The Msh2 mismatch repair (MMR) protein is critical for class switch recombination (CSR) events that occur in mice that lack the Sμ tandem repeat (SμTR) region (SμTR−/− mice). The pattern of microhomology among switch junction sites in Msh2-deficient mice is also dependent on the presence or absence of SμTR sequences. It is not known whether these CSR effects reflect an individual function of Msh2 or the function of Msh2 within the MMR machinery. In the absence of the SμTR sequences, Msh2 deficiency nearly ablates CSR. We now show that Mlh1 or Exo1 deficiencies also eliminate CSR in the absence of the SμTR. Furthermore, in SμTR−/− mice, deficiencies of Mlh1 or Exo1 result in increased switch junction microhomology as has also been seen with Msh2 deficiency. These results are consistent with a CSR model in which the MMR machinery is important in processing DNA nicks to produce double-stranded breaks, particularly in sequences where nicks are infrequent. We propose that double-stranded break paucity in MMR-deficient mice leads to increased use of an alternative joining pathway where microhomologies are important for CSR break ligation. Interestingly, when the SμTR region is present, deficiency of Msh2 does not lead to the increased microhomology seen with Mlh1 or Exo1 deficiencies, suggesting that Msh2 might have an additional function in CSR. It is also possible that the inability to initiate MMR in the absence of Msh2 results in CSR junctions with less microhomology than joinings that occur when MMR is initiated but then proceeds abnormally due to Mlh1 or Exo1 deficiencies.

Antibody gene class switch recombination (CSR)3 occurs in activated B cells and causes a switch from IgM and IgD Ab expression to IgG, IgE, or IgA expression. CSR changes Ab effector functions while maintaining Ag specificity by exchanging the μ C region gene segment (Cμ) in an Ab gene with one of several other downstream C region gene segments. CSR occurs by an intrachromosomal deletional recombination between switch region (S region) sequences located upstream of each Ig heavy chain C region gene (1).

A number of proteins are important for CSR. The activation-induced cytidine deaminase (AID) protein is required for CSR; AID activity has been shown to convert cytosine to uracil in ssDNA (2, 3, 4, 5, 6, 7, 8). AID has known activity hotspots, such as WRC sequence motifs (W = A or T and R = G or A) (7, 9, 10), that are abundant in S region tandem repeat sequences. The deoxyuridine residues introduced by AID into S regions are thought to be substrates for various DNA damage repair proteins that lead to the DNA breaks needed for the recombination events involved in CSR (11, 12, 13). The base excision repair protein uracil DNA glycosylase (UNG) 2 is involved in removing U bases from DNA. Humans and mice having defective UNG expression show substantial reductions in CSR (11, 12). UNG2 activity results in abasic DNA sites that are cleaved by AP endonucleases (APEX1/APEX2) to create single-stranded nicks in the DNA duplex (14).

Several studies have shown that proteins involved in nonhomologous end joining (NHEJ) play an important role in CSR (15, 16). These results suggest that many CSR events represent joining between blunt or nearly blunt dsDNA breaks within two different S regions. AID-induced DNA nicks that are closely spaced, perhaps due to closely spaced WRC AID motifs in S region tandem repeats, could provide the nearly blunt DNA breaks needed for NHEJ and CSR. Widely spaced DNA nicks in S regions would require additional DNA processing to be converted into dsDNA breaks that can be used for CSR.

Mismatch repair (MMR) proteins also appear to have a role in CSR. Mice deficient in the individual MMR proteins Msh2, Msh6, Mlh1, Pms2, or Exo1 all exhibit partial, but significant, reductions in CSR (17, 18, 19, 20, 21). Although each individual MMR protein-deficient mouse has a similar reduction in CSR efficiency, the different proteins have different effects on the length of microhomology at CSR junctions (17, 18, 19, 20, 21, 22). For instance, it has been reported that B cells from Msh2- and Exo1-deficient mice display less switch junction microhomology than wild-type (WT) mice, whereas switch junction microhomology in Mlh1- and Pms2-deficient B cells is increased compared with WT (17, 18, 22). These results suggest that different MMR protein deficiencies can have different effects on the CSR joining process or on the substrates available for recombination. Furthermore, other studies have suggested that Msh2 facilitates S region synapsis (23) and that CSR events that occur in the absence of UNG are dependent on Msh2, although the other MMR proteins have not been tested (13); these findings could indicate CSR roles for the Msh2 protein that may not involve the entire MMR machinery. However, Msh2, Mlh1, or Pms2 deficiency all result in 2- to 5-fold reductions in CSR-associated Sμ DNA double-stranded breaks (24), suggesting that these proteins contribute to the processing of DNA breaks during CSR.

Surprisingly, unlike the partial reductions of CSR observed with MMR protein deficiencies in mice with intact Sμ regions, the Msh2 protein has been found to be critical for CSR in mice that lack the Sμ tandem repeat (SμTR) region (25). B cells that lack either the Msh2 protein or the SμTR sequences have ∼2-fold reductions in CSR, but B cells that lack both the Msh2 protein and the SμTR sequences have an almost complete block in CSR (19, 25, 26). It is not clear why Msh2 has such a strong effect on CSR in SμTR−/− mice. We have proposed that SμTR−/− mice have lower levels of AID-induced DNA nicking sites in the JH-Cμ intron due to the lack of the tandem repeat sequences. Therefore, we hypothesize that Msh2 is required and works as part of the MMR machinery to convert widely spaced DNA nicks into dsDNA breaks for CSR (24, 25, 27). However, it is also possible that Msh2 has unique functions that are needed for CSR in the absence of the SμTR region.

In this study we have used SμTR−/− mice to assess the effects of Mlh1 and Exo1 deficiency on CSR and to determine whether the Msh2, Mlh1, and Exo1 MMR proteins all exhibit similar effects on CSR efficiency and junction microhomology. We find that indeed SμTR−/−:Mlh1−/− and SμTR−/−:Exo1−/− mice show the same large reductions in CSR as seen in SμTR−/−:Msh2−/− mice. This indicates that the entire MMR machinery is important for CSR events that occur in the Sμ regions that lack tandem repeats and supports the hypothesis that the MMR machinery converts widely spaced DNA nicks into nearly blunt DNA breaks. We suggest that MMR-deficient B cells exhibit deficits of the near-blunt ends needed for NHEJ and, therefore, an increased use of an alternative joining pathway that uses microhomologies surrounding DNA nicks to join S regions.

Mlh1−/−, Exo1−/−, and SμTR−/− mice, described previously (17, 22, 26), were crossed to generate SμTR−/−:Mlh1−/− and SμTR−/−:Exo1−/− mice. All animal studies have been reviewed and approved by the Tufts Medical Center Division of Laboratory Animal Medicine (Boston, MA).

Splenic B cells were isolated, cultured, and analyzed by flow cytometry as described (25).

Splenic B cells were purified as referenced above or by negative selection (EasySep mouse B cell enrichment kit; Stem Cell Technologies). B cells were cultured to induce IgG3 or IgG1 switching as previously described (25) or by culturing at 5 × 105 cells/ml supplemented with LPS (25 μg/ml; Sigma-Aldrich) for IgG3 or with LPS and IL-4 (10 ng/ml; PeproTech) for IgG1. DNA was isolated from 4- to 5-day cultures by phenol/chloroform separation and ethanol precipitation.

Sμ-Sγ1 junctions were amplified using two sets of nested PCR primers. One PCR was described previously (25). The second PCR used a nested primer set, SuO.8 (5′-GATGCTGTCTCTATTCAGTTATAC-3′) and Sγ1O.8 (5′-TTCAGTTAGGACTCCCACAC-3′) for the first round and SuI.8 (5′-GAATCTATTCTGGCTCTTCTTAAGC-3′) and Sγ1I.8 (5′-TTCTGCATTACTCCCAACCTC-3′) for the second round of PCR. Sμ-Sγ3 junctions were amplified using three different PCRs, two of which were described previously (22, 25). The third PCR used the nested primer set SuO.8 and Sγ3O.8 (5′-GTCTCCAGTTCTCTCACCTC-3′) for the first round and SuI.8 and Sγ3I.8 (5′-CACATCCTCACTTGCCACTG-3′) for the second round of PCR. Both described rounds of Sμ-Sγ1 and Sμ-Sγ3 PCR amplification performed were 35 cycles of 94°C for 30 s, 60°C for 30 s, and 68°C for 4 min. All PCRs were performed using Expand long template Taq and Pfu polymerase mix (Roche). PCR products were cloned and sequenced as previously described (25). Sequenced switch junctions were aligned to the published sequences MUSIGCD07, MUSIGHANA, MUSIGCD18, D78344, and FJ389571 or to Sμ region sequences accumulated in previous studies (25, 26) and analyzed using the GAP (Accelrys) program. A Student’s two-tailed t test was used to determine the statistical difference in length of microhomology between WT and mutant mouse strains.

The importance of Msh2 for CSR events in regions flanking the SμTR region could reflect a requirement for a specific individual function of Msh2 or the need for Msh2 as part of the entire MMR machinery. To assess whether other MMR proteins are critical for CSR in SμTR flanking sequences, we crossed SμTR−/− mice with either Mlh1−/− or Exo1−/− mice to produce double mutant SμTR−/−:Mlh1−/− and SμTR−/−:Exo1−/− mice and analyzed CSR in these animals. We chose to analyze double mutants lacking Mlh1 or Exo1 because results from previous studies that examined deficiencies in these MMR proteins showed similar effects on CSR levels but different effects on switch junction microhomologies (17, 22).

To quantify CSR in the double-mutant mice, we stimulated purified B cells to undergo CSR for 4 days in culture and then analyzed for cell surface expression of different Ab isotypes. FACS analyses from SμTR−/−:Mlh1−/− B cells stimulated in culture with LPS and IL-4 to induce switching to IgG1 are shown in Fig. 1 A. Double-mutant SμTR−/−:Mlh1−/− B cells showed severe reductions in IgG1 switching when compared with the WT or single mutant controls. Only 5% of WT IgG1 CSR levels could be detected in SμTR−/−:Mlh1−/− B cells. Similar to previous results (20, 26), B cells from control SμTR−/− mice and Mlh1−/− mice exhibit roughly 50% of WT CSR when compared with B cells from WT littermate controls. These results show that the Mlh1 protein is critical for IgG1 CSR events that occur in sequences flanking the SμTR region.

FIGURE 1.

CSR is significantly reduced in SμTR−/−:Mlh1−/− and SμTR−/−:Exo1−/− mice. Purified B cells from the mouse strains indicated were cultured in vitro for 4 days with LPS and cytokines to induce CSR to each isotype and then assayed by flow cytometry to determine the ability of B cells from each mouse strain to undergo CSR. A, Representative FACS dot plots comparing IgG1 CSR in WT, Mlh1−/−, SμTR−/−, and SμTR−/−:Mlh1−/− mice. Cells expressing surface IgG1 are indicated by gated area. B, Comparisons of CSR for various isotypes in the mouse strains are indicated. Average WT levels of CSR for each isotype (IgG1, 24%; IgG2a, 19%; IgG2b, 20%; IgG3, 18%; and IgA, 11%) were set to 100%. CSR levels for mutant strains are shown as an average percentage of WT CSR with error bars representing the range between experiments. Single WT, SμTR−/−, Mlh1−/−, and SμTR−/−:Mlh1−/− mice were analyzed in these experiments. For these mice, multiple independent cultures were assayed and n equals the number of cultures that were used as listed below. For the WT, SμTR−/−, and SμTR−/−:Mlh1−/− mice, n = 3 for all isotypes except IgG2b, where n = 2. For Mlh1−/−, n = 2 for all isotypes except IgG2b, where n = 1. C, Representative FACS dot plots comparing, as in A, IgG1 CSR in WT, Exo1−/−, SμTR−/−, and SμTR−/−:Exo1−/− mice. D, Comparisons of CSR for the mouse strains indicated were, as in B, the average WT levels of CSR for each isotype (IgG1, 29%; IgG2b, 20%; IgG3, 15%; and IgA, 9%) were set to 100% and CSR levels for mutant mouse strains are represented as a percentage of WT CSR. Results are from two experiments that analyzed CSR levels in single cultures from each of two WT, two SμTR−/−, four Exo1−/−, and three SμTR−/−:Exo1−/− mice.

FIGURE 1.

CSR is significantly reduced in SμTR−/−:Mlh1−/− and SμTR−/−:Exo1−/− mice. Purified B cells from the mouse strains indicated were cultured in vitro for 4 days with LPS and cytokines to induce CSR to each isotype and then assayed by flow cytometry to determine the ability of B cells from each mouse strain to undergo CSR. A, Representative FACS dot plots comparing IgG1 CSR in WT, Mlh1−/−, SμTR−/−, and SμTR−/−:Mlh1−/− mice. Cells expressing surface IgG1 are indicated by gated area. B, Comparisons of CSR for various isotypes in the mouse strains are indicated. Average WT levels of CSR for each isotype (IgG1, 24%; IgG2a, 19%; IgG2b, 20%; IgG3, 18%; and IgA, 11%) were set to 100%. CSR levels for mutant strains are shown as an average percentage of WT CSR with error bars representing the range between experiments. Single WT, SμTR−/−, Mlh1−/−, and SμTR−/−:Mlh1−/− mice were analyzed in these experiments. For these mice, multiple independent cultures were assayed and n equals the number of cultures that were used as listed below. For the WT, SμTR−/−, and SμTR−/−:Mlh1−/− mice, n = 3 for all isotypes except IgG2b, where n = 2. For Mlh1−/−, n = 2 for all isotypes except IgG2b, where n = 1. C, Representative FACS dot plots comparing, as in A, IgG1 CSR in WT, Exo1−/−, SμTR−/−, and SμTR−/−:Exo1−/− mice. D, Comparisons of CSR for the mouse strains indicated were, as in B, the average WT levels of CSR for each isotype (IgG1, 29%; IgG2b, 20%; IgG3, 15%; and IgA, 9%) were set to 100% and CSR levels for mutant mouse strains are represented as a percentage of WT CSR. Results are from two experiments that analyzed CSR levels in single cultures from each of two WT, two SμTR−/−, four Exo1−/−, and three SμTR−/−:Exo1−/− mice.

Close modal

Fig. 1 B and supplemental Table I4 summarize additional analyses using a variety of B cell stimulation conditions and show that the important role of Mlh1 in SμTR flanking region CSR events extends across multiple CH gene isotypes. The severe reduction in CSR found in SμTR−/−: Mlh1−/− mice is similar to the reductions previously shown in mice that lack both Msh2 and the SμTR region (25).

We also analyzed isotype switching in mice that lack the MMR protein Exo1. FACS analyses of splenic B cells cultured with LPS and IL-4 are shown for SμTR−/−:Exo1−/− double-mutant mice along with single mutant and WT littermate controls (Fig. 1,C). Consistent with a previous report (17), we found that Exo1 protein deficiency causes a 3-fold reduction in IgG1 switching (Fig. 1,C). Analyses of CSR in the double-mutant SμTR−/−:Exo1−/− mice show a severe reduction in IgG1 switching to ∼5% of WT CSR (Fig. 1,C). Fig. 1 D and supplemental Table II show that the reduction in IgG1 switching found in SμTR−/−:Exo1−/− double mutants extends to multiple isotypes as well. These findings indicate a critical function for both Mlh1 and Exo1 in CSR events occurring in the sequences flanking the SμTR region.

Previous studies have shown that MMR-deficient mice exhibit altered patterns in the structures of switch junctions generated during CSR (17, 18, 22, 25). In WT mice, a majority of switch junctions (94–100%) exhibit a limited amount of microhomology ranging from zero to four nucleotides in length; longer microhomologies of five nucleotides or more are found infrequently (0–3.1%) (18, 22).

Previous studies have also indicated that Msh2 deficiency results in a decrease in microhomology found at switch junctions as compared with switch junctions from WT mice. In Msh2−/− mice it was reported that 89% of the microhomology-based CSR joins displayed only one or two nucleotides of microhomology, and no CSR join could be identified with greater than three nucleotides of microhomology (22). However, the switch junction patterns from SμTR−/−:Msh2−/− mice appear to be distinct from the switch junction patterns of Msh2-deficient mice. Double-deficient SμTR−/−:Msh2−/− mice exhibit a set of Sμ-Sγ3 switch junctions with long stretches of microhomology that are rare in WT mice and undetectable in Msh2−/− mice (22, 25). In SμTR−/−:Msh2−/− mice, 19.1% of Sμ-Sγ3 junctions have five or more nucleotides of microhomology (25). These findings suggest that, for Msh2 deficient mice, the ends available for joining and the joining mechanism used for CSR depend on the presence or absence of SμTR sequences. No significant increase in microhomology was found in SμTR−/−:Msh2−/− Sμ-Sγ1 junctions compared with WT, suggesting the possibility that Msh2 deficiency may affect the CSR joining mechanism used in an isotype-specific manner.

To assess whether removal of SμTR sequences might alter the types of switch junctions observed in mice that lack other MMR proteins, we compared switch junction sequences in SμTR−/−:Mlh1−/− and SμTR−/−:Exo1−/− mice with junctions from WT, Mlh1−/−, Exo1−/−, and SμTR−/− mice. Switch junctions in stimulated B cell populations were amplified by PCR and aligned to corresponding germline S region sequences (see Materials and Methods).

Table I and Fig. 2 summarize the characteristics of switch junctions that we found in WT Mlh1+/+, Mlh1+/−, and Exo1+/− littermate mice (collectively grouped as “WT” in Table I, Fig. 2, and supplemental Fig. 1). All of these control genotypes show similar microhomology frequencies. Consistent with previously published analyses of switch junctions in WT mice (18, 22), these control junctions are mainly blunt or have short microhomologies. Switch junctions with longer stretches of microhomology are detected more frequently in Mlh1−/− and Exo1−/− mice (Table I and supplemental Fig. 1). We found that 13.9% of Exo1−/− Sμ-Sγ3 junctions and 6.3% of Mlh1−/− Sμ-Sγ3 junctions had five or more nucleotides of microhomology as compared with 2.6% in WT controls (Table I). The patterns of Sμ-Sγ3 junction microhomology in Mlh1−/− and Exo1−/− mice are statistically different from those in WT (Table I). In addition to an increase in the percentage of junctions with five or more nucleotides of homology, there was also an increase in the average length of the microhomologies found at the Mlh1−/− (1.7 bp) and Exo1−/− (1.9 bp) switch junctions, compared with WT (0.9 bp). Analogous analyses of Sμ-Sγ1 switch junctions were performed and produced similar results as those obtained for Sμ-Sγ3 switch junctions (Table I and supplemental Fig. 1).

Table I.

Comparison of CSR junction structure in WT and mutant mouse strainsa

MouseSμ-Sγ3 junctions
Blunt (%)1–4 Nucleotides (%)≥5 Nucleotides (%)Insert (%)No. of Junctionsp Valueb
WT 29 47 21 38  
SμTR−/− 25 75 16 0.1839 
Exo1−/− 25 50 14 11 36 0.0245 
SμTR−/−:Exo1−/− 30 50 10 10 20 0.2404c 
Mlh1−/− 17 68 63 0.0084 
SμTR−/−:Mlh1−/− 15 50 25 10 20 0.0076 
SμTR−/−:Msh2−/− 10 57 19 14 21 0.0023 
MouseSμ-Sγ3 junctions
Blunt (%)1–4 Nucleotides (%)≥5 Nucleotides (%)Insert (%)No. of Junctionsp Valueb
WT 29 47 21 38  
SμTR−/− 25 75 16 0.1839 
Exo1−/− 25 50 14 11 36 0.0245 
SμTR−/−:Exo1−/− 30 50 10 10 20 0.2404c 
Mlh1−/− 17 68 63 0.0084 
SμTR−/−:Mlh1−/− 15 50 25 10 20 0.0076 
SμTR−/−:Msh2−/− 10 57 19 14 21 0.0023 
MouseSμ-Sγ1 junctions
Blunt (%)1–4 Nucleotides (%)≥5 Nucleotides (%)Insert (%)No. of Junctionsp Valuec
WT 31 43 25 32  
SμTR−/− 36 59 22 0.1539 
Exo1−/− 26 48 16 10 31 0.0023 
SμTR−/−:Exo1−/− 22 56 17 18 0.0191 
Mlh1−/− 28 57 11 54 0.0251 
SμTR−/−:Mlh1−/− 11 85 27 0.0038 
SμTR−/−:Msh2−/− 37 58 19 0.5544 
MouseSμ-Sγ1 junctions
Blunt (%)1–4 Nucleotides (%)≥5 Nucleotides (%)Insert (%)No. of Junctionsp Valuec
WT 31 43 25 32  
SμTR−/− 36 59 22 0.1539 
Exo1−/− 26 48 16 10 31 0.0023 
SμTR−/−:Exo1−/− 22 56 17 18 0.0191 
Mlh1−/− 28 57 11 54 0.0251 
SμTR−/−:Mlh1−/− 11 85 27 0.0038 
SμTR−/−:Msh2−/− 37 58 19 0.5544 
a

CSR junctions were amplified from 1–5 individual cultures from three WT (mutant mouse littermates), four SμTR−/−:Exo1−/−, one SμTR−/−:Mlh1−/−, seven Exo1−/−, and five Mlh1−/− mice.

b

The significance of the difference in length of microhomology between WT and mutant mouse strains was determined using a Student’s two-tailed t test. Values of p < 0.05 were considered statistically significant.

c

Increases in microhomology are statistically significant for μ-γ 1 Exo1−/− and SμTR−/−:Exo1−/− and μ-γ 3 Exo1−/− CSR junctions. We believe the high p value for the increase in microhomology in SμTR−/−:Exo1−/− μ-γ 3 sequences is coincidental with this sample due to the fact that the increase in longer microhomologies is partially balanced by a concurrent increase in shorter microhomologies.

FIGURE 2.

CSR junctions from MMR-deficient B cells display increased microhomology compared with WT. CSR junctions were amplified and sequenced from B cells activated in vitro to class switch. Percentages of Sμ-Sγ3 (A) and Sμ-Sγ1 (B) junctions that are blunt, have 1–4 nucleotides (nts) of microhomology, 5 or more nucleotides of microhomology, or a nucleotide insert (where the exact junction is undefined) are shown. CSR junctions were amplified from 1–4 in vitro cultures for each mouse analyzed. Three WT (mutant mouse littermates), one SμTR−/−Mlh1−/−, and four SμTR−/−:Exo1−/− mice were used. Results from SμTR−/−:Msh2−/− mice were published previously (25 ).

FIGURE 2.

CSR junctions from MMR-deficient B cells display increased microhomology compared with WT. CSR junctions were amplified and sequenced from B cells activated in vitro to class switch. Percentages of Sμ-Sγ3 (A) and Sμ-Sγ1 (B) junctions that are blunt, have 1–4 nucleotides (nts) of microhomology, 5 or more nucleotides of microhomology, or a nucleotide insert (where the exact junction is undefined) are shown. CSR junctions were amplified from 1–4 in vitro cultures for each mouse analyzed. Three WT (mutant mouse littermates), one SμTR−/−Mlh1−/−, and four SμTR−/−:Exo1−/− mice were used. Results from SμTR−/−:Msh2−/− mice were published previously (25 ).

Close modal

A previous study from Bardwell and colleagues reported switch junctions in Exo1−/− mice (17). The Exo1−/− mouse strain used by Bardwell et al. is the same that we have analyzed. In contrast to our findings, Bardwell at al. concluded that the lack of Exo1 in B cells resulted in a decreased percentage of switch junction sequences with long microhomologies when compared with WT mice (17). We compared the Exo1−/− switch junction data presented by Bardwell et al. (17) to the data that we obtained and found that the amount of microhomology at Exo1−/− switch junctions is the same. In contrast, the two studies differ in the amount of microhomology found in the WT samples. In our study, the lengths of microhomologies for WT samples are shorter than the lengths of the microhomologies found in Exo1−/− mice, whereas Bardwell et al. report that lengths of microhomologies for their WT samples were longer than that for Exo1−/− mice (17). We do not know why the WT data obtained in the two studies are different; possible differences in the genetic backgrounds, the health of the animals used, or the environmental factors of housing may have affected the results. The amount of microhomology in our WT switch junctions (Table I and Fig. 2) closely resembles the amount of microhomology that has been reported for studies investigating proteins other than Exo1 (18, 22).

To assess whether removal of SμTR sequences would alter switch junctions from Mlh1- and Exo1-deficient mice, we amplified switch junctions from SμTR−/−:Mlh1−/− and SμTR−/−:Exo1−/− mice and compared these switch junctions to the junctions amplified from Mlh1−/− and Exo1−/− mice. Similarly as in Exo1−/− and Mlh1−/− mice, we found that an increase in length of microhomology at CSR junctions occurs in SμTR−/−:Exo1−/− and SμTR−/−:Mlh1−/− mice compared with WT (Fig. 2 and supplemental Fig. 1). The amount of microhomologies at SμTR−/−:Exo1−/− mice switch junctions is similar to the amount observed in Exo1−/− mice (Table I). SμTR−/−:Mlh1−/− double deficiency results in a slight increase in the amount of microhomology compared with Mlh1-deficiency alone (Table I). Overall, the Mlh1- and Exo1- deficient mice analyzed all have increased microhomology compared with WT, and removal of the SμTR sequences does not appear to greatly alter the pattern of switch junctions in Mlh1- or Exo1-deficient mice. This differs from the case of Msh2-deficient B cells, which show reduced junctional microhomology when the cells have an intact Sμ region but increased microhomology in cells that lack the SμTRs (Fig. 2 and Refs. 22 and 25). In the absence of SμTR sequences, Msh2, Mlh1, or Exo1 deficiencies all result in a similar increase in switch junction microhomology (Fig. 2 and supplemental Fig. 1). These results suggest that the joining mechanism used for CSR events outside the SμTR sequences is similarly affected by the absence of Msh2, Mlh1, and Exo1.

The results from our studies show that SμTR−/−:Mlh1−/− and SμTR−/−:Exo1−/− mice have severe defects in CSR in vitro. These defects are equivalent to a decreased efficiency of CSR to 5% or less of WT switching for each isotype analyzed. Our studies complement a previous report that analyzed CSR in SμTR−/−:Msh2−/− mice and demonstrate that Msh2, Mlh1, and Exo1 all serve a critical role in CSR events in SμTR flanking sequences. These results show that the critical role for Msh2 in most CSR events in SμTR flanking sequences is not unique to the Msh2 protein, but that these events require multiple MMR proteins. Taken together, the results indicate that the three proteins cooperate in a common function for CSR in SμTR flanking sequences, and we suggest that this function is the processing of widely spaced DNA nicks into double-strand breaks.

MMR proteins work together to repair nucleotide mispairings produced during DNA replication. MMR of single-nucleotide mismatches occurs by the Msh2-Msh6 heterodimer recognizing mismatches, recruiting and forming a complex with the Mlh1-Pms2 heterodimer, and activating mismatch excision mediated by the Exo1 protein, which is initiated at a nearby single-strand break (28). The Exo1 protein hydrolyzes DNA with 5′ to 3′ polarity and therefore can only excise mismatches that lie 3′ to a nick on the same strand (29, 30, 31). The Mlh1-Pms2 heterodimer has endonucleolytic activity and can nick DNA near mismatches to generate a 5′ entry site for Exo1 when a nick is present in the same DNA strand (32). Reconstitution of the MMR process in vitro has demonstrated that repair of nucleotide mismatches can occur in the absence of Mlh1-Pms2 when there is a DNA nick 5′ of the mismatch. However, the Mlh1-Pms2 heterodimer is required for repair when the DNA nick is located 3′ of the mismatch (33). It has also been shown that the Mlh1-Pms2 heterodimer limits the processivity of the Exo1 protein, decreasing the length of excision tracts (34).

We propose that the CSR MMR mechanism involved in processing switch region DNA nicks (Fig. 3) might be similar to the mechanism of MMR during DNA replication (32, 33, 35). In this model, Msh2-Msh6 recognizes U-G mismatches introduced in switch regions by AID and recruits the Mlh1-Pms2 heterodimer to these sites (27). The concomitant activity of the base excision repair proteins UNG2 and APEX1/2 on neighboring U-G mismatches would result in the generation of S region DNA nicks. The presence of S region DNA nicks could allow for additional nicks to be generated surrounding Msh2-Msh6-recognized mismatches by the Mlh1-Pms2 heterodimer. Alternatively, activity at nearby U-G mismatches by the BER protein UNG2 could generate entry sites for Exo1-mediated excision (35). In the absence of Mlh1-Pms2, the remaining MMR factors would need to rely on the activity of other DNA repair proteins, such as UNG2, to create sites that can be used for Exo1 excision. This may lead to longer Exo1 excision tracks if DNA nicks are farther away from Msh2-Msh6-recognized mismatches and also because Exo1 has increased processivity in the absence of the Mlh1-Pms2 heterodimer (34). We propose that when Msh2-Msh6, Mlh1-Pms2, and Exo1 work together during CSR, they could be involved in converting widely spaced S region DNA nicks into blunt or near blunt breaks to be joined by NHEJ proteins. As predicted by this model, fewer double-stranded breaks are detected in MMR-deficient B cells compared with WT, and an even further reduction in Sμ double-stranded breaks is observed in SμTR−/−:Msh2−/− double-deficient B cells than in either SμTR−/− or Msh2−/− B cells (24).

FIGURE 3.

A suggested mechanism for MMR protein function during CSR. In this model, the Msh2-Msh6 heterodimer recognizes U-G mismatches in switch regions and recruits the Mlh1-Pms2 heterodimer to the mismatch. This MMR complex scans the DNA for breaks generated by UNG and APEX1/APEX2 activity. Exo1 is recruited onto the DNA at the nick and excises DNA in the 5′ to 3′ direction to the mismatch. Exo1 is inactivated and displaced from the DNA when it interacts with the Mlh1-Pms2 heterodimer (34 ). Exo1 is repeatedly loaded onto the DNA and excises DNA until the mismatch is removed. Upon mismatch removal, Mlh1-Pms2 interacts with Exo1, inactivating it. This results in a uracil residue on a stretch of ssDNA that can be removed to create a double-strand break. How this ssDNA would be converted into a dsDNA break is currently unknown, but it could be due to a combination of UNG2, APEX1/2, and/or MRE11 activity, all of which possess the ability to act on ssDNA uracil residues (UNG) or AP sites on ssDNA (APEX1/2 and MRE11) (404142 ). This model is based on the proposed mechanism of MMR (31323334 ). Alternatively, MMR activity alone may generate a double-strand break if a nick on the opposite strand is encountered during Exo1 excision, as has been previously proposed (23 ).

FIGURE 3.

A suggested mechanism for MMR protein function during CSR. In this model, the Msh2-Msh6 heterodimer recognizes U-G mismatches in switch regions and recruits the Mlh1-Pms2 heterodimer to the mismatch. This MMR complex scans the DNA for breaks generated by UNG and APEX1/APEX2 activity. Exo1 is recruited onto the DNA at the nick and excises DNA in the 5′ to 3′ direction to the mismatch. Exo1 is inactivated and displaced from the DNA when it interacts with the Mlh1-Pms2 heterodimer (34 ). Exo1 is repeatedly loaded onto the DNA and excises DNA until the mismatch is removed. Upon mismatch removal, Mlh1-Pms2 interacts with Exo1, inactivating it. This results in a uracil residue on a stretch of ssDNA that can be removed to create a double-strand break. How this ssDNA would be converted into a dsDNA break is currently unknown, but it could be due to a combination of UNG2, APEX1/2, and/or MRE11 activity, all of which possess the ability to act on ssDNA uracil residues (UNG) or AP sites on ssDNA (APEX1/2 and MRE11) (404142 ). This model is based on the proposed mechanism of MMR (31323334 ). Alternatively, MMR activity alone may generate a double-strand break if a nick on the opposite strand is encountered during Exo1 excision, as has been previously proposed (23 ).

Close modal

We have also shown that Mlh1−/−, SμTR−/−:Mlh1−/−, Exo1−/−, and SμTR−/−:Exo1−/− mice all exhibit increased frequencies of switch junctions that display longer stretches of microhomology. Increased microhomology at switch junctions in Mlh1−/− and Pms2−/− B cells has been suggested to reflect increased use of an alternative end-joining (AEJ) pathway that is distinct from classical NHEJ and that uses microhomology (18, 22). It has also been shown that CSR occurs in patients and a mouse B cell line with impaired DNA ligase IV function, although this impairment decreases the efficiency of CSR and increases switch junction microhomology (36, 37). This further supports the idea that AEJ pathways can be used for CSR. Such an alternative CSR joining pathway has recently been demonstrated in conditional XRCC4 deficient mice, which lack the ability to join DNA ends by NHEJ but retain the ability to undergo CSR (38, 39). The increased microhomology at switch junctions found in our sequence analyses of Mlh1- and Exo1-deficient B cells suggest that this alternative CSR joining pathway is used at a higher frequency in these cells compared with WT.

The model we outline above proposes that MMR proteins convert widely spaced S region DNA nicks into blunt or near-blunt ends suitable for NHEJ. In the absence of MMR activity, the decreased frequency of DNA breaks that can be joined by NHEJ may lead to increases in CSR through the alternative joining pathway. However, although use of this alternative joining pathway may compensate for part of the defect in CSR during MMR deficiency, it cannot compensate for the CSR defect in mice deficient in both individual MMR proteins and the SμTR sequences. The severe reduction in CSR in the double-deficient mice suggests that removal of the SμTR region, which reduces the number of AID targets, results in an insufficient level of DNA breaks necessary for either CSR joining pathway.

Our studies analyzing CSR in mice deficient in both individual MMR proteins and the SμTR region has revealed an important and collective role of MMR proteins during CSR. We find that Mlh1, Exo1 and Msh2 MMR proteins all have similar effects on CSR events that occur in the SμTR flanking sequences. The double knockout mice, SμTR−/−:Mlh1−/−, SμTR−/−:Exo1−/− and SμTR−/−:Msh2−/−, all have similar CSR junction patterns (see Ref. 25 and Figs. 1 and 2), which suggests that these deficiencies have similar effects on the joining pathways used to complete the few CSR events that occur. Previous reports of switch junction microhomology in MMR protein-deficient mice having intact Sμ regions had raised questions regarding the relationship between MMR proteins and NHEJ and AEJ during CSR. It is difficult to formulate a model that explains why a deficiency of either Mlh1 or Pms2 results in increased switch junction microhomology whereas Exo1, Msh2, or Msh6 deficiencies result in unchanged or decreased switch junction microhomology (17, 18, 19, 21, 22). Our results show that Exo1 deficiency results in an increased switch junction microhomology similar to all other individual MMR deficiencies except those of Msh2 and Msh6. Based on our results and results reported for various other MMR-deficient mice (18, 19, 20, 21) models for possible relationships between MMR proteins and NHEJ- and AEJ-mediated CSR that account for experimental observations are more easily formulated.

An important consideration in any CSR mechanism model is the fact that, in the presence of the SμTR region, switch junctions from Msh2- and Msh6-deficient B cells have similar or decreased microhomologies compared with those of WT (21, 22), which is distinct from the results we and others obtain with Mlh1, Pms2, or Exo1 protein deficiencies (18, 22). Msh2 (or Msh6) deficiency would clearly prevent MMR recognition of CSR DNA damage and, therefore, no MMR processing of mismatches or nicks could occur. This would lead to CSR joining only for blunt or near-blunt CSR-induced DNA breaks by NHEJ.

Because the Msh2-Msh6 heterodimer is present in Mlh1-, Pms2-, and Exo1-deficient B cells, the increases in switch junction microhomology in these mice could suggest that, within the SμTR region, Msh2-Msh6 might be involved in both the processing of widely spaced nicks (in concert with other MMR proteins) and in additional activities that are independent of at least some other MMR proteins. These activities might even result from the absence of MMR proteins that function downstream of mismatch recognition. One possibility is that if Msh2-Msh6 recognizes U-G mismatches but Mlh1, Pms2, or Exo1 are absent, alternative factors may be subsequently recruited to bound but abortive Msh2-Msh6 heterodimers that could result in increased AEJ. Another possibility, although not mutually exclusive, could be that in the absence of MMR processing factors such as Mlh1, Pms2 or Exo1 but in the presence of Msh2-Msh6, MMR might be initiated at mismatches in areas of widely staggered DNA nicks; but, again, the MMR process would be abortive or abnormal. This altered MMR activity might cause nicks/breaks (possibly DNA breaks with long single-stranded tails) to proceed into an alternative joining pathway, leading to CSR junctions with longer microhomologies.

Additional experiments are needed to distinguish the different explanations of CSR junction microhomology results and to better understand the role of MMR proteins during CSR. Production of mouse strains that lack both NHEJ and individual MMR proteins (Msh2, Mlh1, or Exo1) could indicate whether these MMR proteins function predominately to process CSR DNA nicks for NHEJ ligation or whether they might have additional roles in the CSR process.

We thank Drs. Winfried Edelmann and R. M. Liskay for graciously providing the Exo1- and Mlh1-deficient mice. We also thank Dr. Naomi Rosenberg for critically reading the manuscript.

The authors have no financial conflict of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This research was supported by National Institutes of Health Grants AI24465 (to E.S.), AI23283 (to J.S.), and AI65639 (to C.E.S.) and by the Eshe Foundation.

3

Abbreviations used in this paper: CSR, class switch recombination; AEJ, alternative end joining; AID, activation-induced cytidine deaminase; MMR, mismatch repair; NHEJ, nonhomologous end joining; S region, switch region; SμTR, Sμ tandem repeat; UNG, uracil DNA glycosylase; WT, wild type.

4

The online version of this article contains supplemental material.

1
Stavnezer, J., J. E. Guikema, C. E. Schrader.
2008
. Mechanism and regulation of class switch recombination.
Annu. Rev. Immunol.
26
:
261
-292.
2
Bransteitter, R., P. Pham, M. D. Scharff, M. F. Goodman.
2003
. Activation-induced cytidine deaminase deaminates deoxycytidine on single-stranded DNA but requires the action of RNase.
Proc. Natl. Acad. Sci. USA
100
:
4102
-4107.
3
Chaudhuri, J., M. Tian, C. Khuong, K. Chua, E. Pinaud, F. W. Alt.
2003
. Transcription-targeted DNA deamination by the AID antibody diversification enzyme.
Nature
422
:
726
-730.
4
Dickerson, S. K., E. Market, E. Besmer, F. N. Papavasiliou.
2003
. AID mediates hypermutation by deaminating single stranded DNA.
J. Exp. Med.
197
:
1291
-1296.
5
Muramatsu, M., K. Kinoshita, S. Fagarasan, S. Yamads, Y. Shinkai, T. Honjo.
2000
. Class switch recombination and hypermutation require activation-induced cytidine deaminase (AID), a potential RNA editing enzyme.
Cell
102
:
553
-563.
6
Petersen-Mahrt, S. K., R. S. Harris, M. S. Neuberger.
2002
. AID mutates E. coli suggesting a DNA deamination mechanism for antibody diversification.
Nature
418
:
99
-103.
7
Pham, P., R. Bransteitter, J. Petruska, M. F. Goodman.
2003
. Processive AID-catalysed cytosine deamination on single-stranded DNA simulates somatic hypermutation.
Nature
424
:
103
-107.
8
Revy, P., T. Muto, Y. Levy, F. Geissmann, A. Plebani, O. Sanal, N. Catalan, M. Forveille, R. Dufourcq-Labelouse, A. Gennery, et al
2000
. Activation-induced cytidine deaminase (AID) deficiency causes the autosomal recessive form of the hyper-IgM syndrome (HIGM2).
Cell
102
:
565
-575.
9
Brainsteitter, R., P. Pham, P. Calabrese, M. F. Goodman.
2004
. Biochemical analysis of hypermutational targeting by wild type and mutant activation-induced cytidine deaminase.
J. Biol. Chem.
279
:
51612
-51621.
10
Yu, K., F. T. Huang, M. R. Lieber.
2004
. DNA substrate length and surrounding sequences affect the activation-induced deaminase activity at cytidine.
J. Biol. Chem.
279
:
6496
-6500.
11
Imai, K., G. Slupphauq, W. I. Lee, P. Revy, S. Nonoyama, N. Catalan, L. Yel, M. Forveille, B. Kavli, H. E. Krokan, et al
2003
. Human uracil-DNA glycosylase deficiency associated with profoundly impaired immunoglobulin class-switch recombination.
Nat. Immunol.
4
:
1023
-1028.
12
Rada, C., G. T. Williams, H. Nilsen, D. E. Barnes, T. Lindahl, M. S. Neuberger.
2002
. Immunoglobulin isotype switching is inhibited and somatic hypermutation perturbed in UNG-deficient mice.
Curr. Biol.
12
:
1748
-1755.
13
Rada, C., J. M. Di Noia, M. S. Neuberger.
2004
. Mismatch recognition and uracil excision provide complementary paths to both Ig switching and the A/T-focused phase of somatic mutation.
Mol. Cell
16
:
163
-167.
14
Guikema, J. E. J., E. K. Linehan, D. Tsuchimoto, Y. Nakabeppu, P. R. Strauss, J. Stavnezer, C. E. Schrader.
2007
. APE1- and APE2-dependent DNA breaks in immunoglobulin class switch recombination.
J. Exp. Med.
204
:
3017
-3026.
15
Casellas, R., A. Nussenzweig, R. Wuerffel, R. Pelanda, A. Reichlin, H. Suh, X. F. Qin, E. Besmer, A. Kenter, K. Rajewsky, M. C. Nussenzweig.
1998
. Ku80 is required for immunoglobulin isotype switching.
EMBO
17
:
2404
-2411.
16
Manis, J. P., Y. Gu, R. Lansford, E. Sonoda, R. Ferrini, L. Davidson, K. Rajewsky, F. W. Alt.
1998
. Ku70 is required for late B cell development and immunoglobulin heavy chain class switching.
J. Exp. Med.
187
:
2081
-2089.
17
Bardwell, P. D., C. J. Woo, K. Wei, Z. Li, A. Martin, S. Z. Sack, T. Parris, W. Edelmann, M. D. Scharff.
2004
. Altered somatic hypermutation and reduced class-switch recombination in exonuclease 1-mutant mice.
Nat. Immunol.
5
:
224
-229.
18
Ehrenstein, M. R., C. Rada, A. M. Jones, C. Milstein, M. S. Neuberger.
2001
. Switch junction sequences in PMS2-deficient mice reveal a microhomology-mediated mechanism of Ig class switch recombination.
Proc. Natl. Acad. Sci. USA
98
:
14553
-14558.
19
Ehrenstein, M. R., M. S. Neuberger.
1999
. Deficiency in Msh2 affects the efficiency and local sequence specificity of immunoglobulin class-switch recombination: parallels with somatic hypermutation.
EMBO. J.
18
:
3484
-3490.
20
Schrader, C. E., W. Edelmann, R. Kucherlapati, J. Stavnezer.
1999
. Reduced isotype switching in splenic B cells from mice deficient in mismatch repair enzymes.
J. Exp. Med.
190
:
323
-330.
21
Martomo, S. A., W. W. Yang, P. J. Gearhart.
2004
. A role for Msh6 but not Msh3 in somatic hypermutation and class switch recombination.
J. Exp. Med.
200
:
61
-68.
22
Schrader, C. E., J. Vardo, J. Stavnezer.
2002
. Role for mismatch repair proteins Msh2. Mlh1, and Pms2 in immunoglobulin class switching shown by sequence analysis of recombination junctions.
J. Exp. Med.
195
:
367
-373.
23
Larson, E. D., M. L. Duquette, W. J. Cummings, R. J. Streiff, N. Maizels.
2005
. MutSa binds to and promotes synapsis of transcriptionally activated immunoglobulin switch regions.
Curr. Biol.
15
:
470
-474.
24
Schrader, C. E., J. E. Guikema, E. K. Linehan, E. Selsing, J. Stavnezer.
2007
. Activation-induced cytidine deaminase-dependent DNA breaks in class switch recombination occur during G1 phase of the cell cycle and depend upon mismatch repair.
J. Immunol.
179
:
6064
-6071.
25
Min, I. M., C. E. Schrader, J. Vardo, T. M. Luby, N. D'Avirro, J. Stavnezer, E. Selsing.
2003
. The Sμ tandem repeat region is critical for Ig isotype switching in the absence of Msh2.
Immunity
19
:
515
-524.
26
Luby, T. M., C. E. Schrader, J. Stavnezer, E. Selsing.
2001
. The μ switch region tandem repeats are important, but not required, for antibody class switch recombination.
J. Exp. Med.
193
:
159
-168.
27
Stavnezer, J., C. E. Schrader.
2006
. Mismatch repair converts AID-instigated nicks to double-strand breaks for antibody class-switch recombination.
Trends Genet.
22
:
23
-28.
28
Kunkel, E. A., D. A. Erie.
2005
. DNA mismatch repair.
Annu. Rev. Biochem.
74
:
681
-710.
29
Szankasi, P., G. R. Smith.
1992
. A DNA exonuclease induced during meiosis of Schizosaccharomyces pombe.
J. Biol. Chem.
267
:
3014
-3023.
30
Wilson, D. M., III, J. P. Carney, M. A. Coleman, A. W. Adamson, M. Christensen, J. E. Lamerdin.
1998
. Hex1: a new human Rad2 nuclease family member with homology to yeast exonuclease 1.
Nucleic. Acids. Res.
26
:
3762
-3768.
31
Kadyrov, F. A., L. Dzantiev, N. Constantin, P. Modrich.
2006
. Endonucleolytic function of MutLα in human mismatch repair.
Cell
126
:
297
-308.
32
Dzantiev, L., N. Constantin, J. Genschel, R. R. Iyer, P. M. Burgers, P. Modrich.
2004
. A defined human system that supports bidirectional mismatch-provoked excision.
Mol. Cell
15
:
31
-41.
33
Genschel, J., P. Modrich.
2003
. Mechanism of 5′-directed excision in human mismatch repair.
Mol. Cell
12
:
1077
-1086.
34
Zhang, Y., F. Yuan, S. R. Presnell, K. Tian, Y. Gao, A. E. Tomkinson, L. Gu, G. M. Li.
2005
. Reconstitution of 5′-directed human mismatch repair in a purified system.
Cell
122
:
693
-705.
35
Schanz, S., D. Castor, F. Fischer, J. Jiricny.
2009
. Interference of mismatch and base excision repair during the processing of adjacent U/G mispairs may play a key role in somatic hypermutation.
Proc. Natl. Acad. Sci. USA
106
:
5593
-5598.
36
Pan-Hammarstrom, Q., A. M. Jones, A. Lahdesmaki, W. Zhou, R. A. Gatti, L. Hammarstrom, A. R. Gennery, M. R. Ehrenstein.
2005
. Impact of DNA ligase IV on nonhomologous end joining pathways during class switch recombination in human cells.
J. Exp. Med.
201
:
189
-194.
37
Han, L., K. Yu.
2009
. Altered kinetics of nonhomologous end joining and class switch recombination in ligase IV-deficient B cells.
J. Exp. Med.
205
:
2745
-2753.
38
Yan, C. T., C. Boboila, E. K. Souza, S. Franco, T. R. Hickernell, M. Murphy, S. Gumaste, M. Geyer, A. A. Zarrin, J. P. Manis, et al
2007
. IgH class switching and translocations use a robust non-classical end-joining pathway.
Nature
449
:
478
-482.
39
Soulas-Sprauel, P., G. Le Guyader, P. Rivera-Munoz, V. Abramowski, C. Olivier-Martin, C. Goujet-Zalc, P. Charneau, J. de Villartay.
2007
. Role for DNA repair factor XRCC4 in immunoglobulin class switch recombination.
J. Exp. Med.
204
:
1717
-1727.
40
Kavli, B., S. Andersen, M. Otterlei, N. B. Liakakk, K. Imai, A. Fischer, A. Durandy, H. E. Krokan, G. Slupphaug.
2005
. B cells from hyper-IgM patients carrying UNG mutations lack ability to remove uracil from ssDNA and have elevated genomic uracil.
J. Exp. Med.
2001
:
2011
-2021.
41
Larson, E. D., W. J. Cummings, D. W. Bednarski, N. Maizels.
2005
. MRE11/RAD50 cleaves DNA in the AID/UNG-dependent pathway of immunoglobulin gene diversification.
Mol. Cell
20
:
367
-375.
42
Marenstein, D. R., D. M. Wilson, III, G. W. Teebor.
2004
. Human AP endonuclease (APE1) demonstrates endonucleolytic activity against AP sites in single-stranded DNA.
DNA Repair
3
:
527
-533.