WO1999025871A1 - Methods and means for mutagenesis of dna - Google Patents

Abstract

Methods and means for mutagenesis of DNA, particularly site-directed mutagenesis or replacement of DNA segments. Nucleic acid amplification, such as PCR, is used in production of copies of template DNA including one or more mutations. A parental DNA molecule includes template DNA provided within a circular vector. A pair of initial primers is used to generate further primers which are then used to copy the entire parental DNA molecule including template and vector into a form including the desired mutation or mutations. Where only one mutation is to be introduced, only one of the initial primers need be mutagenic. Two distal mutations may be introduced at the same time using two mutagenic primers. The parental DNA including original template is removed by digestion. Parental DNA may be methylated to allow for its digestion with a suitable enzyme, such as DpnI, which does not digest newly synthesized (unmethylated) product DNA. Means for performing the methods are amenable to provision in kit form.

Description

METHODS AND MEANS FOR MUTAGENESIS OF DNA

The present invention relates to mutagenesis of DNA, particularly site-directed mutagenesis or replacement of DNA segments. It relates more specifically to use of nucleic acid amplification, such as PCR, in production of copies of template DNA including one or more mutations, with original template being removed, for instance by digestion. Various methods are disclosed, as are the means for performing the methods, amenable to provision in kit form.

Stratagene™ (LaJolla, California, USA) make and sell under the trademark "QuikChange" a site-directed mutagenesis kit. Details are disclosed in the Stratagene "Strategies" Newsletter, 1996, Volume 9, Number 1, pages 33-35, and the kit has been subject to extensive marketing. An exemplary advertisement is found on the reverse of the title page of Nature Volume 386, issue 6627 dated 24 April 1997. The kit is available from Stratagene under their catalogue number 200518, and includes an instruction manual. Further disclosure is contained in the International Patent Application PCT/US96/19387 , published as O97/20950 (Stratagene and Children's Medical Center Corporation).

To perform the QuikChange approach requires the following:

(i) provide a selected nucleotide sequence as a template to be mutated in a double-stranded DNA plasmid; (ii) denature the plasmid and anneal a pair of mutagenic primers (see below) ;

(iii) synthesize first and second mutagenized strands, each incorporating a member of the mutagenic primer pair, by a linear cyclic amplification reaction sufficient to produce a complete copy of the plasmid;

(iv) digest the template strands;

(v) anneal mutagenized strands to form a double-stranded circular DNA intermediate including a double-strand nick;

(vi) ligate to form closed circular double-stranded DNA containing the desired mutation, preferably by transformation of a competent host cell which repairs the nick.

The Stratagene literature indicates with emphasis that this does not involve use of PCR.

Moreover, the patent application and the instruction manual repeatedly emphasize the essential use of a pair of primers which must be either completely complementary or partially complementary to each other, wherein the mutation site is located within the region of complementarity of both mutagenic primers. See for instance page 4 of O97/20950, under the heading "Summary of the Invention", and page 4 of the QuikChange instruction manual (Catalogue number 200518, Revision 116003), which states: "Both the mutagenic primers must contain the desired mutation and anneal to the same sequence on opposite strands of the plasmid" . The same manual also states that "Primers need not be 5' phosphorylated but must be purified either by fast polynucleotide liquid chromatography (FPLC) or by polyacrylamide gel electrophoresis (PAGE) " . (The emphasis appears in the manual . )

To introduce more than one mutation using the QuikChange kit requires that the method be repeated for each mutation, with a separate pair of mutagenic primers for each mutation, or that, if the location of the desired mutations allows, both or all sites of mutation are included within a single pair of mutagenic primers.

Provision of highly purified pairs of oligonucleotide primers is relatively expensive. Companies such as Stratagene offer custom oligonucleotide synthesis, with current costs of around £40-£50 per oligonucleotide of about 25-30, rising to £100-150 for custom oligonucleotides . Clearly, where it is desired to introduce a number of different mutations which cannot be included within a single pair of oligonucleotides, costs will quickly mount up. Furthermore, having to repeat the process to introduce a second mutation distal to a first essentially doubles the time spent. The present inventor has realised that the QuikChange™ approach can be simplified in a way which may save time and/or cost. Furthermore, it can be adapted for extensive "domain-swapping" mutations of a kind not envisaged in the available literature.

According to a first aspect of the present invention there is provided a method for production of a mutated form of a DNA template sequence (i.e. including one or more desired mutations), the method including:

providing a DNA vector including the DNA template sequence, the vector and template together forming a parental DNA molecule which includes at least a first strand and may be double-stranded including a second strand, wherein if the parental DNA molecule is single-stranded a second strand including sequence complement to at least the template sequence and optionally all or part of the vector is additionally provided;

separating the first and second strands;

annealing a first primer to a sequence within the template sequence in the first or second strand, the first primer including a first desired mutation (i.e. being mutagenic);

annealing a second primer to a sequence in the opposite strand (second or first respectively) , the sequence to which the second primer anneals either being within the template sequence or being within the vector outside the template sequence;

the first and second primers being suitable for performance of a nucleic acid amplification reaction to copy a portion of each of the first and second strands including a portion of the template sequence to provide a first and a second product primer (which may be a "megaprimer" ) each incorporating a copied portion of the template sequence and a first or second primer sequence respectively;

performing a nucleic acid amplification reaction on the parental DNA molecule using the first and second primers to produce said first and second product primers;

performing a second nucleic acid amplification reaction on the parental DNA molecule using the first and second product primers to produce first and second product strands including the template sequence in mutated form (i.e. including the desired mutation or mutations) and vector DNA;

digesting the strands of the parental DNA molecule to leave the first and second product strands.

Where the second primer anneals to a sequence within the template, it may include a second desired mutation (i.e. be mutagenic) . Thus, one or more mutations may be introduced into the template sequence at one or more sites within the sequence, with it being possible to make mutations at two sites in one reaction. The invention is applicable for mutation of any sequence, subject to limits of efficiency of amplification which may be imposed by template size.

The present invention is also useful in making mutations of the "domain swapping" kind, wherein a region within a larger sequence of one form is replaced with a corresponding region from within another related form (e.g. allele) , which region has a different sequence. Examples where sequences exist in related forms include genes with related sequences from different species, different alleles of a given gene in one species, such as MHC molecules, antibodies, T cell receptors, mutated sequences associated with a functional alteration such as associated with one or more disease states or predisposition to one or more disease states, different genes from a species which have arisen from gene duplication, such as in the MHC, actin and so on. Examples of instances where "domain swapping" is useful or desirable include any of these and others, especially where sequence difference is reflected in a difference in function or activity.

According to a further aspect of the present invention there is provided a method of substituting a segment of a first form of a DNA template sequence for a segment of a second form of the template sequence, the method including: providing a DNA vector including the first form of the DNA template sequence, the vector and template together forming a first parental DNA molecule which includes at least a first strand and may be double-stranded including a second strand, wherein if the parental DNA molecule is single-stranded a second strand complement to at least the template sequence and optionally part or all of the vector is additionally provided; separating the first and second strands;

annealing a first primer to a sequence within the first form of the template sequence in the first or second strand, the first primer being suitable for annealing to a corresponding sequence within the second form of the template sequence;

annealing a second primer to a sequence in the opposite (second or first respectively) strand, the sequence to which the second primer anneals either being within the template sequence or being within the vector outside the template sequence, wherein if the second primer anneals to a sequence within the template sequence the second primer is suitable for annealing to a corresponding sequence within the second form of the template sequence;

the first and second primers being suitable for performance of a nucleic acid amplification reaction to copy a portion of each of the first and second strands consisting of said segment of the first form of the template sequence to provide a first and a second product primer (which may be a "megaprimer" ) each incorporating a said copied portion of the template sequence and a first or second primer sequence respectively;

performing a nucleic acid amplification reaction on the first parental DNA molecule using the first and second primers to produce said first and second product primers;

purifying the first and second product primers;

providing a DNA vector including the second form of the DNA template sequence, which vector and second form of the template together form a second parental DNA molecule which includes at least a first strand and may be double-stranded including a second strand, wherein if the second parental DNA molecule is single-stranded a second complementary strand of at least the template sequence is additionally provided;

performing a second nucleic acid amplification reaction on the second parental DNA molecule using the first and second product primers to copy portions of the strands of the second parental DNA molecule including sequence variation to produce first and second product strands including a hybrid form of the template sequence wherein a segment of the first form is substituted for a segment of the second form;

digesting the strands of the second parental DNA molecule to leave the first and second product strands.

Thus, the segment of the first form replaces the corresponding segment of the second form, producing a hybrid sequence .

The vector forming part of the second parental DNA molecule may be the same as the vector forming part of the first parental DNA molecule. Where the second primer is designed to anneal to a sequence within the first vector outside the template, the second primer should be suitable for annealing to a corresponding sequence within the second vector outside the template.

In either of the above aspects of the present invention, one of the first and second DNA product strands is a complete copy of the DNA parent apart from inclusion of the mutation or mutations, or the replacement segment where applicable, and the other product strand is complementary. These newly synthesized mutagenized strands are the desired product. Digestion of the parental template strands allows for efficient recovery, purification and/or use of the double- stranded product .

Following digesting of the parental DNA the product strands generally anneal to form a double-stranded DNA molecule including a "stepped" double-strand nick. A ligation step may be employed to provide a closed circular DNA molecule, as discussed further below, though such ligation may not be required where the product copy of the template can be cloned directly out of the vector DNA even in linear form. Preferably, however, ligation takes place within competent transformed cells (see below) .

Whether a ligation step is utilised or not, the mutated form of the template insert may be manipulated for further use and/or investigation. For example it may be subjected to an amplification reaction, such as PCR. It may be used in production of an encoded polypeptide product by expression.

Nucleic acid amplification is generally performed using the polymerase chain reaction (PCR) . PCR techniques for the amplification of nucleic acid are described in US Patent No.

4,683,195, with references for the general use of PCR techniques including Mullis et al, Cold Spring Harbor Symp .

Quant. Biol., 51:263, (1987), Ehrlich (ed) , PCR technology, Stockton Press, NY, 1989, Ehrlich et al , Science, 252:1643-

1650, (1991) , "PCR protocols; A Guide to Methods and

Applications", Eds. Innis et al , Academic Press, New York,

(1990) .

An oligonucleotide for use in nucleic acid amplification may have about 15-50 nucleotides, preferably 20-30 or about 25. Those skilled in the art are well versed in the design of primers for use in processes such as PCR. Various techniques for synthesizing oligonucleotide primers are well known in the art, including phosphotriester and phosphodiester synthesis methods.

An oligonucleotide primer anneals to or hybridizes with a complementary sequence. A primer may be referred to as being "specific for" a particular sequence, which may be understood as referring to sequence complementarity. A primer is "specific for" a sequence, e.g. on the template nucleic acid, when it is complementary to that sequence or sufficiently complementary to bind that sequence in preference to different sequences under appropriate reaction conditions.

Where a primer is mutagenic it includes at least one nucleotide that differs from the sequence in the parent molecule to which it anneals under the reaction conditions, so that the relevant product strand resulting from performance of the method incorporates that difference as a mutation. Within an oligonucleotide of sufficient length more than one nucleotide substitution may be included without unduly affecting ability of the oligonucleotide to hybridize with parental nucleic acid. One or more point mutations may be included, one or more small substitutions of short lengths of nucleotides, e.g. triplets such as one or more codons within a peptide- or polypeptide-encoding sequence, one or more additions and/or one or more deletions of one or more nucleotides . As is well known, PCR cycles may be designed for particular contexts, depending on the DNA polymerase used, the length and nature (e.g. GC content) of the template, the amount of DNA available, and so on, adjusting temperatures and durations of the denaturation, annealing and extension steps accordingly. As noted in Example 2 below, it may be preferred to provide conditions wherein the annealing and extension steps are combined, particularly in the second PCR stage of the "domain swapping" aspect of the present invention. Temperature ranges and cycle times depend on the enzyme used, but for pfu typical conditions include elongation at 60-68°C, about 2 min/kb of sequence, 10-30 cycles, depending on how much parental template is used.

Any suitable DNA polymerase may be employed. It is preferred that the DNA polymerase is thermostable. It is preferred that the DNA polymerase does not displace primers annealed to the template, facilitating production of mutated strands corresponding in length to the entire strands of the starting material (vector and template) . Taq polymerase may be used but is not preferred because of its relatively high error rate. A higher fidelity polymerase is preferred, such as Pfu polymerase (available from Stratagene)

References for use of "megaprimers" include Barik, S. (1995) Molecular Biotechnology 3(1) : 1-7; Barik et al . , (1991) Bioteciiπigues 10(4): 489-490; Ke et al . , (1997) Nucleic Acids Research 25(16): 3371-3372; Sarkar et al . , (1990) Analytical Biochemistry 186(1): 64-68; Sarkar et al . , (1990) Bi otechniques 8(4) : 404-407.

The DNA molecule which serves as the parent template for performance of the present invention as indicated may be provided initially as single-stranded molecule; the second complementary strand may be provided by means of a nucleic acid amplification reaction such as PCR. This may be an in vitro elongation using a suitable polymerase. Provision of a double-stranded template in a vector as indicated above as a first step in the disclosed methods of the invention may be preceded by an initial step wherein the second strand is provided by elongation from the appropriate primer included in the reaction which then progresses to copying the then double-stranded portion of the template as disclosed. A second strand complement to a single-stranded template may be provided by elongation from a first or second primer (which may be mutagenic) as indicated. If the elongation is for long enough the second strand complement will include vector sequence .

The parent template may be cloned into any suitable vector at an appropriate cloning site . There should be at least one site in the parent template or vector recognised by the enzyme chosen for digestion, e.g. Dpnl (occurs statistically every 256 nucleotides) . Following production of newly synthesized mutagenized DNA product strands, parental template and vector DNA is digested using an appropriate enzyme. Appropriate enzymes for use in this context do not or do not significantly digest newly synthesized DNA. Discriminatory digestion may be as a result of modification to the parental DNA or modification to the newly synthesized DNA. Modification, e.g. of the parent DNA, may take place before synthesis of the mutated copies. For example, in preferred embodiments of the present invention, the modification is methylation of parental DNA. Such methylation may be by means of an enzyme such as dam methylase, dcm methyiase Alu I methylase and so on. Such enzymes are available from New England Biolabs and others. Preferably, where the modification is methylation of parental DNA this is achieved by propagation of cells containing the parental DNA, which cells produce a suitable methylase. Prokaryotic cells are particularly suitable, containing dam methylase. Newly synthesized DNA will not be methylated, allowing for discriminatory digestion of parent DNA using an enzyme such as a restriction endonuclease which digests methylated but not unmethylated DNA, for instance Dpnl . Another possibility is the modification of parental strands by incorporation of uridine residues in place of thymidine . Uridine-containing strands are sensitive to uracil-N- glycosylase. Alternatively, newly-synthesized DNA may be modified to prevent its digestion. Suitable modifications include incorporation of methylated nucleotides, where the parent nucleic acid is unmethylated, for employment with digestion enzymes such as Bell, DpnlI and others which are blocked by overlapping Dam methylation, or Apal , EcoRII and others which are blocked by overlapping Dcm methylation. Further examples are listed on the New England Biolabs 1997 Catalogue in a reference appendix discussing Dam and Dcm methylases .

Following digestion of parental DNA, mutagenized DNA strands anneal under suitable conditions, the design and practice of which are well within the capability of those of ordinary skill in the art, taking into account, for instance, temperature, pH, length of DNA molecules, GC content and so on. The conditions may be such as to allow digestion of parental template and annealing of newly synthesized mutagenized strands to take place at the same time.

The resulting double-stranded mutagenized DNA may include a double-strand nick, requiring ligation of the nucleotide added last to the elongating strand during synthesis to the 5' end of the strand (the 5' nucleotide of the relevant primer as incorporated in the mutated strand) . This ligation may be performed by addition to a reaction medium of a suitable ligase, such as T4 DNA ligase. However, it is preferred that ligation be effected by means of transformation of competent cells, preferably prokaryotic cells such as E. coli , which will repair the nicks.

Transformation may employ any available technique. For eukaryotic cells, suitable techniques may include calcium phosphate transfection, DEAE-Dextran, electroporation, liposome-mediated transfection and transduction using retrovirus or other virus, e.g. vaccinia or, for insect cells, baculovirus . For bacterial cells, suitable techniques may include calcium chloride transformation, electroporation and transfection using bacteriophage . As an alternative, direct injection of the nucleic acid could be employed.

Many known techniques for manipulation of DNA, including transformation of cells, are described in "Molecular Cloning: a Laboratory Manual", 2nd edition, Sambrook et al . , 1989, Cold Spring Harbor Laboratory Press, and Current Protocols in Molecular Biology, Ausubel et al . eds . , John Wiley & Sons, 1992.

The mutated DNA and/or the vector containing it may be recovered from cells using any technique available in the art .

The mutated DNA may be excised from the vector, e.g. using suitable restriction endonucleases . It may be amplified, e.g. using PCR. It may be subject to further mutagenesis, whether involving use of an aspect of the present invention or not. It may be ligated with additional nucleotides, e.g. within an expression vector for production of a transcript and/or an encoded peptide or polypeptide product . A coding sequence may be used in production of the encoded peptide or polypeptide. Transcription may be used in control of gene expression, e.g. using anti-sense or sense regulation (the latter also known as co-suppression) . The transcription product may be a ribozyme, the mutation for example altering its specificity, that its recognition sequence. The invention may be used in mutating vector nucleic acid (e.g. plasmid, cosmid, phage, phagemid) to add, delete or alter one or more restriction sites or other sequences useful in gene cloning and expression.

Reference here to "the mutated DNA" refers not only to the actual DNA molecules created in the steps disclosed above, but also copies created for example using PCR or by propagation in a host such as a bacterium, e.g. E. coli . Thus, one or more intermediate copying steps may be included between mutation of the sequence and downstream use of the mutated sequence .

Reagents for performance of various aspects of the present invention may be provided in kit form. A kit in accordance with the present invention may include instructions for performance of a method disclosed herein. For use in "domain-swapping" , a kit in accordance with the present application may include material for purification of amplified product primers, such as a silica-based purification matrix, e.g. one of those sold by Qiagen or BiolOl. Further aspects and embodiments of the present invention will be apparent to those skilled in the art, as will modifications to the aspects and embodiments specifically disclosed herein.

The present invention will now be illustrated further by means of experimental exemplification, with reference to the following figures:

Figure 1 illustrates performance of a method in accordance with one embodiment of the present invention, making one or two mutations to a parental template sequence within a plasmid vector.

Figure 2 illustrates performance of a method in accordance with another embodiment of the present invention, where an internal portion of a coding sequence in one allelic form of a gene is replaced with a corresponding internal portion with a different sequence from another allelic form of the gene.

EXAMPLE 1 - Si te-specific Mutation of a Sequence

The first step of Figure 1 illustrates two possibilities, depending on whether it is desired to make mutations at one or more than one site within the template sequence. A double-stranded DNA molecule (1) is provided as a double- stranded DNA vector (2) incorporating the template sequence (3) , which may be a coding sequence for a polypeptide of interest .

When it is desired to make a mutation at one site within the template, a first primer (4) is employed which anneals to a sequence in one strand ("the first strand") of the DNA molecule (1) within the template sequence (3) and which includes the desired mutation, and a second primer (5) is employed which anneals to a sequence in the other strand ("the second strand") of the DNA molecule (1), usually outside the template (3) to a sequence of the vector (2) . A primer (5) which anneals to a sequence in the vector (2) may be utilised in making a mutation to any sequence (3) cloned into the vector (2) at a relevant cloning site. It may thus be thought of as a "standard" or "common" primer, and need not be produced specially for each mutation. This saves on cost and time in production over the "QuikChange" approach where two mutagenic primers have to be synthesized and purified for each and every mutation to be made.

Alternatively, when it is desired to use the present invention to make simultaneously mutations at two sites within the template, in conjunction with a first primer (4) which anneals to a sequence in one strand (the first strand) of the DNA molecule (1) to a sequence within the template sequence (3) and which includes a first desired mutation, instead of second primer (5) a second primer (6) is employed which anneals to a sequence in the other strand (the second strand) of the DNA molecule (1) within the template sequence (3) and which includes a second desired mutation.

The primers (4, and 5 or 6) are designed and oriented when annealed to the parental DNA so that performance of a nucleic acid amplification reaction (e.g. PCR) produces copies of part of the template sequence, incorporating the primer sequences at respective ends (the molecules so produced being termed herein "product primers", a complementary pair (7,8) as discussed below) . This first amplification reaction is illustrated in the second step of Figure 1. Exemplary conditions include 9 cycles of denaturation at 95°C for 30 seconds, annealing at 55°C for 1 minute, and elongation at 68 °C for 1 minute. Incorporation of the first primer (4) and its complement at an end of the nucleic acid produced in the amplification reaction incorporates the first mutation. Where a second primer (6) including a second mutation is employed, incorporation of the second primer (6) and its complement at the other end of the nucleic acid produced in the amplification reaction incorporates the second mutation.

The product primers (7,8) produced in the first amplification reaction are used to prime a second amplification reaction in which the remainder of the entire DNA molecule (1) including vector (2) and template (3) is copied. This is illustrated in the third step of Figure 1. The product primers (7,8) may be isolated and purified between the two amplification reactions, but this is avoidable by annealing the product primers (7,8) to the DNA molecule (1) for the second amplification reaction at a higher temperature, e.g. the elongation temperature (such as around 68°C) . At such a higher temperature, the initial short oligonucleotide primers will not anneal to the parent sequence, but the longer product primers will. This favours synthesis of the remainder of the circular parental strands, the resulting products incorporating the mutagenic oligonucleotide (s) . Suitable conditions include 9 cycles of annealing and elongation at 68 °C for 6 minutes preceded in each cycle by denaturation at 95°C for 30 seconds.

Once complete copies of the entire strands of the double- stranded DNA molecule (1) have been made (9,10), including the mutation or mutations, the original DNA molecule (1) is digested, the newly synthesized strands (9,10) annealed together and the ends (11,12) ligated. In the exemplary embodiment, the original DNA molecule which is methylated is digested using Dpnl , which only cuts methylated DNA so the newly-synthesized strands (9,10) carrying the mutation or mutations are left unscathed. Following annealing, repair of the resulting double-strand nick is conveniently achieved by transformation of the annealed newly-synthesized strands into bacteria such as E. coli .

Efficiency of mutagenesis in this example is 70-95%.

This approach has been proved experimentally as follows Green fluorescent protein (GFP) used as an reporter molecule in live cells has allowed tremendous progress to be made in many aspects of biology over the past few years. In addition, high hopes are based on the existence of GFP variants that have distinguishable emission spectra. Of these, BFP is the most useful, but because of its weaker fluorescence, it has so far not been broadly used in mammalian cells.

To obtain an optimised form of BFP for use as a reporter gene in mammalian cells, the brightest available GFP form, EGFP, was mutated at 5 different positions, yielding 8 different mutagenised forms of BFP, in accordance with an embodiment of the present invention, and all the mutants obtained were analysed by flow cytometry of transiently transfected COS 7 cells. The brightest mutant obtained was at least 2 fold brighter than any commercially available form of BFP. This mutagenised form of BFP can be readily detected, even when expressed together with GFP, both by flow cytometry and fluorescence microscopy, and bleaching is no longer a major problem.

Over the past few years, Aequorea victoria ' s green fluorescent protein (GFP) has proven an invaluable tool to study gene expression, protein localisation and cellular compartment dynamics. In addition, the production of variants with altered spectral characteristics has opened the door to the detection of several signals in parallel. Of the variants reported so far, the most useful are derived from the Y66H mutation which leads to emission of blue fluorescence (Heim et al . , Proceedings Of the National Academy of Sciences of the Uni ted States of America 1994, 91: 12501-12504; Heim et al . , Current Biology 1996, 6: 178-182; Rizzuto et al . , Current Biology 1996, 6: 183-188) . The major drawbacks of blue fluorescent protein (BFP) are that its fluorescence is much weaker than that of GFP and the UV required for excitation results in rapid bleaching in fluorescence microscopy. So far, BFP has not been very widely used in mammalian cells as a reporter gene mainly because of its weak signal .

Green fluorescent protein derivatives have been expressed efficiently in many organisms by various modifications. Firstly, efficient protein expression can be affected if the nucleic acid sequence shows a bias towards codons which are rarely used by the organism, as was demonstrated by Haas et al . on expression of HIV-1 glycoprotein, Thy-1 and GFP in human cells (Haas et al . , Current Biology 1996, 6: 315-324) Several FP sequences have now been reported where the whole nucleotide sequence has been synthetically remodelled to accommodate the codon usage of mammalian cells (Haas et al . , Current Biology 1996, 6: 315-324; Zolotukhin et al . , Journal of Virology 1996, 70: 4646-4653; Yang et al . , Nucleic Acids Research 1996, 24: 4592-4593; Zhang et al . , Biochemical and Biophysical Research Communications 1996, 227: 707-711).. Secondly, several mutations resulting in an increase of the protein's stability at 37 ^*C have been reported by different groups (Crameri et al . , Na ture Biotechnology 1996, 14:315-319; Siemering et al . , Current Biology 1996, 6:1653-1663; Cormack et al . , Gene 1996, 173:33-38; Kimata et al . , Biochemical and Biophysical Research Communications 1997, 232:69-73). Indeed, GFP is naturally expressed in sea water environment, which is usually much cooler than 37 ^*C.

In addition to the Y66H mutation necessary for blue fluorescence, 4 other substitutions appeared as strong candidates for improving expression of BFP in mammalian cells. Firstly, Y145F, was reported to double the quantum yield of BFP compared to Y66H alone (Heim et al . , Current Biology 1996, 6:178-18). Secondly, the reversal of Thr 65 to the original Ser found in the natural form of GFP; the S65T mutation in GFP results in a brighter fluorescence and in a red shift of the absorption spectrum which makes it more compatible with the widely available optical filters used with fluorescein (Cormack et al . , Gene 1996, 173:33-38; Kimata et al . , Biochemical and Biophysical Research Communications 1997, 232:69-73; Heim et al . , Nature 1995, 373:663-66), but it was unclear how this mutation would affect the spectrum of BFP. Finally, two mutations, V164A and S175G, were reported to increase the stability of bacterially produced GFP at 37^*C (Siemering et al . , Current Biology 1996, 6:1653-1663). These substitutions have been reported independently from one another, and mainly in bacterially produced GFP, and it has not previously been proven that they would all be beneficial for a blue fluorescent protein synthesised in mammalian cells.

To obtain an optimised version of BFP to use in parallel with GFP in mammalian cells, a commercially available version of GFP called EGFP (Clontech) was chosen. This is reputedly the brightest version of GFP available for expression in mammalian cells (Yang et al . , Nucleic Acids Research 1996, 24:4592-4593; Zhang et al . , Biochemical and Biophysical Research Communications 1996, 227 :707^'-711; Bierhuizen et al . , Biochemical and Biophysical Research Communications 1997 , 234:371-375) . The EGFP sequence has been adapted for mammalian codon usage, carries a canonical Kosak sequence (Kozak M Journal Of Molecular Biology 1987, 196:947-950), and an F64L substitution which increases the solubility and/or stability of GFP at 37^βC (Cormack et al . , Gene 1996, 173:33-38; Miyawaki et al . , Nature 1997, 388:882-887).

The five mutations described above were introduced into the EGFP sequence .

Materials and methods :

The coding sequence for EGFP was initially derived from the pEGFPl vector purchased from Clontech (accession U55761) . The EGFP gene, excised by BamHI/NotI was then subcloned into these same sites in pCATCH-NLS (Georgiev et al . , Gene 1996, 168:165-167). This pcDNAl/Amp-based eukaryotic expression plasmid (Invitrogen) , has been designed to tag proteins with a FLAG epitope, a heart muscle kinase phosphorylation site and a nuclear localisation site. Frame continuity between the CATCH-NLS tag and the EGFP sequence was subsequently obtained by filling of the BamHI site with Klenow polymerase. Untagged EGFP under the control of the same CMV promoter was obtained by excising the CATCH-NLS tag with HindiII and BamHI and ligating the blunted overhangs.

For ER-localisation, SBFP and SGFP were amplified by PCR using pfu DNA polymerase (Stratagene) . The upstream oligonucleotide 5 ' -GCCTGCAGAGCAAGGGCGAGGAGCTGTTC-3 ' introduced a Pstl site (shown in bold) , whilst the downstream oligonucleotide 5 ' -CGTCTAGAGCTCGTCCTTGTACAGCTCGTCCATGC-3 ' created a C-terminal KDEL ER-retention sequence and introduced an Xbal site (in bold) immediately after the stop codon. The PCR amplified fragments were digested with Pstl and Xbal and cloned into the same sites in pS8DXMl. In this plasmid, a 170 bp Hindlll/PstI fragment encodes for an immunoglobulin leader peptide. The ER-tagged sequence was then subcloned back into pcDNAI/Amp with the enzymes HindiII and BamHI .

For cell surface localisation, the KDEL sequences in ER-SBFP and ER-SGFP were replaced by the transmembrane and intracytoplasmic domain of a rat MHC class I molecule, RT1-A^U (Joly et al . , Immunogenetics 1995, 41:326-328) (ace. X82106) . For this, the BsrGI site situated just upstream of the stop codon in EGFP was used, and filled in with klenow. After phenol extraction, the KDEL-coding sequences were then excised using Xbal, and the Stul/Xbal 350 bp fragment from pCMU-A^u (Joly et al . , Immunogenetics 1995, 41:326-328) was cloned in their place. This resulted directly in continuous reading frames between SBFP or SGFP and the MHC class I sequence .

Site directed mutagenesis was carried out using an embodiment of the present invention which enabled introduction of one or two mutations at once, and required only one oligonucleotide per mutation. The 50 μl reaction mixes were assembled as directed in the Stratagene QuickChange™ kit's instructions, using 50 ng of template plasmid. The oligonucleotides, however (12.5 pmole of each per reaction), rather than being complementary, were always chosen so as to yield a short PCR fragment during the first phase of the amplification reaction. The thermal cycler was then programmed as follows: 9x(95^*C for 30"- 55'C for 1' - 68^*C for 1') followed by 9x(95^*C for 30" - 68'C for 6'). In the final cycle, extension time was increased to 16' . 10 U of Dpnl were then added to each reaction, and the digestion carried out at 37 "C for 1 hour. 1 ml of this was then used to transform 10-20 μl of competent XLlblue bacteria (provided with the kit) . Plasmid DNAs from the recovered colonies were prepared by standard alkaline lysis minipreparation, including a phenol -chloroform extraction.

The various oligonucleotides used had the following sequences :

T7 : 5 ' -TAATACGACTCACTATAGGG-3 ' Y66H : 5 ' -ACCACCCTGACCCACGGCGTGCAGT-3 ' Y145F : 5 ' -TTGTGGCTGTTGAAGTTGTACTCCAG-3 ' LTH>LSH : 5 ' -CTCGTGACCACCCTGTCCCACGGCGTGCAGTGC-3 ' VI64A : 5 ' -GGATCTTGAAGTTCGCCTTGATGCCG-3 ' S175G: 5' -GGCGAGCTGCACGCCGCCGTCCTCGATG-3 '

Transfection of COS 7 cells was performed as described previously (Gonzalez et al . , Trends In Genetics 1995,

11:216-217) . When many plasmids were analysed, transfections were carried out in multiwell tissue culture dishes. In such case, the low C0₂ concentration required for optimal transfection efficiency was obtained by turning off the incubator's C0₂ input during the 2 hour incubation with the DNA/DEAE-dextran transfection mixture.

For FACS analyses, cells were harvested by trypsinisation 48 or 72 hours after transfection, washed once with PBS, and fixed in PBS/1% formaldehyde. Fixation of cells with formaldehyde does not noticeably affect the levels of fluorescence recorded from BFP or GFP, but it results in a slight increase of autofluorescence of the cells in the green channel, and in a 2 fold decrease in the blue channel. This proved very useful for the detection of weak blue fluorescent signals such as those from plasmids 3 and 28, since the overall effect is an increase of the signal/noise ratio for blue fluorescence. Cells could be kept at 4°C in PBS/1% formaldehyde for over 2 weeks without any noticeable decrease of the fluorescence intensity of GFP or BFP. FACS analyses were performed with CellQuest software on a Becton Dickinson FACS Vantage, using two different Spectra Physics 2020 lasers on two independent excitation lines. For GFP analyses, a tuneable laser was used on 488 nm/400 mW, and emissions recorded through the standard 530/30 fluorescein filter. For BFP analyses, a 351-363 nm UV laser was used at 300mW, and emissions recorded through a 460/40 filter obtained from Ealing Optics (ref. 35-5024). 10⁵ cells, gated for FSC/SSC, were collected for each sample.

For observation by fluorescence microscopy, cells were trypsinised and seeded onto sterile cover slips 48h after transfection, and analysed the following day. After washing once with PBS, cells were either fixed in PBS/1% formaldehyde, or mounted directly in PBS. Pictures were obtained with an Axiophot microscope (Zeiss) using a 63x oil immersion objective and Kodak 400 ASA slide film. For GFP, the fluorescein filter cube was used (450-490, FT510, LP520) , whilst BFP was observed with a filter cube designed for DAPI stains (G365, FT395, LP420) . This latter set of filters is however suboptimal for co-detection of GFP and BFP, since green light emitted by GFP is not eliminated by the 420nm LP filter. The possibility of assembling a more appropriate filter set is currently under investigation.

Spectral characteristics of the various constructs were analysed on a Perkin-Elmer LS3 fluorimeter using transiently transfected COS cells in PBS-1% formaldehyde. For plasmids n°3 and 28, the analysis was however not possible because the signals were indistinguishable from that of the untransfected negative control. Constant wavelengths used were 380nm and 460nm for BFPs, and 460 and 540 for GFPs .

Resul ts

After Dpnl digestion, transformation with lμl of the 50 μl reaction mixture consistently produced between 10 and 200 colonies (Table 1) . This was about 5-fold lower than expected from the original QuickChange™ protocol, but still ample to recover mutagenised plasmids. Of the plasmids recovered and analysed, >85 % had the expected structure and still expressed of a fluorescent protein. Of these, 60 -100 % had incorporated both mutations, and 10 -20 % had either one mutation or none. Apart from gross structure alterations detected by restriction digest analysis in <10% of plasmids, no undesired mutation was found outside the mutagenic oligonucleotides in all the sequences that were checked by automated sequencing (i.e. > 8 kb in total) .

EXAMPLE 2 - "Domain Swapping"

The term "domain swapping" is used to refer to replacement of a portion of one form of a sequence (e.g. allele of a gene) with a corresponding portion from a related sequence (e.g. another allele of the gene) . There may be no actual "swapping" since often only one hybrid sequence is produced in which part of the original sequence has been replaced.

The steps for achieving this in accordance with an embodiment of the present invention are illustrated in Figure 2 and are similar to those described in Example 1 for site-directed mutagenesis .

A double-stranded DNA molecule (101) is provided as a double- stranded DNA vector (102) incorporating the template sequence (103), which may be a coding sequence for a polypeptide of interest. The template (103) is one allelic form of a sequence which exists in at least one other allelic form. The experimental exemplification below employs portions of genes encoding rat MHC class I molecules, which are highly polymorphic molecules containing conserved and variable residues. The variable residues influence the specificity of the peptide-binding groove in the membrane distal part of the molecule.

There are two possible approaches illustrated in Figure 2 in accordance with embodiments of the present invention, depending on whether it is desired to transfer/replace a sequence at one end of the template sequence (which will in a polypeptide- or peptide-encoding sequence encode the N- terminus or C-terminus of the polypeptide or peptide) , or a sequence which is an internal portion of the template sequence .

If the former is desired, a first primer (104) is employed which anneals to a sequence in one strand ("the first strand") of the DNA molecule (101) within the template sequence (103) , and a second primer (105) is employed which anneals to a sequence in the other strand ("the second strand") of the DNA molecule (101) outside the template (103) to a sequence of the vector (102) . Primer (105) may be utilised in making a mutation to any sequence (103) cloned into vector (102) at a relevant cloning site. It may thus be thought of as a "standard" or "common" primer, and need not be produced specially for each mutation.

Alternatively, when it is desired to use the present invention to replace an internal portion of the template sequence, in conjunction with a first primer (104) which anneals to a sequence in one strand (the first strand) of the DNA molecule (101) to a sequence within the template sequence (103) , instead of second primer (105) which anneals outside of the template (103) to a sequence of the vector (102) a second primer (106) is employed which anneals to a sequence in the other strand (the second strand) of the DNA molecule (101) within the template sequence (103) .

Both possibilities are illustrated in the first step of Figure 2. The primers (104, and 105 or 106) are designed to anneal to a sequence within the template that is sufficiently conserved between both sequences of interest to allow primer-directed amplification on both sequences. They are oriented so that performance of a nucleic acid amplification reaction (e.g. PCR) produces resultant molecules (107,108) which each include a copy of part of the template sequence and incorporate primer sequences at respective ends. The molecules so produced are termed herein "product primers", a complementary pair (107,108) as discussed below. Though not of primary interest in "domain swapping experiments", one or more of the primers, particularly the primer identified as "104" and, where used, the primer identified as "106", may incorporate one or more mutations. Generally, however, for domain swapping the primers (104, and 105 or 106) are not mutagenic .

The first amplification reaction is illustrated in the second step of Figure 2. Exemplary conditions include 20 cycles of denaturation at 95°C for 30 seconds, annealing at 55°C for 1 minute, and elongation at 68 °C for 1 minute.

The product primers (107,108) produced in the first amplification reaction are purified from the reaction milieu, for instance using agarose gel electrophoresis, fast polynucleotide liquid chromatography (FPLC) or polyacrylamid gel electrophoresis (PAGE) or other technique available to those skilled in the art. This step is represented as the third step in Figure 2.

The purified product primers (107,108) are used to prime a second amplification reaction on a second DNA molecule (101') which is provided as a second double-stranded DNA vector

(102') incorporating a template sequence (103') which is an allelic form of the original template sequence (103) . The second DNA vector (102') may or may not be the same as the original vector (102) incorporating the original template (103) , though it is preferred that it is when a primer (105) is used (as above) which anneals to a sequence within the vector.

In the second amplification reaction, the elongation step is sufficiently long to allow for the product primers (107,108) be extended into complete copies of the entire DNA molecule (101') including vector (102') and template (103'), save for replacement of a portion of the template sequence (103') with the product primer sequences (107,108). This is illustrated in the fourth step of Figure 2. Suitable conditions include 20 cycles of annealing and elongation at 68°C for 6 minutes preceded in each cycle by denaturation at 95°C for 30 seconds .

Once complete copies of the entire strands of the double- stranded DNA molecule (101') have been made (109,110), including the product primer sequences (107,108), the original DNA molecule (101') is digested (illustrated in the fifth step of Figure 2), the newly synthesized strands (109,110) annealed together and the ends (111,112) ligated. In the exemplary embodiment, the original DNA molecule (101') which is methylated is digested using Dpnl, which only cuts methylated DNA so the newly-synthesized strands (109,110) carrying the hybrid template sequence including the replacement portion are left unscathed. Following annealing, repair of the resulting double-strand nick is conveniently achieved by transformation of the annealed newly-synthesized strands into bacteria such as E. coli .

This approach has been proved experimentally as explained and reported in the following section.

A method in accordance with an embodiment of the present invention has been used successfully to create two different hybrid sequences between two rat MHC class I molecules, RT1-A^U and RTl-AlA

When closely related genes differ in their particular function, comparison of their respective sequences can often assist to map that function to a specific portion of the gene, or even to a single point mutation. In most cases, however, such sequences will differ at several positions along the gene's sequence, and swapping of DNA sequences is usually the quickest way to initiate the mapping of the function under study. Domain swapping can also be used to generate protein with altered structures and/or immunogenicity . For example, for all the molecules of the immunoglobulin superfamily, many monoclonal antibodies have been generated that recognise the constant domain of a molecule from one particular species. Replacing this constant domain with that from another species can provide very valuable reagents.

For swapping of sequences between closely related genes, the most direct and simple approach is to look for convenient restriction sites. If these are not found in the original sequences, they can often be introduced by means of site-directed mutagenesis, but this is a lengthy and expensive process, even if PCR is used. In the embodiment of the present invention exemplified here, the only absolute requirement is that the two sequences are sufficiently close so as to allow for the design of oligonucleotides that will allow priming on both genes on either side of the region to be swapped .

As discussed in more detail above, Figure 2 is a schematic representation of such a method. In the first part of an exemplary protocol, the region to be swapped is amplified from the first gene in a standard PCR reaction, using a high fidelity enzyme to limit the error rate to a minimum, and produce DNA fragments devoid of undesirable mononucleotide 3' overhangs. It is also preferable to start from relatively high amounts of template, so as to limit the number of PCR cycles required. Different mixes of thermostable polymerases and optimized protocols allow for efficient amplification of DNA fragments superior to lOkb in length and the size of the swapped sequence should therefore not be an unsurmountable obstacle for most applications. This PCR-amplified double stranded DNA fragment is then purified away from the oligonucleotides and original template in the reaction mix. For this, agarose gel electrophoresis followed by purification of the band using a silica-based matrix was chosen, but any other purification method is equally suitable. The two strands from this purified DNA fragment, which carry the sequence of the first gene, are then used to amplify the remainder of the sequence using the other version of the gene cloned in a plasmid as a template.

The design of oligonucleotides is important. The sequences chosen for exemplification were sufficiently close that 24 nucleotides-long stretches of perfect identity could easily be identified. This might not always be the case. Longer oligonucleotides may be needed. Oligonucleotides may be designed to be homologous to the first gene towards their 3' end, and homologous to the second gene towards their 5' end. Indeed, in the PCR-amplified megaprimers, the 3' end will consist of what has been copied from the 5' end of the original oligonucleotides. Such oligonucleotides, however, would not be suitable for reciprocal exchanges between the two sequences, whereas oligonucleotides picked in regions of perfect homology would. The second part of the protocol relies of linear amplification of the DNA, where each new copy is copied from the original template sequence. The overall size of the plasmid is a potential limiting factor but, since the newly synthesized DNA is always copied off the original template, a more processive, lower fidelity, mix of thermostable enzymes may be used to increase the efficiency of the method without incorporating unwanted mutations at an unacceptable frequency. After this linear amplification, the parental DNA template plasmid, containing the second gene, may be digested with Dpnl . Dpnl cuts GATC DNA sequences only when they are methylated (i.e. G^mATC) , and therefore leaves the newly synthesized DNA strands unscathed. On the other hand, the DNA strands from the parental plasmid, amplified in a dam+ bacterial strain, will be cut in many fragments

(statistically every 256 bp) . After Dpnl digestion, the mutagenised plasmids are recovered by simply transforming the reaction mixture into competent bacteria.

As experimental exemplification, it was desired to exchange the upstream part of the cDNA coding for the RT1-A^U molecule (Joly et al . (1995) Immunogenetics, 41(5), 326-328) (accession X82106) for the corresponding sequence in the RT1-A1^C cDNA (Joly et al . , (1996) Journal of Immunology, 157(4), 1551-1558) (accession X90370) . The two cDNAs are >90% homologous, but the MHC class I molecules they encode are recognised differently by Natural Killer cells (NK cells) (Naper et al . , (1996) International Immunology, 8(11), 1779-1785) . The aim was to map the part of the MHC class I molecules recognised by NK cells by swapping portions of one molecule with that found in the other allele.

To achieve this, 'reverse' oligonucleotides were designed which annealed on regions of the DNA sequence that were identical in the two sequences in the regions encoding for aa 33-26 (F33 back: 5 ' -GAACTCCGTGTCGTCCACGTAGCC-3 ' ) and aa 61-54 (E61 back: 5' -CTCCCAATACTCCGGCCCCTCCCG-3 ' ) . A single upstream oligonucleotide, pCMU5 ' , was chosen in the vector outside of the cloning site (5 ' -CCGCGCCCCAAGCATAAACCCTGG-3 ' ) . These 3 oligonucleotides were purified by HPLC before using. The initial PCR reactions were performed in 50 μl using 2.5 U of Pfu polymerase (Stratagene) , the provided 10 x Pfu buffer, 50 μM each dNTP, 50 ng of pCMU-Al^c, and 1 μM of each oligonucleotide. PCR amplifications were performed using the following program: 95 ^*C for 90", 20 x (62 'C for 1', 68 ^*C for 1', 95 ^*C for 30"), 62 ^*C for 1', 68 ^*C for 11'. The whole PCR reactions were loaded onto a 2% agarose TAE gel, revealing DNA fragments of the expected size (i.e. 290 bp for the pair pCMU5'/F33 back, and 390 bp for the pair pCMU5'/E61 back) . These DNA fragments were purified using the QiaEx kit (Qiagen) , and eluted in water, with final concentrations of 50 ng/μl (as estimated by analysis on another agarose gel) .

For the second phase of the protocol, 350 ng of these fragments were used as megaprimers, with 50 ng of pCMU-A^u as a template. As previously, the 50 μl amplification reactions were assembled with 2.5 U Pfu polymerase and 50 μM dNTPs , and the amplification program used was: 95 "C for 90", 20 x (68'C for 10', 95'C for 30"), 68°C for 20" . Since the melting temperature of primers > 100 nucleotides is superior to the elongation temperature, there is no longer a need for an annealing step. Following this amplification, digestion with 10 U of Dpnl was performed at 37°C for 90', before being precipitated with 100 ml ethanol . After rinsing with 70% ethanol, the precipitated DNA was resuspended in 3 ml of water. This was then used to transform XLl-blue supercompetent bacteria (Stratagene) .

The results of the analysis of the plasmids recovered from this experiment are shown in Table 2. In both cases, plasmids with duly swapped sequences were recovered, although at a relatively low frequency (33% for the experiment using the F33 back primer, and 50% for the E61 primer) . Analysis of the recovered plasmids was done by sequencing, and no point mutations were identified in any of the sequenced plasmids. Regarding the relative low efficiency obtained, it should be noted that amplification conditions were not optimized at all. For example, rather than Pfu on its own, a more processive mix of thermostable enzymes may be used

The results show clearly that this novel approach can rapidly yield hybrid DNA molecules between closely related genes without recourse to any specific restriction site. Table 1 : Efficiency of mutagenesis using a modified QuickChange™ approach .

Two mutagenic oligonucleotides were used for the experiments presented on the first 3 lines of the table. In the other 7 experiments, only one of the oligonucleotides was mutagenic. When the mutation (s) introduced resulted in a clear difference in fluorescence, proper mutagenesis could be scored (column 7) , but not for the S175G mutation since this mutation does not noticeably influence levels of fluorescence (see text) . Sequences for the FPs in all the plasmids listed in the last column were checked by automated sequencing.

Table 2 : Resul ts of sequence swapping experiments between two rat MHC class I cDNAs, RTl -Au and RTl -Alc .

For each experiment, the whole 1.1 kb cDNAs were sequenced, and no unexpected mutation was found in any of these 11 plasmids .

All documents mentioned herein are incorporated by reference.

Table 1

-C- r

Tabled

Claims

CLAIMS :

1. A method of producing a mutated form of a DNA template sequence, the method including: providing a DNA vector including the DNA template sequence, the vector and template together forming a parental DNA molecule which includes at least a first strand and may be double-stranded including a second strand, wherein if the parental DNA molecule is single-stranded a second strand including a sequence complement of at least the template sequence and optionally all or part of the vector is additionally provided; separating the first and second strands; annealing a first primer to a sequence within the template sequence in the first or second strand, the first primer including a first desired mutation; annealing a second primer to a sequence in the opposite strand, the sequence to which the second primer anneals either being within the template sequence or being within the vector outside the template sequence; the first and second primers allowing for performance of a nucleic acid amplification reaction to copy a portion of each of the first and second strands including a portion of the template sequence to provide a first and a second product primer each incorporating a copied portion of the template sequence and a first or second primer sequence respectively; performing a nucleic acid amplification reaction on the parental DNA molecule using the first and second primers to produce said first and second product primers; performing a second nucleic acid amplification reaction on the parental DNA molecule using the first and second product primers to produce first and second product strands including the template sequence in mutated form and vector DNA; digesting the strands of the parental DNA molecule to leave the first and second product strands.

2. A method according to claim 1 wherein the second primer anneals to a sequence within the vector outside the template sequence .

3. A method according to claim 1 wherein the second primer anneals to a sequence within the template sequence.

4. A method according to claim 2 wherein the second primer includes a second desired mutation.

5. A method of producing a mutated form of a DNA template sequence by substituting a segment of a first form of a DNA template sequence for a segment of a second form of the template sequence, the method including: providing a DNA vector including the first form of the DNA template sequence, the vector and template together forming a first parental DNA molecule which includes at least a first strand and may be double-stranded including a second strand, wherein if the parental DNA molecule is single- stranded a second strand complement to at least the template sequence and optionally part or all of the vector is additionally provided; separating the first and second strands; annealing a first primer to a sequence within the first form of the template sequence in the first or second strand, the first primer being able to anneal to a corresponding sequence within the second form of the template sequence; annealing a second primer to a sequence in the opposite strand, the sequence to which the second primer anneals either being within the template sequence or being within the vector outside the template sequence, wherein if the second primer anneals to a sequence within the template sequence the second primer able to anneal to a corresponding sequence within the second form of the template sequence; the first and second primers allowing for performance of a nucleic acid amplification reaction to copy a portion of each of the first and second strands consisting of said segment of the first form of the template sequence to provide a first and a second product primer each incorporating a said copied portion of the template sequence and a first or second primer sequence respectively; performing a nucleic acid amplification reaction on the first parental DNA molecule using the first and second primers to produce said first and second product primers; purifying the first and second product primers; providing a DNA vector including the second form of the DNA template sequence, which vector and second form of the template together form a second parental DNA molecule which includes at least a first strand and may be double- stranded including a second strand, wherein if the second parental DNA molecule is single-stranded a second complementary strand of at least the template sequence is additionally provided; performing a second nucleic acid amplification reaction on the second parental DNA molecule using the first and second product primers to copy portions of the strands of the second parental DNA molecule including sequence variation to produce first and second product strands including a hybrid form of the template sequence wherein a segment of the first form is substituted for a segment of the second form; digesting the strands of the second parental DNA molecule to leave the first and second product strands.

6. A method according to claim 5 wherein the second primer anneals to a sequence within the first vector outside the template and a corresponding sequence within the second vector outside the template.

7. A method according to claim 6 wherein the vector forming part of the second parental DNA molecule is the same as the vector forming part of the first parental DNA molecule.

8. A method according to any one of claims 1 to 7 wherein the parental DNA molecule is modified to allow for its discriminatory digestion compared with the product strands.

9. A method according to any one of claims 1 to 7 wherein the product strands are modified to allow for discriminatory digestion of the parental DNA molecule compared with the product strands .

10. A method according to any one of claims 1 to 9 wherein the first and second product strands contain nicks and are ligated to provide a closed circular DNA molecule.

11. A method according to claim 10 wherein ligation is effected by transformation of competent cells which repair the nicks.

12. A method according to any one of claims 1 to 11 wherein the first and second product strands are subjected to a further nucleic acid amplification reaction.

13. A method according to any one of claims 1 to 12 wherein the first and second product strands are recovered or isolated.

14. A method according to any one of claims 1 to 13 wherein one of the first and second product strands includes a polypeptide coding sequence.

15. A method according to claim 14 wherein the first and second product strands are employed in an expression system to produce the polypeptide.

16. A kit including reagents for use in a nucleic acid amplification reaction and instructions for performance of a method according to any one of claims 1 to 15.