METHOD FOR IN VITRO MOLECULAR EVOLUTION OF PROTEIN FUNCTION
The present invention relates to a method for in vitro creation of molecular libraries evolution of protein function. Particularly, it relates to variability and modification of protein function by shuffling polynucleotide sequence segments. A protein of desired characteristics can be obtained by incorporating variant peptide regions (variant motifs) into defined peptide regions (scaffold sequence). The variant motifs can be obtained from parent DNA which has been subjected to mutagenesis to create a plurality of differently mutated derivatives thereof or they can be obtained from in vivo sequences. These variant motifs can then be incorporated into a scaffold sequence and the resulting coded protein screened for desired characteristics. This method is ideally used for obtaining antibodies with desired characteristics by isolating individual CDR DNA sequences and incorporating them into a scaffold which may, for example, be from a totally different antibody.
The present invention relates to a method for in vitro molecular evolution of protein function. Particularly, but not exclusively, it relates to the shuffling of polynucleotide sequence segments within a coding sequence.
BACKGROUND OF THE INVENTIONProtein function can be modified and improved in vitro by a variety of methods, including site directed mutagenesis (Moore et al, 1987) combinatorial cloning (Huse et al, 1989; Marks et al, 1992) and random mutagenesis combined with appropriate selection systems (Barbas et al, 1992).
The method of random mutagenesis together with selection has been used in a number of cases to improve protein function and two different strategies exist. Firstly, randomization of the entire gene sequence in combination with the selection of a variant (mutant) protein with the desired characteristics, followed by a new round of random mutagenesis and selection. This method can then be repeated until a protein variant is found which is considered optimal (Moore et al, 1996). Here, the traditional route to introduce mutations is by error prone PCR (Leung et al, 1989) with a mutation rate of ≈0.7%.
Secondly, defined regions of the gene can be mutagenized with degenerate primers, which allows for mutation rates up to 100% (Griffiths et al, 1994; Yang et al, 1995). The higher the mutation rate used, the more limited the region of the gene that can be subjected to mutations.
Random mutation has been used extensively in the field of antibody engineering. In vivo formed antibody genes can be cloned in vitro (Larrick et al, 1989) and random combinations of the genes encoding the variable heavy and light genes can be subjected to selection (Marks et al, 1992). Functional antibody fragments selected can be further improved using random mutagenesis and additional rounds of selections (Hoogenboom et al, 1992).
The strategy of random mutagenesis is followed by selection. Variants with interesting characteristics can be selected and the mutagenized DNA regions from different variants, each with interesting characteristics, are combined into one coding sequence (Yang et al, 1995). This is a multi-step sequential process, and potential synergistic effects of different mutations in different regions can be lost, since they are not subjected to selection in combination. Thus, these two strategies do not include simultaneous mutagenesis of defined regions and selection of a combination of these regions. Another process involves combinatorial pairing of genes which can be used to improve e.g. antibody affinity (Marks et al, 1992). Here, the three CDR-regions in each variable gene are fixed and this technology does not allow for shuffling of individual CDR regions between clones.
Selection of functional proteins from molecular libraries has been revolutionized by the development of the phage display technology (Parmley et al, 1987; McCafferty et al, 1990; Barbas et al, 1991). Here, the phenotype (protein) is directly linked to its corresponding genotype (DNA) and this allows for directly cloning of the genetic material which can then be subjected to further modifications in order to improve protein function. Phage display has been used to clone functional binders from a variety of molecular libraries with up to 1011 transformants in size (Griffiths et al, 1994). Thus, phage display can be used to directly clone functional binders from molecular libraries, and can also be used to improve further the clones originally selected.
Random combination of DNA from different mutated clones is a more efficient way to search through sequence space. The concept of DNA shuffling (Stemmer, 1994) utilises random fragmentation of DNA and assembly of fragments into a functional coding sequence. In this process it is possible to introduce chemically synthesised DNA sequences and in this way target variation to defined places in the gene which DNA sequence is known (Crameri et al, 1995). In theory, it is also possible to shuffle DNA between any clones. However, if the resulting shuffled gene is to be functional with respect to expression and activity, the clones to be shuffled have to be related or even identical with the exception of a low level of random is mutations. DNA shuffling between genetically different clones will generally produce non-functional genes.
SUMMARY OF THE INVENTIONAt its most general the present invention provides a method of obtaining a polynucleotide sequence encoding a protein of desired characteristics comprising the steps of incorporating at least one variant nucleotide region (variant motif) into defined nucleotide regions (scaffold sequence) derived from a parent polynucleotide sequence. The new assembled polynucleotide sequence may then be expressed and the resulting protein screened to determine its characteristics.
The present method allows protein characteristics to be altered by modifying the polynucleotide sequence encoding the protein in a specific manner. This may be achieved by either a) replacing a specified region of the nucleotide sequence with a different nucleotide sequence or b) by mutating the specified region so as to alter the nucleotide sequence. These specified regions (variant motifs) are incorporated within scaffold or framework regions (scaffold sequence) of the original polynucleotide sequence (parent polynucleotide sequence) which when reassembled will encoded a protein of altered characteristics. The characteristics of the encoded protein are altered as a result of the amino acid sequence being changed corresponding to the changes in the coding polynucleotide sequence.
Rather than modifying a sequence at random and then relying on extensive screening for the desired coded protein, the present inventors have found it desirable to provide a method which modifies selected segments (variant motifs) of a protein while maintaining others.
The variant motifs may be segments of nucleotide sequence that encode specified regions of a protein. For example, functional regions of a protein (e.g. loops) or CDR regions in an antibody.
The scaffold sequence may be segments of nucleotide sequence which it is desirable to maintain, for example they may encode more structural regions of the protein, e.g. framework regions in an antibody.
The variant motifs may be nucleotide segments which originated from the same polynucleotide sequence as the scaffold sequence, i.e. the parent polynucleotide sequence, but which have been mutated so as to alter the coding sequence from that in the parent. For example, the parent polynucleotide sequence may encode an antibody. The nucleotide sequences encoding the CDR regions of the antibody (variant motifs) may be selected from the remaining coding sequence of the parent polynucleotide, mutated and then reassembled with scaffold sequence derived from the remaining coding sequence. The expressed antibody will differ from the wild type antibody expressed by the parent polynucleotide in the CDR regions only.
Alternatively, the variant motif may be derived from a polynucleotide sequence encoding a protein sequentially related to the protein encoded by the parent polynucleotide sequence. For example, the CDR regions from one antibody (antibody A) may be replaced by the CDR regions of another antibody (antibody B).
In each case the resulting expressed protein can be screened for desired characteristics. Desirable characteristics may be changes in the biological properties of the protein. For example, the tertiary structure of the protein may be altered. This may affect its binding properties, the ability for it to be secreted from cells or into cells or, for enzymes, its catalytic properties. If the protein is an antibody or part thereof it may be desirable to alter its ability to specifically bind to an antigen or to improve its binding properties in comparison to the parent antibody.
According to one aspect of the present invention, there is provided a method of obtaining a protein of desired characteristics by incorporating variant peptide regions (variant motifs) into defined peptide regions (scaffold sequence), which method comprises the steps of:
(a) subjecting parent polynucleotide sequence encoding one or more protein motifs to mutagenesis to create a plurality of differently mutated derivatives thereof, or obtaining parent polynucleotide encoding a plurality of variant protein motifs of unknown sequence,
(b) providing a plurality of pairs of oligonucleotides, each pair representing spaced-apart locations on the parent polynucleotide sequence bounding an intervening variant protein motif, and using each said pair of oligonucleotides as amplification primers to amplify the intervening motif;
(c) obtaining single-stranded nucleotide sequence from the thus-isolated amplified nucleotide sequence; and
(d) assembling nucleotide sequence encoding a protein by incorporating nucleotide sequences derived from step (c) above with nucleotide sequence encoding scaffold sequence.
The method may further comprise the step of expressing the resulting protein encoded by the assembled nucleotide sequence and screening for desired properties.
Preferably the parent polynucleotide sequence is DNA from which is derived DNA sequences encoding the variant motifs and scaffold sequences.
Preferably the pairs of oligonucleotides are single-stranded oligonucleotide primers. One of said pair may be linked to a member of a specific binding pair (MSBP). The MSBP is preferably biotin, whose specific binding partner could for example be streptavidin. By using the specific binding pair the amplified nucleotide sequences may be isolated.
Random mutation can be accomplished by any conventional method; but a suitable method is error-prone PCR.
The protein in question could, for example, be an antibody or antibody fragment having desirable characteristics. Example of antibody fragments, capable of binding an antigen or other binding partner, are the Fab fragment consisting of the VL, VH, Cl and CHI domains; the Fd fragment consisting of the VH, and CHl domains; the Fv fragment consisting of the VL and VH domains of a single arm of an antibody; the dAb fragment which consists of a VH domain; isolated CDR regions and F(ab′)2 fragments, a bivalent fragment including two Fab fragments linked by a disulphide bridge at the hinge region. Single chain Fv fragments are also included.
In one approach, after randomly mutating DNA encoding the antibody, or a portion of that DNA (eg that which encodes the Fab regions or variable regions), oligonucleotide primers could be synthesised corresponding to sequences bounding the CDRs (the variant motifs), so that DNA encoding the CDRs are amplified, along with any mutations that may have occurred in the CDRs. These can be incorporated in the reassembly of the antibody coding sequence, using the amplified CDR DNA sequences and the unmutated scaffold framework (FR) DNA sequences, resulting in the expression of an antibody which has a novel combination of CDRs, and potentially having altered properties which can be selected or screened for in conventional manner.
In another approach, rather than mutate CDRs and reassembling them back into an antibody which will be closely related to the parent antibody from which the CDRs were derived, the CDRs may be taken from one or more existing antibodies, but be of unknown sequence. Using oligonucleotide primers representing sequences bounding the various CDRs, the individual CDRs can be amplified, isolated and assembled into a predetermined scaffold.
Of course, combinations of the foregoing approaches could be used, with CDRs taken from one or more parent antibodies, and assembled into a scaffold to produce a completely new, secondary antibody, then, after screening to obtain a secondary antibody with desired characteristics, the DNA encoding it could be mutated, the CDRs amplified and isolated, and then reassembled with unmutated non-CDR (scaffold) DNA from the secondary antibody, to produce variants of the secondary antibody which are mutated in the CDRs, and which can be screened for improved properties with respect to the originally selected secondary antibody.
The present invention allows a novel way for the isolation of DNA sequences from genetically related clones that are functionally different. Genetically related clones are those that belong to a particular structural class, for example immunoglobulins or alpha-beta-barrels. The invention allows for both isolation and random combination into a given DNA sequence of functional sequences from these related clones. These functional sequences may be loops that perform binding or catalysis.
The concept of the invention is demonstrated using antibody molecules where CDR-regions from different germline sequences can be isolated and randomly combined into a defined framework sequence. The invention expands the complexity of the molecular libraries that can be selected using phage display. The concept of the invention is also demonstrated by the affinity maturation of antibody fragments by the isolation and random combination of mutated CDR-regions.
It is not possible to use the DNA shuffling concept (Stemmer, 1994) to isolate specific sequences and randomly combine these into a given gene sequence, as it is not possible to amplify individual DNA regions formed in vivo using DNA shuffling. Combination of entire gene sequences is possible, but here defined regions cannot be shuffled. Rather all the DNA is shuffled. Thus, DNA sequences from genetically related clones that are functionally different, eg proteins that belong to structural classes like immunoglobulins or alpha-beta-barrels, cannot be shuffled in such a way that specific regions are kept constant and other regions are shuffled.
The system provided by the present invention offers a simple way to randomly combine functional regions of proteins (eg loops) to a defined (specifically selected) scaffold, ie shuffling of loops to a given protein tertiary structure in order to find new protein functions. Furthermore, the DNA shuffling technology introduces mutations at a rate of 0.7% (Stemmer, 1994). Thus, the known DNA shuffling technology (Stemmer, 1994) does not allow for shuffling of unmutated regions, since the process itself introduces mutations at random positions, including the scaffold regions.
In contrast, the invention allows for mutagenesis of defined DNA-sequences together with shuffling and assembly of these pieces of DNA into a coding region, and will allow for mutagenesis of defined regions and subsequent selection of these regions in combination.
The invention allows for different regions of DNA from different sequences (clones) to be shuffled and randomly combined. This increases the genetic variation from which functional antibody fragments are selected and will thus increase the probability of selecting proteins with the desired characteristics. It can be realised that by randomly shuffling as few as a hundred CDRs at each position in the VH and VL of an fragment, as many as 1012 combinations may be obtained thereby extending the variability normally found in the immune system.
The invention provides amplification of defined regions from eg a cDNA library using two primers of which one is biotinylated. Using the MSBP, e.g. biotin, group, single stranded DNA can be isolated and used in the gene assembly process. The present inventors have demonstrated this with the amplification of diverse CDR regions from an antibody gene library and the combination of these CDR regions randomly to a given framework region. Thus, defined regions of DNA (framework regions) can be interspaced by random regions of DNA (CDR regions), which have an in vivo origin or can be chemically synthesized.
The present invention also provides polynucleotide sequences and the proteins they encoded produced by the method described above. There is also provided vectors incorporating the polynucleotide sequences and host cell transformed by the vectors.
The present invention also provides a polynucleotide library comprising polynucleotides created by the method described above which may be used for phage display.
Aspects and embodiments of the present invention will now be illustrated, by way of example, with reference to the accompanying figures. Further aspects and embodiments will be apparent to those skilled in the art. All documents mentioned in this text are incorporated herein by reference.
Under these improved experimental conditions, essentially all clones had different restriction patterns/fragments sizes. All clones were different from the original scFv-B11 gene (lane 8, gel 1). Moreover, the groups of clones which appeared similar in
In conclusion, these experiments suggest that the library contains high variability.
DETAILED DESCRIPTION AND EXEMPLIFICATION OF THE INVENTIONOne aspect of the DNA shuffling procedure can be illustrated by the following steps in
A: A gene coding for a protein of interest is divided into overlapping oligonucleotides.
B: The oligonucleotides are assembled using PCR into a full length gene-sequence.
C: The gene sequence is subjected to mutagenesis, eg by error-prone PCR.
D: Pairs of oligonucleotides are synthesized, each pair covering a region defined by one of the oligonucleotides in step A above, except for a region located in the middle of the step A oligonucleotide. This uncovered region is the DNA sequence that can be shuffled after PCR amplification. These two synthesised oligonucleotides can thus be used as amplification primers to amplify the uncovered region.
E: One of these amplification primers is biotinylated and the double-stranded PCR product can then be isolated using well-known strepavidin systems.
F: From the thus isolated amplified oligonucleotides can be obtained a single-stranded DNA sequence containing DNA from the uncovered region mentioned above, which can then be used as oligo-nucleotide in a new assembly of the gene sequence as described in step A.
G: If DNA sequences from different clones and from different regions of the mutated gene sequence are amplified and made single-stranded, they will combine randomly in the PCR process of gene assembly. This random combination is the basis for in vitro molecular evolution.
EXAMPLESThe present inventors have demonstrated the concept of shuffling of defined DNA in different experimental settings. Firstly, the shuffling of in vitro mutated CDR regions in an antibody fragment for affinity maturation purposes (example 1 and 2) is exemplified and secondly the shuffling of in vivo formed CDRs for creation of a highly variable antibody library (example 3 and 4) is exemplified.
1. Affinity MaturationA model system was developed, based on the scFv-B11 antibody fragment which binds to FITC. The full-length gene encoding this scFv was assembled from a set of 12 oligonucleotides (
All six CDR regions were amplified and a new gene was assembled using six oligonucleotides selected from the first assembly of 12 oligonucleotide (see above) (these were not mutagenized) and six from the amplification of mutagenized CDR regions. Selection of functional antibody fragments that bound FITC was carried out using phage display. 50% of the clones bound FITC with different dissociation-rates than did the original scFv-B11, as measured in the BIAcore biosensor (
Of the 16 clones identified to bind FITC in BIAcore (
Clone #11 exhibited an affinity 2.8 times higher than the original scFv-B11 antibody fragment. This increase is based on a slower off-rate. One clone (#27) showed 2 times increase in association-rate. However, the overall affinity of this clone was similar to the original FITC-B11 clone due to a faster dissociation-rate. The distribution of different association and dissociation-rates among the clones was considered a source for CDR-reshuffling for further improvement of affinities.
Three clones were sequenced. In the VH region (ie half of the scFv-B11 and carrying three CDR regions) the mutations found were all in the CDR regions as expected, since these were the only regions mutagenized and amplified using the amplification primers. Interestingly, all the CDR regions were different and carried different mutations (
Furthermore, the mutation rates were found to be in between 2% and 4%, as determined from the base changes in the 90 bp long sequence built up from three CDR regions together. This is more than the error-prone PCR mutation rate, and indicates that there is combination of individual CDR regions from different clones.
2. Affinity Maturation-ReshufflingIn order to perform a second shuffling (reshuffling), clones selected for their binding affinity to FITC were used in an additional round of CDR-amplification and library construction. In theory, the reshuffled library will contain mutated shuffled CDR-regions, selected for improved binding to FITC. In this way, new combinations of CDR-regions, improved with respect to binding, could be constructed and the library subjected to selection for binders with improved affinities.
The pool of all clones obtained from the selection procedure (as detailed in example 1) were used as template for CDR amplifications. One amplification was carried out for each CDR using primers listed in Table 2.
The amplification was performed according to following parameters: 100 ng template (1.6×108 CFU bacteria grown for 6 h), 60 pmol each primer, 5 Units PFU polymerase (Stratagene), 1×PFU buffer, 500 μM dNTPs, reaction volume 100 μl, preheat 96° C. for 10 minutes, 96° C. for 1 minute: 68° C. for 1 minute: 72° C. for 1 minute for 25 cycles, 72° C. for 10 minutes. This procedure was essentially the same as for CDR amplification in Example 1. The amplified CDR were used for assembly into VH and VL encoding sequence according to
The VH and VL were then assembled into a scFv encoding sequence according to standard procedures (Griffiths et al 1994). The resulting library was subjected to panning so as to select binders with improved affinities to FITC. The selection procedure for the reshuffled library was essentially the same as for the initially shuffled library. The total number of clones obtained after selection was 510. Six clones (B) were chosen from this new pool and were tested and compared to 6 clones (A) from the first pool, originating from the shuffled library (Table 4).
Two clones from the reshuffling experiments (22B and 31B) exhibited substantially slower dissociation-rates, indicating that the reshuffling process yielded binders with improved affinities.
3. Cloning and Shuffling of Defined DNA RegionsIn our system it is possible to amplify defined regions from a cDNA library using two primers of which one is biotinylated. Using the biotin group, single stranded DNA can be isolated an used in the gene assembly process (
A gene library was constructed encoding scFv antibody fragments. The strategy used for this library is based on the assembly of a set of oligonucleotides into a sequence encoding VH and VL antibody domains (
The PCR parameters for CDR and framework region amplification were essentially the same as described in example 2. The PCR parameters for assembly of genes encoding VH and VL are described in Table 7.
The assembled VH and VL gene sequences were assembled into a scFv coding sequence using standard protocols (Griffiths et al 1994). A library of 1.1×109 members were constructed out of the 40 clones tested all 40 contained an insert of the right size as determined by PCR agarose gel electrophoresis. In order to test the variability in the library, PCR amplified and purified inserts were subjected to cleavage by BsTN1 and BamH1. Clones showed different restriction patterns, as determined by agarose gel electrophoresis and compared to the control scFv-B11 (
In order to estimate the frequency of clones able to express scFv antibody fragments, clones from the library containing the FLAG sequence (Hopp et al, 1989), as well as control bacteria with and without FLAG sequence, were plated at low density on Luria broth-plates containing 100 μg/ml ampicillin, 25 μg/ml tetracycline and 1% glucose. The plates were grown at 37° C. over night and lifted to nitrocellulose filters by standard methods (Sambrook et al 1989). In order to induce synthesis of the scFv genes in the bacteria, filters were incubated for 4 hrs on plates containing 0.5 mM isopropyl-thio-β-D-galactoside (IPTG) but without glucose. Bacteria were then lysed by lyzosyme/chloroform treatment, the filters were washed and incubated with anti-FLAG M2 antibody (Kodak) followed by anti-mouse peroxidase conjugated second antibody (P260 Dakopatts) and detected by DAB 3,3′-diaminobenzidine tetrahydroklorid, Sigma) (Table 8).
The anti-FLAG antibody detects a FLAG sequence situated downstream of the scFv gene in the library constructs as well as in the control vector pFAS5cHis scFvB11, but not in the original vector pFAB5cHis. Clones, to which the anti-FLAG antibody binds, therefore contains an intact open reading frame of the scFv gene.
REFERENCES
- Barbas, C F et al: Proc Natl Acad Sci USA, 88:7978-82 (1991)
- Barbas, C F et al: Proc Natl Acad Sci USA, 89:4457-61 (1992)
- Crameri, A et al: Biotechniques, 18:194-196 (1995)
- Griffiths, A D et al: EMBO J, 13:3245-3260 (1994)
- Hoogenboom, H R et al: J Mol Biol, 227:381-8 (1992)
- Hopp, T. P. et al: Biotechniques 7: 580-589 (1989)
- Huse, W D et al: Science, 246:1275-81 (1989)
- Larrick, J W et al: Biochem Biophys Res Commun, 160:1250-6 (1989)
- Leung, D W et al: Technique, 1:11-15 (1989)
- Marks, J D et al: Biotechnology, 10:779-83 (1992)
- McCafferty, J et al: Nature, 348:552-4 (1990)
- Moore, J C et al: Nature Biotechnology, 14:458-467 (1996)
- Parmley, S F et al: Gene, 73:305-318 (1988)
- Roberts, S et al: Nature, 328:731-4 (1987)
- Sambrook, J et al: Molecular cloning. A laboratory manual. Cold spring Harbor Laboratory Press 1989.
- Stemmer, W P: Nature, 370:389-391 (1994)
- Yang, W P et al: J Mol Biol, 254:392-403 (1995)
Claims
1-55. (canceled)
56. A method of generating an assembled polynucleotide sequence encoding a protein of desired characteristics comprising the steps of:
- a) providing at least one polynucleotide sequence comprising one or more variant polynucleotide sequences encoding one or more in vivo formed variant protein motifs;
- b) providing one or more pairs of defined oligonucleotides, each pair representing spaced apart locations on the at least one polynucleotide sequence of step a), each pair binding adjacent to a variant polynucleotide sequence encoding an in vivo formed variant protein motif;
- c) using the pairs of defined oligonucleotides as amplification primers for PCR to amplify the variant polynucleotide sequences encoding the in vivo formed variant protein motifs of the at least one polynucleotide sequence of step a) and performing PCR amplification on the at least one polynucleotide sequences of step a);
- d) obtaining one or more single-stranded polynucleotide sequences from the amplified polynucleotide sequences in step c);
- e) providing one or more unmutated, specifically selected, scaffold polynucleotide sequences encoding one or more mutated peptide regions; and
- f) annealing said one or more single-stranded polynucleotide sequences from step d) with the unmutated, specifically selected, scaffold polynucleotide sequences from step e) such that annealed polynucleotides with one or more gaps is formed, and filling the one or more gaps present in the annealed polynucleotides, thereby generating one or more assembled polynucleotide sequences.
Type: Application
Filed: Apr 13, 2007
Publication Date: May 8, 2008
Inventors: Ulf Hans Eskil Soderlind (Lund), Carl Arne Krister Borrebaeck (Lund)
Application Number: 11/735,213
International Classification: C40B 10/00 (20060101); C07H 21/00 (20060101); C07H 21/04 (20060101); C12N 15/63 (20060101); C12N 1/00 (20060101); C12P 21/00 (20060101); C40B 40/08 (20060101); C07K 16/00 (20060101); C40B 50/06 (20060101); C40B 30/00 (20060101);