WO1993001285A1 - Process for modifying proteins by random mutation - Google Patents

Abstract

A process is described for the modification of a protein product of recombinant DNA technology comprising subjecting DNA coding for the protein product, or coding for at least that portion of the protein product to be modified, to random mutation in the presence of at least one chemical or physical mutagenising agent, expression of the mutated DNA in a host organism and screening of the expression products with at least one antibody so as to select the expression product containing the desired modification.

Description

PROCESS FOR MODIFYING PROTEINS BY RANDOM MUTATION

This invention relates to the preparation of modified proteins and is particularly concerned with the modification of proteins obtained by reco binant DNA technology.

Proteins and polypeptides produced by recombinant DNA technology have in recent years found wide applicability in many areas of vaccine technology, therapy, diagnostics and enzymology. Such synthetically created proteins rely for their effectiveness on the inclusion of critical regions of activity, such as epitopes or binding sites, of which one wishes to make optimum use, but they generally contain other regions of a less critical nature or which are positively harmful to the intended use of the protein.

Thus, for example, if such a protein is intended for use as a vaccine, while one must conserve the epitope or epitopes capable of eliciting the desired immune response, it is in many cases desirable to eliminate other epitopes which hinder the potential usefulness of the vaccine. Such a situation occurs, for example, where an unwanted epitope cross-reacts with a self-molecule and is potentially capable of giving rise to autoaπtibodies when used for immunization, a situation illustrated by the known streptococcus cross-reaction with heart tissue, giving rise to rheumatic fever. As a further example, it can also be desirable to remove immunodominant epitopes which might otherwise prevent or diminish a protective immune response to other epitopes on the antigen of interest (e.g. HIV gp 120).

One solution to such a problem is to selectively mutate the gene which codes, for the protein of interest. However, this approach requires knowledge of the precise sequence of amino acids contributing to the epitope so that the specific corresponding gene sequence can be mutated. As the majority of epitopes recognised by antibodies are discontinuous and rely on their tertiary structure for their effectiveness, such an approach is of very limited application. An alternative approach is provided by EP-A-341444 which aims to identify and characterise epitopes of cell surface proteins such as CD2 and CD4 by first preparing a series of oligonucleotides, covering the substitution of all possible base pairs encoding the extracellular domain of the protein under investigation, separate transformation of each of the oligonucleotide-modified recombinant plasmids in E. coli to yield a series of primary mutants which were used to transfect mammalian cells, followed by selection of the desired expression product. The method is complex and applicable only to cell surface proteins. Winter and Milstein (Nature, 349, 1991, 293-299) propose in vitro imitation of antibody "hypermutation" by site specific genetic engineering techniques. Such methods are comparatively complicated and are directed against predetermined restricted sites in the molecule.

There is therefore clearly a need for a straightforward means whereby a wide range of protein products (e.g. whole native proteins or portions thereof) of recombinant DNA technology can be tailored to make them better adapted to the task for which they are intended.

Accordingly the present invention provides a process for the modification of a protein product of recombinant DNA technology comprising subjecting DNA (suitably cDNA or genomic DNA) coding for the protein product, or coding for at least that part of the protein product to be modified, to random mutation in the presence of at least one chemical or physical mutagenising agent, expression of the mutated DNA in a host organism and screening of the expression products with at least one antibody so as to select the expression product containing the desired modification.

When the random mutation is carried out on DNA coding for less than the complete protein product, for example when coding only for the specific portion of the protein product to be modified, it will be necessary in order to obtain the desired modified protein product to employ genetic engineering techniques to introduce the remaining protein product gene sequence or sequences following the random mutation. Such introduction may be before expression of the DNA in the host organism or even downstream of the screening procedure.

The invention also includes modified proteins obtained by this process and vectors such as plasmids capable of expressing the modified proteins in a host organism.

The starting DNA corresponding to the protein product to be modified may be obtained by standard genetic engineering techniques including screening of cDNA and genomic libraries with ol igonucleotide probes based on partial amino acid sequence, or screening of expression libraries of PCR-amplified DNA using antibodies. The DNA is suitably incorporated in a vector, suitably a carrier plasmid prior to the random mutation step, the plasmid selected being dependent on the DNA sequence to be ligated therewith and the host organism in which the DNA is to be expressed.

The modification desired may be the deletion of an unwanted immunoactive epitope or binding site, in which case the screening of the expression products comprises a two stage selection which involves detecting with a suitable monoclonal antibody the absence of the unwanted site and detecting with a suitable mono- and/or polyclonal antibody the presence of the activity desired in the modified protein product. Alternatively, if a modification such as enhancement of an active site of the protein product is desired, the screening may comprise a single selection step employing a reagent reacting with the active site, such as a ligand, a receptor, or an enzyme substrate.

It will be appreciated that the process of the invention allows one to tailor the protein itself to retain or enhance the epitope of interest while overcoming problems resulting from unwanted sites in the protein. The invention therefore exploits the unique ability of proteins to fold and form epitopes (as distinct from synthetic peptides⁾ but allows one to eliminate unwanted features of the proteins.

The chemical or physical mutagenising agent may be any such agent that is suitable for effecting random mutations in a nucleotide sequence. One suitable example of a chemical agent is hydroxylamine ⁽which is known to induce AT to GC transitions and vice versa) and which can cause mutation directly of a plasmid containing the desired DNA sequence in the absence of a host in which the plasmid is expressed. A typical strategy for use with hydroxylamine as mutagenising agent comprises incubating plasmid containing the cDNA corresponding to the protein to be modified with hydroxyl mine (suitably as its hydrochloride) at a concentration ranging from 0.5 to 2M at 37°C for a time period from 8 to 24 hours, preferably about 18 hours. The DNA is precipitated and the mixed mutated and unmutated plasmids transformed in E . col i grown for a few generations, followed by plating onto the growth media to obtain single colonies, blotting onto a medium such as nitrocellulose and subsequent screening of the expressed proteins with appropriate antibodies. The desired mutated plasmid DNA is recovered from the corresponding bacterial colony and can be used to express the desired modified protein.

An alternative chemical mutagenising agent is N-methyl-N'- nitro-N-nitrosoguanidine (MNNG) which is known to induce transitions and transversions, together with small deletions, at a low rate in the presence of a host in which the plasmid is expressed. A typical strategy for use with MNNG as mutagenising agent comprises growing a host such as E. coli transformed with a plasmid containing the cDNA corresponding to the protein to be modified, exposing an aliquot of recombinant plasmid-containing bacteria to MNNG in a buffer such as sodium acetate, removal of aliquots at suitable time intervals ⁽for example up tc 20 minutes), growth of the individual aliquots through several generations, plating onto growth media-containing agar plates to obtain single colonies and application of suitable selection procedures as for hydroxylamine mutagenesis. Other possible agents include 0-methyl hydroxylamine, nitrous acid and ethyl methane sulphonate, or DNA polymerase used in a polymerase chain reaction (PCR) method. Alternatively, ultra violet irradiation may be used as a physical mutagenising agent.

As described above, the products of mutagenesis (i.e. a mixture of mutated and unmutated plasmids) are expressed in a suitable host. The host may be a well understood bacterial host such as E. coli K12, in which case a known plasmid such as pEX2 (described by Stanley and Luzio, 1984, EMBO 0., 3, 1429), or any other vector which allows the expression of cloned DNA in a host such as £. coli . may be employed to carry the DNA of interest. However, it is a^'lso contemplated that expression may be carried out in other host systems such as yeast cells and mammalian cells.

As described above, the screening of the expression products depends on whether it is desired to produce an "epitope loss" protein product or a "locally modified" protein product. For screening to obtain a desired epitope loss product, following expression for example in E . coli . the bacterial colonies are separated and examined with at least two different antibodies to identify the epitope loss mutants. Those colonies which fail to react with a monoclonal antibody to the epitope to be removed are deemed to have lost that epitope, while ability to react with a polyclonal antiserum and/or a monoclonal antibody to the epitope it is desired to retain is evidence that the desired antlgenic structure of the protein has remained intact.

For screening to obtain a "locally modified" product, for example to improve the affinity of an antibody combining site for a particular antigen, a single screening can be employed, using for example a low concentration of antigen and determining those colonies which give an expression product with a high affinity reaction.

It will be appreciated that any convenient analytical technique known in the art can be employed in the screening procedures, such as expression screening, enzyme-linked im unosorbent assay, Western blotting etc. Where two different antibody reagents are employed, they can each be separately labelled with distinct fluorophors and used as a single mixture, the fluorescence emissions being analysed by appropriate filters.

The above described process is simple and believed to be applicable to any protein produced by recombinant DNA technology without the necessity of knowing the exact sequence or position of the epitopes or local structures to be removed, retained or modified. Iterative mutations can be carried out to sequentially improve upon the desired modification.

Tne process of the invention is intended for a wide variety of uses. These uses include destroying unwanted epitopes in im unogens being investigated as potential vaccines, e.g. bacterial surface antigens, toxins (for toxoid production), parasites, viruses and also contraceptive vaccines based upon modified immunogenic hormone. Further uses include improving antibody affinity, both for intact monoclonal antibodies and genetically engineered (for example single domain) antibodies. This is particularly relevant to human monoclonal antibodies which are difficult to obtain with high affinity. The process of the invention is contemplated for use in improving the efficiency of catalytic antibodies (known as "abzymes") by randomly altering the protein structure and selecting variants with enhanced function or altered specificity, as well as use in refining the specificity of conventional enzymes. Another application is for the identification of amino acids contributing to the active site of epitopes, hormones and cytokines by further sequencing procedures, and use is also contemplated in the development of hormone and cytokine antagonists and antagonists for cell adhesion molecules made by mutating natural ligands, for example for use as novel anti-inflammatory agents. Use of the process is also contemplated as a powerful technique in the identification of contact residues of discontinuous B cell epitopes and protein hormones.

The invention will now be described by way of example with reference to the preparation of epitope loss mutants of the bacterial β-galactosidase protein which have been found no longer to react with a given monoclonal antibody but to retain the ability to bind to polyclonal antisera against the native molecule. Example

A plasmid pEX2 containing the lacZ gene (encoding β-galactosidase) having the structure shown in Figure 1 was utilised (Stanley and Luzio, 1984, EMBO 0., 3, 1429). pEX2 is a plasmid about 5.8kb in length, designed for expression of cDNA fused at the 3' terminus of the lacZ gene. The amino-terminal part of the lacZ gene has been replaced with some sequences from the bacteriophage λ cro gene and the E . col i lad gene. The bacteriophage λ p_R promoter is used, which allows expression of the fusion protein to be regulated by the bacteriophage λ clts857 repressor. A polycloning site is present at the 3' terminus of the lacZ gene, followed by the translation stop codons (Stop) and a transcription terminator (Term) from the bacteriophage fd.

The plasmid was randomly mutated with hydroxylamine hydrochloride (HA) using the following procedure. Twenty μg samples of plasmid DNA in lOOμl lOmM Tris I M EDTA pH 8.0 (TE) were incubated with 2M, 1.5M, 1M or 0.5M HA at 37°C for 18 hrs. After the addition of 0.3M sodium acetate the DNA was precipitated with 2.5 volumes ethanol . Following resuspension in 50μl TE an aliquot of the mutated ⁽and control unmutated) plasmids were inserted into competent E. col i DH5α. The bacteria were then grown for 16 hrs. in order to replicate mutations into both strands of the DNA. Plasmid was then isolated from these bacteria and again used to transform competent DH5α using a ratio of DNA to bacteria which should give only one plasmid per bacterial cell. Transformed cells were spread onto L-Broth (lOg/l Bacto-tryptone, 5g/l yeast extract, lOg/l sodium chloride) agar plates containing 50μg/ml ampicillin. After growth at 30°C the colonies were 'lifted' onto nitrocellulose membranes and the filters duplicated. The filters were then grown on agar plates for 2 hrs. at 42° to induce protein expression. Control (uninduced) filters were maintained at 30°C.

Bacteria on the filters were lysed by incubation for 25 min. at 95°C on 3MM paper soaked in 57. sodium dodecyl sulphate (SDS). The SDS residue was then removed and the protein fixed to the nitrocellulose membrane by electrophoresis for 1 hr. (buffer 0.2M Tris base containing 1.6M glycine). Filters were washed three times in 0.1% Triton X-100, 0.5% gelatin in phosphate buffered saline for 5 min. each, the second wash additionally containing lOμg/ml DNAse.

Prior to reaction with antibodies, the filters were washed in TNT (lOmM Tris pH 8.0 150mM NaCl ; 0.05% Tween 20) for 20 min. and then transferred to 3% bovine serum albumin (BSA) in TNT in order to block remaining protein binding sites on the nitrocellulose. Filters were incubated with pre-determined optimal dilutions of primary antibody (1:1000 - 1:5000 of mouse monoclonal or rabbit polyclonal anti-β-galactosidase) for 150 min. at 20°C in blocking buffer and then washed three times for 10 min. in 0.1% BSA in TNT. The middle wash buffer also contained 0.1% Nonidet P40 (detergent supplied by BDH). Filters were then incubated for 90 min. at 20^βC with a 1:2000 dilution of alkaline phosphatase- conjugated goat anti-mouse IgG or anti-rabbit IgG antibody followed by washing as previously. Reactivity of the antibodies was detected by incubation of the filters with nitroblue tetrazoliu (NBT)/ 5-bromo-4-chloro-3 indolyl phosphate (BCIP) in the dark for 45 min. followed by two washes in double distilled water.

Confirmation of specificity was obtained by the inclusion of irrelevant antibodies such as anti-thyroglobulin. Furthermore, the protein product of the bacteria was analysed by SDS- polyacrylamide gel electrophoresis followed by Western blotting using anti-β-galactosidase antibodies.

There were clearly identified plasmids which no longer had the ability to express protein reacting with the monoclonal antibody and had therefore lost the corresponding epitope, while those same proteins retained the ability to react with anti-β- galactosidase polyclonal antibodies.

Claims

1. A process for the modification of a protein product of recombinant DNA technology comprising subjecting DNA coding for the protein product, or coding for at least that portion of the protein product to be modified, to random mutation in the presence of at least one chemical or physical mutagenising agent, expression of the mutated DNA in a host organism and screening of the expression products with at least one antibody so as to select the expression product containing the desired modification.

2. A process according to claim 1 wherein the DNA subjected to random mutation is cDNA.

3. A process according to claim 1 wherein tne DNA subjected to random mutation is genomic DNA.

4. A process according to any one of claims 1 to 3 wherein the DNA codes for the entire protein product.

5. A process according to any one of claims 1 to 3 wherein the DNA codes for less than the complete protein product and the remaining protein product gene sequence or sequences are introduced following the random mutation.

6. A process according to any one of the preceding claims wherein the DNA is incorporated in a carrier plasmid prior to mutation.

7. A process according to any one of the preceding claims wherein the host organism is E. coli .

8. A process according to any one of the preceding claims wherein the mutagenising agent is hydroxylamine.

9. A process according to any one of claims 1 to 3 wherein the mutagenising agent is N-methyl-N'-nitro-N-nitrosoguanidine.

10. A process according to any one of claims 1 to 7 wherein the mutagenising agent is ultraviolet irradiation.

11. A process according to any one of the preceding claims wherein the modification desired is epitope loss and the screening comprises selection procedures comprising identifying modified proteins which do not react with a monoclonal antibody to the epitope whose loss is desired but do react with a polyclonal and/or monoclonal antibody to the epitope to be retained in the modified protein.

12. A modified protein when obtained by the process of any one of the preceding claims.

13. A vector comprising a mutated DNA coding for a modified protein obtainable by the process of any one of claims 1 to 12.