Abstract

It is generally accepted that longer microsatellites mutate more frequently in defective DNA mismatch repair (MMR) than shorter microsatellites. Indeed, we have previously observed that the A10 microsatellite of transforming growth factor beta type II receptor (TGFBR2) frameshifts −1 bp at a faster rate than the A8 microsatellite of activin type II receptor (ACVR2), although both genes become frameshift-mutated in >80% of MMR-defective colorectal cancers. To experimentally determine the effect of microsatellite length upon frameshift mutation in gene-specific sequence contexts, we altered the microsatellite length within TGFBR2 exon 3 and ACVR2 exon 10, generating A7, A10 and A13 constructs. These constructs were cloned 1 bp out of frame of EGFP, allowing a −1 bp frameshift to drive EGFP expression, and stably transfected into MMR-deficient cells. Subsequent non-fluorescent cells were sorted, cultured for 7–35 days and harvested for EGFP analysis and DNA sequencing. Longer microsatellites within TGFBR2 and ACVR2 showed significantly higher mutation rates than shorter ones, with TGFBR2 A13, A10 and A7 frameshifts measured at 22.38 × 10−4, 2.17 × 10−4 and 0.13 × 10−4, respectively. Surprisingly, shorter ACVR2 constructs showed three times higher mutation rates at A7 and A10 lengths than identical length TGFBR2 constructs but comparably lower at the A13 length, suggesting influences from both microsatellite length as well as the sequence context. Furthermore, the TGFBR2 A13 construct mutated into 33% A11 sequences (−2 bp) in addition to expected A12 (−1 bp), indicating that this construct undergoes continual subsequent frameshift mutation. These data demonstrate experimentally that both the length of a mononucleotide microsatellite and its sequence context influence mutation rate in defective DNA MMR.

INTRODUCTION

Microsatellites are repetitive DNA sequences consisting of nucleotide units ranging from 1 to 6 bp and are ubiquitous throughout the genome (1). The majority of microsatellites are located in non-coding regions of the genome, but a minority are present in coding regions (exons) of key growth regulatory genes such as transforming growth factor beta type II receptor (TGFBR2), activin type II receptor (ACVR2), phosphatase and tensin (PTEN) homolog and Bcl-2-associated X protein (BAX) (2–6). When DNA mismatch repair (MMR), the principal enzyme system that directs repair of slippage mistakes at microsatellite sequences, is defective, frameshift mutations within alleles of these genes are accumulated and their protein functions lost (7,8). This mode of gene mutation is thought to drive the pathogenesis of colorectal cancer and other tumors with microsatellite instability (MSI) (7–9). For example, an A10 microsatellite within exon 3 of TGFBR2 and an A8 microsatellite within exon 10 of ACVR2 become frameshift-mutated (A9 for TGFBR2 and A7 for ACVR2) at both alleles in 70–90% of colorectal cancers with MSI (2,4,10). These biallelic frameshift mutations inactivate the TGFBR2- and ACVR2-encoded protein receptors and allow tumors to escape the growth-suppressive effects that are mediated by these two receptors and their intracellular SMAD signaling (4,9,11). We have measured frameshift mutation rates of the TGFBR2 exon 3 A10 microsatellite and the ACVR2 exon 10 A8 microsatellite and demonstrated that both accumulate −1 bp frameshift mutations in human colorectal cancer cells with DNA MMR defects (hMLH1−/− and hMSH6−/−) after forming heteroduplexes at a constant rate of formation (12). We also observed a higher frameshift mutation rate at the A10 microsatellite of TGFBR2 exon 3 compared with the A8 microsatellite of ACVR2 exon 10 in MMR-deficient cells (12), which is presumed to be due to the longer microsatellite within TGFBR2. However, ACVR2 has identical length A8 polyadenine tracts in exon 3 and exon 10 of its gene, and only its exon 10 microsatellite sequence is mutated in colonic tumors with MSI, indicating exonic selectivity of ACVR2 for frameshift mutation (4). We replicated this observation in a cell model (13) and demonstrated that this exonic selectivity is partially caused by the immediate flanking DNA sequences surrounding both coding microsatellites of ACVR2 (13). Our observations suggest a role of the flanking sequence context for microsatellite mutation with defective DNA MMR in addition to the general observation that longer length microsatellites are more likely to undergo frameshift mutation (8), although this last point has not been experimentally tested for human genes.

Several characteristics of microsatellites have been shown to influence the extent of their instability with defective DNA MMR, using non-coding sequences in mostly yeast or bacterial models, and include: repeat-unit length [e.g. An versus (GA)n versus (GAC)n], length of the microsatellite (e.g. A7 versus A13), base composition (e.g. An versus Gn), sequence context and the degree of ‘perfection’ of the repeated microsatellite; pure repeats are less stable than mixed repeats (e.g. A17 is less stable than A8GA8) (14–21). However, there are no data regarding frameshift mutation rates on microsatellite length within actual human genes with DNA MMR deficiency, largely because a good model to measure human mutation was not available compared with fast-growing bacteria or yeast. We developed a model in which to test the hypothesis that longer microsatellite lengths within TGFBR2 and ACVR2 coding sequences would undergo more rapid frameshift mutation in DNA MMR deficiency. To test this, we changed the length of microsatellite sequences within TGFBR2 exon 3 and ACVR2 exon 10 and generated three different constructs (A7, A10 and A13). These constructs were cloned into an EGFP plasmid 1 bp out of frame (OF) such that a −1 bp frameshift mutation could be detected by EGFP expression. Observed in real time, we demonstrate that longer microsatellites within TGFBR2 and ACVR2 coding sequences undergo higher frameshift mutation rates in DNA MMR deficiency. We further demonstrate that TGFBR2 containing an A13 microsatellite sequence is the most prone to the frameshift mutation when DNA MMR is defective, even over the A13 microsatellite constructed for ACVR2. These findings showed experimentally that the length of a coding microsatellite greatly determines its mutation rate with defective DNA MMR. Additionally, by directly comparing identical length ACVR2 and TGFBR2 constructs, we showed that other factors such as the sequence context play a role in determining frameshift mutation rates.

RESULTS

Longer microsatellites within TGFBR2 and ACVR2 coding sequences mutate more frequently in hMLH1−/− cells

To test whether coding microsatellite length influences frameshift mutation in DNA MMR deficiency, we constructed TGFBR2 and ACVR2 plasmids containing different microsatellite lengths (A7, A10 and A13) within exon 3 of human TGFBR2 and exon 10 of human ACVR2 sequences by modifying previously constructed plasmids (12) that were designed to detect a −1 bp frameshift mutation (Table 1) as described in Materials and Methods. In our cell model, a −1 bp frameshift mutation at the microsatellite is detected by EGFP expression because the TGFBR2 and ACVR2 sequences are cloned 1 bp OF of EGFP immediately after the start codon of the EGFP gene in pIREShyg2-EGFP. Mutation resistant (MR) in frame (IF) of the EGFP plasmids and MR OF plasmids were constructed as positive and negative controls, respectively, for each length of microsatellite within each gene (TGFBR2 and ACVR2) by interrupting the microsatellite sequence to prevent frameshift mutation. Positive and negative control plasmids and corresponding experimental plasmids were transfected into hMLH1-deficient (completely DNA MMR-defective) HCT116 cells, and stably transfected cell lines (Supplementary Material, Table S1) were established by hygromycin B selection for mutation analysis. Sequences containing 9, 11 or 12 polyadenines were not used for the establishment of stable cell lines due to the predicted generation of a new in-frame stop codon or a new out-frame start codon which would confound mutation analysis by EGFP expression. At 2 weeks after selection, the proportion of fluorescent cells for each cell line was measured by flow cytometry. All MR IF cell lines containing the various microsatellite lengths within TGFBR2 or ACVR2 showed fluorescence between 93 and 100% (median 99%), demonstrating robust EGFP expressional selection efficiency of the MR IF stable cell lines. The hMLH1−/− cells containing TGFBR2 (A7), TGFBR2 (A10) or TGFBR2 (A13) OF sequences or ACVR2 (A7), ACVR2 (A10) or ACVR2 (A13) OF sequences revealed newly fluorescent cells ranging between 0.01 and 4.18% (median 0.31%) net fluorescence over matched counterpart cell lines containing MR OF sequences.

Table 1.

Sequences of TGFBR2 exon 3 and ACVR2 exon 10 containing different lengths of microsatellites used for cloning into pIREShyg2-EGFP plasmid

EGFP plasmids Microsatellite sequences
 
 A7 A10 A13 
MR TGFBR2 IF ◂A2GA2CA▸ ◂A2CA2GA2CA▸ ◂A2GA2CA2GA2CA▸ 
MR TGFBR2 OF ◂A2GA2CA▹ ◂A2CA2GA2CA▹ ◂A2GA2CA2GA2CA▹ 
TGFBR2 OF ◂A7 ▹ ◂A10 ▹ ◂A13 ▹ 
MR ACVR2 IF ⇐A2GA2CA ⇒ ⇐A2CA2GA2CA ⇒ ⇐A2GA2CA2GA2CA ⇒ 
MR ACVR2 OF ⇐A2GA2CA⇒ ⇐A2CA2GA2CA ⇒ ⇐A2GA2CA2GA2CA ⇒ 
ACVR2 OF ⇐ A7⇒ ⇐ A10⇒ ⇐ A13⇒ 
EGFP plasmids Microsatellite sequences
 
 A7 A10 A13 
MR TGFBR2 IF ◂A2GA2CA▸ ◂A2CA2GA2CA▸ ◂A2GA2CA2GA2CA▸ 
MR TGFBR2 OF ◂A2GA2CA▹ ◂A2CA2GA2CA▹ ◂A2GA2CA2GA2CA▹ 
TGFBR2 OF ◂A7 ▹ ◂A10 ▹ ◂A13 ▹ 
MR ACVR2 IF ⇐A2GA2CA ⇒ ⇐A2CA2GA2CA ⇒ ⇐A2GA2CA2GA2CA ⇒ 
MR ACVR2 OF ⇐A2GA2CA⇒ ⇐A2CA2GA2CA ⇒ ⇐A2GA2CA2GA2CA ⇒ 
ACVR2 OF ⇐ A7⇒ ⇐ A10⇒ ⇐ A13⇒ 

The various TGFBR2 exon 3 and ACVR2 exon 10 sequences were inserted immediately after the translation initiation codon of EGFP gene, being IF of the EGFP gene or 1 bp OF of the EGFP gene in the pIRES hyg2-EGFP plasmid. MR plasmids were constructed by interrupting microsatellite sequences to prevent any frameshift mutation. MR IF and MR OF plasmids were used as positive and negative controls for EGFP expression, respectively. OF plasmids are the experimental plasmids to detect a −1 bp frameshift mutation by EGFP expression.

◂, ▸, ▹, ⇐, ⇒ and ⇒ represent partial sequences of TGFBR2 exon 3 and ACVR2 exon 10. ◂, CTGCTTCTCCAAAGTGCATTATGAAGG; ▸, GCCTGGTGAGACTTTCT; ▹, GCCTGGTGAGACTTTCTT; ⇐, GACCTGTAGATGAATACATGTTGCCATTTGAGGAGGAAATTGGCCAGCATCCATCTCTTGAAGACATGCAGGAAGTTGTTGTGCAT; ⇒, GAGGCCTG TTT; ⇒, GAGGCCTG TTTT.

To compare mutation frequencies and rates of cell lines containing TGFBR2 or ACVR2 sequences with different-length microsatellites in DNA MMR deficiency, non-fluorescent cells containing either MR TGFBR2 (A7) OF, TGFBR2 (A7) OF, MR TGFBR2 (A10) OF, TGFBR2 (A10) OF, MR TGFBR2 (A13) OF, TGFBR2 (A13) OF, MR ACVR2 (A7) OF, ACVR2 (A7) OF, MR ACVR2 (A10) OF, ACVR2 (A10) OF, MR ACVR2 (A13) OF or ACVR2 (A13) OF sequences were sorted and exponentially grown for 7–35 days. At specific time points (days 7, 14, 21, 28 and 35), two cultures of each cell line were analyzed in parallel for EGFP expression by using flow cytometry to detect a −1 bp frameshift mutation. Details of flow cytometry analysis are described previously (12). Representative EGFP histograms at day 21 are shown in Figure 1 for mutation analysis that compares EGFP expression in hMLH1−/−TGFBR2 and hMLH1−/−ACVR2 cells containing different-length microsatellites (A7, A10 and A13). hMLH1−/− MR TGFBR2 OF cell lines demonstrated approximately three to four times less EGFP expression (0.02–0.08%) compared with hMLH1−/− MR ACVR2 OF cell lines, which indicates a higher background of EGFP expression of hMLH1−/−ACVR2 cell lines. Except for hMLH1−/−TGFBR2 OF and ACVR2 OF cell lines containing the A7 microsatellite, all other OF cell lines revealed significantly higher EGFP expression (0.5–6.6%) compared with each MR counterpart cell line (P < 0.05). The longer the microsatellite length was within both TGFBR2 and ACVR2, the higher the proportion of EGFP expression was observed (A13 > A10 > A7) with 46-, 24- and 1.4-fold higher EGFP expression in TGFBR2 OF cell lines and 18-, 15- and 1.2-fold higher EGFP expression in ACVR2 OF cell lines compared with each MR counterpart cell line (Fig. 1). As we observed previously in hMLH1−/−TGFBR2 (A10) OF and ACVR2 (A8) OF cells (12), two distinct EGFP-expressing cell populations, M1 and M2, were observed from hMLH1−/−TGFBR2 OF and ACVR2 OF cells containing A10 or A13 microsatellite sequences, in which the M2 cells showed brighter EGFP expression compared with the M1 cells (data not shown). However, there was no statistical increase in the proportion of the M2 population in hMLH1−/−TGFBR2 (A7) OF and ACVR2 (A7) OF cells when compared with each MR counterpart.

Figure 1.

Mutation analysis of hMLH1−/− cells containing varied microsatellite sequences within TGFBR2 exon 3 and ACVR2 exon 10 by flow cytometry. Non-fluorescent cells were sorted and exponentially grown for 7–35 days. At specific time points, cells were harvested, and 200 000 cells were analyzed for EGFP expression (identifying a −1 bp frameshift mutation) by flow cytometry. Gated live cells were analyzed on an EGFP histogram and two distinct EGFP populations were plotted. The population with dim EGFP expression was named M1, and the population with bright EGFP expression was named M2. EGFP histograms of MR TGFBR2 OF, TGFBR2 OF, MR ACVR2 OF and ACVR2 OF cells containing different lengths of microsatellites (A7, A10 and A13) in the hMLH1−/− background at day 21 were shown as representatives of mutation analysis. Note that different EGFP gates were set to properly distinguish the M1 and M2 populations in TGFBR2 and ACVR2 cells and that scaling of cell counts in each EGFP histogram is different.

Figure 1.

Mutation analysis of hMLH1−/− cells containing varied microsatellite sequences within TGFBR2 exon 3 and ACVR2 exon 10 by flow cytometry. Non-fluorescent cells were sorted and exponentially grown for 7–35 days. At specific time points, cells were harvested, and 200 000 cells were analyzed for EGFP expression (identifying a −1 bp frameshift mutation) by flow cytometry. Gated live cells were analyzed on an EGFP histogram and two distinct EGFP populations were plotted. The population with dim EGFP expression was named M1, and the population with bright EGFP expression was named M2. EGFP histograms of MR TGFBR2 OF, TGFBR2 OF, MR ACVR2 OF and ACVR2 OF cells containing different lengths of microsatellites (A7, A10 and A13) in the hMLH1−/− background at day 21 were shown as representatives of mutation analysis. Note that different EGFP gates were set to properly distinguish the M1 and M2 populations in TGFBR2 and ACVR2 cells and that scaling of cell counts in each EGFP histogram is different.

The −1 bp frameshift mutation frequency at each time point was expressed as a fold change using the following formula: (EGFP-positive cells/total live cells from TGFBR2 OF or ACVR2 OF cells)/(EGFP-positive cells/total live cells from MR TGFBR2 OF or MR ACVR2 OF cells). As we observed previously (12), the M2 population from the TGFBR2 (A10), TGFBR2 (A13), ACVR2 (A10) and ACVR2 (A13) sequences in the hMLH1−/− cells dramatically accumulated over time, whereas the M1 population showed relatively little change over time (Fig. 2), indicating that the M1 and M2 are distinct populations. The M1 and M2 populations were plotted separately for analysis of mutation frequency. Regarding the M1 population (Fig. 2A), there was no consistent overall increase in the mutation frequency for any construct, but the cells containing longer microsatellites showed a relatively higher mutation frequency in both TGFBR2 and ACVR2 sequences compared with MR constructs (Fig. 2A, insert). Comparing TGFBR2 and ACVR2, the TGFBR2 (A10) sequence showed significantly higher mutation frequency than ACVR2 (A10) sequence after day 7 (P < 0.05) in the M1 population, but there was no significant difference in mutation frequency of the M1 population between TGFBR2 and ACVR2 sequences containing A7 and A13 microsatellites. For the M2 population, the hMLH1−/− cells containing longer microsatellites within both TGFBR2 and ACVR2 sequences showed a significantly higher mutation frequency (P < 0.05); at day 35, we observed a 501-, 90- and 3-fold change in TGFBR2 cell lines and a 429-, 129- and 14-fold change in ACVR2 cell lines at A13, A10 and A7 microsatellites, respectively (Fig. 2B). However, overall, there was no significant difference in mutation frequency between TGFBR2 and ACVR2 sequences with A10 and A13 microsatellites for the M2 population. Although the total percentages of EGFP-positive cells are low (<0.03%) in the M2 population, hMLH1−/−ACVR2 (A7) OF cells showed significantly higher mutation frequency compared with hMLH1−/−TGFBR2 (A7) OF cells (P < 0.05).

Figure 2.

Mutation frequencies of TGFBR2 exon 3 and ACVR2 exon 10 containing different-length microsatellites in the hMLH1−/− background. Non-fluorescent cells were analyzed for EGFP expression by flow cytometry at 7, 14, 21, 28 and 35 days after being sorted and cultured, and EGFP analysis was performed at each day as shown in Figure 1. Mutation frequency at each time point was expressed as a fold change using the following formula: (EGFP-positive cells/total live cells from TGFBR2 OF or ACVR2 OF cells)/(EGFP-positive cells/total live cells from MR TGFBR2 OF or MR ACVR2 OF cells). The M2 population from the TGFBR2 (A10), TGFBR2 (A13), ACVR2 (A10) and ACVR2 (A13) sequences in the hMLH1−/− background greatly accumulated over time, whereas the M1 population showed relatively little change over time, indicating that the M1 and M2 are distinct populations. Overall, longer microsatellites within both TGFBR2 and ACVR2 sequences showed a higher mutation frequency at the microsatellite in both M1 and M2 populations (A13 > A10 > A7). Cell lines showing lower mutation frequencies (<30-fold change) were separately plotted in the right panel using a smaller y-axis scale. Data are means from two independent experiments at each time point.

Figure 2.

Mutation frequencies of TGFBR2 exon 3 and ACVR2 exon 10 containing different-length microsatellites in the hMLH1−/− background. Non-fluorescent cells were analyzed for EGFP expression by flow cytometry at 7, 14, 21, 28 and 35 days after being sorted and cultured, and EGFP analysis was performed at each day as shown in Figure 1. Mutation frequency at each time point was expressed as a fold change using the following formula: (EGFP-positive cells/total live cells from TGFBR2 OF or ACVR2 OF cells)/(EGFP-positive cells/total live cells from MR TGFBR2 OF or MR ACVR2 OF cells). The M2 population from the TGFBR2 (A10), TGFBR2 (A13), ACVR2 (A10) and ACVR2 (A13) sequences in the hMLH1−/− background greatly accumulated over time, whereas the M1 population showed relatively little change over time, indicating that the M1 and M2 are distinct populations. Overall, longer microsatellites within both TGFBR2 and ACVR2 sequences showed a higher mutation frequency at the microsatellite in both M1 and M2 populations (A13 > A10 > A7). Cell lines showing lower mutation frequencies (<30-fold change) were separately plotted in the right panel using a smaller y-axis scale. Data are means from two independent experiments at each time point.

TGFBR2 is more susceptible to frameshift mutation at an A13 microsatellite sequence compared with ACVR2

We performed DNA sequencing analysis to confirm that EGFP expression from the M1 and M2 populations was caused by a −1 bp frameshift mutation at the microsatellites of TGFBR2 OF and ACVR2 OF. At day 21 after being plated as non-fluorescent cells, the M1 and/or M2 cell populations of hMLH1−/− MR TGFBR2 (A7) OF, TGFBR2 (A7) OF, MR TGFBR2 (A10) OF, TGFBR2 (A10) OF, MR TGFBR2 (A13) OF, TGFBR2 (A13) OF, MR ACVR2 (A7) OF, ACVR2 (A7) OF, MR ACVR2 (A10) OF, ACVR2 (A10) OF, MR ACVR2 (A13) OF and ACVR2 (A13) OF cells were sorted and cultured for ∼2 weeks for DNA sequencing analysis. DNA from each cell line was amplified by PCR and then sequenced to assess for frameshift mutation at the microsatellites (Fig. 3). No frameshift mutation was observed in any MR TGFBR2 OF or MR ACVR2 OF cell lines as we expected (data not shown), indicating that EGFP fluorescence in these cells is the background fluorescence. As we observed in hMLH1−/−TGFBR2 (A10) OF and ACVR2 (A8) OF cells previously (12), the M2 population of hMLH1−/−TGFBR2 and hMLH1−/−ACVR2 OF cell lines revealed −1 bp frameshifted microsatellite sequences (A7 to A6, A10 to A9 and A13 to A12) except hMLH1−/−TGFBR2 (A7) OF cells that lacked enough M2 population cells to be sorted for DNA sequencing (Fig. 3). The M2 populations of hMLH1−/−TGFBR2 (A10) OF, ACVR2 (A7) OF, ACVR2 (A10) OF and ACVR2 (A13) OF cells revealed −1 bp frameshifted microsatellite without a significant overlap with each corresponding non-frameshifted microsatellite sequence, indicating full mutant (A6, A9 and A12 paired with T6, T9 and T12, respectively) formation that induces bright EGFP expression as observed previously (12). Furthermore, the M2 population of hMLH1−/−TGFBR2 (A13) OF cells showed a more extensive frameshift mutation pattern. It showed a higher proportion of −1 bp frameshifted microsatellite sequence (A12) overlapping with other shortened microsatellite sequences (indicated by arrows in Fig. 3), suggesting multiple heteroduplex formation with further frameshifts in the M2 population. The TGFBR2 (A13) OF sequence was the only sequence to show evidence of continuing mutation beyond a −1 bp frameshift full mutation.

Figure 3.

DNA sequencing analysis demonstrating frameshift mutations at different-length microsatellites within TGFBR2 exon 3 and ACVR2 exon 10 sequences in the hMLH1−/− background. Cells from the M1 and/or M2 populations of hMLH1−/−TGFBR2 exon 3 OF and hMLH1−/−ACVR2 exon 10 OF cells containing different microsatellite sequences were sorted and cultured for DNA sequencing analysis. DNA from each cell line was amplified by PCR, and the anti-sense strand was sequenced to assess for frameshift mutation at the microsatellites of hMLH1−/−TGFBR2 exon 3 OF and hMLH1−/−ACVR2 exon 10 OF cells. The M1 populations of cells containing A10 or A13 microsatellite sequences revealed overlapping sequences, indicating heteroduplex formation. The M1 cells containing TGFBR2 (A10) OF or TGFBR2 (A13) OF sequences revealed −1 bp frameshifts (A9 and A12) as dominant sequences, whereas cells containing ACVR2 (A10) OF or ACVR2 (A13) OF sequences revealed A10 and A13 sequences as dominant. No frameshift mutation was observed in the M1 populations of hMLH1−/−TGFBR2 (A7) OF and hMLH1−/−ACVR2 (A7) OF cells. The M2 population containing TGFBR2 (A10) OF, ACVR2 (A7) OF, ACVR2 (A10) OF and ACVR2 (A13) OF sequences showed −1 bp frameshifted sequence with minimal overlap with the WT sequence, suggesting stable full mutant formation. However, the M2 population of TGFBR2 (A13) OF cells revealed an A12 microsatellite sequence as dominant but overlapped with further frameshifted microsatellite sequences (arrows), suggesting additional heteroduplex formation with continuing frameshifts.

Figure 3.

DNA sequencing analysis demonstrating frameshift mutations at different-length microsatellites within TGFBR2 exon 3 and ACVR2 exon 10 sequences in the hMLH1−/− background. Cells from the M1 and/or M2 populations of hMLH1−/−TGFBR2 exon 3 OF and hMLH1−/−ACVR2 exon 10 OF cells containing different microsatellite sequences were sorted and cultured for DNA sequencing analysis. DNA from each cell line was amplified by PCR, and the anti-sense strand was sequenced to assess for frameshift mutation at the microsatellites of hMLH1−/−TGFBR2 exon 3 OF and hMLH1−/−ACVR2 exon 10 OF cells. The M1 populations of cells containing A10 or A13 microsatellite sequences revealed overlapping sequences, indicating heteroduplex formation. The M1 cells containing TGFBR2 (A10) OF or TGFBR2 (A13) OF sequences revealed −1 bp frameshifts (A9 and A12) as dominant sequences, whereas cells containing ACVR2 (A10) OF or ACVR2 (A13) OF sequences revealed A10 and A13 sequences as dominant. No frameshift mutation was observed in the M1 populations of hMLH1−/−TGFBR2 (A7) OF and hMLH1−/−ACVR2 (A7) OF cells. The M2 population containing TGFBR2 (A10) OF, ACVR2 (A7) OF, ACVR2 (A10) OF and ACVR2 (A13) OF sequences showed −1 bp frameshifted sequence with minimal overlap with the WT sequence, suggesting stable full mutant formation. However, the M2 population of TGFBR2 (A13) OF cells revealed an A12 microsatellite sequence as dominant but overlapped with further frameshifted microsatellite sequences (arrows), suggesting additional heteroduplex formation with continuing frameshifts.

On the other hand, no frameshift mutation was observed in the M1 population of either hMLH1−/−TGFBR2 (A7) OF or hMLH1−/−ACVR2 (A7) OF cells, which was anticipated from their lower mutation frequencies compared with the other OF cells containing longer microsatellites (Figs 1 and 2). As observed previously (12), the M1 population of hMLH1−/−TGFBR2 (A10) OF cells revealed a −1 bp frameshifted microsatellite sequence (A9) without significant overlap with the WT microsatellite sequence, but the M1 population of hMLH1−/−ACVR2 (A10) OF cells revealed a higher proportion of the WT microsatellite sequence overlapping with an A9 microsatellite sequence (Fig. 3). Additionally, the M1 population of hMLH1−/−ACVR2 (A13) OF cells revealed a higher proportion of A13 microsatellite sequence overlapping with A12 microsatellite sequence. Different from hMLH1−/−ACVR2 (A13) OF cells, the M1 population of hMLH1−/−TGFBR2 (A13) OF cells revealed a higher proportion of −1 bp frameshifted microsatellite sequence (A12) overlapped with multiple shortened microsatellite sequences (arrows in Fig. 3). This type of sequence data indicate heteroduplex formation in the M1 population of hMLH1−/−TGFBR2 (A13) OF cell line as observed in its M2 population, although the proportion of its overlapped sequences seems to be less compared with the M2 population.

To dissect the overlapping sequences in the M1 and M2 populations of hMLH1−/−TGFBR2 (A13) OF and hMLH1−/−ACVR2 (A13) OF cells, DNA from each cell line was amplified by PCR, subcloned into a TA cloning vector, and single-cell clones were individually sequenced. As suggested by the data in Figure 3, single-DNA clones from the M1 and M2 populations of TGFBR2 (A13) OF cells revealed different microsatellite frameshifts (Table 2). In the M1 population of TGFBR2 (A13) OF cells, the majority of DNA clones (52%) showed a −1 bp frameshift sequence (A12), and 22% (5/23) of clones showed an A13 sequence. Interestingly, A10 (−3 bp frameshifted, 13%) and A11 (−2 bp frameshifted, 13%) sequences were also observed in the M1 population of TGFBR2 (A13) OF cells, as indicated in the pooled sequences (arrows in Fig. 3). The M2 population of TGFBR2 (A13) OF cells also revealed a −1 bp frameshift (A12) as the dominant microsatellite (50%); the M2 population showed much less non-frameshifted microsatellite (A13) (4.2%) and more −2 bp frameshifted microsatellite (A11) sequences (33.3%) compared with the M1 population, indicating continuing further deletion frameshift mutation on top of the −1 bp frameshift mutation in the M2 population of TGFBR2 (A13) OF cells (these cells were sorted with the −1 bp frameshifts expressing fluorescence). The TGFBR2 (A13) OF also showed an A10 microsatellite sequence (4.2%), and it revealed A14 and A15 microsatellite sequences (4.2% each) that were not observed in the M1 population. This indicates frameshift expansion in the TGFBR2 (A13) OF cells, a finding not observed in the A7 and A10 sequences.

Table 2.

Frameshift mutations at A13 microsatellite within TGFBR2 exon 3 and ACVR2 exon 10 coding sequences in the hMLH1−/− background

Gene TGFBR2
 
ACVR2
 
EGFP intensity M1 M2 M1 M2 
A9/T9 clones 
A10/T10 clones 
A11/T11 clones 
A12/T12 clones 12 12 14 
A13/T13 clones 13 
A14/T14 clones 
A15/T15 clones 
%Mutated 78 96 31.5 84 
Gene TGFBR2
 
ACVR2
 
EGFP intensity M1 M2 M1 M2 
A9/T9 clones 
A10/T10 clones 
A11/T11 clones 
A12/T12 clones 12 12 14 
A13/T13 clones 13 
A14/T14 clones 
A15/T15 clones 
%Mutated 78 96 31.5 84 

Cells from the M1 and M2 populations of hMLH1−/−TGFBR2 (A13) OF and hMLH1−/−ACVR2 (A13) OF cells were sorted and cultured. DNA from each cell line was amplified by PCR, subcloned and all single-cell clones were individually sequenced to assess for frameshift mutation. The M1 population of both cell lines indicates heteroduplex formation (a mixture of mutant and WT), but TGFBR2 (A13) OF cells revealed more −1 bp frameshifts when compared with ACVR2 (A13) OF cells. In the M2 population, ACVR2 (A13) OF cells revealed −1 bp frameshifts (74%) dominantly, suggesting full mutant formation (A12/T12), whereas TGFBR2 (A13) OF cells indicated continuous heteroduplex formation (dominantly A12/T11) passing through full mutant formation (A12/T12). This suggests that TGFBR2 is more susceptible to frameshift mutation at the A13 microsatellite in the hMLH1−/− background compared with ACVR2.

Conversely, the majority of DNA clones in the M1 population of hMLH1−/−ACVR2 (A13) OF cells revealed the A13 microsatellite sequence (68.5%) with small proportions of A10, A12 and A14 sequences (10.5%), whereas the majority of DNA clones in the M2 population of ACVR2 (A13) OF cells revealed a −1 bp frameshifted sequence (A12, 74%) with 16% of A13 microsatellite and 5% of A11 sequences. A rare A9 sequence (5%) was also observed.

The sequencing analysis indicates that TGFBR2 and ACVR2 sequences mutate differently at longer microsatellites (A13). The M1 populations of both TGFBR2 (A13) OF and ACVR2 (A13) OF sequences indicate heteroduplex formation, but TGFBR2 (A13) OF cells revealed more −1 bp frameshifts. Furthermore, the M2 population of TGFBR2 (A13) OF cells suggests a continuing transition from a full-mutant population to further heteroduplex populations, generating additional shortened frameshifted sequences. These observations suggest that hMLH1−/−TGFBR2 (A13) OF cells are the most prone to frameshift mutation at the microsatellite compared with other OF cell lines in this study.

The −1 bp frameshift mutation rates at the microsatellite within TGFBR2 and ACVR2 coding sequences are dependent on the length of the microsatellite

The M2 population was used to calculate the mutation rates at microsatellites of TGFBR2 and ACVR2 sequences by the ‘method of the mean’ as described previously (12). The mutation rate at the microsatellite of TGFBR2 (A13) sequence in hMLH1 deficiency was highest: 22.38 × 10−4 ± 2.50 × 10−4, but it was not significantly different from that of the ACVR2 (A13) sequence (18.32 × 10−4 ± 0.52 × 10−4) (Table 3). The lack of significant difference may be caused by a smaller proportion of −1 bp frameshifts in the M2 population of TGFBR2 (A13) OF cells compared with ACVR2 (A13) OF cells (Table 2; 50 versus 79%). As predicted in Figures 1 and 2, significantly higher mutation rates were observed in the hMLH1−/− cells containing longer microsatellites within both TGFBR2 and ACVR2 compared with shorter microsatellites (P< 0.05) (Table 3). The mutation rate at the microsatellite of TGFBR2 (A10) sequence is ∼10 times lower than that of TGFBR2 (A13) sequence, whereas it is ∼16 times higher than that of TGFBR2 (A7) sequence (Table 3). Mutation at the microsatellite of ACVR2 (A13) sequence is ∼2.6 times more frequent than mutation at the microsatellite of ACVR2 (A10) sequence, which is ∼16 times higher compared with that at the microsatellite of ACVR2 (A7) sequence. At the A7 microsatellite sequence, ACVR2 showed approximately three times higher −1 bp frameshift mutation rate compared with TGFBR2 (A7) sequence (P < 0.05). Different from the mutation frequency observed in Figure 2B, ACVR2 showed approximately three times higher −1 bp frameshift mutation rate compared with TGFBR2 at A10 microsatellite sequence (P < 0.05). Overall, the longer the length of the coding microsatellite, the higher its mutation rate.

Table 3.

Calculated −1 bp frameshift mutation rates at the microsatellites within TGFBR2 and ACVR2 coding sequences in the hMLH1−/− background

Gene Microsatellite Rate for mutation 
TGFBR2 A13 22.38 × 10−4 ± 2.50 × 10−4(a) 
TGFBR2 A10 2.17 × 10−4 ± 0.11 × 10−4(b) 
TGFBR2 A7 0.13 × 10−4 ± 0.05 × 10−4(c) 
ACVR2 A13 18.32 × 10−4 ± 0.52 × 10−4(a) 
ACVR2 A10 7.11 × 10−4 ± 0.42 × 10−4(d) 
ACVR2 A7 0.44 × 10−4 ± 0.08 × 10−4(e) 
Gene Microsatellite Rate for mutation 
TGFBR2 A13 22.38 × 10−4 ± 2.50 × 10−4(a) 
TGFBR2 A10 2.17 × 10−4 ± 0.11 × 10−4(b) 
TGFBR2 A7 0.13 × 10−4 ± 0.05 × 10−4(c) 
ACVR2 A13 18.32 × 10−4 ± 0.52 × 10−4(a) 
ACVR2 A10 7.11 × 10−4 ± 0.42 × 10−4(d) 
ACVR2 A7 0.44 × 10−4 ± 0.08 × 10−4(e) 

Data from the M2 cell populations at each time point between day 21 and day 35 were used for −1 bp frameshift mutation rate analysis. Single-mutation rates were calculated by combining and averaging time-specific mutation rates. Rates are expressed as mutations at microsatellite sequence per cell per generation. Data shown are mean ± SEM.

a, b, c, d and e represent significant difference in the mutation rate (P < 0.05). a: TGFBR2 A13 and ACVR2 A13 versus TGFBR2 A10 and A7, and ACVR2 A10 and A7; b: TGFBR2 A10 versus TGFBR2 A13 and A7, and ACVR2 A13, A10 and A7; c: TGFBR2 A7 versus TGFBR2 A13 and A10, and ACVR2 A13, A10 and A7; d: ACVR2 A10 versus TGFBR2 A13, A10 and A7, and ACVR2 A13 and A7; e: ACVR2 A7 versus TGFBR2 A13, A10 and A7, and ACVR2 A13 and A10.

DISCUSSION

In this study, we measured and compared mutation rates at varying lengths of microsatellites within actual human coding sequences: TGFBR2 exon 3 and ACVR2 exon 10 in hMLH1−/− colorectal cancer cells. Other studies have shown that the propensity for DNA slippage increased with the length of microsatellite by studying non-coding microsatellite sequences (14,17,20). In addition, by utilizing coding microsatellites within human genes, we demonstrated that the mutation rate at the A10 microsatellite of TGFBR2 exon 3 is significantly higher than the A8 microsatellite of ACVR2 exon 10 in DNA MMR-defective cells, which was hypothesized to be caused by the longer microsatellite length within TGFBR2 (12). Besides the length of a microsatellite, other factors may influence mutation rates at microsatellite sequences such as repeat-unit length, base composition of the microsatellite and sequence context (14–20). Exonic selectivity of ACVR2 for frameshift mutation has been observed in colorectal cancers with MSI (4,10) as well as in DNA MMR-deficient human colorectal cells (13) even though both microsatellites within ACVR2 are identical. Furthermore, we demonstrated that the ACVR2 exonic selectivity is partially caused by the flanking DNA sequences surrounding each microsatellite within ACVR2 (13). This suggests that the sequence context may be a key factor that allows frameshift mutation independent of microsatellite length. The effect of microsatellite length upon frameshift mutation rates has not previously been tested for coding microsatellites. We had hypothesized that longer microsatellites within TGFBR2 and ACVR2 would undergo more frameshift mutation in DNA MMR-deficient cells. To examine this hypothesis, we utilized the DNA MMR-deficient hMLH1−/− cell line in which we transfected TGFBR2 exon 3 and ACVR2 exon 10 sequences containing different-length microsatellites (A7, A10 and A13) with the sequences 1 bp OF to EGFP, which allowed us to observe −1 bp frameshift mutations occurring in real time with the onset of EGFP fluorescence. This experimental design allowed us to compare mutation rates of different-length microsatellites within the same gene, as well as the same microsatellite length across two different genes (i.e. two different sequence contexts).

As we observed in TGFBR2 exon 3 and ACVR2 exon 10 OF cells previously (12), OF cells containing A10 and A13 microsatellites revealed two distinct EGFP populations of mutant cells, an M1 population expressing dim fluorescence and an M2 population expressing bright fluorescence (Fig. 1). As we noted previously, the M2 population accumulated over time, but the M1 population showed little change (Fig. 2). We observed significantly higher mutation frequencies in hMLH1−/− cells containing longer microsatellites within TGFBR2 and ACVR2 in both M1 and M2 populations (Fig. 2). For the M1 population, which represents heteroduplex intermediates prior to full mutation, we observed a significantly higher mutation frequency in hMLH1−/− cells containing TGFBR2 (A10) sequence compared with ACVR2 (A10) sequence after day 7 (P < 0.05) (Fig. 2A), which was confirmed by DNA sequencing (Fig. 3). The M1 population of hMLH1−/−TGFBR2 (A10) OF cells revealed a −1 bp frameshift (A9) with a small proportion of WT microsatellite sequence (A10), whereas hMLH1−/−ACVR2 (A10) OF cells revealed a higher proportion of A10 microsatellite sequence overlapping with the A9 microsatellite sequence. For the M2 population, hMLH1−/−ACVR2 (A7) OF cells showed significantly higher mutation frequency than hMLH−/−TGFBR2 (A7) OF cells (P < 0.05) (Fig. 2B). Except for these two observations, we did not find any significant difference in mutation frequency for the same length of microsatellite between TGFBR2 exon 3 and ACVR2 exon 10 sequences in either M1 or M2 populations.

DNA sequencing analysis revealed different frameshift mutation patterns in the cells containing A10 and A13 microsatellites (Fig. 3). As observed previously (12), the M2 populations of TGFBR2 and ACVR2 containing the A10 microsatellite demonstrated that they are full mutants, revealing a −1 bp frameshifted sequence (A9 paired with T9). The M2 populations of TGFBR2 and ACVR2 containing the A13 microsatellite revealed a mixture of frameshift mutations and indicated that the A13 microsatellite sequence is subjected to a continuous and accelerated pattern of frameshift mutation in DNA MMR deficiency. DNA sequencing analysis supports that TGFBR2 (A13) OF cells are most susceptible to frameshift mutation at this length of microsatellite in the hMLH1−/−background compared with other OF cell lines.

Mutation rates calculated for −1 bp frameshift mutation within the M2 population confirmed our hypothesis that longer microsatellites within TGFBR2 and ACVR2 coding sequences have significantly higher mutation rates in DNA MMR deficiency (Table 3). Compared with the native TGFBR2 (A10) coding sequence, the modified TGFBR2 (A13) and TGFBR2 (A7) sequences showed ∼10 times higher and ∼16 times lower mutation rates, respectively. Modified ACVR2 (A13) and ACVR2 (A7) sequences showed ∼3 times higher and ∼16 times lower mutation rates compared with the modified ACVR2 (A10) sequence, respectively. The TGFBR2 (A13) mutation rate is not significantly different from the ACVR2 (A13) sequence, and the failure to detect a difference is likely caused by characteristics of our cell-model system. Our system does not allow detection of −2 bp frameshifts [33% of the M2 population from the TGFBR2 (A13) sequence], and thus we are unable to calculate such mutation rates. Because of this, we believe that the mutation rate for the TGFBR2 (A13) sequence is underestimated. In contrast, shorter microsatellite sequences (A10 and A7) allowed us to detect significant differences in mutation rate for identical length microsatellites between TGFBR2 and ACVR2 sequences, likely because there was no ongoing frameshift mutation beyond the −1 bp full mutants. In any event, these differences detected with identical length microsatellites suggest other factors that influence frameshift mutation, the most likely being sequence context, which was shown to be a partial cause of exonic selectivity of ACVR2 for frameshift mutation in defective DNA MMR (13). In hMLH1 deficiency, substitution of flanking nucleotides surrounding exon 10 microsatellite of ACVR2 with those surrounding the exon 3 microsatellite of ACVR2 greatly diminished heteroduplex (A7 paired with T8) and full (A7 paired with T7) frameshift mutation, whereas substitution of flanking nucleotides from exon 3 with those from the exon 10 enhanced frameshift mutation (13). Additionally, when the immediate surrounding DNA sequences of the A10 microsatellite of TGFBR2 exon 3 were swapped with DNA sequences from ACVR2 exon 3, we observed a significant decrease in frameshift mutation at the A10 microsatellite of native TGFBR2 in the hMLH1−/− background (Supplementary Material, Fig. S1). All these observations suggest that in addition to microsatellite length, the surrounding sequence context of microsatellite likely regulates frameshift mutation rate in DNA MMR deficiency. As suggested in Figure 4, the sequence context may have more influence on frameshift mutation at shorter microsatellite lengths. At longer microsatellite lengths, the microsatellite length may be the dominant factor influencing frameshift mutation.

Figure 4.

Schematic of the effect of ‘the microsatellite length’ and ‘sequence context’ on frameshift mutation at varied microsatellite lengths within TGFBR2 exon 3 and ACVR2 exon 10 coding sequences in hMLH1−/− cells. In both TGFBR2 and ACVR2 sequences, longer microsatellites mutated faster (distinguished by the thickness of arrows). The ACVR2 sequence frameshifted −1 bp faster at A10 and A7 microsatellite sequences compared with the TGFBR2 sequence (red versus black arrows), suggesting that another factor such as sequence context might regulate frameshift mutation at shorter microsatellite sequences. On the other hand, TGFBR2 and ACVR2 containing A13 microsatellite sequences frameshifted −1 bp at similar rates with the TGFBR2 sequence continuing to frameshift further, suggesting that microsatellite length may be the key regulator for frameshift mutation at longer microsatellite sequences (A13). Dotted arrows mean that mutation rates were not measured (our cell model can calculate only −1 bp frameshift mutation) but were estimated from DNA sequencing analysis.

Figure 4.

Schematic of the effect of ‘the microsatellite length’ and ‘sequence context’ on frameshift mutation at varied microsatellite lengths within TGFBR2 exon 3 and ACVR2 exon 10 coding sequences in hMLH1−/− cells. In both TGFBR2 and ACVR2 sequences, longer microsatellites mutated faster (distinguished by the thickness of arrows). The ACVR2 sequence frameshifted −1 bp faster at A10 and A7 microsatellite sequences compared with the TGFBR2 sequence (red versus black arrows), suggesting that another factor such as sequence context might regulate frameshift mutation at shorter microsatellite sequences. On the other hand, TGFBR2 and ACVR2 containing A13 microsatellite sequences frameshifted −1 bp at similar rates with the TGFBR2 sequence continuing to frameshift further, suggesting that microsatellite length may be the key regulator for frameshift mutation at longer microsatellite sequences (A13). Dotted arrows mean that mutation rates were not measured (our cell model can calculate only −1 bp frameshift mutation) but were estimated from DNA sequencing analysis.

In summary, using our cell model in which we measure frameshift mutation in real time, we demonstrate that longer microsatellites within TGFBR2 exon 3 and ACVR2 exon 10 coding sequences undergo higher frameshift mutation rates in DNA MMR deficiency than shorter microsatellites. Our observations further suggest that frameshift mutation is influenced by both microsatellite length and its sequence context. Our study has implications for the mutation rates at other coding microsatellite sequences and for how DNA MMR deficiency drives frameshift mutation.

MATERIALS AND METHODS

Cloning of pIREShyg2-TGFBR2-EGFP and pIREShyg2-ACVR2-EGFP plasmids containing different-length microsatellites

Plasmids pIREShyg2-TGFBR2-EGFP containing portions of exon 3 of TGFBR2 and pIREShyg2-ACVR2-EGFP containing portions of exon 10 of ACVR2 were constructed previously (12). Details of plasmid cloning were previously described (12). With these plasmids containing native coding microsatellite sequences (A10 for TGFBR2 and A8 for ACVR2), pIREShyg2-TGFBR2-EGFP and pIREShyg2-ACVR2-EGFP plasmids containing varying lengths of microsatellites were constructed by changing only microsatellite sequences (A10 to A7 and A13 for TGFBR2 and A8 to A7, A10 and A13 for ACVR2) by using a QuikChange II site-directed mutagenesis kit (Stratagene, La Jolla, CA, USA) (Table 1). For experimental plasmids, TGFBR2 or ACVR2 sequences were cloned 1 bp OF in pIREShyg2-EGFP and thus a −1 bp frameshift mutation at the microsatellite would shift the EGFP gene into the proper reading frame to allow EGFP expression. MR counterpart plasmids were constructed by interrupting microsatellites (A7 to A2GA2CA, A10 to A2CA2GA2CA and A13 to A2GA2CA2GA2CA) using a QuikChange II site-directed mutagenesis kit (Stratagene) to prevent any frameshift mutation. The MR plasmids were placed OF (1 bp) and IF to be used as negative and positive controls for EGFP expression, respectively. Positive colonies were screened, and the correct sequences of TGFBR2 and ACVR2 with different microsatellite lengths were confirmed by DNA sequencing in an ABI 3700 analyzer.

Cell lines, transfection and selection

The human colon cancer cell line, HCT116 (hMLH1−/−), was obtained from American Type Culture Collection (Rockville, MD, USA) and maintained in Iscove's modified Dulbecco's medium (Invitrogen Corp., Carlsbad, CA, USA) with 10% fetal bovine serum and penicillin (100 U/ml)/streptomycin (100 μg/ml) (Invitrogen Corp.) as supplements. Cells were transfected with the different pIREShyg2-TGFBR2-EGFP and pIREShyg2-ACVR2-EGFP plasmids (Table 1) by using Nucleofector Kit V (Amaxa, Cologne, Germany) following the manufacturer's instructions. Selection with hygromycin B (Invitrogen Corp.) commenced at 24 h after nucleofection to generate stable cell lines. After selection, colonies from each cell line were initially pooled and cultured for mutation analysis. All stable cell lines were confirmed by DNA sequencing.

Mutation analysis by flow cytometry

Details of analysis of mutant cells by flow cytometry were previously described (12). Briefly, 5000 non-fluorescent cells expressing MR TGFBR2 (A7) OF, TGFBR2 (A7) OF, MR TGFBR2 (A10) OF, TGFBR2 (A10) OF, MR TGFBR2 (A13) OF, TGFBR2 (A13) OF, MR ACVR2 (A7) OF, ACVR2 (A7) OF, MR ACVR2 (A10) OF, ACVR2 (A10) OF, MR ACVR2 (A13) OF or ACVR2 (A13) OF were sorted into 24-well plates on a FACS ARIA by using Diva software [Becton Dickinson Immunocytometry Systems (BDIS), San Jose, CA, USA]. During a 7–35 day analysis period, cultures were expanded as required to keep cells in exponential growth. Cells were trypsinized, washed in PBS and resuspended in a total volume of 200 μl of PBS/0.5 μg/ml of propidium iodide (PI) and 3% BSA. Cell suspensions were analyzed on a FACSCalibur with CELLQUEST acquisition and analysis software (BDIS). At specified time points, two cultures were analyzed in parallel. To identify EGFP-positive cells, region (R)1, R2 and R3 were set according to cell size, fluorescence and live cells, respectively. Cells from the gated R1 and R3 and R2 were further plotted on a fluorescence intensity histogram. The population displaying no fluorescence and two distinct EGFP-positive populations were separated. The population with dim EGFP fluorescence was designated M1, and the population with bright EGFP fluorescence was designated M2. The counts of M1 and M2 cells were expressed as percentages of R3 (total live cell number). Different EGFP gates were set to properly distinguish the M1 and M2 populations of hMLH1−/−TGFBR2 and hMLH1−/−ACVR2 cell lines. A lower EGFP gate was set for hMLH1−/−TGFBR2 cell lines compared with hMLH1−/−ACVR2 cell lines.

PCR and DNA sequencing

Total cellular DNA from stable cell lines and M1 and M2 cell populations were PCR-amplified by specific primers 5′-GCGTCGTTTAAACCTGCTTCTCCAAAGTGCATTATG-3′ and 5′-TGCCGTCGTCCTTGAAGAAGA-3′ for TGFBR2 and 5′-GATCCGCCACCATGTTTAAACGAC-3′ and 5′-GCTGTTGTAGTTGTACTCCAGCTTG-3′ for ACVR2 in a reaction containing the primers, buffer, DNA template, deoxynucleotides and Pfu ultra high-fidelity DNA polymerase (Stratagene). The PCR products were used for DNA sequencing to identify stable cell lines and frameshift mutations at microsatellites of TGFBR2 and ACVR2. In addition, we subcloned PCR-amplified TGFBR2 OF (A13) and ACVR2 OF (A13) DNA fragments from the M1 and M2 cell populations by utilizing a TA cloning vector (Invitrogen Corp.) as per the manufacturer's protocol. DNA clones were then individually sequenced to determine the prevalence of mutated and A13 microsatellite sequences of TGFBR2 and ACVR2.

Determination of −1 bp frameshift mutation rates at the microsatellite of TGFBR2 and ACVR2 in MMR-deficient cells

Mutation rates were calculated by the method of the mean developed by Luria and Delbruck (22) as described previously (12). The formula used in the computation is forumla, where forumla is the mean number of mutants in a culture, C the number of parallel cultures, μ the mutation rate and N the number of cells at risk of undergoing a mutation, which Luria–Delbruck assumed to be equal to the final number of cells in a culture. Two parallel cultures were used, and forumla was estimated as the mean of the number of mutants across the two cultures. The total number of cells, N, was based on averaging two cultures. The mutation rates of the M2 cell population (full mutants) were calculated using data from flow cytometry analysis at each time point between day 21 and day 35. Single-mutation rates were then calculated by combining and averaging time-specific mutation rates to minimize the variance of the estimate. Data were expressed as mean ± the standard errors of mean (SEM).

Statistical analysis

Mutation frequencies and rates of cell lines were compared by t-test or one-way ANOVA.

SUPPLEMENTARY MATERIAL

Supplementary Material is available at HMG online.

FUNDING

This work was supported by the United States Public Health Service (DK067287 to J.M.C.) and the UCSD Digestive Diseases Research Development Center (DK080506).

ACKNOWLEDGEMENTS

We thank the support of the following shared resources: Moores UCSD Cancer Center Flow Cytometry Shared Resource and Moores UCSD Cancer Center DNA Sequencing Shared Resource.

Conflict of Interest statement. None declared.

REFERENCES

1
Thibodeau
S.N.
Bren
G.
Schaid
D.
Microsatellite instability in cancer of the proximal colon
Science
 , 
1993
, vol. 
260
 (pg. 
816
-
819
)
2
Parsons
R.
Myeroff
L.L.
Liu
B.
Willson
J.K.
Markowitz
S.D.
Kinzler
K.W.
Vogelstein
B.
Microsatellite instability and mutations of the transforming growth factor beta type II receptor gene in colorectal cancer
Cancer Res.
 , 
1995
, vol. 
55
 (pg. 
5548
-
5550
)
[PubMed]
3
Mori
Y.
Yin
J.
Rashid
A.
Leggett
B.A.
Young
J.
Simms
L.
Kuehl
P.M.
Langenberg
P.
Meltzer
S.J.
Stine
O.C.
Instabilotyping: comprehensive identification of frameshift mutations caused by coding region microsatellite instability
Cancer Res.
 , 
2001
, vol. 
61
 (pg. 
6046
-
6049
)
[PubMed]
4
Jung
B.
Doctolero
R.T.
Tajima
A.
Nguyen
A.K.
Keku
T.
Sandler
R.S.
Carethers
J.M.
Loss of activin receptor type 2 protein expression in microsatellite unstable colon cancers
Gastroenterology
 , 
2004
, vol. 
126
 (pg. 
654
-
659
)
[PubMed]
5
Guanti
G.
Resta
N.
Simone
C.
Cariola
F.
Demma
I.
Fiorente
P.
Gentile
M.
Involvement of PTEN mutations in the genetic pathways of colorectal cancerogenesis
Hum. Mol. Genet.
 , 
2000
, vol. 
9
 (pg. 
283
-
287
)
6
Rampino
N.
Yamamoto
H.
Ionov
Y.
Li
Y.
Sawai
H.
Reed
J.C.
Perucho
M.
Somatic frameshift mutations in the BAX gene in colon cancers of the microsatellite mutator phenotype
Science
 , 
1997
, vol. 
275
 (pg. 
967
-
969
)
7
Woerner
S.M.
Kloor
M.
Mueller
A.
Rueschoff
J.
Friedrichs
N.
Buettner
R.
Buzello
M.
Kienle
P.
Knaebel
H.P.
Kunstmann
E.
, et al.  . 
Microsatellite instability of selective target genes in HNPCC-associated colon adenomas
Oncogene
 , 
2005
, vol. 
24
 (pg. 
2525
-
2535
)
8
Grady
W.M.
Carethers
J.M.
Genomic and epigenetic instability in colorectal cancer pathogenesis
Gastroenterology
 , 
2008
, vol. 
135
 (pg. 
1079
-
1099
)
9
Jung
B.
Smith
E.J.
Doctolero
R.T.
Gervaz
P.
Alonso
J.C.
Miyai
K.
Keku
T.
Sandler
R.S.
Carethers
J.M.
Influence of target gene mutations on survival, stage and histology in sporadic microsatellite unstable colon cancers
Int. J. Cancer
 , 
2006
, vol. 
118
 (pg. 
2509
-
2513
)
10
Hempen
P.M.
Zhang
L.
Bansal
R.K.
Iacobuzio-Donahue
C.A.
Murphy
K.M.
Maitra
A.
Vogelstein
B.
Whitehead
R.H.
Markowitz
S.D.
Willson
J.K.
, et al.  . 
Evidence of selection for clones having genetic inactivation of the activin A type II receptor (ACVR2) gene in gastrointestinal cancers
Cancer Res.
 , 
2003
, vol. 
63
 (pg. 
994
-
999
)
[PubMed]
11
Markowitz
S.
Wang
J.
Myeroff
L.
Parsons
R.
Sun
L.
Lutterbaugh
J.
Fan
R.S.
Zborowska
E.
Kinzler
K.W.
Vogelstein
B.
, et al.  . 
Inactivation of the type II TGF-beta receptor in colon cancer cells with microsatellite instability
Science
 , 
1995
, vol. 
268
 (pg. 
1336
-
1338
)
12
Chung
H.
Young
D.J.
Lopez
C.G.
Le
T.A.
Lee
J.K.
Ream-Robinson
D.
Huang
S.C.
Carethers
J.M.
Mutation rates of TGFBR2 and ACVR2 coding microsatellites in human cells with defective DNA mismatch repair
PLoS ONE
 , 
2008
, vol. 
3
 pg. 
e3463
 
13
Chung
H.
Lopez
C.G.
Young
D.J.
Lai
J.F.
Holmstrom
J.
Ream-Robinson
D.
Cabrera
B.L.
Carethers
J.M.
Flanking sequence specificity determines coding microsatellite heteroduplex and mutation rates with defective DNA mismatch repair (MMR)
Oncogene
 , 
2010
, vol. 
29
 (pg. 
2172
-
2180
)
[PubMed]
14
Schlotterer
C.
Tautz
D.
Slippage synthesis of simple sequence DNA
Nucleic Acids Res.
 , 
1992
, vol. 
20
 (pg. 
211
-
215
)
15
Bichara
M.
Schumacher
S.
Fuchs
R.P.
Genetic instability within monotonous runs of CpG sequences in Escherichia coli
Genetics
 , 
1995
, vol. 
140
 (pg. 
897
-
907
)
[PubMed]
16
Sia
E.A.
Kokoska
R.J.
Dominska
M.
Greenwell
P.
Petes
T.D.
Microsatellite instability in yeast: dependence on repeat unit size and DNA mismatch repair genes
Mol. Cell. Biol.
 , 
1997
, vol. 
17
 (pg. 
2851
-
2858
)
[PubMed]
17
Wierdl
M.
Dominska
M.
Petes
T.D.
Microsatellite instability in yeast: dependence on the length of the microsatellite
Genetics
 , 
1997
, vol. 
146
 (pg. 
769
-
779
)
[PubMed]
18
Eckert
K.A.
Yan
G.
Mutational analyses of dinucleotide and tetranucleotide microsatellites in Escherichia coli: influence of sequence on expansion mutagenesis
Nucleic Acids Res.
 , 
2000
, vol. 
28
 (pg. 
2831
-
2838
)
19
Harfe
B.D.
Jinks-Robertson
S.
Sequence composition and context effects on the generation and repair of frameshift intermediates in mononucleotide runs in Saccharomyces cerevisiae
Genetics
 , 
2000
, vol. 
156
 (pg. 
571
-
578
)
[PubMed]
20
Yamada
N.A.
Smith
G.A.
Castro
A.
Roques
C.N.
Boyer
J.C.
Farber
R.A.
Relative rates of insertion and deletion mutations in dinucleotide repeats of various lengths in mismatch repair proficient mouse and mismatch repair deficient human cells
Mutat. Res.
 , 
2002
, vol. 
499
 (pg. 
213
-
225
)
[PubMed]
21
Boyer
J.C.
Hawk
J.D.
Stefanovic
L.
Farber
R.A.
Sequence-dependent effect of interruptions on microsatellite mutation rate in mismatch repair-deficient human cells
Mutat. Res.
 , 
2008
, vol. 
640
 (pg. 
89
-
96
)
[PubMed]
22
Luria
S.E.
Delbruck
M.
Mutations of bacteria from virus sensitivity to virus resistance
Genetics
 , 
1943
, vol. 
28
 (pg. 
491
-
511
)
[PubMed]