Skip to main content
NCBI home page
Nucleic Acids Research logoLink to Nucleic Acids Research
. 2003 Aug 15;31(16):e97. doi: 10.1093/nar/gng099

Ligation of high-melting-temperature ‘clamp’ sequence extends the scanning range of rare point-mutational analysis by constant denaturant capillary electrophoresis (CDCE) to most of the human genome

Andrea S Kim 1,*, William G Thilly 1
PMCID: PMC169989  PMID: 12907749

Abstract

Mutations cause or influence the prevalence of many diseases. In human tissues, somatic point mutations have been observed at fractions at or below 4/10 000 and 5/100 000 in mitochondrial and nuclear DNA, respectively. In human populations, fractions for the multiple alleles that code for recessive deleterious syndromes are not expected to exceed 5 × 10–4. Both nuclear and mitochondrial point mutations have been measured in human cells and tissues at fractions approaching 10–6 using constant denaturant capillary electrophoresis (CDCE) coupled with high-fidelity PCR (hifiPCR). However, this approach is only applicable to those target sequences (∼100 bp) juxtaposed with a ‘clamp’, a higher-melting-temperature sequence, in genomic DNA; such naturally clamped targets represent ∼9% of the human genome. To open up most of the human genome to rare point-mutational analysis, a high-efficiency DNA ligation procedure was recently developed so that a clamp could be attached to any target of interest. We coupled this ligation procedure with prior CDCE/hifiPCR and achieved a sensitivity of 2 × 10–5 in human cells for the first time using an externally attached clamp. At this sensitivity, somatic mutations, each representing an anatomically distinct cluster of cells (turnover unit) derived from a mutant stem cell, may be detected in a series of tissue samples, each containing as many as 5 × 104 turnover units. Additionally, rare inherited mutations may be scanned in pooled DNA samples, each derived from as many as 105 persons.

INTRODUCTION

Point mutations cause or influence the prevalence of many diseases (1–4). Databases in which inherited or somatic mutations associated with human disease are compiled reveal that when genes carry mutations leading to disease, the condition is invariably multi-allelic (2–4). For some phenotypes such as proto-oncogene activation, a small set of specific missense mutations (e.g. ras mutations) is found during tumor growth (1). Conversely, genes involved in tumor initiation (e.g. APC and VHL) are found to have mostly gene inactivating mutations scattered throughout the coding regions (2,3).

Figure 1 illustrates the distribution of reported germline and somatic mutations in the APC gene of human cancer. A mutational spectrum is defined by a quantitative distribution of mutations in a given DNA sequence, and the APC mutational spectra in Figure 1 show that mutations are non-randomly distributed. These spectra are typical of those observed for all such gene/disease associations.

Figure 1.

Figure 1

Open in a new tab

Distribution of reported APC germline and somatic mutations in human cancer. Source: http://p53.curie.fr; (2).

Mutational spectra are critical in the search for the causes of mutations. For example, analysis of these spectra can be used as a tool to determine the effects of environmental agents on human mutation. Furthermore, such analysis offers a means to discover if a person carries an inherited genetic risk for a disease.

In human tissues, somatic point mutations have been observed at fractions at or below 4 × 10–4 and 5 × 10–5 in mitochondrial and nuclear DNA, respectively (5–12; X.C. Li-Sucholeiki, personal communication). In human populations, fractions for the multiple alleles that code for recessive deleterious syndromes are not expected to exceed 5 × 10–4. Thus, methods of detecting mutations at fractions below 5 × 10–5 are necessary to analyze somatic and inherited point-mutational spectra in human tissues and pooled DNA samples, respectively.

Methods that detect mutations based on genotype, rather than phenotype, are applicable to a larger pool of target genes and tissues. Phenotype-based methods are limited to selectable genes, and most of these methods can be applied only to cultured cells and to certain tissue types, the cells of which can be grown in vitro (e.g. blood) (13–16). For this reason, genotype-based methods are the methods of choice.

Three genotype-based methods have been applied to analysis of point mutations at fractions below 5 × 10–5 in human genomic DNA (5–12). Among them, allele-specific PCR (17) and restriction fragment length polymorphism (RFLP)/PCR (18) offer a target size of 1–6 bp. For analysis of mutational spectra in a larger target size of ∼100 bp, a third method employing constant denaturant capillary electrophoresis (CDCE) and high-fidelity PCR (hifiPCR) (19,20) is the method of choice.

CDCE (21), derived from denaturing gradient gel electrophoresis (DGGE) (22), is a method by which single point mutations in a 100 bp target can be spatially separated from the numerically predominant wild-type. The separation is based on the cooperative melting equilibria of DNA molecules under partially denaturing conditions, and this use of partial denaturation requires that a target domain be juxtaposed with a domain of a higher melting temperature called a ‘clamp’. For this reason, CDCE/hifiPCR developed for analysis of rare mutations (19,20) remains limited to those target sequences with a naturally occurring clamp (natural clamp) in genomic DNA. Only 9% of the human genome is represented by such naturally clamped sequences, many of which do not include gene-coding regions (23).

To increase the scanning range, we have developed high-efficiency DNA ligation by which a clamp can be attached/ligated to any 100 bp sequence in the human genome (23). Additionally, we have proposed the addition of clamp ligation to CDCE/hifiPCR for analysis of mutations at fractions as low as 10–6 in DNA sequences without a natural clamp (23,24). The use of clamp ligation with a demonstrated clamp-attachment efficiency of >95% (23) is essential in such analysis because lower efficiencies can leave a majority of individual mutant sequences without a clamp and generate false negatives. However, whether the proposed method is able to analyze mutations at such low fractions remains to be discovered. Although a previously developed method of clamp attachment by PCR had been successfully applied to DGGE/CDCE analysis of mutations in DNA sequences without a natural clamp, only those mutations at fractions at or above 10–3 could be analyzed (25–27).

To discover the sensitivity and accuracy of the clamp ligation-added method, we directly compared point-mutational analysis using a ligated clamp to that using a natural clamp (Fig. 2). For this comparison, we used the same target, the 3′ end of exon 3 in the human hypoxanthine-guanine phosphoribosyl transferase (HPRT) gene (cDNA bp 222–318, see Fig. 3).

Figure 2.

Figure 2

Open in a new tab

Flow diagram of rare point-mutational analysis by CDCE/hifiPCR: natural versus ligated clamp. The copy numbers of the wild-type and of a mutant added at an initial fraction of 5 × 10–5 are shown after each step. Restriction digestion by BstNI and DraI liberates the HPRT target-embedded fragment of 438 bp from genomic DNA. The lines in this fragment indicate the positions of the G to A transition and of the G to T transversion carried by the internal standards of the 438 bp PCR fragment and of the HPRTMunich cells, respectively. The restriction-recognition sites of AhdI, ApoI and HinfI are also indicated by the lines. The open and filled bars indicate the positions of the probes (Probe 1 and Probe 2) used for target isolation and of the PCR primers (P3 and P1), respectively.

Figure 3.

Figure 3

Open in a new tab

Restriction digestion by AhdI and HinfI generates the human HPRT target (cDNA bp 222–318) with a natural clamp (cDNA bp 141–217). Restriction digestion by ApoI, followed by ligation of a GC-base-rich clamp to the ApoI restriction end (HPRT cDNA bp 218–221), generates the target with a ligated clamp. The filled bars indicate the positions of the PCR primers (P3, L1 and P1). The solid and dotted lines represent the melting profiles of the target wild-type with the natural and ligated clamps, respectively. These profiles were constructed using WinMelt™ 2.0 (Medprobe, Norway).

MATERIALS AND METHODS

Overview

Figure 2 illustrates a flow diagram of rare point-mutational analysis by CDCE/hifiPCR using either a natural or ligated clamp. For the analysis using a ligated clamp, an extra step attaching/ligating a clamp to a target sequence of interest is necessary prior to pre-PCR mutant enrichment by CDCE.

After isolating genomic DNA from human cells, the HPRT target-embedded fragment of 438 bp was liberated by BstNI and DraI restriction endonucleases. This fragment was then isolated in single strands using probe–target hybridization coupled with a biotin–streptavidin capture system; target isolation was followed by renaturation. To generate the HPRT target with a natural clamp (Fig. 3), restriction digestion by AhdI and HinfI was accomplished with an aliquot of the target-renatured sample. Using the same aliquot, ApoI restriction digestion, clamp ligation and HinfI restriction digestion were performed to generate the same target with a ligated clamp (Fig. 3). A GC-base-rich clamp with an end complementary to the restriction end of ApoI was used for this ligation. For the HPRT target, whether with the natural or ligated clamp, pre-PCR enrichment of mutant sequences against the excess wild-type was accomplished by CDCE, followed by capillary electrophoresis (CE). HifiPCR, post-PCR mutant enrichment by CDCE, and a second round of hifiPCR were then performed prior to displaying mutants by CDCE. These mutants were isolated individually by CDCE for DNA sequencing.

Human wild-type genomic DNA

Human B cells, TK6, had been derived from a clonal isolate, maintained in exponential growth by daily dilution, and frozen after 15.5 generations (28). Using a spontaneous HPRT mutation rate of 2.7 × 10–7/doubling (28), a total mutant fraction of 4.2 × 10–6 (2.7 × 10–7/doubling × 15.5 doublings) was estimated. Assuming a target size of 1000 bp for the HPRT gene, any hotspot mutations that had occurred 10 and 100 times more frequently than expected by chance were estimated to appear at fractions of 4.2 × 10–8 {[(4.2 × 10–6) ÷ 1000] × 10} and 4.2 × 10–7, respectively. Thus, genomic DNA isolated from the TK6 cells represented mutant-free DNA and served as the wild-type when creating control mixtures with known mutant fractions >10–6 for reconstruction experiments.

Mutant cell line and DNA sequence as internal standards

Two mutations in the HPRT target, cDNA bp 222–318, served as internal standards when creating control mixtures of known mutant fractions for reconstruction experiments. One internal standard that carries a G to T transversion at HPRT cDNA bp 312 was provided by HPRTMunich, a human mutant cell line (29). This internal standard is an ideal one because it may be added prior to genomic DNA isolation.

Because a cell line carrying a mutation in a target of interest is not always available, a second form of internal standard was constructed via PCR using mutant primers (19,24). This second internal standard carries a G to A transition at HPRT cDNA bp 309 and is compatible with the 438 bp fragment of BstN I and Dra I restriction digestion (Fig. 2). It may be added subsequent to genomic DNA isolation and prior to target isolation.

Control sample preparation and genomic DNA isolation

Samples, each containing 5 × 108 TK6 cells, were admixed with 0, 1.5 × 104, 4.5 × 104 and 1.35 × 105 HPRTMunich cells to create control mixtures of 0, 3 × 10–5, 9 × 10–5 and 2.7 × 10–4 mutant fractions, respectively. Each mutant fraction was prepared in duplicate.

Genomic DNA was isolated from these mixtures as described (19,24). A DNA yield of >90% and a ratio of A260 to A280 in the range of 1.6–2.0 were estimated by a UV spectrophotometer.

Restriction digestion by BstNI and DraI

Restriction digestion by BstNI and DraI (New England Biolabs, Beverly, MA) was performed with the samples of genomic DNA to liberate the HPRT target-embedded fragment of 438 bp suitable for target isolation by probe-target hybridization coupled with a biotin–streptavidin capture system (20). Digestion was accomplished at 37°C for 4 h using an enzyme/DNA ratio of 1 U/µg.

A copy number of ∼5 × 108 in each sample of restriction digestion was estimated for the HPRT target by quantitative PCR, followed by CDCE (19,23,24). According to this estimation, the 438 bp internal standard, constructed via PCR, was added at a fraction of 5 × 10–5 (∼2.5 × 104 copies).

Target isolation and renaturation

The HPRT target-embedded fragment of 438 bp was isolated from the genomic DNA digests of BstNI and DraI to reduce sample size. This procedure was necessary because DNA loading capacities >10 µg (>1.5 × 106 cells) were determined for capillaries employed for CDCE (30). Target isolation was accomplished as described (20) using two PAGE-purified and 5′-end biotin-labeled probes: Probe 1 (5′-TGTGGAAGTT TAATGACTAA GAGGTGTTTG-3′) and Probe 2 (5′-GACGTTCAGT CCTACAGAAA TAAAATCAGG-3′) (Synthetic Genetics, San Diego, CA, USA). Additionally, the isolation procedure was modified to minimize sample exposure to heat and was optimized for the HPRT target: target denaturation for 1 min; probe-target hybridization at 60°C; the binding of beads to probe-target hybrids at 50°C for 30 min; the washing of probe-bound beads at room temperature; and an additional washing using a buffer containing 50 mM NaCl and 1 mM Tris–HCl (pH 7.4).

The yield and efficiency of target isolation were estimated to be 72 ± 2% (n = 8) and 700 ± 210 fold (n = 8), respectively. The yield was estimated as described (20), and the efficiency was estimated by a UV spectrophotometer. The total DNA concentrations measured before (AI) and after (AF) the procedure were normalized by the target copy number to estimate the efficiency (normalized AI ÷ normalized AF).

After target isolation, renaturation was accomplished as described (20) to generate the isolated target suitable for restriction digestion by AhdI and HinfI or by ApoI. During target renaturation, mutant sequences formed heteroduplexes with the wild-type by mass action. Subsequent desalting was performed by drop dialysis (31), and this procedure generated a DNA yield of ≥80%.

Using ∼5 × 108 target copies in the genomic DNA digests of BstNI and DraI, ∼3 × 108 copies [(5 × 108) × 0.72 (yield of target isolation) × 0.8 (yield of dialysis)] were estimated prior to restriction digestion by AhdI and HinfI or by ApoI. Additionally, a reduced sample size of ∼4 µg genomic DNA ([3.3 mg ÷ 700 (efficiency of target isolation)] × 0.8 (yield of dialysis)) was estimated using an initial size of ∼3.3 mg.

Restriction digestion by AhdI and HinfI (natural clamp)

To generate the HPRT target with the natural clamp suitable for CDCE analysis (Fig. 3), restriction digestion by AhdI and HinfI (New England Biolabs) was performed with 1/20 of the target-renatured samples (i.e. ∼1.5 × 107 target copies). By measuring the copy numbers of the HPRT target-embedded fragment of 438 bp before (CB) and after (CA) restriction digestion by AhdI and HinfI, a digestion efficiency of ∼80% was estimated (100 – [(CA ÷ CB) × 100]). Subsequent desalting was accomplished by drop dialysis (31) to prevent high salt concentration from interfering with loading of DNA into capillaries, and ∼107 copies [(1.5 × 107) × 0.8 (efficiency of digestion) × 0.8 (yield of dialysis)] were estimated for the restriction fragment of AhdI and HinfI.

ApoI restriction digestion and clamp ligation (ligated clamp)

To generate the HPRT target with the ligated clamp suitable for CDCE analysis (Fig. 3), restriction digestion by ApoI (New England Biolabs) was followed by clamp ligation. Digestion was performed with 1/20 of the target-renatured samples, and a digestion efficiency of ∼80% was estimated.

Employing an approach described previously (23), the number of ApoI restriction ends in the samples of restriction digestion was estimated to be 3.5 × 1011 per 1.5 × 107 HPRT targets (∼200 ng genomic DNA). According to this estimation, a 10-fold higher number of clamps (3.5 × 1012 copies) was used for ligation to achieve a clamp-ligation efficiency of >95% by mass action (23). A GC-base-rich clamp of synthesized and hybridized oligonucleotides (23) was used for this ligation.

Clamp ligation was followed by restriction digestion by HinfI (New England Biolabs) to generate the 3′ end of the clamp-ligated target compatible to that of the target with the natural clamp. Subsequent desalting was achieved by drop dialysis (31), and ∼107 copies [(1.5 × 107) × 0.8 (efficiency of ApoI digestion) × 0.95 (efficiency of clamp ligation) × 0.8 (yield of dialysis)] were estimated for the HPRT target with the ligated clamp.

Pre-PCR mutant enrichment by CDCE and by CE

With the samples of AhdI and HinfI restriction digestion (natural clamp), as well as with the clamp-ligated samples of HinfI restriction digestion (ligated clamp), CDCE, followed by CE, was performed as described (20,24) to achieve pre-PCR mutant enrichment. Briefly, mutant sequences in mutant/wild-type heteroduplexes were separated from the wild-type homoduplex by CDCE using a temperature and water-jacket (temperature-regulated zone) length of ∼66°C and 15 cm, respectively. Elution of these heteroduplexes resulted in mutant enrichment. Subsequent CE was performed to separate the double-stranded target of an expected length from other target-containing species (e.g. single strands). Elution of this double-stranded target resulted in further mutant enrichment.

As described (20), the efficiency of mutant enrichment was estimated by comparing the wild-type copies loaded into capillaries (CT) to those eluted (CE): CT ÷ CE. Using ∼107 wild-type copies, reduced copies of ∼1.7 × 105 and ∼5 × 105 were estimated using the natural and ligated clamps, respectively. Consequently, mutant-enrichment efficiencies of 61 ± 8.3 fold (n = 4) and 19 ± 4 fold (n = 8) were estimated using the natural and ligated clamps, respectively.

HifiPCR

Target copies ranging from 2 × 105 to 5 × 105 in the samples of pre-PCR mutant enrichment were increased to ∼1012 by hifiPCR using Pfu Turbo DNA polymerase (Stratagene, La Jolla, CA). During this increase, mutant sequences formed heteroduplexes with the wild-type by mass action as the primers were depleted. PCR was performed as described (23), and primers were synthesized and PAGE-purified by Synthetic Genetics (San Diego, CA).

A PCR efficiency of ∼60% was observed when using P3 [5′FITC-ACGTCTTGCT CGAGATGTGA-3′, labeled with fluorescein (FITC) at the 5′end] and P1 (5′-CATATATTAA ATATACTCAC-3′) primers to amplify the target with the natural clamp. The same observation was made when using L1 (5′FITC-CGCCCGCCGC GCCCCGCGCC CGTCCCGCCG CCCCCGCCCG ATAATAAC-3′) and P1 primers to amplify the target with the ligated clamp. When using the L1 and P1 primers, additional post-PCR incubation was performed with 0.05 U/µl Exo– Pfu DNA polymerase (Stratagene) at 72°C for 5 min to reduce PCR by-products.

Post-PCR mutant enrichment by CDCE

To further enrich for mutant sequences in the PCR-amplified samples, post-PCR mutant enrichment by CDCE was accomplished as described (19,20,24) using a 15-cm long water-jacket. A wild-type copy number of ∼108 was loaded into capillaries, and ∼7 × 106 copies were eluted. Consequently, a mutant-enrichment efficiency of ∼15-fold was estimated using the natural and ligated clamps.

Post-PCR mutant enrichment was followed by a second round of hifiPCR to generate ∼1012 target copies.

Mutant display and isolation by CDCE

To display mutant sequences in mutant/wild-type heteroduplexes for quantitative analyses, the PCR-amplified samples of post-PCR mutant enrichment were separated by CDCE. These mutants were isolated individually by CDCE, followed by DNA sequencing (MIT Biopolymers Laboratory, Cambridge, MA). Using a 15-cm long water-jacket, mutant display and isolation were achieved as described (20,24,32).

RESULTS AND DISCUSSION

Pre-PCR mutant enrichment

Mutant enrichment by CDCE prior to PCR is required for analysis of mutations at fractions as low as 10–6 (11,12,19,20,33), and thus a PCR-based method of clamp attachment (25) cannot be applied to such analysis. Conversely, clamp attachment by ligation (23) allows for pre-PCR mutant enrichment.

On average, total pre-PCR mutant-enrichment efficiencies of 19- and 61-fold were estimated using the ligated and natural clamps, respectively. To discover the cause of this difference, we investigated the effects of variables introduced by clamp ligation on the mutant-enrichment efficiency. Three variables investigated were a 104-fold higher number of non-target ApoI restriction ends compared to that of target restriction ends in the samples of clamp ligation; impure clamp oligonucleotides created during the synthesis; and the sequence context of the ligated clamp. Among these variables, the third was determined as the cause.

The ligated and natural clamps, respectively, consist of 83% and 53% GC bases. To investigate the effects of this difference on the efficiency of pre-PCR mutant enrichment, a ligated-control clamp comprising the sequence of the natural clamp was used. Approximately a 3-fold increased efficiency was estimated using the ligated-control clamp compared to that using the ligated clamp.

We speculate that secondary structures resulting from the GC-base-rich sequence of the ligated clamp altered the expected electrophoretic mobility of the clamp-ligated target during pre-PCR mutant enrichment by CDCE. As a result, a larger fraction of the wild-type co-migrated/eluted with mutant/wild-type heteroduplexes, and the 3-fold decreased efficiency was estimated using the ligated clamp compared to that using the natural clamp.

HifiPCR and post-PCR mutant enrichment

Pre-PCR mutant enrichment by CDCE was followed by hifiPCR using Pfu DNA polymerase. When amplifying the HPRT target with the ligated clamp, up to 20% of the total products appeared as by-products, thought to be caused by the high GC content of the clamp. These by-products also appeared when amplifying other target sequences with a high GC-content clamp (data not shown) but disappeared when amplifying the same HPRT target with the natural clamp (Fig. 4). While we were able to reduce the amount of the by-products by changing PCR variables such as template-primer annealing temperature and primer extension time, incubating the amplified samples with Exo– Pfu DNA polymerase was observed to be the most effective means, generating up to a 75% reduction.

Figure 4.

Figure 4

Open in a new tab

Mutant display by CDCE: natural versus ligated clamp. Natural clamp 1 (A) and Natural clamp 2 (B) are independent experimental results. The Roman numeral I represents the HPRT target wild-type in homoduplex; II represents the PCR by-products; III represents the 438 bp internal standard added at an initial fraction of 5 × 10–5; and IV represents a region where the majority of the background noise/mutants appear.

The by-products, discovered to be forms of the wild-type, migrated slower than the wild-type homoduplex but faster than mutant/wild-type heteroduplexes during CDCE (Fig. 4). To prevent co-elution of these by-products from interfering with the generation of the expected post-PCR mutant- enrichment efficiency, we had to carefully determine the experimental conditions. First, we used a water-jacket, where the differential migration took place, three times longer than that typically used for naturally clamped sequences. This change drove mutant/wild-type heteroduplexes further away from the by-products. Second, we defined precisely, rather than approximately, when to start eluting these heteroduplexes.

Using these conditions, we were able to achieve the same post-PCR mutant-enrichment efficiency, 15-fold, for the HPRT target with the ligated and natural clamps. This efficiency is in close agreement with those, 20- and 25-fold, observed for other target sequences with a natural clamp (19,20).

Mutational analysis

Pre- and post-PCR mutant enrichment resulted in total efficiencies of 285 (19 × 15) fold and 900 (60 × 15) fold using the ligated and natural clamps, respectively. Based on these efficiencies, a mutant at an initial fraction of 5 × 10–5 was expected at increased fractions of 1.4 × 10–2 (ligated clamp) and 4.5 × 10–2 (natural clamp) (Fig. 2). Consequently, detection of this mutant was expected above a CDCE-detection limit of 10–3. Figure 4 illustrates that our expectations were confirmed. Additionally, Figures 5 and 6 show that the accurate enumeration of mutant sequence copies was achieved using the ligated clamp. To our knowledge, ours is the first study demonstrating analysis of rare point mutations at initial fractions below 3 × 10–4 using an externally attached clamp.

Figure 5.

Figure 5

Open in a new tab

Quantitative reconstruction experiments using the ligated clamp. The Roman numeral I represents the HPRT target wild-type in homoduplex; II represents the PCR by-products; III represents a region where the majority of the background noise/mutants appear; IV represents the HPRTMunich internal standard added at initial fractions of 0, 3 × 10–5, 9 × 10–5 and 2.7 × 10–4; and V represents the 438 bp internal standard added at an initial fraction of 5 × 10–5.

Figure 6.

Figure 6

Open in a new tab

Quantitative reconstruction experiments using the ligated clamp (quantitative analysis of Fig. 5). Each set of triangles represents duplicate experimental results. Each output was estimated by comparing the area under the CDCE peak of the HPRTMunich internal standard (A1) to the area of the 438 bp internal standard (A2) that had been added at an initial fraction of 5 × 10–5 (MF): (A1 ÷ A2) × MF. Using the six independent sets of inputs and outputs, the least squares regression line, with the error bars (95% confidence interval) for the fitted mutant fractions, was generated. The average background mutant fractions in the individual samples are represented by the squares; each mutant fraction was estimated by comparing the average area of the background mutants (B) to the area of A2: (B ÷ A2) × MF.

Detection, enumeration and isolation of individual mutants by CDCE were previously generated in mutant homoduplexes for those target sequences with a natural clamp (11,12,19, 20,33). Achieving the same analyses in mutant/wild-type heteroduplexes is more cumbersome because the heteroduplexes that belong to one mutant appear as two CDCE peaks and do not always migrate consecutively (Figs 7 and 8). However, the PCR by-products, created during the amplification of the target with the ligated clamp, co-migrated with the majority of mutant homoduplexes during CDCE and interfered with mutation detection and enumeration. For this reason, we had to analyze mutants in mutant/wild-type heteroduplexes, the majority of which could be separated from the by-products.

Figure 7.

Figure 7

Open in a new tab

(A) After PCR amplification of the HPRT target with the ligated clamp, the wild-type was purified by CDCE. Displayed is the purified wild-type. (B) A PCR-control sample was generated by co-amplifying the wild-type from (A) and an internal standard (MIS) at a fraction of 10–3. Post-PCR mutant enrichment and PCR were then performed prior to mutant display by CDCE. (C) The sample of human cells from Figure 4(C). *Normalized based on the peak height of the wild-type.

Figure 8.

Figure 8

Open in a new tab

Positions and kinds of predominant mutations generated by Pfu DNA polymerase in the HPRT target (cDNA bp 222–318) during PCR. The group of bases in bold (cDNA bp 218–221) represents the ApoI restriction end to which the clamp was ligated. The letters in parentheses are from Figure 7.

Given that mutants in the PCR-amplified samples of post-PCR mutant enrichment can be both separated from the PCR by-products and detected above the CDCE-detection limit, analysis of these mutants is limited by the level of background noise. Equivalently, this level determines mutation-detection sensitivity. To estimate the sensitivity, the average level of background noise/mutants (MB) in each sample was compared to that of an internal standard (MI) added at a known initial fraction (F): (MB ÷ MI) × F.

A sensitivity of 6 × 10–6 was estimated for the HPRT target with the natural clamp (Fig. 4). This estimation differs by 6-fold from that, 10–6, observed in a previous study analyzing another target with a natural clamp (20). Such a difference is expected because the sensitivity is directly related to the efficiency of pre-PCR mutant enrichment and to the fidelity of Pfu DNA polymerase used for PCR. Any two given sequences can generate slightly different efficiencies and fidelities, thereby introducing variations in sensitivity.

A sensitivity of 2 × 10–5 was estimated for the HPRT target with the ligated clamp (Fig. 4). This estimation, equivalent to detection of as few as 200 copies of each mutant in the presence of 107 copies of the wild-type, was achieved empirically for the first time using an externally attached clamp.

The sensitivity estimated using the ligated clamp differs by 3-fold from that estimated using the natural clamp. This difference is assumed to be caused by the 3-fold decreased efficiency of pre-PCR mutant enrichment since the fidelity of Pfu is expected to be the same for the HPRT target with the ligated and natural clamps. We determined the sequence context of the ligated clamp, a variable introduced by clamp ligation, as the cause of the decreased efficiency.

The GC content and melting temperature of the ligated clamp are 83% and >90°C, respectively (Fig. 3). Reducing the GC content can increase the efficiency of pre-PCR mutant enrichment, thereby increasing the sensitivity. However, we sacrificed the sensitivity to make the clamp ligation-added method general to the human genome. Given that the melting temperature of a clamp must be higher than that of a sequence to be analyzed by CDCE, a clamp with such a high GC content is necessary to analyze all 100 bp sequences in the human genome.

Source of background noise

We investigated two possible sources of the background noise: heat applied to the samples prior to the first PCR and Pfu DNA polymerase used for PCR.

Heat induces DNA modifications at different rates, depending on temperature, DNA conformation, pH and buffer composition (34–37). Such modifications can contribute to background noise. For example, deamination of cytosine, forming uracil, can cause C to T transition. When adenine is converted to hypoxanthine as a result of deamination, this lesion forms a more stable base pair with cytosine than with thymine and can cause A to G transition (38,39). Although apurinic/apyrimidinic sites can block DNA synthesis (39,40), bypassing of these sites can be mutagenic (41–45). Using all the published rates of modifications at the given temperatures (35–37,46–48), we investigated our procedures prior to the first hifiPCR. As a result, target renaturation was identified as the procedure that generates the highest modification fraction.

Target isolation was followed by renaturation using an established method: incubation at 55°C for 16 h (20). After this incubation, deaminated cytosine was estimated to appear in a given target at an accumulated fraction of >5.8 × 10–5; this fraction is greater than the sensitivity of 2 × 10–5 estimated using the ligated clamp. To investigate the effects of heat on the level of background noise, the noise level of the established method was compared to that of an alternative method. The alternative method was performed with 40% dimethyl sulfoxide (DMSO) at 28°C for 16 h because the use of 40% DMSO had been shown to decrease the substrate DNA’s melting temperature by 27°C (49). The two methods did not generate any difference, and the possibility that heat-induced lesions appeared as the background noise was ruled out.

Pfu DNA polymerase was identified as the source of the background noise in a previous study analyzing rare mutations by CDCE using a natural clamp (20). We confirmed this identification using an externally attached clamp. As Figure 7 illustrates, the individual background noise/mutants in the PCR-control sample (B) were observed to co-migrate with those in the sample of human cells (C). By DNA sequencing, each of the co-migrating mutants was verified to be the same with regard to position and kind; as Figure 8 summarizes, four G to T transversions, one G to A transition and one A to G transition were observed. These six mutations represented ∼60% of the total Pfu-generated mutations and appeared at an average fraction equivalent to ∼2 × 10–5 in the sample of human cells.

Furthermore, we theoretically estimated the levels of background noise using a Pfu fidelity of 10–6/bp/doubling (20,50–55). We applied ∼23 doublings (∼107-fold target amplification) to each sample of pre-PCR mutant enrichment, and any Pfu-generated hotspot mutations that had occurred 10 times more frequently than expected by chance were estimated to appear at a fraction of 2 × 10–4 (10–6/bp/doubling × 23 doublings × 10) in the amplified target. This mutant fraction is equivalent to Pfu-generated background-noise levels of 10–5 [(2 × 10–4) ÷ 19 (pre-PCR mutant-enrichment efficiency)] and 3 × 10–6 in the samples of human cells, the mutational analyses of which were achieved using the ligated and natural clamps, respectively. These theoretically estimated noise levels are in close agreement with the sensitivities of 2 × 10–5 and 6 × 10–6 empirically estimated in our study. Moreover, increasing the number of target-amplification doublings with the samples of pre-PCR mutant enrichment resulted in an increase in the noise levels (data not shown). This observation indicates that the fidelity of Pfu determines the level of background noise.

To date, a few studies are available to explain why Pfu, not heat, is the source of the background noise. Pfu was shown to recognize the presence of uracil in single-stranded DNA and to stall DNA synthesis (56,57). As a result, the major primer-extended product was observed to be significantly shorter than the expected full-length product (57). This characteristic of Pfu may explain why the sensitivity of 2 × 10–5 empirically estimated using the ligated clamp did not match the background-noise level of >5.8 × 10–5 theoretically estimated for deaminated cytosine induced by heat.

Pfu offers the highest fidelity among the thermostable DNA polymerases commercially available to date (52,53,55) and thus is the polymerase of choice when analyzing rare mutations. However, the fidelity of Pfu limited the sensitivity of our mutational analysis by CDCE. Studies on mechanisms by which Pfu generates mutations in an amplified target may suggest a means to increase the fidelity, thereby increasing the sensitivity.

Applications

Each mutant sequence must be represented by sufficient copies for it to be measured with adequate precision. While 100 copies are measured with ±20% precision (80–120 copies, 95% confidence limits) (19), 10 copies are expected to be measured within 2-fold of the true quantity (58). We used an initial sample size of ∼108 human cells to measure mutants at fractions as low as 10–6 with ±20% precision or better. During the course of this study, however, we learned that the limit of detection is ∼10–5 using the ligated clamp. For this reason, we loaded an HPRT target copy number of ∼107 into a capillary for pre-PCR mutant enrichment by CDCE. This copy number is equivalent to an initial sample size of 2.6 × 107 cells based on an accumulated target yield of 39% prior to pre-PCR mutant enrichment: genomic DNA isolation (90% target yield), target isolation (72%), clamp ligation (>95%) and dialysis (≥80% × 2, after target renaturation and before pre-PCR mutant enrichment). The sample size should be reduced by half when analyzing non-X-linked sequences of single-copy genes.

For analysis of somatic mutational spectra in human tissues, it is feasible to obtain a sample size of ∼107 cells. Dissection of lung has yielded up to 107 epithelial cells from each bifurcation section of bronchial tree airways (12), and 2 × 107 copies of the p53 gene have been derived from liver (5). Conversely, the same sample size may not be feasible for other tissue types, such as skin; approximately 4.5 × 106 keratinocytes per cm2 have been estimated (59,60), and 3 cm2 of skin is necessary. In such cases, the 10-fold reduced sample size can be used.

Mutant-rich clusters, each representing a clonally-derived mutant turnover unit, have been observed in tissue samples (11,12,17,61). A turnover unit consists of a single stem cell and its descendent transition and terminal cells in close proximity, and the size of a turnover unit is expected to vary depending on tissue type. Sizes of ∼32 and ≥60 cells have been observed in human lung epithelium (H.Sudo, MIT, personal communication) and skin (61), respectively. A size of ∼128 cells has been observed in rat mammary epithelium (17). Using a turnover unit size of 128 cells, 5 × 103 [(6.4 × 105) ÷ 128] and 5 × 104 turnover units are expected for sample sizes of ∼106 and ∼107 cells, respectively. Thus, an anatomically distinct mutant turnover unit may be detected by our clamp ligation-added method with statistical significance in a series of tissue samples, each containing as many as 5 × 104 turnover units.

For analysis of inherited mutational spectra, pooled samples, each representing 105 persons’ blood can be obtained by taking 100 cells from each person to generate 107 cells. Such an approach permits the rapid testing of the relationship of genes to diseases and the rapid discovery of causative alleles over the entire gene in the human population. Assuming total gene-inactivating allele fractions in the human population are 3 × 10–1 and 5 × 10–3 for non-deleterious and recessive deleterious genes, respectively, 1% hotspot alleles that belong to recessive deleterious genes are expected to appear at a fraction of 5 × 10–5. Even at this fraction, alleles are represented by 500 copies in the pooled sample of 105 persons and can be measured by our method with better than ±20% precision.

An additional application includes analysis of pooled tumors to discover unknown tumor suppressor genes and proto-oncogenes.

CDCE/hifiPCR developed for analysis of rare point mutations (19,20) is applicable to only those 100 bp sequences with a natural clamp, representing 9% of the human genome (23). To increase the scanning range to most of the human genome, we added clamp ligation (23) to previous CDCE/hifiPCR. Furthermore, we modified the previous method to prevent variables introduced by clamp ligation from interfering with rare mutational analysis. As a result, we achieved a sensitivity of 2 × 10–5 and accurate enumeration of individual mutant sequences at fractions below 3 × 10–4, both of which were demonstrated for the first time in human cells using an externally attached clamp.

Two percent of the human genome is represented by DNA sequences of any length with melting temperatures ≥80°C (23) and cannot be analyzed because, based on our experience, mutations in such high-melting-temperature sequences cannot be separated from the wild-type by CDCE. Thus, while CDCE/hifiPCR alone is applicable to 9% of the human genome, our clamp ligation-added method can be applied to an additional 89%. Furthermore, our method opens up >90% of the exons in known human genes (http://pubgeneserver.uio.no/PubGene/tools/MeltMap/index.cgi) to rare mutational analysis. In such an extended pool of target sequences, any and all point mutations can be uncovered by our method, whether analyzing a series of tissue sectors or pooled DNA samples. Rare mutations, together, generate a significant fraction of the total mutational events in a gene of analysis while hotspot mutations are expected to be distinguished by their high level of occurrence (Fig. 1).

Mutational spectra are critical in the search for the causes of mutations, and analysis of these spectra offers a means to discover if a person carries an inherited genetic risk for a disease. Using our method, somatic and inherited point-mutational spectra can be analyzed in human tissues and pooled DNA samples, respectively. Such analysis in any 100 bp sequence of interest is expected to be expedited by an ongoing study in which 10 000 micrometer-scale capillaries are used per electrophoresis run and CDCE procedures are automated to generate a high-throughput mutational spectrometer (HTMS) (http://www.peoplesgenetics.com). Instead of analyzing one sequence by the conventional approach using a single centimeter-scale capillary per run, analysis of a large number of exons with a ligated clamp can be accomplished simultaneously by the HTMS.

Acknowledgments

ACKNOWLEDGEMENTS

We thank Dr X.C. Li-Sucholeiki (Peoples Genetics) for her technical advice and Ms P. Siska (MIT) for her editorial advice during the course of this study and of this manuscript preparation. We also thank Ms J. Goodluck-Griffith (MIT) and Dr A. Tomita-Mitchell (Peoples Genetics) for providing the TK6 and HPRTMunich cells. We would like to acknowledge Dr T. Soussi, (University of Pierre & Marie Curie, France) for sharing the APC data while updating his web site, http://p53.curie.fr. This study was supported by grants from the National Institute of Environmental Health Sciences (P01-ES07168 and 5P30ES02109).

REFERENCES

  • 1.Bos J.L. (1989) ras oncogenes in human cancer: a review. Cancer Res., 49, 4682–4689. [PubMed] [Google Scholar]
  • 2.Laurent-Puig P., Beroud,C. and Soussi,T. (1998) APC gene: database of germline and somatic mutations in human tumors and cell lines. Nucleic Acids Res., 26, 269–270. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Beroud C., Collod-Beroud,G., Boileau,C., Soussi,T. and Junien,C. (2000) UMD (Universal mutation database): a generic software to build and analyze locus-specific databases. Hum. Mutat., 15, 86–94. [DOI] [PubMed] [Google Scholar]
  • 4.Krawczak M., Ball,E.V., Fenton,I., Stenson,P.D., Abeysinghe,S., Thomas,N. and Cooper,D.N. (2000) Human gene mutation database—a biomedical information and research resource. Hum. Mutat., 15, 45–51. [DOI] [PubMed] [Google Scholar]
  • 5.Aguilar F., Harris,C.C., Sun,T., Hollstein,M. and Cerutti,P. (1994) Geographic variation of p53 mutational profile in nonmalignant human liver. Science, 264, 1317–1319. [DOI] [PubMed] [Google Scholar]
  • 6.Nakazawa H., English,D., Randell,P.L., Nakazawa,K., Martel,N., Armstrong,B.K. and Yamasaki,H. (1994) UV and skin cancer: specific p53 gene mutation in normal skin as a biologically relevant exposure measurement. Proc. Natl Acad. Sci. USA, 91, 360–364. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Ouhtit A., Ueda,M., Nakazawa,H., Ichihashi,M., Dumaz,N., Sarasin,A. and Yamasaki,H. (1997) Quantitative detection of ultraviolet-specific p53 mutations in normal skin from Japanese patients. Cancer Epidemiol. Biomarkers Prev., 6, 433–438. [PubMed] [Google Scholar]
  • 8.Ouhtit A., Nakazawa,H., Armstrong,B.K., Kricker,A., Tan,E., Yamasaki,H. and English,D.R. (1998) UV-radiation-specific p53 mutation frequency in normal skin as a predictor of risk of basal cell carcinoma. J. Natl Cancer Inst., 90, 523–531. [DOI] [PubMed] [Google Scholar]
  • 9.Wilson V.L., Wei,Q., Wade,K.R., Chisa,M., Bailey,D., Kanstrup,C.M., Yin,X., Jackson,C.M., Thompson,B. and Lee,W.R. (1999) Needle-in-a-haystack detection and identification of base substitution mutations in human tissues. Mutat. Res., 406, 79–100. [DOI] [PubMed] [Google Scholar]
  • 10.Wilson V.L., Yin,X., Thompson,B., Wade,K.R., Watkins,J.P., Wei,Q. and Lee,W.R. (2000) Oncogenic base substitution mutations in circulating leukocytes of normal individuals. Cancer Res., 60, 1830–1834. [PubMed] [Google Scholar]
  • 11.Khrapko K., Coller,H.A., Andre,P.C., Li,X.C., Hanekamp,J.S. and Thilly,W.G. (1997) Mitochondrial mutational spectra in human cells and tissues. Proc. Natl Acad. Sci. USA, 94, 13798–13803. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Coller H.A., Khrapko,K., Torres,A., Frampton,M.W., Utell,M.J. and Thilly,W.G. (1998) Mutational spectra of a 100-base pair mitochondrial DNA target sequence in bronchial epithelial cells: a comparison of smoking and nonsmoking twins. Cancer Res., 58, 1268–1277. [PubMed] [Google Scholar]
  • 13.Albertini R.J., Castle,K.L. and Borcherding,W.R. (1982) T-cell cloning to detect the mutant 6-thioguanine-resistant lymphocytes present in human peripheral blood. Proc. Natl Acad. Sci. USA, 79, 6617–6621. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Morley A.A., Trainor,K.J., Seshadri,R. and Ryall,R.G. (1983) Measurement of in vivo mutations in human lymphocytes. Nature, 302, 155–156. [DOI] [PubMed] [Google Scholar]
  • 15.Albertini R.J., O’Neill,J.P., Nicklas,J.A., Heintz,N.H. and Kelleher,P.C. (1985) Alterations of the hprt gene in human in vivo-derived 6-thioguanine-resistant T lymphocytes. Nature, 316, 369–371. [DOI] [PubMed] [Google Scholar]
  • 16.Janatipour M., Trainor,K.J., Kutlaca,R., Bennett,G., Hay,J., Turner,D.R. and Morley,A.A. (1988) Mutations in human lymphocytes studied by an HLA selection system. Mutat. Res., 198, 221–226. [DOI] [PubMed] [Google Scholar]
  • 17.Cha R.S., Thilly,W.G. and Zarbl,H. (1994) N-nitroso-N-methylurea-induced rat mammary tumors arise from cells with preexisting oncogenic Hras1 gene mutations. Proc. Natl Acad. Sci. USA, 91, 3749–3753. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Chiocca S.M., Sandy,M.S. and Cerutti,P.A. (1992) Genotypic analysis of N-ethyl-N-nitrosourea-induced mutations by Taq I restriction fragment length polymorphism/polymerase chain reaction in the c-H-ras1 gene. Proc. Natl Acad. Sci. USA, 89, 5331–5335. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Khrapko K., Coller,H., Andre,P., Li,X.C., Foret,F., Belenky,A., Karger,B.L. and Thilly,W.G. (1997) Mutational spectrometry without phenotypic selection: human mitochondrial DNA. Nucleic Acids Res., 25, 685–693. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Li-Sucholeiki X.C. and Thilly,W.G. (2000) A sensitive scanning technology for low frequency nuclear point mutations in human genomic DNA. Nucleic Acids Res., 28, e44. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Khrapko K., Hanekamp,J.S., Thilly,W.G., Belenkii,A., Foret,F. and Karger,B.L. (1994) Constant denaturant capillary electrophoresis (CDCE): a high resolution approach to mutational analysis. Nucleic Acids Res., 22, 364–369. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Fischer S.G. and Lerman,L.S. (1983) DNA fragments differing by single base-pair substitutions are separated in denaturing gradient gels: correspondence with melting theory. Proc. Natl Acad. Sci. USA, 80, 1579–1583. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Kim A.S., Holmquist,G.P. and Thilly,W.G. (2002) High-efficiency DNA ligation for clamp attachment without polymerase chain reaction. Anal. Biochem., 310, 179–185. [DOI] [PubMed] [Google Scholar]
  • 24.Kim A.S., Li-Sucholeiki,X.C. and Thilly,W.G. (2001) Applications of constant denaturant capillary electrophoresis and complementary procedures. Measurement of point mutational spectra. Methods Mol. Biol., 163, 175–189. [DOI] [PubMed] [Google Scholar]
  • 25.Sheffield V.C., Cox,D.R., Lerman,L.S. and Myers,R.M. (1989) Attachment of a 40-base-pair G + C-rich sequence (GC-clamp) to genomic DNA fragments by the polymerase chain reaction results in improved detection of single-base changes. Proc. Natl Acad. Sci. USA, 86, 232–236. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Bjorheim J., Lystad,S., Lindblom,A., Kressner,U., Westring,S., Wahlberg,S., Lindmark,G., Gaudernack,G., Ekstrom,P., Roe,J., Thilly,W.G. and Borresen-Dale,A.L. (1998) Mutation analyses of KRAS exon 1 comparing three different techniques: temporal temperature gradient electrophoresis, constant denaturant capillary electrophoresis and allele specific polymerase chain reaction. Mutat. Res., 403, 103–112. [DOI] [PubMed] [Google Scholar]
  • 27.Ekstrom P.O., Borresen-Dale,A.L., Qvist,H., Giercksky,K.E. and Thilly,W.G. (1999) Detection of low-frequency mutations in exon 8 of the TP53 gene by constant denaturant capillary electrophoresis (CDCE). Biotechniques, 27, 128–134. [DOI] [PubMed] [Google Scholar]
  • 28.Tomita-Mitchell A. (1999) The spontaneous and MNNG-induced mutational spectra in the human HPRT gene in vitro. PhD thesis, MIT, Cambridge, MA. [Google Scholar]
  • 29.Wilson J.M. and Kelley,W.N. (1984) Human hypoxanthine-guanine phosphoribosyltransferase. Structural alteration in a dysfunctional enzyme variant (HPRTMunich) isolated from a patient with gout. J. Biol. Chem., 259, 27–30. [PubMed] [Google Scholar]
  • 30.Li X.C. and Thilly,W.G. (1996) Use of wide-bore capillaries in constant denaturant capillary electrophoresis. Electrophoresis, 17, 1884–1889. [DOI] [PubMed] [Google Scholar]
  • 31.Marusyk R. and Sergeant,A. (1980) A simple method for dialysis of small-volume samples. Anal. Biochem., 105, 403–404. [DOI] [PubMed] [Google Scholar]
  • 32.Khrapko K., Coller,H.A., Li-Sucholeiki,X.C., Andre,P.C. and Thilly,W.G. (2001) High resolution analysis of point mutations by constant denaturant capillary electrophoresis (CDCE). Methods Mol. Biol., 163, 57–72. [DOI] [PubMed] [Google Scholar]
  • 33.Marcelino L.A., Andre,P.C., Khrapko,K., Coller,H.A., Griffith,J. and Thilly,W.G. (1998) Chemically induced mutations in mitochondrial DNA of human cells: mutational spectrum of N-methyl-N′-nitro-N-nitrosoguanidine. Cancer Res., 58, 2857–2862. [PubMed] [Google Scholar]
  • 34.Lindahl T. and Nyberg,B. (1972) Rate of depurination of native deoxyribonucleic acid. Biochemistry, 11, 3610–3618. [DOI] [PubMed] [Google Scholar]
  • 35.Lindahl T. and Karlstrom,O. (1973) Heat-induced depyrimidination of deoxyribonucleic acid in neutral solution. Biochemistry, 12, 5151–5154. [DOI] [PubMed] [Google Scholar]
  • 36.Lindahl T. and Nyberg,B. (1974) Heat-induced deamination of cytosine residues in deoxyribonucleic acid. Biochemistry, 13, 3405–3410. [DOI] [PubMed] [Google Scholar]
  • 37.Ehrlich M., Norris,K.F., Wang,R.Y., Kuo,K.C. and Gehrke,C.W. (1986) DNA cytosine methylation and heat-induced deamination. Biosci. Rep., 6, 387–393. [DOI] [PubMed] [Google Scholar]
  • 38.Lindahl T. (1979) DNA glycosylases, endonucleases for apurinic/apyrimidinic sites and base excision-repair. Prog. Nucleic Acid Res. Mol. Biol., 22, 135–192. [DOI] [PubMed] [Google Scholar]
  • 39.Lindahl T. (1993) Instability and decay of the primary structure of DNA. Nature, 362, 709–715. [DOI] [PubMed] [Google Scholar]
  • 40.Sagher D. and Strauss,B. (1985) Abasic sites from cytosine as termination signals for DNA synthesis. Nucleic Acids Res., 13, 4285–4298. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Shearman C.W. and Loeb,L.A. (1979) Effects of depurination on the fidelity of DNA synthesis. J. Mol. Biol., 128, 197–218. [DOI] [PubMed] [Google Scholar]
  • 42.Schaaper R.M., Kunkel,T.A. and Loeb,L.A. (1983) Infidelity of DNA synthesis associated with bypass of apurinic sites. Proc. Natl Acad. Sci. USA, 80, 487–491. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Kunkel T.A. (1984) Mutational specificity of depurination. Proc. Natl Acad. Sci. USA, 81, 1494–1498. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Loeb L.A. (1985) Apurinic sites as mutagenic intermediates. Cell, 40, 483–484. [DOI] [PubMed] [Google Scholar]
  • 45.Loeb L.A. and Preston,B.D. (1986) Mutagenesis by apurinic/apyrimidinic sites. Annu. Rev. Genet., 20, 201–230. [DOI] [PubMed] [Google Scholar]
  • 46.Karran P. and Lindahl,T. (1980) Hypoxanthine in deoxyribonucleic acid: generation by heat-induced hydrolysis of adenine residues and release in free form by a deoxyribonucleic acid glycosylase from calf thymus. Biochemistry, 19, 6005–6011. [DOI] [PubMed] [Google Scholar]
  • 47.Wang R.Y., Kuo,K.C., Gehrke,C.W., Huang,L.H. and Ehrlich,M. (1982) Heat- and alkali-induced deamination of 5-methylcytosine and cytosine residues in DNA. Biochim. Biophys. Acta, 697, 371–377. [DOI] [PubMed] [Google Scholar]
  • 48.Frederico L.A., Kunkel,T.A. and Shaw,B.R. (1990) A sensitive genetic assay for the detection of cytosine deamination: determination of rate constants and the activation energy. Biochemistry, 29, 2532–2537. [DOI] [PubMed] [Google Scholar]
  • 49.Escara J.F. and Hutton,J.R. (1980) Thermal stability and renaturation of DNA in dimethyl sulfoxide solutions: acceleration of the renaturation rate. Biopolymers, 19, 1315–1327. [DOI] [PubMed] [Google Scholar]
  • 50.Lundberg K.S., Shoemaker,D.D., Adams,M.W., Short,J.M., Sorge,J.A. and Mathur,E.J. (1991) High-fidelity amplification using a thermostable DNA polymerase isolated from Pyrococcus furiosus. Gene, 108, 1–6. [DOI] [PubMed] [Google Scholar]
  • 51.Barnes W.M. (1994) PCR amplification of up to 35-kb DNA with high fidelity and high yield from lambda bacteriophage templates. Proc. Natl Acad. Sci. USA, 91, 2216–2220. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Flaman J.M., Frebourg,T., Moreau,V., Charbonnier,F., Martin,C., Ishioka,C., Friend,S.H. and Iggo,R. (1994) A rapid PCR fidelity assay. Nucleic Acids Res., 22, 3259–3260. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Cline J., Braman,J.C. and Hogrefe,H.H. (1996) PCR fidelity of pfu DNA polymerase and other thermostable DNA polymerases. Nucleic Acids Res., 24, 3546–3551. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Smith J. and Modrich,P. (1996) Mutation detection with MutH, MutL and MutS mismatch repair proteins. Proc. Natl Acad. Sci. USA, 93, 4374–4379. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Andre P., Kim,A., Khrapko,K. and Thilly,W.G. (1997) Fidelity and mutational spectrum of Pfu DNA polymerase on a human mitochondrial DNA sequence. Genome Res., 7, 843–852. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.Lasken R.S., Schuster,D.M. and Rashtchian,A. (1996) Archaebacterial DNA polymerases tightly bind uracil-containing DNA. J. Biol. Chem., 271, 17692–17696. [DOI] [PubMed] [Google Scholar]
  • 57.Greagg M.A., Fogg,M.J., Panayotou,G., Evans,S.J., Connolly,B.A. and Pearl,L.H. (1999) A read-ahead function in archaeal DNA polymerases detects promutagenic template-strand uracil. Proc. Natl Acad. Sci. USA, 96, 9045–9050. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 58.McKinzie P.B., Delongchamp,R.R., Heflich,R.H. and Parsons,B.L. (2001) Prospects for applying genotypic selection of somatic oncomutation to chemical risk assessment. Mutat. Res., 489, 47–78. [DOI] [PubMed] [Google Scholar]
  • 59.Bergstresser P.R., Pariser,R.J. and Taylor,J.R. (1978) Counting and sizing of epidermal cells in normal human skin. J. Invest. Dermatol., 70, 280–284. [DOI] [PubMed] [Google Scholar]
  • 60.Weinstein G.D., McCullough,J.L. and Ross,P. (1984) Cell proliferation in normal epidermis. J. Invest. Dermatol., 82, 623–628. [DOI] [PubMed] [Google Scholar]
  • 61.Jonason A.S., Kunala,S., Price,G.J., Restifo,R.J., Spinelli,H.M., Persing,J.A., Leffell,D.J., Tarone,R.E. and Brash,D.E. (1996) Frequent clones of p53-mutated keratinocytes in normal human skin. Proc. Natl Acad. Sci. USA, 93, 14025–14029. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Nucleic Acids Research are provided here courtesy of Oxford University Press

RESOURCES