PCR and Site-Directed Mutagenesis


An example of a complex mutation constructed by oligonucleotide mutagenesis is shown below (Data is from Carter, P. 1991. Mutagenesis facilitated by the removal or introduction of unique restriction sites. In M. J. McPherson (Ed.), Directed Mutagenesis: A Practical Approach. Oxford University Press, NY).

The top strand shown is the primer drawn in the 3' to 5' orientation, and the bottom strand shown is the template drawn in the complementary 5' to 3' orientation. The primer was a 105-mer (that is, it was 105 nucleotides long) that was complementary to the template except for numerous mismatches (indicated with an *) and a 21-nucleotide insertion (indicated by - in the template strand). The restriction sites present in the double-stranded DNA after incorporation of the mutant oligonucleotide are shown above the primer, and the restriction sites present in the double-stranded template DNA before incorporation of the mutant oligonucleotide are shown below the template strand. Note that these restriction sites are not present anywhere else on the plasmid.

  1. How could you use restriction enzymes to select against plasmids that did not incorporate the mutant oligonucleotide?

    ANSWER: The site-directed mutagenesis procedure requires that you move the plasmids into cells after you have made the desired construct (this can be done by transformation or electroporation). However, recall that potent cellular exonucleases (like the RecBCD enzyme) can rapidly degrade linear DNA but not closed circular DNA. If you incorporate the indicated oligonucleotide and digest the resulting plasmid with EcoRI (or KpnI or SmaI), the mutant derivatives will not be linearized but the wild-type plasmids will be linearized. Therefore, upon transformation into a recipient cell, the linearized wild-type plasmids will be destroyed and the circular mutant plasmids will survive.

  2. How could you use restriction enzymes to confirm that the plasmid you isolated did contain the mutant oligonucleotide?

    ANSWER: If the mutant oligonucleotide was incorporated into the plasmid, it would be cut by BglII which does not cut the wild-type plasmid.

  3. If a 70-mer that was 20 nucleotides shorter at the 3' end and 15 nucleotides shorter at the 5' end was used as the primer instead of the 105-mer, the frequency of mutagenesis would have been much lower. Explain why.

    ANSWER: Because for this technique to work the mutant oligonucleotide has to hybridize reasonably well with the plasmid DNA. Efficient hybridization requires at least a short stretch of pairing between the two single-stranded DNAs. If the shorter primer was used, there would be substantial mispairing at both ends which would substantially decrease the hybridization.


The leader sequence for the E. coli phoA gene (which encodes alkaline phosphatase) is shown below (Chang et al. 1986. Gene 44:121-125). Note that the predicted start codon is GTG which usually encodes Val, not ATG the common met codon. The protein is cleaved between the Ala and Arg to produce mature alkaline phosphatase.

  1. How could you change a single codon in the leader sequence by site-directed mutagenesis (using the mutS method) to determine whether translation of the leader sequence sequence is essential? [Note this question asks you to change a single codon, not a single nucleotide!]

    ANSWER: There are several ways you could answer this question depending upon which mutation you choose to construct. A generic approach to answering this question is outlined below, but the specific primer sequence and amino acid mutated will differ for different mutant constructs.

  2. How you would confirm that you had constructed the desired mutant?

    ANSWER: Reversion! The reversion should include the compensating wild-type phoA leader oligonucleotide and the opposite antibiotic selection nucleotides (AmpS and TetR).

  3. How could you determine if the Lys at position 20 is required for proper function of the leader sequence? If this residue was changed to AGA and the resulting leader was non-functional, would this prove that Arg cannot substitute for Lys at this position? Why or why not?

    ANSWER: No. Note that the Lys codon AAA is a very common codon that is used frequently and the Arg AGA codon is a very rare codon. This could result is substantially reducted translation of this sequence -- potentially making too little of the product to do it's normal job in the cell. Thus, this would not be an adequate test for whether Arg can or cannot be tolerated at that particular position.

  4. Which of the following would be the best substitution for the AAA at this position to determine whether or not Lys is required? Explain your answers.

  5. How would you design a site directed mutagenesis experiment to determine whether the GTG codon is read as a rare Met start codon or if it actually encodes Val?

    ANSWER: The answer for question b directly addresses this question.


To fold into a functional enzyme, alkaline phosphatase must form disulfide bonds between two Cys residues in the protein. Given the region of phoA shown below, design a site-directed mutagenesis experiment to determine whether the Cys residue shown is essential for the structure and function of alkaline phosphatase? If you were restricted to a single nucleotide substitution mutation, what amino acid codon would you substitute for Cys?

ANSWER: Ala would be an excellent choice because (a) it is a small nonpolar amino acid like cysteine and (b) because it tends to be less disruptive to protein structure than many other amino acid substitutions.


You just finished cloning and sequencing a gene that encodes a multifunctional enzyme, but you do not know anything else about the structure and function of the gene. You would like to find mutations that identify each of the functional domains of the protein. Under what circumstances might it be smart to jump right into doing site-directed mutagenesis on your gene? Under what circumstances would it probably be premature to begin doing site-directed mutagenesis on your gene?

ANSWER: If your gene showed substantial sequence similarity to other genes in the database that were well understood, you might want to test some of the "computer" predictions directly by site-directed mutagenesis. However, it is often wiser to try some more general mutations before jumping into site-directed mutagenesis.


If you wanted to test the function of many different amino acid substitutions at a single site in a protein, you might begin by converting that codon to UAG by site directed mutagenesis. How would this help you solve the problem?

ANSWER: If you made the amber mutation at a particular site, you could test the function of multiple amino acids at that site by moving the mutant into multiple isogenic strains that carry different amber suppressor mutations -- for example, supE would insert Gln, supD would insert Ser, etc. A very important consideration when doing this type of experiment is the efficiency of suppression -- if suppression is inefficient than failure to observe activity of the gene product could be simply due to insufficient amounts of the gene product, not necessarily a problem with the particular amino acid substitution.


The Nac regulatory protein binds to the wild-type put control region from Klebsiella aerogenes. Over 100 mutants that prevented Nac binding were isolated by random, in vivo mutagenesis. The base alterations in the put control region of the mutants shown below. (The sequences are aligned for comparison and the mutations are underlined. Each mutation was isolated multiple times.)

  1. Note that certain mutations were not found in certain positions. Suggest two possible explanations for this result.

    ANSWER: (i) The unmutated positions may not be essential for Nac binding. (ii) Failure to find mutations at those sites may simply be due to chance or sensitivity of those residues to the mutagens used.

  2. How could you use the MutS method of site-directed mutagenesis to obtain a wide variety of base substitutions in the TTT and GCG sequences? [Describe any oligonucleotides you would use and draw a diagram showing how you would do the experiment.]

    ANSWER: Several approaches could be used. One example follows. Synthesize a 19 nucleotide DNA fragment (a "19-mer") corresponding to the above sequence except with random nucleotides incorporated at positions 6-8 and 13-15, as shown below:

    Use the resulting population of oligonucleotides for site-directed mutagenesis of the Nac binding-site cloned on the AmpR TetS plasmid. The mutagenesis should also include an oligonucleotide that will mutate AmpR to AmpS and an oligonucleotide that will mutate TetS to TetR. After hybridization of the oligonucleotides, synthesis of the new strand, ligation, and transformation, select for TetR colonies and screen for AmpS. Then test the resulting plasmids for Nac binding.

  3. How could you use site-directed mutagenesis to obtain a deletion mutation that removes the sequence TATTT? [Draw a diagram showing any oligonucleotides you would use and how you would do the experiment.]

    ANSWER: Synthesize an oligonucleotide missing the bases to be deleted (5' TATTT 3'). For example use the following oligonucleotide for site-directed mutagenesis with either the dut ung or mutS approach.


You just finished cloning and sequencing a gene that encodes a multifunctional enzyme, but you do not know anything else about the structure and function of the gene. You would like to find mutations that identify each of the functional domains of the protein. Under what circumstances might it be smart to jump right into doing site-directed mutagenesis on your gene? Under what circumstances would it probably be foolish to jump right into doing site-directed mutagenesis on your gene?

ANSWER: This is right out of my harangue from lecture. If your gene showed substantial sequence similarity to other genes in the database that were well understood, you might want to test some of the "computer" predictions directly by site-directed mutagenesis. However, in almost every case it is wiser to try some more general mutations before jumping into site-directed mutagenesis.


If you wanted to test the function of many different amino acid substitutions at a single site in a protein, you might begin by converting that codon to UAG by site directed mutagenesis. How would this help you solve the problem?

ANSWER: If you made the amber mutation at a particular site, you could test the function of multiple amino acids at that site by moving the mutant into multiple isogenic strains that carry different amber suppressor mutations -- for example, supE would insert Gln, supD would insert Ser, etc. A very important consideration when doing this type of experiment is the efficiency of suppression -- if suppression is inefficient than failure to observe activity of the gene product could be simply due to insufficient amounts of the gene product, not necessarily a problem with the particular amino acid substitution.


The sefD gene encodes a subunit of the SEF fimbriae from Salmonella enteritidis. The sefD gene was colned onto a multicopy plasmid as shown below. How could you use PCR to construct a derivative of this plasmid clone with a deletion in sefD and with an EcoR1 site at the join point? Indicate the location and DNA strand where any primers you use would hybridize on the original clone.

ANSWER: See figure below.


Return to Microbial Genetics Problems list.

Please send comments, suggestions, or questions to smaloy@sciences.sdsu.edu
Last modified October 31, 2003