INTRODUCTION PROTOCOLS 1 Preparation of Uracil-containing Single-stranded Bacteriophage M13 DNA 13.11 2 Oligonucleotide-directed Mutagenesis of Single-stranded DNA 13.15 3. In Vitro Mutagenesis Using Double-stranded DNA Templates: Selection of Mutants 13.19 with Dpal 4 Oligonucleotide-directed Mutagenesis by Elimination of a Unique Restriction Site 13.26 (USE Mutagenesis) 5 Rapid and Efficient Site-directed Mutagenesis by the Single-tube Megaprimer PCR 13.31 Method 6 Site-specific Mutagenesis by Overlap Extension 13.36 7 Screening Recombinant Clones for Site-directed Mutagenesis by Hybridization to 13.40 Radiolabeled Oligonucleotides ‘* Alternative Protocol: Screening Phagemid-containing Bacterial Colonies by 13.47 Hybridization to Radiolabeled Oligonucleotides * Alternative Protocol: Detection of Defined Mutants by PCR 13.48 8 Detection of Mutations by Single-strand Conformational Polymorphism and 13.49 Heteroduplex Analysis 9 Generation of Sets of Nested Deletion Mutants with Exonuclease III 13.57 10 Generation of Bidirectional Sets of Deletion Mutants by Digestion with BAL 31 13.62 Nuclease INFORMATION PANELS. BAL 31 13.68 Exonuclease II 13.72 inker-scanning Mutagenesis 13.75 Random Mutagenesis: 13.78 ‘Alanine-scanning Mutagenesis 13.81 Mutagenic Oligonucleotides 13.82 Selecting against Wild-type DNA in Site-directed Mutagenesis 13.84 N’-methyladenine, Dam Methylase, and Methylation-sensitive Restriction Enzymes 13.87 Commercial Kits for Site-directed Mutagenesis, 13.89 Glycerol 13.90 Mutation Detection 13.91 13.113.2 Chapter 13: Mutagenesis l N VITRO MUTAGENESIS IS USED TO CHANGE THE BASE SEQUENCE of a segment of DNA. The may be localized or general, random or targeted. More catholic and less specific m mutagenesis are better suited to analysis of regulatory regions of genes, whereas more types of mutagenesis are used to understand the contribution of individual amino acit groups of amino acids, to the structure and function of a target protein. Both methods s Virtue of generating mutants in vitro, without phenotypic selection. MUTAGENESIS OF REGULATORY REGIONS Most of the cis-acting elements that control expression of mammalian genes were originally i tified and localized by analyzing nested sets of deletions, generated by in vitro mutagenesis, penetrate the region of interest for different distances. ‘To make these deletions, the target DNA digested from a fixed point upstream or downstream from the region of interest with an such as BAL 31, which progressively shortens a double-stranded DNA fragment, or with exo clease III, which digests double-stranded DNA from 3° termini. For further details on thi enzymes, please see the information panels on BAL 31 and EXONUCLEASE III at the end of this ter. Nested deletions allowed definition of the outer borders of the regulatory domains and op. the way to more precise mapping and analysis of internal subdomains by linker scanning (pl see the information panel on LINKER-SCANNING MUTAGENESIS) and site-directed mutagenesis, Work of this type was central to our current understanding of mammalian gene control led to the description of binding sites for many transcription factors and to the identification other types of cis-acting elements. Productive as this work may have been, it was certainly for mulaic and dreary! ‘To our great relief, nested deletions and linker-scanning mutagenesis pass their zenith sometime in the late 1980s and are now used only rarely (please see Protocols 9 10). Regulatory elements are nowadays more easily identified by computer searches for consen= sus sequences in regulatory regions, whereas mutations in cis-acting elements are more efficient- ly constructed by polymerase chain reaction (PCR)-mediated methods (please see Protocols 5 and 6). MUTAGENESIS OF CODING SEQUENCES Several different types of mutagenesis are used to understand the contribution of specific amino acids and groups of amino acids to the structure and function of proteins. Saturation Mutagenesis Saturation mutagen used to generate mutations at many sites in a particular coding sequence. Every effort is made to introduce mutations in an unbiased fashion; preconceptions and knowledge about the functions of individual amino acids in the wild-type sequence are dis- regarded. The aim is to gather information about the entire “sequence space,” ie., about the rela- tionship between the amino acid sequence and the three-dimensional structure of the protei Saturation mutagenesis is usually carried out on small segments of DNA that encode an individual structural domain. At its best, the method can provide catalogs of amino acids or com- binations of amino acids that are tolerated within a domain without deleterious effect on struc- ture and function. Studies of the bacteriophage 2. repressor, for example, have shown that a large number of combinations of amino acids can satisfy the structural and functional requirementsIntroduction 13.3 of the hydrophobic core and ot helices of the molecule (Reidhaar-Olson and Sauer 1988; Lim and Sauer 1989). For further details about saturation mutagenesis, please see the information panel on RANDOM MUTAGENESIS. Mutagenesis Alanine-scanning mutagenesis is used to analyze the function(s) of particular amino acid residues on the surface of a protein. The charged residues that normally dapple the surface of proteins are not usually required for structural integrity, but they are generally involved in ligand binding, oligomerization, or catalysis. Systematic replacement of charged amino acids with alanine residues eliminates side chains beyond the B-carbon and disrupts the functional interactions of the amino acids without changing the conformation of the main chain of the protein. Alanine scanning is therefore a powerful method for assigning functions to particular regions of the pro- tein surface (Cunningham and Wells 1989) (please see the information panel on ALANINE-SCAN- NING MUTAGENESIS). In an extension of this approach — cysteine-scanning mutagenesis — unpaired cysteine residues are used to replace individual amino acid residues at particular sites in the protein. Unpaired cysteine residues are of average size, uncharged, and hydrophobic. Because they react efficiently with modifying reagents such as N-ethylmaleimide, cysteine residues introduced by scanning mutagenesis can be used as biochemical tags to verify the topology of transmembrane proteins and to measure the accessibility of residues to modifying reagents in the aqueous or lipid phases (e.g., please see Akabas et al. 1992; Dunten et al. 1993; Karz et al. 1995; Frillingos and Kaback 1996, 197; He et al. 1996; Frillingos et al. 1997a,b, 1998). nucleotide-directed Mutagene: Oligonucleotide-directed mutagenesis is used to test the role of particular residues in the struc- ture, catalytic activity, and ligand-binding capacity of a protein. In the absence of a three-dimen- sional structure, this type of protein engineering relies on informed guesses concerning the struc- ture of the protein and the contribution of individual residues to protein stability and function. A major problem is distinguishing mutations that affect local structures from those that have pro- found and deleterious effects on the folding or stability of the entire protein. Consider a typical experiment in which a number of point mutations have been generated at various sites in a gene coding for an enzyme. When the activities of these mutants are assayed, some of them show a reduction in catalytic function and others do not. In the absence of any other data, it is not pos- sible to draw firm conclusions about the structure of the enzyme from this result. There is no way to know whether the substitution of one amino acid for another has affected only the function of the active site or whether it has had more global effects. The problem would remain even if the three-dimensional structure of the wild-type enzyme were known, No algorithms have yet been devised that accurately predict the perturbations in protein structure caused by the substitution, addition, or deletion of amino acid residues. However, these difficulties can be alleviated by devel- oping independent assays for the folding of the protein of interest. Such assays commonly include the ability of the protein to react with mono- or polyclonal antibodies that are specific for native or unfolded epitopes, the proper movement and posttranslational modification of the protein within a cell, the retention of catalytic or ligand-binding functions, and the sensitivity or resis- tance of the mutant protein to digestion with proteases. If reliable assays are available to confirm that the mutagenized protein is correctly folded, oligonucleotide-directed mutagenesis becomes an analytical technique with both exquisite speci- ficity and extraordinary breadth, Mutations that could never be found in nature can now be13.4 METHODS OF OLIGONUCLEOTIDE-DIRECTED MUTAGENESIS DESIGN OF MUTAGENIC OLIGONUCLEOTIDES Chapter 13: Mutagenesis placed precisely in the target gene; functions of proteins can be mapped to specific str domains; undesirable activities of enzymes can be eliminated and their desirable catalytic physical properties can be enhanced, In short, oligonucleotide-directed mutagenesis has becor the genetic engineer's alchemy. The scientific literature on oligonucleotide-directed mutagenesis is both highly redundant unnecessarily complex. In fact, the hundreds of methods that have been described during the two decades are all based on a simple concept (Zoller and Smith 1982, 1983, 1984; for r please see Smith 1985). ‘A synthetic oligonucleotide encoding the desired mutation is annealed to the target region’ the wild-type template DNA where it serves as a primer for initiation of DNA synthesis i vitro. Extension of the oligonucleotide by a DNA polymerase generates a double-stranded DNA t carries the desired mutation. ¢ The mutated DNA is then inserted at the appropriate location of the target gene, and. mutant protein is expressed. Figure 13-1 shows one way in which this general scheme has been adapted for use with a plasmid vector and a DNA polymerase such as bacteriophage T4 DNA polymerase. Many elabo- rations and improvements have been made to this scheme over the years, particularly since the advent of PCR, but the basic principles have not changed. ‘A crucial step in site-directed mutagenesis is the design of the mutagenic oligonucleotide. By def inition, mutagenic oligonucleotides must contain at least one base change, but they may inor= porate far more complicated mutations including insertions, deletions, and compound substitu: tions. The minimum length of the mutagenic oligonucleotide is defined by the complexity of the mutation, Simple single-base substitutions can be accomplished with oligonucleotides ~25 bases in length. More complicated mutations may require oligonucleotides 80 bases or more in length, which is close to the practical limit of most automated synthesizers. Once the length of the oligonucleotide is defined, other properties such as base sequence, base composition, melting temperature, propensity to form secondary structures, and specificity of annealing must be brought into balance in order to maximize the efficiency of mutagenesis. For the principles that guide these design features, please see the information panel on MUTAGENIC OLIGONU- CLEOTIDES. CLASSICAL SITE-DIRECTED MUTAGENESIS ‘The feasibility of introducing specific changes at defined locations in DNA was first recognized in the early 1970s from work aimed at mapping the locations of mutations on the single-stranded genome of the small bacteriophage $X174, When fragments of denatured wild-type bacterio- phage DNA were transfected into susceptible bacteria together with intact single-stranded bacte-Introduction 13.5 ‘mutagenic oligonucleotide primer — ‘single-stranded recombinant M13 DNA containing a copy of wild-type target sequence ‘extend mutagenic oligo- nucleotide primer using DNA polymerase and gNTPs Verity mutation by ae ae as fitter se 8 T8 a” egghaccnanegneccoongs so FIGURE 13-1 General Scheme for Oligonucleotide-directed Mutagenesis riophage DNA carrying an amber mutation, “marker rescue” was observed, i.e., bacteriophages carrying wild-type genomes were generated. Marker rescue occurred in the transfected bacteria because the fragment of wild-type DNA annealed to the corresponding sequence of the amber mutant, forming a mismatched heteroduplex that was converted by host-specified mismatch- repair systems into a full-length wild-type genome. It was quickly realized that this process could also be used in reverse, i., that specific mutations could be introduced into wild-type DNA using mutated double-stranded fragments of viral DNA (Weisbeek and van de Pol 1970; Hutchison and Edgell 1971). Later, when pioneering work in DNA chemistry led to the routine synthesis of oligonucleotides (Letsinger and Lunsford 1976; Khorana 1979; Caruthers et al. 1983), and when the availability and quality of DNA polymerases and DNA ligases had improved, Smith and col- leagues developed in vitro techniques for oligonucleotide-directed DNA mutagenesis. The first methods used synthetic oligonucleotides that were completely homologous to single-stranded13.6 Chapter 13: Mutagenesis bacteriophage 6X174 DNA except for a single base change that, if incorporated into the bacteti phage genome, would generate a selectable phenotype. The oligonucleotides were annealed single-stranded bacteriophage 6X174 DNA and used as primers for DNA synthesis catalyzed vitro by the Klenow fragment of Escherichia coli DNA polymerase I. When the resulting eroduplexes were transfected into bacteria, a dramatic increase was observed in the frequency: bacteriophages displaying the desired phenotype (Hutchison et al. 1978; Razin et al, 1978). Many of the methods used for oligonucleotide-mediated mutagenesis differ little in pri ple from the marker rescue techniques used originally by Clyde Hutchison and his colleagues Chapel Hill, North Carolina. Throughout the 1980s, protocols for oligonucleotide-m mutagenesis typically involved the following steps: Design and synthesis of mutagenic oligonucleotides. «Hybridization of mutagenic oligonucleotides to single-stranded target DNA cloned in a bi riophage or phagemid vector. Use of such single-stranded templates eliminates competiti between the mutagenic oligonucleotide and a complementary strand of DNA. ‘¢ Extension of the hybridized oligonucleotide by DNA polymerase in the presence of all deoxynucleoside triphosphates (dNTPs). ‘® Formation of closed circular DNA by ligation with DNA ligase. ‘© Transfection of susceptible bacteria, Screening of clones by hybridization (e.g., to the mutagenic oligonucleotide primer) for carrying the desired mutation. ‘Preparation of single-stranded DNA from the mutagenized clones. ‘* Confirmation by DNA sequencing that the target DNA in the mutagenized clones carries desired mutation and no other. ‘© Recovery of the mutated fragment of DNA. © Substitution of the mutagenized fragment for the corresponding segment of wild-type DI ‘Throughout the 1980s, a succession of improvements to virtually all of the technical aspe of the procedure included: ‘© Development of a versatile set of bacteriophage M13 and phagemid vectors, equipped polycloning sites. ‘© Use of two oligonucleotide primers, one of which is a standard universal primer and the ‘a mutagenic primer (Norris et al. 1983; Zoller and Smith 1984, 1987). The use of two pri in the extension reaction eliminated the need to isolate covalently closed circular DNA. ‘* Use of two primers and a double-stranded plasmid or phagemid DNA as template (Schold: al, 1984). This procedure, although less efficient in terms of yield of mutants, eliminated need to subclone target DNA into a bacteriophage M13 vector. ‘* Improvement of techniques to screen bacteriophage M13 plaques for the desired mutation oligonucleotide hybridization (Wallace et al. 1979, 1981; Suggs et al. 1981). ‘© Use of highly accurate DNA polymerases encoded by bacteriophage T4 or 17, which st the time required to synthesize the heteroduplex from 12 hours to 1 or 2 hours (Nossal I! McClary et al. 1989; Schena 1989). These polymerases have the additional advantage of causing displacement of the 5’ end of the mutagenic oligonucleotide primer.Introduction 13.7 In classical methods of site-directed mutagenesis, the proportion of clones containing the desired mutation varies from 0.1% to >50% depending on the efficiency of the mutagenic oligonucleotide. Several methods have been developed to select against clones that contain non- mutant DNAs, including (1) selective digestion of template DNAs by removal of modified bases (Kunkel 1985; Kunkel et al. 1991), restriction enzymes (Hofer and Kiihlein 1993), or exonucleas- es (Taylor et al. 1985); (2) restoration of an antibiotic resistance gene or an origin of replication (Hashimoto-Gotoh et al. 1995); and (3) elimination of a restriction site from mutant DNAs (Deng and Nickoloff 1992). For a more detailed discussion of these methods, please see the infor- mation panel on SELECTING AGAINST WILD-TYPE DNA IN SITE-DIRECTED MUTAGENESIS, Because of these improvements, the classic method of oligonucleotide-mediated site-direct- ed mutagenesis over the years became extremely efficient and reliable. However, like many other elegant and inventive techniques, it was quickly rendered obsolete by PCR and is no longer the first choice to create mutations in a target segment of DNA. Most site-directed mutations these days are more easily generated by one of the many variants of PCR than by the classic methods. |EDIATED SITE-DIRECTED MUTAGENESIS Most of the PCR-based methods of mutagenesis in current use are direct descendants of tech- niques originally described in the late 1980s, soon after the introduction of thermostable DNA polymerases to PCR (Higuchi et al. 1988; Ho et al. 1989; Kadowaki et al. 1989; Vallette et al. 1989). At that time, it was already known that centrally located single-base mismatches in hybrids between oligonucleotide primers and the target DNA did not affect the efficiency of amplifica- tion. Early investigators showed that these mismatches could be converted into mutations locat- ed in the primer regions of PCR products, which, in turn, could be substituted for the homolo- gous wild-type segment in, for example, a recombinant plasmid. The following are important advantages of PCR-based methods for site-directed mutagene: © High rates of recovery of mutants. In many cases, the yield of mutants is so high that selection against nonmutant templates is not required. * Ability to use double-stranded DNA templates and to introduce mutations at almost any site. ‘© Use of high temperatures to reduce the ability of the template DNA to form secondary struc- tures that lower the efficiency of the extension reaction on single-stranded DNA templates. ‘* Development of methods in which all of the reactions are carried out in one tube (e.g. please ee Marini et al. 1993; Picard et al. 1994; Ke and Madison 1997). * Availability of commercial kits. © Speed and ease. No more cloning in bacteriophage M13 vectors! ‘The following are potential disadvantages of PCR-based methods: ‘© Relatively high rate of errors in PCR products that often contain unwanted mutations in addi- tion to desired alterations. This problem can be minimized by (1) limiting the number of amplification cycles and (2) using, in preference to Taq, thermostable DNA polymerases, such as Pfu and Vent, which carry a 3’-5° exonuclease activity and have an editing ability. © Introduction of unwanted nucleotides at the 3” termini of amplified DNAs. This problem can also be solved by using Pfir DNA polymerase rather than Tag.13.8 METHODS OF PCR-BASED MUTAGENESIS Chapter 13: Mutagenesis ‘© Large number of primers and amplification reactions required for each mutagenesis ment by some PCR-based methods. For example, some of the “megaprimer” methods. cussed below require three or four primers and/or three sequential PCRs. # Requirement to optimize the conditions for PCR for each new set of primers and/or temy Without optimization, the DNA products. of the PCR-based methods may be heterogeneous if size and may migrate through agarose gels as a smear, rather than as a discrete band. ‘© High frequency of unmutagenized clones resulting from contamination of the PCR-ampli DNA with the parental, wild-type DNA used as template in the PCR. This problem is solved by using a restriction enzyme such as Dpnl to digest the plasmid selectively (e.g. pl see Weiner et al. 1994; Ansaldi et al. 1996; Li and Wilkinson 1997) (please see the information panel on SELECTING AGAINST WILD-TYPE DNA IN SITE-DIRECTED MUTAGENESIS). © Inefficiency in amplifying DNA fragments longer than 2-3 kb by standard PCR, Potent solutions to this problem are discussed in Chapter 8, Protocol 13. ‘This list of disadvantages may seem long, but it should not be seen as formidable. Most of the problems can be avoided entirely with a little forethought and planning; solutions to others are available either in Chapter 8 of this manual, in the literature that accompanies commercially available mutagenesis kits, on manufacturers’ Internet sites, o in the scientific literature. Of the many published variants of PCR-based mutagenesis, two stand out for their durability and robustness: overlap extension mutagenesis and megaprimer mutagenesis. In overlap extension mutagenesis (Higuchi et al. 1988; Ho et al. 1989), two overlapping DNA fragments are amplified in separate PCRs (please see Protocol 6). The mutation of interest is constructed in the region of overlap and is present in both amplified fragments. The overlap- ping fragments are mixed and, in a third PCR, are amplified into a full-length DNA using two primers that bind to the extremes of the two initial fragments. The method is surprisingly effec- tive, but it requires two mutagenic primers, two flanking oligonucleotides, and three PCRs to con- struct a mutation. In some cases, a simpler version of the method can be used (one mutagenic primer and two sequential PCRs) if strategically placed restriction sites are available to clone the segment of amplified DNA containing the mutation (Aiyar et al. 1996). This is not always the case. ‘The megaprimer method introduced by Kammann et al. (1989) and subsequently modified by Sarkar and Sommer (1990, 1992), Giebel and Spritz (1990), and Landt et al, (1990) is the sim- plest and most cost-effective method of PCR-based mutagenesis currently available, The method. involves two rounds of PCR that employ two flanking primers and one internal mutagenic primer containing the desired base substitutions. The flanking primers can be complementary to sequences in the cloned gene or to adjacent vector sequences. The mutagenic primer can, in the- ory, be oriented toward either of the flanking primers. In practice, however, the mutagenic primer is always oriented toward the nearer of the two flanking primers so that the length of the megaprimer is kept to a minimum. In the original protocols (please see Figure 13-2), the first round of PCR is performed with the mutagenic primer, the nearer of the flanking primers, and a wild-type DNA template. The product of this first reaction, a double-stranded megaprimer, is purified and used in a second PCR, together with the reverse flanking primer. The product is a double-stranded DNA that con tains the mutation and whose size is equal to the distance between the two outside flankingIntroduction 13.9 —_—n0 —<—<—<$_____s, ss Cee ae Af eee ea ae ee megaprimer ee 2 —— 3 5 ertendabie and and os ate eae 3 3 8 | Por 2 | extension Fr a (fr ee ee. cere ees ae Roe gfe erento Wh are een, ——— ene, Early method Modtied method FIGURE 13-2 The Basic Megaprimer Method PCR 1 is performed with a mutagenic forward primer (M) and a reverse primer (R1) to generate and ampli- fy a double-stranded megaprimer (shown in colon). (Left) In the early method, the megaprimer is purified and used with an additional forward primer (F2) in PCR 2 to obtain the desired full-length mutant. (Right) In the modified method, the megaprimer is extended on the original template to form mutant:wild-type heteroduplexes, which are used as templates in a second PCR (PCR 2), primed by oligonucleotides R2 and F2, (Modified, with permission, from Ling and Robinson 1997 [Academic Press}.) primers. If the flanking primers contain restriction sites for ease of cloning, the product of the second-round PCR can be used to reconstruct a mutagenized version of the target gene. In the second and more recent method, a purified or unpurified megaprimer is extended on a wild-type DNA template (Marini et al. 1993). Because only one strand of the megaprimer can be-extended, the product of the reaction is a full-length single-stranded DNA that carries the desired mutation(s). This DNA is then amplified using the two flanking primers (Picard et al. 1994; Ling and Robinson 1997). In our hands, however, this protocol is tricky and works well only when the extended megaprimer is purified and used in high concentration in a second PCR whose parameters have been carefully optimized. The chief problem with both variants of the megaprimer method is the time-consuming and laborious purification step, which is usually accomplished by agarose gel electrophoresis and lution. Purification is required to remove residual primers from the amplified megaprimer syn- thesized in the first PCR. An ingenious and simple solution to this difficulty is to use flanking primers with significantly different melting temperatures (T,,) to prime the two PCRs (Ke and Madison 1997). The first flanking primer is just 15 or 16 bases long and has a calculated T,,, of13.10 Chapter 13: Mutagenesis ~459C, The first PCR is therefore programmed with a low annealing temperature, At the end the first PCR, the second flanking primer, 25-30 bases in length and with a calculated T,, 72-80°C, is added directly to the reaction tube. The second PCR is then carried out using a high= temperature annealing step (usually 72-80°C). The final, full-length DNA product is therefore generated by priming with the megaprimer and the second flanking primer. Because annealing of the first flanking primer is suppressed at high temperature, the wild-type template DNA is not amplified efficiently in the second PCR. The product of the second PCR therefore consists almost entirely of mutated DNA. § METHODS USED I} THIS CHAPTER TABLE 13-1 Site- rected Mutagenesis Protocot TEMPLATE METHOD USED TO ENRICH FOR MUTANTS Land 2 single-stranded DNA selection against uracil-substituted DNA 3 double-stranded DNA selection of mutants with Dpnl 4 double-stranded DNA li on of a restriction site (USE mutagenesis) 5 double-stranded DNA megaprimer 6 double-stranded DNA splice overlap extension ‘The methods of site-directed mutagenesis described in the first six protocols of this chapter are summarized in Table 13-1. Protocol 7 describes screening of bacteriophage plaques for mutants by radiolabeled oligonucleotides, whereas Protocol 8 describes two methods to detect mutations: SSCP and heteroduplex analysis, For further details of screening, scanning, and identification of mutations, please see the information panel on MUTATION DETECTION and its accompanying table that summarizes methods commonly used to detect mutations. In addition, we describe methods to generate nested unidirectional (Protocol 9) and bidirectional deletions (Protocol 10) using exonuclease III and BAL 31 nuclease, respectively. Finally, as is increasingly the trend for many methods in molecular biology, a wide variety of kits are available from various manufacturers (please see the information panel on COMMER- CIAL KITS FOR SITE-DIRECTED MUTAGENESIS). Any sufficiently advanced technology is indistinguishable from magic. Arthur C. Clarke