Identification of frame-shift intermediate mutant cells

Abstract

Frame-shift mutations at microsatellites occur as a time-dependent function of polymerase errors followed by failure of postreplicational mismatch repair. A cell-culture system was developed that allows identification of intermediate mutant cells that carry the mutation on a single DNA strand after the initial DNA polymerase errors. A plasmid was constructed that contained 13 repeats of a poly(dC-dA).poly(dG-dT) oligonucleotide immediately after the translation initiation codon of the enhanced GFP (EGFP) gene, shifting the EGFP gene out of its proper reading frame. The plasmid was introduced into human mismatch repair-deficient (HCT116, hMLH1-mutated) and mismatch repair-proficient (HCT116+chr3, hMLH1 wild type) colorectal cancer cells. After frame-shift mutations occurred that restored the EGFP reading frame, EGFP-expressing cells were detected, and two distinct fluorescent populations, M1 (dim cells) and M2 (bright cells), were identified. M1 cell numbers were stable, whereas M2 cells accumulated over time. In HCT116, single M2 cells gave rise to fluorescent colonies that carried a 2-bp deletion at the (CA)(13) microsatellite. Twenty-eight percent of single M1 cells, however, gave rise to colonies with a mixed fluorescence pattern that carried both (CA)(13) and (CA)(12) microsatellites. It is likely that M1 cells represent intermediate mutants that carry (CA)(13).(GT)(12) heteroduplexes. Although the mutation rate in HCT116 cell clones (6.2 x 10(-4)) was 30 times higher than in HCT116+chr3 (1.9 x 10(-5)), the proportion of M1 cells in culture did not significantly differ between HCT116 (5.87 x 10(-3)) and HCT116+chr3 (4.13 x 10(-3)), indicating that the generation of intermediate mutants is not affected by mismatch-repair proficiency.

Figure 1

pIREShyg2-EGFP/CA13 plasmid. The microsatellite (CA)₁₃ was inserted immediately after the EGFP start codon, thereby shifting the downstream portion of the gene out of its reading frame. Deletion of 2 bp or insertion of 4 bp restores the proper reading frame.

Figure 2

Microsatellite PCR of HCT116 and HCT116+chr3 single-cell clones. A 227-bp product including the (CA)₁₃ microsatellite of the transfected plasmid was amplified with radiolabeled primers and separated by PAGE. In all HCT116 clones, a second band, 2 bp shorter, was observed. In HCT116+chr3 clones, a single 227-bp band was found. HCT116 cell clones display microsatellite instability by giving rise to a mutated subpopulation (which is likely to express EGFP). HCT116+chr3 clones are microsatellite-stable.

Figure 3

Analysis of mutation rate by flow cytometry. One thousand nonfluorescent HCT116 cells had been sorted into a 24-well plate, and cells were cultured for 13 days, harvested, and analyzed by flow cytometry. Regions R1 and R2 were set according to cell size (A) and fluorescence (B). When cells of R1 and R2 were analyzed on an EGFP histogram (C), two distinct populations were retrieved. The population displaying no fluorescence was designated M0, the population with low EGFP intensity was designated M1, and the population with high EGFP intensity was designated M2.

Figure 4

M1 and M2 are distinct populations. Sorted HCT116 cells were analyzed after 5, 7, 10, and 13 days of culture, and EGFP histograms (A) were generated as described (Fig. 3). At days 5 and 7, the M1 population was larger than M2. At day 10, both populations were approximately equal in size, and at day 13 the M1 population was smaller than M2 (note the different scale of the y axis). M1 and M2 data were expressed further as percentage of R1 and plotted for each time point (B) and for each of three different HCT116 cell clones (A1.3, circles; A2.1, squares; A2.3, triangles). The M1 population showed little change (average proportion of fluorescent cells: 0.51%; 95% confidence interval 0.47–0.55%), and the M2 population accumulated over time. Each data point represents the mean of quadruplicate wells.

Figure 5

Index sorting of M1 and M2 populations. Single cells in the M1 (dim fluorescence, gated FL1 intensity 126–813) or M2 (strong fluorescence, gated FL1 intensity 1,263–9,910) population were sorted into 96-well plates and cultured for 7 days. (A) The colony phenotype (positive, negative, or mixed colonies, final magnification ×100) was determined by fluorescence microscopy. Index data [plate location, forward scatter (FSC), side scatter (SSC), FL1 (green), and FL2 (red)] that had been collected at the time of cell sorting were linked later to the colony phenotype. M1 cells that gave rise to positive colonies were labeled “positive M1 cells” (n = 10), and M1 cells that gave rise to negative or mixed colonies were labeled “negative M1 cells” (n = 26) or “mixed M1 cells” (n = 12), respectively. M2 cells gave rise only to positive colonies and were designated “positive M2 cells” (n = 23). (B) At the time of cell sorting, negative M1 cells expressed significantly less EGFP than positive M1 cells or mixed M1 cells (see text). (C) When comparing cell size and granularity (FSC/SSC), the mixed M1 cells colocalized within the other populations, indicating that they are single cells.

Figure 6

M1 and M2 populations in HCT116+chr3 cells. Sorted HCT116+chr3 cells (containing a wild-type copy of hMLH1) were analyzed after 10, 17, and 32 days of culture, and the M1 and M2 populations (percent cells of R1) were plotted for each time point and for each of three different clones (A3.1, circles; A3.3, squares; A3.7, triangles). The M1 population showed little change over time and was similar in size as seen in HCT116 cell clones (average size 0.45%; 95% confidence interval 0.34–0.55%). Although M2 cells were rare events (note different scales on the y axes), they also showed an increase over time. The M2 population of the A3.3 clone, however, had dropped at day 32, possibly due to dilution errors during serial passage of the cells.

Figure 7

Pathway of postreplicational frame-shift mutations. After the initial polymerase error, one of the two newly synthesized DNA strands loses 2 bp, (CA)₁₃ to (CA)₁₂ or (GT)₁₃ to (GT)₁₂. If this event occurs on the antisense strand [(GT)₁₃ to (GT)₁₂], the intermediate mutant of generation one starts to express EGFP and the cell becomes dimly fluorescent. If the mutation is not repaired, the cell divides into a wild-type cell and an intermediate mutant cell of generation two. The wild-type cell loses and the intermediate mutant cell gains fluorescence. If the mutation is not repaired, the intermediate mutant cell of generation two gives rise to another wild-type cell, and a mutant cell now carriers two mutant strands at the microsatellite sequence [(CA)₁₂⋅(GT)₁₂].