Patent Application
20160207986
The present invention relates to the immobilisation of the antigen binding domains (VHH) of antibodies naturally devoid of light chains or functional equivalents thereof at solid surfaces with retention of binding activity. More particularly, the invention relates to the preparation and use of immunoactive materials comprising a VHH domain, without any peptide extensions, inmmobilised on a solid surface by means of covalent interactions.
Processes directed towards immobilising proteins on to a solid surface are of considerable commercial interest. Functionalisation of surfaces with immunological materials, such as antibodies or antibody fragments, in particular, forms the basis of immunoadsorption techniques such as immuno-affinity purification processes which are increasingly being applied to the recovery or purification of a range of commercially important materials.
Commonly, attachment of proteins to solid surfaces, such as chromatography media, has been brought about by exposing the surface to a solution of the protein such that the protein is adsorbed onto the solid surface via non-specific binding mechanisms. Methods for immobilising proteins on chromatography media are well established in the literature, see for example, in Protein Immobilisation, R. F. Taylor ed., Marcel Dekker, Inc., New York, 1991. Where the solid surface is provided by a hydrophobic material such as polystyrene, for example, then attachment is generally brought about by adsorption of hydrophobic regions of the protein onto the hydrophobic surface.
Adsorption onto the solid surface is usually accompanied by significant conformational disruption with partial unfolding and denaturation of the protein concerned. The concomitant loss of protein activity detracts from the overall usefulness of the process. Commonly, for example, adsorption of antibodies onto a hydrophobic surface is accompanied by the loss of in the order of greater than 95% of binding activity as well as diminished binding specificity. Where smaller antibody fragments are involved, the degree of specific binding affinity retained upon adsorption onto a solid surface can be even lower, as described by Molina-Bolivar et al, J. Biomaterials Science-Polymer Edition, 9, 1103-1113, 1998.
Alternatives to, or improvements upon, the method of adsorption of proteins in the preparation of immobilised protein surfaces have been considered.
One alternative approach is to use chemical cross-linking of residues in the protein for covalent attachment to an activated solid surface using conventional coupling chemistries, for example as described in Bioconjugate Techniques, G. T. Hermanson, ed. Academic Press, Inc., San Diego, Calif., USA. Amino acid residues incorporating sulphydryl groups, such as cysteine, may be covalently attached using a bispecific reagent such as succinimidyl-maleimidophenylbutyrate (SMPB), for example. Alternatively, lysine groups located at the protein surface may be coupled to activated carboxyl groups at the solid surface by conventional carbodiimide coupling using 1, ethyl-3-[3-dimethyl aminopropyl]carbodiimide (EDC) and N-hydroxysuccinimide (NHS). A disadvantage of this approach is that cross-linking residues in the protein can interfere with the functionality of the protein. The smaller the antibody fragment to be immobilised, the greater the probability that cross-linking will involve a residue in, or close to, the antigen binding site and, consequently, interfere with binding activity. This has been noted specifically for the case of isolated classical VH domains, where cross-linking was found to lead to substantial losses in binding affinity and specificity (see, for example, Berry et al, Journal of Chromatography, 597, 239-245 (1992), which suggests that classical VH domains are more readily inactivated than Fv fragments during immobilisation). As shown by Spitznagel et al, Bio/Technology, 11, 825-829, (1993), in addition to decreased specific activities, immobilised antibodies typically exhibit lower binding affinities than their soluble counterparts.
By providing the antibody fragment with an extension in the form of a peptide tail or an additional protein domain, attachment of the protein to the surface can be brought about by non-covalent adsorption or using conventional chemical cross-linking agents at a site remote from the main body of the protein. In this way, the immobilisation process itself is less likely to interfere with the functionality of the protein.
Rational coupling procedures to control the orientation of the antibody fragment can also be employed (Rao et al, Mikrochimica Acta, 128 (3-4), 127-143, 1998). Furthermore, increasing the length of the chemical linker used to attach the immobilised molecule to the solid support can reduce the conformational stresses on the protein, allowing it to exist in a more natural conformation.
EP 0434317 (Joseph Crosfield & Sons) discloses the use of improved affinity purification media which employ small specific binding agents, especially Fv antibody fragments. These optionally have a hydrophobic tail, with a particularly preferred linking group being the myc epitope (Munro, S., and Pelham, H. R. (1986) Cell 46, 291-300). Such a tail is primarily intended to facilitate immobilisation of the binding agent by non-covalent attachment onto a hydrophobic surface although as the myc group contains a lysine residue, it could also be used for covalent attachment onto surfaces.
WO 91/08482 (Unilever) describes in Example 6 an anti-lysozyme VH with no tail, immobilised on a solid phase; from the results presented, it is clear that antibody binding activity is adversely and severely affected compared to VH fragments immobilised at a solid surface via a peptide tail.
There remains a continuing need to improve the ability to prepare immnunoactive surfaces by immobilisation of antibody fragments with retention of their binding affinity and specificity, preferably without the need to elaborate them with tail sequences.
A class of antibodies of particular interest in relation to biotechnological applications comprises the Heavy Chain antibodies such as found in camelidae, like the camel or the llama. The binding elements of these antibodies consist of a single polypeptide domain, namely the variable region of the heavy chain polypeptide (VHH). These antibodies are naturally devoid of light-chains with the heavy chain variable domain forming the complete antigen-binding site. In contrast, classical antibodies (murine, human, and so on), have binding elements comprising two polypeptide domains (the variable regions of the heavy chain (VH) and the light chain (VL)). The lack of dependence on interaction with a light chain variable domain for maintaining structural and functional integrity gives these VHH domains a substantial advantage over other small antibody fragments, in terms of ease of production and behaviour in solution. In particular, VHH fragments are the preferred types of molecules for immuno-affinity purification, because of their unusual stability and their ability to refold efficiently after complete denaturation, which frequently occurs during elution of antigen.
Procedures to obtain heavy chain immunoglobulins from Camelidae, or (functionalised) fragments thereof, have been described in WO 94/04678 (Casterman et al) and WO 94/25591 (Frenken et al). VHH fragments can be produced through recombinant DNA technology in a number of microbial hosts (bacterial, yeast, mould), as described in WO 94/29457 (Frenken et al). Alternatively, binding domains can be obtained by modification of the VH fragments of classical antibodies by a procedure termed “camelisation”, described by Davies et al, Bio/Technology, 13, 475-479 (1995). Hereby the classical VH fragment is mutated, by substitution of a number of amino acids present in the VH/VL interface, into a VHH-like fragment, whereby its binding properties are retained.
The immobilisation of a VHH fragment with a His-tail on a carboxylated dextran layer using EDC/NHS coupling chemistry is described by Muyldermans et al, EMBO Journal, 17 (13), p 3512-3520, 1998. Polypeptide tail-mediated immobilisation of VHH fragments on solid surfaces is also proposed in WO 99/23221 (Unilever).
The invention describes the immobilisation of single-domain antigen-binding fragments from antibodies naturally devoid of light chains (VHH), or protein domains functionally equivalent thereto, onto a solid surface through covalent cross-links while still retaining sufficient biological activity to enable the fragments to function (e.g. as in an affinity purification matrix). The fragments themselves are immobilised directly on to the solid surface; they are devoid of any appended polypeptide group through which immobilisation could otherwise be mediated. The VHH fragments described are typically less than 20 kDa in molecular weight and contain only the amino acid sequence necessary to form a complete functional VHH fragment or a proportion thereof (truncated fragment).
The present invention may be more fully understood with reference to the following description when read together with the accompanying drawings in which:
Right panel: purification using immobilised, tagged VHH#1. (lane 1: molecular weight markers; lane 2: mouse serum; lane 3: flow through (fraction 3); lane 4: elution (fraction 10).
Molecular weight markers: Marker GibcoBRL BenchMark (Lot. No.: 1080925)
Lane 1, 5, 9: markers.
Lanes 2-4: purification using immobilised, untagged VHH#1 (lane 2: mouse serum; lane 3: flow through (fraction 3—see
Lanes 6-8: purification using immobilised, tagged VHH#1 (lane 6: mouse serum; lane 7: flow through (fraction 3—see
Molecular weight markers: Marker GibcoBRL BenchMark (Lot. No.: 1080925).
The invention is based on the finding that cloned VHH fragments can be attached to a solid phase matrix surface (and still retain sufficient activity to function as the affinity ligand of a purification matrix) in the absence of any appended polypeptide group such as peptide tails or extra (non-variable heavy domain) protein domains.
This contrasts surprisingly with the situation with classical VH fragments where, in order to retain binding activity, it has been found to be necessary to provide the fragment with a polypeptide tail through which immobilisation can be mediated, such that attachment to the solid surface occurs at a location remote from the antigen binding site. The conventional rationale for this is that all residues in such a small binding domain are likely to be important either for antigen binding or for ensuring structural integrity and therefore that cross-linking at any point in the molecule will tend to result in its inactivation. Given that VHH and VH domains are similar in size, it would generally have been expected that they both would be subject to similar inactivating perturbations upon immobilisation at a surface.
The present inventors have clearly shown that immobilisation of VHH domains, devoid of any peptide extensions, at a surface through covalent cross-links does not destroy their ability to bind antigen with high affinity and selectivity. Further, it has been shown that provision of a cross-linkable peptide extension improves neither the efficiency of the cross-linking reaction nor the antigen-binding capacity or selectivity of the bound VHH domains.
The ability to attach functional antibody fragments directly to a surface, rather than via a fused tag sequence, is extremely advantageous because it is well-known that such fusion proteins are highly susceptible to proteolysis, with consequent loss of the immobilised antibody from the surface; see for example McCafferty, J. et al (1990) Nature 348, 552-554.
A VHH domain is a heavy chain variable domain derived from an immunoglobulin naturally devoid of light chains, such as may be obtained from camelids as described above. The antigen-binding capacity and specificity is located naturally and exclusively in the heavy chain variable domain; that is, the heavy chain variable domain forms the complete antigen-binding site.
The terms “imnunoglobulin” and “antibody” are used synonymously throughout this specification unless indicated otherwise.
An advantage of using immunoglobulins or heavy chain variable domains naturally devoid of light chains is that they can be readily and conveniently produced economically on a large scale, for example, using a transformed lower eukaryotic host as described in WO 94/25591, mentioned above. Further, as mentioned above, the absence of dependence on a light chain for ensuring structural and functional integrity renders isolated VHH domains more stable and easier to handle than other small antibody fragments such as classical VH domains. This is especially advantageous for the preparation of reusable affinity matrices.
By functionally equivalent is meant any protein which has the same or similar antigen binding properties located in a single binding domain. It will be appreciated that proteins modified to enable them to function as binding domains in the same way as heavy chain immunoglobulin variable domains from camelids (see Davies et al, Bio Technology, 13, 475-479, (1995)) may also suitably be used according to the invention.
According to one embodiment of the invention, the VHH domain or functionally equivalent protein is attached to a solid surface by covalent cross-linking using conventional coupling chemistries.
It will be appreciated that since the VHH domain or functional equivalent is attached to a solid surface by covalent coupling, the surface must be capable of being covalently coupled to the VHH domain. The solid surface may naturally comprise cross-linkable residues suitable for covalent attachment or it may be coated or derivatised to introduce suitable cross-linkable groups according to methods well known in the art.
The solid surface onto which immobilisation according to the invention takes place may be provided by a variety of materials and may suitably be any solid phase carrier material conventionally used in immobilising proteins.
The invention is applicable to any solid phase material that is amenable to the immobilisation of proteins or protein fragments, either directly or after pre-treatment. The carrier materials may be particulate (e.g. beads or granules, generally used in extraction columns) or in sheet form (e.g. membranes or filters, glass or plastic slides, microtitre assay plates, dipstick, capillary fill devices or such like) which can be flat, pleated, or hollow fibres or tubes. The following matrices are given as examples and are not exhaustive, such examples could include silica (porous amorphous silica), i.e. the FLASH series of cartridges containing 60 Å irregular silica (32-63 μm or 35-70 μm) supplied by Biotage (a division of Dyax Corp.), agarose or polyacrylamide supports, for example the Sepharose range of products supplied by Amersham Pharmacia Biotech, or the Affi-Gel supports supplied by Bio-Rad. In addition there are macroporous polymers, such as the pressure-stable Affi-Prep supports as supplied by Bio-Rad. Other supports that could be utilised include; dextran, collagen, polystyrene, methacrylate, calcium alginate, controlled pore glass, aluminium, titanium and porous ceramics. Alternatively, the solid surface may comprise part of a mass dependent sensor, for example, a surface plasmon resonance detector. Further examples of commercially available supports are discussed by Taylor, R. F. (1991), referred to above. Conveniently, an array comprising plurality of individual antigen-binding fragments bound to a solid surface is provided.
Advances in molecular biology, particularly through site-directed mutagenesis, enable the mutation of specific amino acid residues in a protein sequence. The mutation of a particular amino acid (in a protein with known or inferred structure) to a lysine or cysteine (or other desired amino acid) can provide a specific site for covalent coupling, for example. It is also possible to reengineer a specific protein to alter the distribution of surface available amino acids involved in the chemical coupling (Kallwass et al, Biotechnol. Lett., 15 (1), 29-34, 1993), in effect controlling the orientation of the coupled protein. A similar approach can be applied to antibody fragments, specifically VHH fragments, so providing a means of oriented immobilisation without the addition of extra-VHH peptide tails or domains containing either natural or unnatural amino acids. Introduction of mutations in the framework region of the antibody fragment is preferred, minimising disruption to the antigen-binding activity of the VHH fragment. A particular region of the VHH fragment suitable for mutagenesis is in the portion of the molecule that is in proximity to the constant heavy domain, which occurs in naturally produced heavy chain antibodies.
A particular advantage of the invention is that sufficient biological activity of the antibody domain, more specifically the VHH fragment, is retained following immobilisation to allow the coupled fragment to function as an affinity ligand. In one particular embodiment, sufficient functionality is retained following direct covalent coupling to the desired matrix via a reactive moiety that does not contain a chemical spacer arm.
Materials prepared according to the invention may be used in any process where it is useful to bind a molecule to an immobilised antibody fragment. Suitable applications will readily suggest themselves to the average skilled person in the art. Conveniently, the immobilised materials may be used in immunoadsorption processes such as immunoassays, for example ELISA, or immunoaffinity purification processes by contacting a material according to the invention with a test sample according to standard methods conventional in the art. Alternatively, an assay comprising a plurality of individual antigen-binding fragments bound to a solid surface according to the invention may be used, for example, to test for presence of one or more specific binding partners.
The following examples are provided by way of illustration only, are not exhaustive and should not be considered as limiting the scope of the invention. Techniques used for the manipulation and analysis of nucleic acid materials can be performed as described in Sambrook et al, (1989), unless otherwise indicated.
The following protocol is taken as an example of how specific VHH fragments can be isolated, cloned, expressed, then coupled onto the desired matrix:
Although the particular VHH fragments described in this example were derived from an immune repertoire, they also could have been selected from a synthetic/semi-synthetic naive VHH library (see WO 00/43507, Unilever).
A llama was immunised with Fc fragments prepared from polyclonal IgG's from mouse serum or with an Anti-Freeze Peptide (AFP) from Lolium perenne (grass AFP) (Sidebottom et al. (2000), Nature 406, 256). The llama was boosted several times following the initial immunisation (one month between injections) to increase the specificity of the immune response. DNA encoding specific VHH fragments can then be isolated using similar methods to those described in WO 94/04678 (Casterman et al). If required an immune repertiore of VHH fragments can be selected against the desired antigen as described in WO 94/18330 (Frenken et al). Alternatively selection methods based on phage display can be used to isolate the anti-Fc VHH or anti-AFP VHH producing clones from immune repertoires.
The encoded amino acid sequence of the isolated anti-Fc VHH fragment used in this example is presented below:
(cloned in E. coli as a MYC-His6-tagged VHH in production plasmid pUR1490 and as untagged VHH in production plasmid pUR1491):
The amino acid sequence of the anti-AFP VHH used in this example is as follows:
(cloned in E. coli as a MYC-His6-tagged VHH in production plasmid pUR1492 and as untagged VHH in production plasmid pUR1493):
The tags used are C-MYC (shown in bold below), recognized by monoclonal antibody 9E10 (Munro, S., and Pelham, H. R. (1986) Cell 46, 291-300), followed by a 12-mer peptide encoding an in vivo biotinylation signal (shown in bold and underlined) and the hexahistidine tail (shown in italics) for purification with IMAC (Hochuli, E., Bannwarth, W., Döbeli, H., Gentz, R., and Stuber, D. (1988) Biotechnology 6, 1321-1325); the complete sequence fused to the carboxy-terminus of the VHH is presented below: EQKLISEEDLN GAA LRSIFEAQKMEW HHHHHH
For expression in Escherichia coli using the production plasmids indicated in example 1, transformants were cultivated in 400 ml 2TY medium and induced with isopropyl-β-D-thiogalactopyranoside as described (Skerra and Pluckthun (1991) Prot. Eng. 4, 971-979). Cells were harvested by centrifugation and lysed in the French Press. Insoluble protein was removed by centrifugation and from the soluble protein fraction the tagged VHH was purified.
Isolated VHH fragments may alternatively be expressed in a transformed lower eukaryote such as Saccharomyces cerevisiae or Pichia pastoris as described in WO 94/25591 (Frenken et al). Tagged VHH may be purified by IMAC using TALON column material according to the instructions of the supplier (Clontech). Expressed VHH fragments without tags may be isolated by either Protein-G or Protein-A affinity chromatography for example using HiTrap rProtein-A columns (Pharmacia Biotech) according to the manufacturer's instructions.
Alternatively, if the VHH fragments cannot be purified by Protein-G or Protein-A then ion exchange chromatography may be used as an alternative, for example using an AKTAexplorer chromatography system (Pharmacia Biotech) according to the manufactures' instructions. Otherwise matrices containing coupled antigen could be used to purify the fragments.
Purification of the non-tagged VHH antibody fragments (anti-grass AFP and anti-mouse Fc) was performed using a Centriplus-50 spin concentrator (Amicon); the concentrator was centrifuged at 3,000 g (Sorvall RC5B). Purification of the tagged anti-Mouse Fc VHH fragment was performed with IMAC using TALON column material according to the instructions of the supplier (Clontech).
Between 200 μg and 2 mg of VHH fragment was coupled to CNBr Sepharose 4B (approx. 0.3 g, Amersham Pharmacia Biotech, Product code 17-0430-01) following the manufacturer's instructions. Prior to coupling a buffer exchange was performed by gel filtration on PD10 columns (Amersham Pharmacia Biotech), so that the VHH fragments would be in a buffer compatible with the immobilisation procedure. The coupled material was packed into a 1 cm diameter column, to give a bed volume of approximately 1.5 ml. The column was then washed with buffer A (10 mM Na2HPO4/NaH2PO4 150 mM NaCl pH 7.4) and pre-eluted with elution buffer B (10 mM Na2HPO4/NaH2PO4 150 mM NaCl pH 2.1), then washed with approximately 5 column volumes buffer A.
Four different fragments were coupled to a CNBr activated matrix in this way, the anti-mouse Fc VHH#1 and the anti-grass AFP VHH#2; both the tagged and the untagged versions.
Analysis of the fractions before and after coupling on a Coomassie stained gel showed that the degree of coupling of the tagged as well as the untagged version of the two different VHH fragments was high and revealed no difference in coupling efficiency between the tagged and untagged VHH (see
This result demonstrates that VHH fragments can be efficiently coupled to a solid surface via covalent bonds even when no peptide tail is present to provide additional sites for cross-linking.
To analyse the performance of the matrices prepared according to example 3, several purifications were performed. After coupling of both the tagged and untagged Anti-mouse Fc VHH, complete mouse serum was loaded onto the columns. The buffers for the purification were, buffer A: 10 mM Na2HPO4/NaH2PO4 150 mM NaCl pH 7.4, and elution buffer B 10 mM Na2HPO4/NaH2PO4 150 mM NaCl with an addition of 12 mM HCl, final pH 2. Fraction collection was done manually, fractions were immediately neutralised with 10% (v/v) 0.2 M Tris.HCl buffer, not pH adjusted. In
The eluted fractions were analyzed on Coomassie stained gel and western blot using anti-mouse Ig alkaline phosphatase conjugate for detection according to the instructions of the supplier (Promega)(see
When using complete mouse serum for affinity purification a large fraction of the polyclonal antibodies are bound and eluted, whereas no contaminating protein such as serum albumin was found in the eluted fractions. Only a small proportion of antibodies was found in the non-bound fraction, which are probably derived from other isotypes of Ig (such as IgM, IgD or otherwise).
The results obtained with immobilised, tagged VHH and immobilised, untagged VHH are essentially the same, demonstrating that the provision of a peptide extension is not necessary to ensure that the immobilised VHH retains its antigen-binding capacity.
After coupling of the anti-grass AFP VHH, a broth containing grass AFP was loaded onto the column. The buffers for the purification were, buffer A: 10 mM Na2HPO4/NaH2PO4, 150 mM NaCl pH 7.4, and elution buffer B 10 mM Na2HPO4/NaH2PO4, 150 mM NaCl with an addition of 12 mM HCl, final pH 2. Fraction collection was carried out manually; fractions were immediately neutralised with 10% (v/v) 0.2 M Tris.HCl buffer, not pH adjusted. In
The untagged version of VHH#2 purified in the same way as the tagged version.
This example demonstrates that the absence of a peptide tail does not affect the capability of surface-bound VHH to bind antigen.
| Number | Date | Country | Kind |
|---|---|---|---|
| 99309516.5 | Nov 1999 | EP | regional |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 10111851 | Oct 2002 | US |
| Child | 14997924 | US |
