Method to Overcome Inefficiencies in Site-Directed Mutagenesis of A/T-Rich DNA

Abstract

Site-directed mutagenesis (SDM) is an invaluable technique that enables the manipulation of DNA and therefore the primary structure and function of any encoded gene products. Commercial protocols for SDM have been optimized for Escherichia coli and mean A/T content but may hinder generation of desired products using other templates. Mutagenesis of A/T-rich DNA is often hindered by low oligodeoxynucleotide (oligo)-annealing temperatures, requiring oligos longer than manufacturer protocol recommendations. However, longer oligos can result in primer dimer formation and decreased SDM efficiencies. Commercially available kits proved inefficient at generating AT-rich mutants. We sought to generate a modified protocol that generated SDM products detectable using gel electrophoresis and that did not require an apparent limit on oligo length.

INTRODUCTION

PCR-mediated site-directed mutagenesis (SDM) is a versatile molecular biologic technique used for the generation of mutant DNA. SDM can be used to specifically introduce insertions, deletions, or substitutions to DNA, facilitating the investigation of downstream structural and functional effects.¹ Most SDM protocols generate linear products that require a host recombination mechanism or an additional circularization event.² Linear product can be transformed directly into Escherichia coli (E. coli) and recombined using the host RecA system.^3–5 SDM kit–based protocols are most effective in DNA with typical A/T content because most are optimized using E. coli DNA (∼50% A/T). Our work focuses on Borrelia burgdorferi (B. burgdoferi), which is an A/T-rich organism. A/T-rich DNA is often more difficult to manipulate molecularly and can lead to attenuated product formation following standard SDM protocols. Both QuikChange II and Q5 SDM protocols were investigated for their abilities to generate our A/T-rich SDM products.

QuikChange II is a widely used, commercially available PCR-mediated SDM kit from Agilent Technologies (La Jolla, CA, USA) that utilizes complementary oligodeoxynucleotides (oligos) designed with a central targeted mutation.⁶ QuikChange II recommends using oligo lengths between 25 and 45 nt and indicates that longer oligos may predominantly lead to primer dimers.⁷ Our intended B. burgdoferi target sequences required oligo lengths exceeding 45 nt to reach the annealing temperature range suggested in the QuikChange II SDM protocol.

Q5 SDM (New England BioLabs, Ipswich, MA, USA) utilizes Q5 DNA polymerase. Unlike QuikChange II, nonoverlapping oligos are utilized and designed with only 1 oligo containing the mutagenic sequence, flanked by an additional oligo to amplify the plasmid. Treatment with a proprietary enzyme mix [kinase, ligase, and DpnI (KLD)] generates circular products and removes template DNA. Neither of the commercially available kits generated the desired mutations using the B. burgdorferi DNA template. We modified SDM parameters, adjusting oligo concentration, annealing temperature, template DNA concentration, and PCR reagents, resulting in development of a protocol that generated our desired mutations.

MATERIALS AND METHODS

Chemically competent Escherichia coli

Initially, chemically competent E. coli purchased from Agilent Technologies was used, but chemically competent cells were then generated in the laboratory using the Agilent strain XL-Gold E. coli and the transformation and storage solution (TSS) protocol⁸: 12.5 mM polyethylene glycol (8 kDa), 50 mM magnesium chloride hexahydrate, 5% DMSO, and Luria broth (LB) to a final volume of 100 ml. XL-Gold cells were grown in LB to an optical density 0.400-0.600 measured at 600 nm and chilled at 4°C for 20 min. The chilled culture was pelleted at 3000 RCF at 4°C for 10 min and placed on ice. The pellet was resuspended with 5 ml of chilled TSS reagent (4.5 ml of TSS plus 500 μl of 100% autoclaved glycerol). The cell suspensions were dispensed into 50-µl aliquots, flash frozen, and stored at −80°C.

QuikChange II protocol

The QuikChange II protocol uses a complementary oligo design with the mutagenic sequence located within the center of each oligo pair. The control reaction (50 µl) was performed as per the manufacturer’s directions.

Sample reactions contained the maximum concentration of template DNA, 50 ng/5 μl; 5 µl of 10X QuikChange reaction buffer; 2.5 µl each of forward and reverse oligos (125 ng each); 1 µl of deoxynucleoside triphosphate mix; 33 µl of nuclease-free water; and 1 µl of PfuUltra high-fidelity DNA polymerase. The sample PCR protocol was performed using an initial denaturation step of 95°C for 30 s, a melting step of 95°C for 30 s, an annealing step of 55°C for 1 min, and an extension step of 68°C for 8 min. The melting, annealing, and extension steps were cycled an additional 14 times (15 cycles total), which was followed by a final extension step. The sample PCR protocol was performed as described for the control reaction, adjusting the extension time for the added base pairs of our vector [pGTE-p66-gent (7.5 kb)].⁹

All samples were digested by adding 1 µl (50 U) of the restriction enzyme DpnI (methylation-dependent) to the total PCR product and incubating at 37°C for 1 h to digest the parental double-stranded DNA (methylated). Two microliters of digested PCR product and 50 µl of chemically competent XL-Gold E. coli (Agilent Technologies) were incubated on ice for 10 min, heat-shocked at 43.5°C for 30 s, and incubated on ice for 5 min. All digested samples were transformed along with the following controls: no DNA, parental plasmid alone, and parental plasmid digested with DpnI. Cells were recovered in 1 ml of SOC medium for 1 h at 37°C. Three hundred microliters of each recovered cell culture was spread onto LB agar with antibiotic supplements (controls were spread onto medium supplemented with 100 µg/ml ampicillin, and our samples were spread onto 10 µg/ml gentamicin) and incubated at 37°C for 16 h.

Q5 SDM protocol

The Q5 SDM protocol uses a noncomplementary oligo design with the mutation located in only one of the oligos. The control reaction was performed as per manufacturer’s directions. The sample reactions contained the maximum recommended template DNA concentration of 25 ng/5 μl, 12.5 µl of the 2X Q5 master mix, 2 μl (10 nM) of each forward and reverse oligo, and 3.5 μl of nuclease-free water. The sample PCR protocol was performed using individualized annealing temperatures determined by the New England Biolabs (NEB)-recommended primer design software, NEBaseChanger (www.NEBaseChanger.com). The PCR protocol used an initial polymerase activation step of 98°C for 30 s, a melting step of 98°C for 10 s, a variable annealing step with temperature dependent on the oligos for 20 s, an extension step of 72°C for 4 min, and a final extension step of 72°C for 2 min. The melting step through the extension step was repeated for an additional 24 times (25 cycles in total).

All samples were treated with the provided KLD reagent using 1 μl of the PCR reaction, 1 μl of the 10X KLD enzyme mix, 3 μl of nuclease-free water, and 5 μl of 2X KLD reaction buffer. The KLD reactions were incubated at room temperature for 5 min. All digested samples were transformed into chemically competent XL-Gold E. coli using 5 µl of the KLD-treated PCR reaction. Heat shock, recovery, and plating were performed as described in the QuikChange II SDM protocol.

A/T-rich SDM protocol

The A/T-rich protocol uses overlapping oligos and the mutation located within the center of each oligo. All reactions were prepared using a final volume of 50 µl and included 5 µl of template DNA (10 ng/µl), 2 µl each of the forward and reverse oligos (4 μl/120 ng of oligos total), 16 µl of nuclease-free water, and 25 µl of high-fidelity, hot start Q5 master mix (New England BioLabs). The A/T-rich protocol uses a 2-part PCR with an initial polymerase activation step of 98°C for 30 s, a melting step of 98°C for 30 s, a step-down annealing temperature with a range of 65°C–55°C for 1.5 min, and an extension step of 68°C for 30 s/kb of the template plasmid. The melting step through the extension step was cycled for an additional 4 times (5 cycles in total), with the annealing temperature decreasing by 2°C per cycle. The second part of the PCR protocol continued with a melting step of 98°C for 30 s, an annealing step of 65°C for 1.5 min, an extension step of 68°C for 30 s/kb of the parental plasmid repeated for an additional 14 times (15 cycles in total), and a final extension step of 68°C for 2 min. PCR product was digested with DpnI and transformed into chemically competent XL-Gold E. coli.

DNA extraction, sequencing, and alignment

E. coli transformant colonies generated by the SDM protocols were purified by restreaking on LB agar plates supplemented with gentamicin. Single colonies were inoculated in LB supplemented with 10 μg/ml gentamicin and grown for 18 h at 37°C. The 6-ml cultures were pelleted, and plasmid DNA was extracted using the Wizard Plus SV Minipreps DNA Purification kit (Promega, Madison, WI, USA). DNA was quantified using a NanoDrop 2000 (Thermo Fisher Scientific, Waltham, MA, USA) spectrophotometer, diluted to 100 ng/μl, and requantified. All clones were sequenced by MCLABS (San Francisco, CA, USA) using 26 oligos that covered the 4-kb region of B. burgdorferi DNA cloned in the plasmid.⁹ The sequencing oligos were a mix of both forward and reverse oligos to generate overlapping sequences of both DNA strands. DNA sequences were aligned to the template DNA sequence (nonmutated) using the Lasergene SeqMan Pro software (DNASTAR Inc., Madison, WI, USA).

RESULTS AND DISCUSSION

SDM can be used to test a number of hypotheses experientially and is an invaluable tool for generating site-specific mutations. The protocol described here was initially designed to generate 16 different mutations for use in protein topology studies. Our SDM mutants introduce variations in a gene encoding a Borrelia burgdorferi integral outer-membrane protein, P66 (BB_0603). For simplicity, we will describe 3 representative mutagenized constructs selected because the required oligos had different A/T contents: K103C = ∼75% A/T, T559C = ∼65%, and K591 = ∼55% (approximate values) (Fig. 1 and Table 1).

FIGURE 1.

Oligodeoxynucleotide design and utilization in the site-directed mutagenesis protocols. Three methods of SDM were utilized to establish the ability to produce PCR products for site-directed mutagenesis. Noncomplementary oligos contain the mutation on the forward oligo with the reverse oligo binding adjacent to the 5′ start of the forward oligo. Gentamicin resistance was conferred by a gentamicin acetyltransferase gene, aacC1. The targeted mutations reported here were selected to represent A/T contents ranging from ∼75% to ∼54%. Oligos are not drawn to scale.

TABLE 1.

Oligodeoxynucleotide properties

Oligo designator	Direction	Sequence (5′—3′)	Length (bp)	% A/T	Tm (°C)	NEB Tm (°C)
K103C C	Forward	CAAATCAATATGACGATTTTTTTATTTGTATAAGTACTATGACAGATTTTTGAC	54	84	59.8	69
T559C C	Forward	AACAGGACTTAGCCTTGAAAAACTAATAAGATTTTGCACAATTTCTCTTGGATGGGA	57	65	66	77
K591C C	Forward	TGCTATGGAAGTGCTTTCTTGCAATTCTTGCATAGCCTACAGCGGAAG	48	54	67.9	88
K103C NO	Forward	TAGTACTTATACAAATAAAAAAATCG	26	81	56	58
	Reverse	TGACAGATTTTGACTTTAATAAAG	24	75	56	60
T559C NO	Forward	GAGAAATTGTGCAAAATCTTATTAG	25	72	58	62
	Reverse	TTGGATGGGATTCAAATAAC	20	65	57	62
K591C NO	Forward	TGTAGGCTATGCAGAATTGCAAG	23	57	66	70
	Reverse	GCGGAAGCTAACAGCAAAAG	20	50	65	70

All oligos used in this study were purchased from Integrated DNA Technologies (Skokie, IL, USA) with standard desalting and no further purification. These oligos are denoted by the substitution mutation they encode (K103C/T559C/K591C) within p66 and by their design (C, complementary; NO, nonoverlapping). The information provided by this table includes: the oligo sequence, length in base pairs (bp), the percent A/T content, and the calculated melting temperature (Tm). For the complementary oligos, only the forward sequence and corresponding data are included because the reverse oligos are complementary in sequence and identical in the other parameters. The complementary oligos were selected for their range in A/T content and were designed using the QuikChange Primer Design software (https://www.genomics.agilent.com/primerDesignProgram.jsp). The nonoverlapping primers were used for the NEB Q5 SDM protocol and were designed by the NEBaseChanger program [v1.2.7] (https://www.nebasechanger.neb.com). The NEB Tm column contains the suggested melting temperatures for using a given oligo specifically in the Q5 master mix (https://tmcalculator.neb.com).

Our lab previously used the QuikChange II SDM protocol with success,⁹ but we experienced difficulties in generating this set of mutants. Our yield of synthesized product was poorly visualized by gel electrophoresis using a template DNA concentration of 50 ng/reaction (Fig. 2A). Absolute product concentrations of the pretreated and posttreated (DpnI digestion or KLD treatment) SDM products were estimated by gel electrophoresis coupled with densitometry and the Image Lab software (Bio-Rad, Hercules, CA, USA) (Fig. 2D). Although product bands at ∼7000 bp were observed, they were faint and not discernable after DpnI digestion. This suggests the visualized bands were largely template DNA. The quantity of SDM product did not increase when template DNA was increased to 200 ng/reaction (4 times the QuikChange II SDM protocol max) (unpublished data). Agilent recommends proceeding with transformation even in the absence of visualized products. Transformation of our PCR product yielded no colony-forming units (CFUs) for the intended mutants, whereas the kit’s positive control generated CFUs too numerous to count.

FIGURE 2.

Electrophoretic separation, visualization, and relative concentration of site-directed mutagenesis products. SDM products were visualized following electrophoretic separation with 1.1% Tris, acetate, and EDTA (TAE) agarose gels. Loading controls of known linearized plasmid concentrations were used for densitometry to obtain relative concentrations of product. The SDM product lanes are unaltered PCR reactions; the treated product lanes are DpnI-digested or KLD-treated PCR reactions (KLD treatment is specific to Q5 SDM reactions only). Each triplicate series was visualized and electrophoresed independently from other triplicate series. Agarose gels were stained in an ethidium bromide bath for 45 min and destained for 30 min in deionized water. A) The QuikChange II protocol utilizing complementary oligos. B) The Q5 SDM protocol utilizing nonoverlapping oligos. C) The A/T-rich SDM protocol utilizing complementary oligos as described for the QuikChange II protocol. D) Relative concentrations of product generated determined using the Image Lab software from Bio-Rad and the known loading controls. Control reactions for the QuikChange II protocol were not resolvable by gel electrophoresis. Results are reported as a mean of 3 trials with sd. The same standards were used to determine product concentrations of all 3 trials. Concentrations were estimated using densitometry values and dividing by 15 (the volume of each reaction loaded).

We then tested the efficacy of a different SDM kit, Q5 SDM (New England BioLabs), because of its utilization of an alternate DNA polymerase with higher efficiency and an ability to accurately synthesize DNA that is A/T- or G/C-rich. Like the QuikChange II protocol, the NEB protocol generated low levels of potential product visualized using gel electrophoresis and no CFUs (Fig. 2B, D and Table 2). Template DNA concentrations were increased up to 100 ng/reaction (4 times the Q5 SDM protocol max) and, as with the QuikChange II protocol, no CFUs were observed (unpublished data).

TABLE 2.

Site-directed mutagenesis results

Oligo	Total CFU/ml	Colonies sequenced	Desired mutations	Undesired mutations	No mutations
QuikChange II SDM	0	0	N/A	N/A	N/A
NEB Q5 SDM	0	0	N/A	N/A	N/A
A/T-rich SDM
K103C	3.83 × 102	12	10	1	2
T559C	5.60 × 101	10	10	2	0
K591C	4.40 × 101	8	6	3	1

To generate CFU counts and sequencing data, 500 μl of the SOC recovery was plated onto selective agar as described in the Materials and Methods section. Neither the QuikChange II nor the NEB Q5 SDM protocol generated CFUs after 16 or 24 h of incubation at 37°C. The positive controls for both of these protocols generated CFUs too numerous to count on traditional media. The A/T-rich SDM protocol generated colonies that were then prepared for sequencing as described in the Materials and Methods (n = 29, ∼10 per mutant construct). In addition to transformation efficiency, the sequencing data for the desired and any undesired mutations are also reported. Some clones contained no mutations, others contained >1. The no-DNA control yielded no CFUs, which was identical to the DpnI-digested or KLD-treated 50-ng template DNA controls.

Because neither commercially available SDM kit produced our desired mutants, we developed a protocol that is efficient in generating A/T-rich SDM products. This protocol design utilizes complementary oligos, decreased oligo concentrations, a step-down annealing temperature gradient, an increase in template DNA concentration, and the use of Q5 DNA polymerase. Of the selected mutants, our protocol greatly increased the amount of product that was formed and detected using gel electrophoresis and quantified using densitometry (Fig. 2C, D). Additionally, our protocol had a positive correlation (R² = 0.954) between product formed and the percent A/T content of the primers utilized (Fig. 3). Our A/T-rich protocol allowed us to generate the additional 13 targeted mutants with oligos of various A/T contents and annealing temperatures. Data for these mutations are not shown because they were not performed in triplicate or with standards for quantification.

FIGURE 3.

Correlation between A/T content and product formation. The correlation of product formed reported as the DNA concentration (nanograms per microliter) to the percent A/T content of the 3 primers used throughout this communication. Data are reported in triplicate and were fit using linear regression (solid line). The correlation coefficient (R²) between the variables is 0.954.

Sequencing data for the individual clones obtained after transformation of the DNA generated from the A/T-rich SDM protocol were aligned to the parental plasmid sequence [pGTE-p66-gent (7.5 kb)].⁹ Twenty-six sequencing oligos were utilized to enhance read depth which was necessary because of the length of the B. burgdorferi DNA insert (∼4 kb) and the overall A/T content.⁹ The A/T-rich protocol was the only tested protocol able to generate clones with the desired mutations (Table 2). Unintended mutations are defined as any mutations found within the sequencing reads that do not match the consensus, are accurate (after interrogating the sequencing files), are confirmed by at least 50% of the total reads, and are not found within repetitive tracts of DNA base pairs.

All 16 of our targeted mutants were generated using our A/T-rich SDM protocol. Increased template DNA was required for some but not all mutations. The annealing temperature stepdown generated visible SDM products. Although the step-down protocol is not unique to PCR, the ability for it to increase product generation coupled with the changes we describe increases the versatility of this protocol and allows the production of 16 unique substitution constructs.