore the dissociation of all constructs from the chitin beads was investigated under elution conditions (used for the anti-his tag scFv fragment).

The same amount of fusion protein was bound to the beads for all constructs. After shaking the beads for one hour under elution conditions, the supernatant containing the dissociated molecules was analyzed by an anti-FLAG blot (Fig. 2B). The results suggest that both mutations result in weaker binding of the single CBDs, but there is almost no dissociation detectable for the scFv-CBD-CBD and the miniAb-CBD-CBD fusion proteins. Even though miniantibody-CBD fusions contain two CBDs as do the scFv-CBD-CBD constructs, they showed more leakage from the chitin beads. This

may be due to some dissociation of the miniantibody over time. The scFv-C30A-C30A and the scFv-C30S-C30S fusion proteins show the highest expression level and sufficiently stable binding to the column material. Based on homology to other non- cysteine containing CBDs (Ikegami et al., 2000) the scFv-C30A-C30A construct was used as affinity ligand for further experiments.

To demonstrate the general applicability of the method, scFv-C30A-C30A fusion proteins of six further scFv fragments were constructed. All of them were selected out of the Human Combinatorial Antibody Library (HuCAL ; Knappik et al. 2000), a na'fve fully synthetic scFv-library with a diversity of 2 109, by two rounds of phage display (Krebs et al., 2001). Four scFv fragments with different frameworks for VH and VL as well as different CDR3s were selected as recognizing the E coli protein FkpA (Bothmann & Pluckthun, 2000; Ramm & Pluckthun, 2000). Using similar procedures, an scFv fragment specifically recognizing either the B-phage coat protein gpD (Forrer & Jaussi, 1998; Yang et al., 2000) or the homologous protein SHP from phage 21 were selected.

Affinity purification Preparation of crude extracts for chromatography experiments Cell pellets expressing the fusion proteins (scFv-C30A-C30A constructs) used for chromatography experiments were resuspended in TBST buffer using 5 ml of buffer per 1 g of cells. After addition of DNase I, cell disruption was achieved by French- Press lysis. The suspension was clarified by centrifugation at maximum speed for 60 min and 4 °C and filtration through a 0.22 um filter. The same procedure was performed with antigen cell pellets, but using MBS buffer (20 mM Mes/NaOH pH 6.5,

500 mM NaCI, 0.1 mM EDTA).

Column chromatography Chromatography experiments were performed using SPE columns, usually mounted in a positive pressure manifold. Empty SPE-columns were filled with 1.5 ml ethanol suspension of chitin beads (corresponding to approximately 1 ml settled beads) and after sedimentation a porous PTFE-filter disc was placed on top of the settled column bed. Columns were equilibrated with 20 ml of TBST buffer. For equilibration a pressure of 0.4 bar was applied to the columns. Next, between 1.5 ml and 6 ml of scFv-C30A-C30A crude extract was loaded containing between 0.4 mg and 2 mg of fusion protein. No pressure was applied to the columns during loading of CBD- fusions containing crude extracts. The columns were washed at 0.2 bar with 2 ml (2 column volumes, cv) of TBST buffer and equilibrated with 8 ml (8 cv) of MBS buffer also at 0.2 bar. In the next step 2-3 ml crude extract containing the antigen was loaded without applying any pressure. The columns were washed with 20 ml (20 cv) MBS at 0.2 bar. For the anti-FkpA fusions an additional washing step at pH 3.2 (100 mM glycine/HCI pH 3.2, 500 mM NaCl, 0.1 mM EDTA; 10 ml of buffer at 0.2 bar) was included. Elution of the different antigens was performed with 2 ml of elution buffer as follows : Elution from the anti-his tag columns could be achieved at pH 10 (50 mM Caps/NaOH pH 10,500 mM NaCl, 0.1 mM EDTA). All 4 different anti-FkpA columns were eluted at pH 2 (100 mM glycine/HCI pH 2,500 mM NaCl, 0.1 mM EDTA). SHP could be either eluted at pH 3.2 or pH 10, while the elution of GpD or GpD-fusion proteins was more effective at pH 3.2. Eluted fractions at pH 3.2 were neutralized with 17 ul 1 M Tris per ml, for the fractions eluted at pH 2 100 pl of 1 M Tris was required per mi.

Immobilized metal ion affinity chromatography For the second purification step of the his-tagged proteins Ni-NTA Superflow was used. As for the IAC runs, all IMAC experiments were performed using the positive pressure manifold. The reservoirs were filled with 0.5 ml sedimented Ni-NTA material. The gel bed was covered with a PTFE disc and equilibrated with 20 ml of binding buffer (50 mM Tris/HCI pH8,500 mM NaCl, 20 mM imidazole). The pHlO- eluted fractions of the anti-his tag affinity columns were directly loaded. The columns were washed with 20 ml of binding buffer. Elutions were carried out with 2 ml of elution buffer (50 mM Tris/HCI pH 8,500 mM NaCl, 200 mM imidazole).

Most of the chromatography experiments were carried out using a positive pressure manifold. Sample processing without any detector became possible after some preliminary tests to determine buffer volumes necessary for washing and elution of the columns. The low overpressure applied to the columns lead to a more uniform flow rate than under gravity flow.

To demonstrate the performance of the immobilized scFv-C30A-C30A fusion proteins as affinity ligands, three different series of chromatography experiments were shown. One series demonstrates purifications with the anti-gpD and anti-SHP ligands (Fig. 3A). gpD by itself and the his tagged kinase fusion protein gpHD-cCrk (Forrer & Jaussi, 1998) bind to the anti-gpD column and can be eluted either at pH 3.2 or pH 10, whereas for this scFv fragment elution at pH 3.2 is more effective. SHP binds to its scFv fragment and can be eluted at pH 3.2 and pH 10 with the same recovery.

In a second series of experiments the performance of four different anti-FkpA scFv fragments 6B1, 7B2,9B3 and 11B4 was investigated. The screening for elution conditions showed that elution of the antigen was only possible at pH 2,

independent of the scFv immobilized. This tight binding made it possible to introduce an additional washing step at pH 3.2, removing some further contaminants. The same defined amount of the scFv fusion proteins (0.4 mg) was immobilized on the columns and overloaded them with FkpA. After washing the columns most FkpA from the 7B2 and 11B4 columns could be eluted. While there was one additional band visible when the 11B4 scFv fragment was used all other columns yielded pure protein (Fig. 3B). The experiments in Fig. 3A and 3B also show that the antibody itself is the most crucial determinant of purification quality and that with suitable antibodies, highly pure protein is obtained and no significant contaminants are introduced by the immobilization strategy.

The above examples have been carried out to investigate the range of applications of the immobilization concept described, the influence of the scFv chosen and to demonstrate that IAC can be easily standardized for parallel purification of different samples. A model system-anti-his tag scFv and his-tagged <BR> <BR> GFP (Crameri et al., 1996) -has been used to develop the protocol at different scales and in particular optimize the buffers applied for the purifications described above. Binding experiments with the anti-his tag-C30A-C30A fusion protein in different buffers showed that the fusion proteins bound completely in most buffers, including pH-values between 3 and 10, NaCl concentrations between 0 M and 1 M, and some additives such as 1 % Triton X-100,0. 4 M arginine, 0.1 % SDS. These results are in accordance with results described in the literature (Chong et al., 1997) and suggest that the equilibrium binding constant is not influenced much by the buffer components mentioned above. As the CBD is a very hydrophobic domain (Ikegami et al., 2000), membrane components associated with the CBDs initially clogged the columns and appeared as contaminants in the eluted fractions. This

problem could be reduced to a minimum by addition of 1 % Triton X-100 to the buffer. Additionally, a high concentration of NaCl (> 1 M) reduces unspecific ionic interactions. In conclusion, TBST buffer was found as the optimal binding buffer for all scFv-C30A-C30A fusion proteins. Further experiments showed that in the antigen loading step Triton X-100 was not necessary and that 500 mM NaCl was sufficient to reduce protein binding to chitin.

In addition the capacity of the column for scFv-C30A-C30A fusion proteins and for the desired antigen was determined. Batch binding experiments suggested that the capacity of chitin for CBD fusions is extremely high. Indeed, 100 u of sedimented chitin beads bound approximately 5 mg of the anti-his tag-C30A-C30A fusion protein. Nevertheless, dynamic conditions are different, since some binding sites will not be accessible under flow conditions because of diffusion limitation, and bound molecules will dissociate from their binding sites and will be transported by the buffer flow resulting in a slow migration of the ligands over the column. Therefore, loading more than 2 mg of scFv-C30A-C30A fusion proteins per 1 ml of settled chitin beads is not recommended. This"dynamic capacity"is still in the upper range described for conventional affinity columns with monoclonal antibodies (Jack & Beer, 1996) 2. To determine the capacity of antigen, batch binding experiments with the GFP/anti-his tag model system were carried out, which demonstrated that the ratio of bound GFP to the scFv-C30A-C30A fusion is approximately 1: 1.

20 mg his-tagged GFP were purified on 17 ml beads with 27 mg scFv-fusion (Fig.

3C). This procedure should be directly saleable to multi-gram amounts with any scFv fragment.

Two-step purification of his-tagged proteins.

The above results demonstrate that IAC with the scFv fragments described could provide a tool for the one-step purification of the respective antigen to a high degree of purity (Fig. 3), which might be sufficient for many applications. However, a generic technology to obtain high purity proteins with as simple a procedure as plasmid purification should also be provided. Thus, a coupled anti-his tag/IMAC procedure was provided, in which the same tag is recognized twice but by totally different physical principles, and different contaminants should thus be removed. To avoid any dialysis steps and allow direct coupling of the columns, the elution conditions without imidazole for the anti-his tag affinity column were investigated, and found that elution of his-tagged GFP was complete at pH 10 and resulted in a sharp peak. This convenient elution behavior is the result of this scFv fragment recognizing a protonated, C-terminally located his tag (Lindner et al., 1997), which becomes deprotonated when the pH value is increased. To demonstrate the generality of the method, the following proteins were tested: GroES-his (Lindner et al., 1997; Nieba et al., 1997) (7 his tags), the scFv fragment 4D5-his (Carter et al., 1992) (1 his tag), citrate synthase (Lindner et al., 1997) (CS-his, 2 his tags), GFP-his (1 his tag) and E coli alkaline phosphatase (Inouye et al., 1981) (PhoA-his, 2 his tags). Even though there are a different numbers of his tags on the different antigens, no differences in the elution behavior of the antigens could be detected, as most antigen was found in the first 2 ml fraction of the elution. The eluted fractions again show a very high degree of purity (Fig. 4, lanes A). They were directly loaded on the Ni-NTA column equilibrated with binding buffer containing 20 mM imidazole to reduce specific binding. With the coupled IMAC step the contaminants could be further reduced (Fig. 4, lanes B).

Figure legends Fig. 1. Schematic representation of the fusion proteins containing one, two or four CDBs. Each of the constructs was cloned with wt-CBDs and with CBDs, where the single cysteine was mutated to serine (C30S) or alanine (C30A), resulting in 12 different constructs. The scFv-CBD constructs have one CBD, while the dimeric molecules miniAb-CBD bind via two CBDs. The scFv-CBD-CBD constructs represent a tandem CBD. The combination of the dimerization motif and the tandem CBDs results in functional units possessing four CBDs,"miniantibodies, miniAbs".

Fig. 2. Anti-FLAG blot showing the different expression levels and binding characteristics of the 12 different anti-his tag scFv-CBD fusion proteins. A.

Comparison of the expression level. S denotes the soluble fraction of the fusion protein; P the pellet fraction. B. Binding of the different constructs to chitin beads.

Blots detecting the dissociated molecules are shown. Constructs suited as affinity ligands should therefore contain no band or only a weak band in the supernatant shown here. An independent blot (not shown) had verified that equal amounts of the constructs were initially incubated with the beads. Stable binding was found for all scFv-CBD-CBD and all miniAb-CBD-CBD constructs. The bands that are visible in the corresponding lanes (7-12) are proteolytic digestion products carrying only one CBD (of the size of the molecules in Fig. 1), as can also be seen from Fig. 2A.

Fig. 3. SDS-PAGE analysis demonstrating the performance of purifications with

different antibodies and elution conditions. The gels (15 %) were run under reducing conditions and stained with Coomassie brilliant blue R250. A. Eluted fractions of the anti-gpD and anti-SHP affinity columns: (M) molecular weight marker, (1) SHP eluted at pH 3.2, (2) SHP eluted at pH 10, (3) His-tagged gpD (gpHD) fusion protein eluted at pH 3.2 (gpHDL-cCrk), (4) gpHD eluted at pH 3.2, (5) gpHD eluted at pH 10. B.

Purification of FkpA with four different scFv fragments 6B1, 7B2,9B3 and 11B4. (M) molecular weight marker, (1) wash at pH 3.2 where some further contaminants can be eluted, (2) elution at pH 2, (3) affinity beads after the elution of bound FkpA. The affinity bead fractions show that approximately the same amount of each scFv fragment was bound to the column. Therefore, the different amounts of eluted FkpA are the result of different binding characteristics of the scFv fragments. C.

Purification of his-tagged GFP at different scales. (1) Small scale (2 mg) using the positive pressure manifold and (2) large scale purification, where approximately 20 mg have been purified using a FPLC system. Both purifications show the same degree of purity. The smaller band observed in the GFP-his purification (about 25 kDa) is a N-terminal proteolytic digestion product of GFP, still containing the C- terminal his tag.

Fig. 4. Two-step purification of different his-tagged proteins. Lanes A show the eluted fractions of the anti-his tag affinity column, lanes B the fractions after the anti-his tag affinity column coupled with IMAC. (M) molecular weight marker, (1) scFv 4D5-his, (2) GroES-his, (3) PhoA-his, (4) CS-his and (5) GFP-his.

Fig. 5. Starting plasmid pKB2scFvCBD for cloning and expression of scFv-CBD fusion proteins. The plasmid is a derivative of the pAK-series (Krebber et al., 1997): The

plasmid has a/ac-promoter and the RBS of pAK400 (Krebber et al., 1997). Fusion proteins are expressed in the periplasm coexpressing the periplasmic folding factor Skp (Bothmann & Pluckthun, 1998) regulated by its own promotor. Also represented are the restriction sites used for cloning of the dHLX-fragment (Pack & Pluckthun, 1992)-EcoRI-and for the second linker-CBD fragment-NheI.

References Arnold-Schild, D. et al. One-step single-chain Fv recombinant antibody-based purification of gp96 for vaccine development. Cancer Res. 60,4175-4178 (2000). <BR> <BR> <P>Ausubel, F. M. , Brent, R. , Kingston, R. E. , Moore, D. D., Sedman, J. G. , Smith, J. A. , & Struhl, K. eds. (1999). Current Protocols in Molecular Biology. New York: John Wiley and Sons. <BR> <BR> <P>Bass, S. , Gu, Q. & Christen, A. Multicopy suppressors of prc mutant Escherichia coli include two HtrA (DegP) protease homologs (HhoAB), DksA, and a truncated RIpA. J. Bacteriol. 178,1154-1161 (1996).

Berry, MJ. & Pierce, J. J. Stability of immunoadsorbents comprising antibody fragments. Comparison of Fv fragments and single-chain Fv fragments. J.

Chromatogr. 629,161-168 (1993).

Berry, M. J., Davies, J., Smith, C. G. & Smith, I. Immobilization of Fv antibody fragments on porous silica and their utility in affinity chromatography. J.

Chromatogr. 587,161-169 (1991). <BR> <BR> <P>Bird, R. E. , Hardman, K. D., Jacobson, J. W. , Johnson, S. , Kaufman, B. M. , Lee, S.<BR> <P>M. , Lee, T. , Pope, S. H. , Riordan, G. S. & Whitlow, M. (1988). Single-chain antigen-binding proteins. Science 242,423-426. <BR> <BR> <P>Blake, M. S. , Johnston, K. H., Russel-Jones, GJ. & Gotschlich, E. C. (1984). A rapid, sensitive method for detection of alkaline phosphatase-conjugated antiantibody on Western blots. Anal. Biochem. 136,175-179.

Bothmann, H. & Pluckthun, A. Selection for a periplasmic factor improving phage display and functional periplasmic expression. Nat. Biotechnol. 16,376-380.

(1998).

Bothmann, H. & Pluckthun, A. The periplasmic Escherichia coli peptidylprolyl cis, trans-isomerase FkpA. I. Increased functional expression of antibody fragments with and without cis-prolines. J. Biol. Chem. 275,17100-17105.

(2000).

Brinkmann, U. , Reiter, Y., Jung, S. , Lee, B. & Pastan, 1. (1993). A recombinant immunotoxin containing a disulfide-stabilized Fv fragment. Proc. Natl. Acad. Sci.

U. S. A. 90,7538-7542. <BR> <BR> <P>Carrard, G., Koivula, A. , Soderlund, H. & Beguin, P. Cellulose-binding domains promote hydrolysis of different sites on crystalline cellulose. Proc. Natl. Acad.

Sci. U. S. A. 97, 10342-10347 (2000).

Carter, P. et al. High level Escherichia coli expression and production of a bivalent humanized antibody fragment. Biotechnology. (N. Y). 10,163-167. (1992).

Chong, S. et al. Single-column purification of free recombinant proteins using a self- cleavable affinity tag derived from a protein splicing element. Gene 192,271- 281 (1997). <BR> <BR> <P>Crameri, A., Whitehorn, E. A. , Tate, E. & Stemmer, W. P. Improved green fluorescent protein by molecular evolution using DNA shuffling. Nat. Biotechnol. 14,315- 319 (1996).

Cutler, P. Affinity chromatography. Methods Mol. Biol. 59,157-168 (1996). <BR> <BR> <P>Ford, C. F. , Suominen, I. & Glatz, C. E. Fusion tails for the recovery and purification of recombinant proteins. Protein Expression & Purification 2,95-107 (1991).

Forrer, P. & Jaussi, R. High-level expression of soluble heterologous proteins in the cytoplasm of Escherichia coli by fusion to the bacteriophage lambda head protein D. Gene 224,45-52 (1998).

Ge, L., Knappik, A. , Pack, P. , Freund, C. & Pluckthun, A. Expressing antibodies in Escherichia coli. In Antibody Engineering, 2nd. Edn. (ed. Borrebaeck, C. A. K.) 229-266 (Oxford University Press, New York, NY; 1995). <BR> <BR> <P>Glockshuber, R. , Malia, M. , Pfitzinger, I. & Pluckthun, A. (1992). A comparison of strategies to stabilize immunoglobulin Fv-fragments. Biochemistry 29,1362- 1367. <BR> <BR> <P>Hanes, J., Schaffitzel, C. , Knappik, A. & Pluckthun, A. Picomolar affinity antibodies from a fully synthetic naïve library selected and evolved by ribosome display.

Nat. Biotechnol. 18,1287-1292. (2000).

Hochuli, E. , Bannwarth, W., Dobeli, H. , Gentz, R. & Stuber, D. Genetic approach to facilitate purification of recombinant proteins with a novel metal chelate adsorbent. Biotechnology 6,1321-1325 (1988).

Hopp, T. P. , Prickett, K. S. , Price, V. L., Libby, R. T. , March, CJ., Cerretti, D. P., Urdal, D. L. & Conlon, P. J. (1988). A short polypeptide marker sequence useful for recombinant protein identification and purification. Bio/Technology 6,1204- 1210. <BR> <BR> <P>Huston, J. S., Levinson, D. , Mudgett-Hunter, M. , Tai, M. S., Novotny, J., Margolies,<BR> M. N. , Ridge, R. J., Bruccoleri, R. E. , Haber, E. , Crea, R. & Oppermann, H.

(1988). Protein engineering of antibody binding sites: recovery of specific activity in an anti-digoxin single-chain Fv analogue produced in Escherichia coli.

Proc. Natl. Acad. Sci. USA 85,5879-5883.

Ikegami, T. et al. Solution structure of the chitin-binding domain of Bacillus circulans WL-12 chitinase A1. J. Biol. Chem. 275,13654-13661 (2000). <BR> <BR> <P>Inouye, H., Michaelis, S. , Wright, A. & Beckwith, J. Cloning and restriction mapping of the alkaline phosphatase structural gene (phoA) of Escherichia coli and generation of deletion mutants in vitro. J. Bacteriol. 146,668-675. (1981).

Jack, G. W. & Beer, D. J. Immunoaffinity chromatography. Methods Mol. Biol. 59,187- 196 (1996).

Jones, C. et al. Current trends in molecular recognition and bioseparation. J.

Chromatogr. A 707,3-22 (1995). <BR> <BR> <P>Kleymann, G. , Ostermeier, C. , Ludwig, B. , Skerra, A. & Michel, H. Engineered Fv fragments as a tool for the one-step purification of integral multisubunit membrane protein complexes. Bio/Technology 13,155-160 (1995).

Knappik, A. & Pluckthun, A. An improved affinity tag based on the FLAG peptide for the detection and purification of recombinant antibody fragments.

Biotechniques 17,754-761. (1994).

Knappik, A. et al. Fully synthetic human combinatorial antibody libraries (HuCAL) based on modular consensus frameworks and CDRs randomized with trinucleotides. J. Mol. Biol. 296,57-86. (2000).

Krebber, A. et al. Reliable cloning of functional antibody variable domains from hybridomas and spleen cell repertoires employing a reengineered phage display system. J. Immunol. Methods 201,35-55. (1997).

Krebs, B. et al. High-throughput generation and engineering of recombinant human antibodies. J. Immunol. Methods 254,67-84. (2001).

Lanio, T., Jeltsch, A. & Pingoud, A. Automated purification of His6-tagged proteins allows exhaustive screening of libraries generated by random mutagenesis.

Biotechniques 29,338-342. (2000).

Linder, M. , Nevanen, T., Soderholm, L., Bengs, 0. & Teeri, T. T. Improved immobilization of fusion proteins via cellulose-binding domains. Biotechnology & Bioengineering 60,642-647 (1998).

Linder, M. , Salovuori, I. , Ruohonen, L. & Teeri, T. T. Characterization of a double cellulose-binding domain. Synergistic high affinity binding to crystalline cellulose. J. Biol. Chem. 271,21268-21272 (1996).

Lindner, P. et al. Purification of native proteins from the cytoplasm and periplasm of Escherichia coli using IMAC and histidine tails : a comparison of proteins and protocols. Methods: A Companion to Methods in Enzymology 4,41-56. (1992).

Lindner, P. et al. Specific detection of his-tagged proteins with recombinant anti-His tag scFv-phosphatase or scFv-phage fusions. Biotechniques 22,140-149.

(1997).

Molloy, P. , Brydon, L. , Porter, A. J. & Harris, W. J. Separation and concentration of bacteria with immobilized antibody fragments. J. Appl. Bacteriol. 78,359-365 (1995).

Mössner, E. , Koch, H. & Pluckthun, A. Fast selection of antibodies without antigen purification: adaptation of the protein fragment complementation assay to select antigen-antibody pairs. J. Mol. Biol. 308,115-122. (2001).

Nieba, L. et al. BIACORE analysis of histidine-tagged proteins using a cheating NTA sensor chip. Anal. Biochem. 252,217-228. (1997).

Nilsson, J., Stahl, S. , Lundeberg, J., Uhlen, M. & Nygren, P. A. Affinity fusion strategies for detection, purification, and immobilization of recombinant proteins. Protein Expression & Purification 11,1-16 (1997).

Pack, P. & Pluckthun, A. Miniantibodies: use of amphipathic helices to produce functional, flexibly linked dimeric Fv fragments with high avidity in Escherichia coli. Biochemistry 31,1579-1584. (1992). <BR> <BR> <P>Pack, P. , Kujau, M. , Schroeckh, V., Kupfer, U. , Wenderoth, R. , Riesenberg D. & Pluckthun, A. (1993). Improved bivalent miniantibodies, with identical avidity as whole antibodies, produced by high cell density fermentation of in Escherichia coli. Bio/Technology 11, 1271-1277.

Pack, P. (1994). Mini-Antikörper : Bivalente, tetravalente und bispezifische Immunglobuline aus E. coli. Ph. D. thesis, Ludwig-Maximilians-Universitat München.

Pluckthun, A. , Schaffitzel, C. , Hanes, J. & Jermutus, L. In vitro selection and evolution of proteins. Adv. Protein Chem. 55,367-403 (2000).

Ramirez, C. et al. A bifunctional affinity linker to couple antibodies to cellulose.

Bio/Technology 11, 1570-1573 (1993).

Ramm, K. & Pluckthun, A. The periplasmic Escherichia coli peptidylprolyl cis, trans- isomerase FkpA. II. Isomerase-independent chaperone activity in vitro. J. Biol.

Chem. 275,17106-17113. (2000).

Shpigel, E. et al. Immobilization of recombinant heparinase I fused to cellulose- binding domain. Biotechnology & Bioengineering 65,17-23 (1999). <BR> <BR> <P>Shpigel, E. , Elias, D. , Cohen, I. R. & Shoseyov, 0. Production and purification of a recombinant human hsp60 epitope using the cellulose-binding domain in Escherichia coli. Protein Expression & Purification 14,185-191 (1998).

Sii, D. & Sadana, A. Bioseparation using affinity techniques. J. Biotechnol. 19,83-98.

(1991).

Skerra, A. & Pluckthun (1988). Assembly of a functional immunoglobulin Fv fragment in Escherichia coli. Science 240,1038-1041.

Smith, G. P. & Scott, J. K. Libraries of peptides and proteins displayed on filamentous phage. Methods Enzymol. 217,228-257 (1993).

Turkova, J. Oriented immobilization of biologically active proteins as a tool for revealing protein interactions and function. Journal of Chromatography. B, Biomedical Sciences & Applications 722,11-31 (1999).

Vaughan, T. J. et al. Human antibodies with sub-nanomolar affinities isolated from a large non-immunized phage display library. Nat. Biotechnol. 14,309-314 (1996).

Watanabe, T. et al.-The roles of the C-terminal domain and type III domains of chitinase Al from Bacillus circulans WL-12 in chitin degradation. J. Bacteriol.

176,4465-4472 (1994).

Weiss, E., Chatellier, J. & Orfanoudakis, G. In vivo biotinylated recombinant antibodies: construction, characterization, and application of a bifunctional Fab- BCCP fusion protein produced in Escherichia coli. Protein Expression & Purification 5,509-517 (1994). <BR> <BR> <P>Winter, G. , Griffiths, A. D. , Hawkins, R. E. & Hoogenboom, H. R. Making antibodies by phage display technology. Annu. Rev. Immunol. 12,433-455 (1994).

Yang, F. et al. Novel fold and capsid-binding properties of the lambda-phage display platform protein gpD. Nat. Struct. Biol. 7,230-237. (2000).