PMS2 Endonuclease Activity Has Distinct Biological Functions and Is Essential for Genome Maintenance

PMS2 endonuclease activity has distinct biological functions and is essential for genome maintenance

Johanna M. M. van Oersa,1, Sergio Roaa,1, Uwe Werlinga, Yiyong Liub,c, Jochen Genschelb, Harry Hou, Jr.d, Rani S. Sellerse, Paul Modrichb,c, Matthew D. Scharffa,2, and Winfried Edelmanna,2

Departments of aCell Biology and ePathology, and dCancer Center, Albert Einstein College of Medicine, Bronx, NY 10461; and bDepartment of Biochemistry and cHoward Hughes Medical Institute, Duke University Medical Center, Durham, NC 27710

Contributed by Matthew D. Scharff, June 16, 2010 (sent for review May 14, 2010) − − The DNA mismatch repair protein PMS2was recently found to encode nation (CSR) of Ig genes, because Pms2 / mice (14, 15) and a novel endonuclease activity. To determine the biological functions patients with deleterious homozygous mutations in PMS2 (16) of this activity in mammals, we generated endonuclease-deficient have a significant reduction in isotype switching. Despite numer- E702K EK/EK Pms2 knock-in mice. Pms2 mice displayed increased genomic ous reports (16–19), however, the role of PMS2 in somatic hypermutation rates and a strong cancer predisposition. In addition, class mutation (SHM), if any, has not been resolved. switch recombination, but not somatic hypermutation, was impaired To determine the significance of the PMS2 endonuclease for in Pms2EK/EK B cells, indicating a specificroleinIgdiversity.Incontrast these processes, we generated a mouse line with an E702K mis- − − EK/EK to Pms2 / mice, Pms2 male mice were fertile, indicating that this sense mutation in the metal-binding motif of PMS2. This mutation activity is dispensable in spermatogenesis. Therefore, the PMS2 en- corresponds to the E705K mutation in human PMS2 that com- donuclease activity has distinct biological functions and is essential pletely inactivates the endonuclease activity, and also has been for genome maintenance and tumor suppression. found in a patient with Turcot syndrome (4). The analysis of Pms2E702K mice (termed Pms2EK) showed that the loss of this DNA mismatch repair | B cell lymphoma | class switch recombination | activity leads to a dramatic increase in genomic mutation fre- somatic hypermutation | spermatogenesis quencies and tumor incidence, and a significant decrease in CSR of EK/EK the Ig locus. Strikingly, Pms2 male mice were fertile, as op- GENETICS −/− MS2 is an essential component of DNA mismatch repair posed to the sterile Pms2 male mice, indicating that the endo- P(MMR) complexes, which play an important role in maintain- nuclease function of PMS2 is dispensable in spermatogenesis ing genetic stability. MMR functions primarily in the detection and and that an unknown function of the protein is required for this repair of mismatched bases that result from erroneous replication, biological process. and also plays important roles in DNA damage response, genetic Results and Discussion recombination, control of meiotic progression, and generation of Pms2EK/EK EK antibody diversity (1). During MMR, heterodimeric MutS homo- Generation of Mutant Mice. To generate Pms2 mutant log (MSH) complexes, consisting of either MSH2 and MSH6 mice, a gene-targeting vector was designed that introduces the (MutSα) or MSH2 and MSH3 (MutSβ), bind to mismatched bases. murine equivalent of the human E705K mutation (E702K) into On binding, a conformational change in the MutS complexes exon 12 of the mouse gene (Fig. 1Aand Fig. S1). This mutation activates downstream events and leads to the recruitment of MutL disrupts the DQHA(X)2E(X)4E metal-binding motif, which is homolog (MLH) complexes consisting of either MLH1 and PMS2 essential for the endonuclease function of PMS2 (4). RT-PCR and α γ α qPCR confirmed that the mutant allele was expressed at WT RNA (MutL ) or MLH1 and MLH3 (MutL ). MutL interacts with the fi MutS complexes via MLH1, and this interaction is essential for levels in both mouse embryonic broblasts (MEFs) and primary B cells of Pms2EK/EK mice (Fig. 1 B and C). The presence of the coordination of downstream repair events, including excision of PMS2E702K protein in Pms2EK/EK mice was confirmed in splenic the mismatch carrying strand and its resynthesis (2, 3). tissue by Western blot analysis (Fig. 1D); however, expression was Important insights into the role of MutLα in MMR came from reduced to ≈25% compared with that seen in WT mice. Similar biochemical studies of reconstituted human MMR (4, 5), which results were obtained in brain, testis, and MEF cell lines. The re- showed that PMS2 is a latent endonuclease that introduces addi- duced expression of PMS2E702K protein in Pms2EK/EK mice might tional nicks into the discontinuous DNA strand of nicked het- be due to increased degradation of the mutant protein and is in line eroduplex substrates. This activity is dependent on the integrity of − − with previous results showing that transfecting Pms2 / MEFs with a highly conserved metal-binding motif found in many MutL E705K hPMS2 leads to lower protein levels compared with cells homologs, including MLH3. MutLα appears to play an essential transfected with WT hPMS2 (8). Consistent with previous studies role in 3′-directed MMR, where the bias for incision on the distal ′ α (20), MLH1 protein levels were not affected by the reduced level side of the mismatch results in a 5 entry site for MutS -activated E702K fi ′ of PMS2 protein (Fig. 1D), indicating either that suf cient Exo1. For 5 -directed MMR, however, several studies in both the mutant PMS2E702K protein is present to form a stable complex reconstituted human MMR system and MutLα-deficient cells have EK/EK α fi with MLH1 (in Pms2 mice) or that MLH1 is stabilized by shown that MutL is not essential, and that a puri ed system formation of complexes with other proteins, such as MLH3 or consisting of MutSα, Exo1, and RPA is sufficient to support 5′- directed MMR (6, 7). Studies in yeast have further delineated roles of the PMS2 endonuclease activity in DNA repair, recombination, Author contributions: J.M.M.v.O., S.R., P.M., M.D.S., and W.E. designed research; J.M.M.v.O., and the DNA damage response (5, 8, 9). S.R., U.W., Y.L., J.G., and H.H. performed research; J.M.M.v.O., S.R., R.S.S., P.M., M.D.S., and −/− Pms2 mice (10) showed an increase in base substitution and W.E. analyzed data; and J.M.M.v.O., S.R., M.D.S., and W.E. wrote the paper. insertion/deletion mutation frequencies at both a reporter locus The authors declare no conflict of interest. and microsatellite sequences, with a higher proportion of frame- Freely available online through the PNAS open access option. shift mutations compared with other MMR knockout models, such 1J.M.M.v.O. and S.R. contributed equally to this work. −/− −/− −/− as Msh2 and Mlh1 mice (11, 12). As a result, Pms2 mice 2 −/− To whom correspondence may be addressed. E-mail: [email protected] developed lymphomas and sarcomas (13). Moreover, Pms2 or [email protected]. male mice are sterile, suggesting an important role for PMS2 in This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. spermatogenesis. PMS2 also plays a role in class switch recombi- 1073/pnas.1008589107/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1008589107 PNAS Early Edition | 1of6 Downloaded by guest on September 30, 2021 EK/EK Genomic Instability in Pms2 Mutant Mice. The loss of PMS2 leads to increased microsatellite instability (MSI) in the genome of − − Pms2 / mice (13). To study whether this increase in MSI was specifically caused by the loss of the endonuclease function, we analyzed MSI at two dinucleotide markers in mouse genomic DNA. At both markers, Pms2EK/EK mice displayed elevated levels of unstable alleles compared with WT mice: 12% versus 0.8% for marker D7Mit91 and 19% versus 5% for marker D17Mit123 (Fig. 2B). These MSI levels were similar to those observed in Pms2−/− mice (14% and 17%, respectively) and slightly lower than − − those reported in Mlh1 / mice (15% and 28%, respectively) (21). Pms2+/EK mice showed no increase in MSI. To investigate the effect of the loss of PMS2 endonuclease function on mutagenesis in more detail, we determined the mutation rate at the cII reporter locus in 10-wk-old Pms2EK/EK, − − Pms2+/EK, Pms2 / , and WT littermates. DNA was isolated from spleen, liver, and small intestine. In all three organs, mutation frequencies were significantly increased in Pms2EK/EK mice compared with WT and Pms2+/EK mice (Fig. 2C). The mutation fre- − − quencies tended to be slightly higher in Pms2 / mice than in Pms2EK/EK mice, but the differences were not significant. As re- − − ported previously for Pms2 / mice (22), there was a greater increase in the number of insertions and deletions in Pms2EK/EK mice − − compared with WT mice (Fig. 2D). Pms2EK/EK and Pms2 / mice Fig. 1. Generation of Pms2EK/EK mice. (A) PMS2 domains and location of showed a significant increase in mutations at A:T in spleen and E702K. (B) RT-PCR analysis of total RNA from MEFs. The PCR product was small intestine (P < 0.001) (Fig. 2E). A similar increase in A:T − − digested with SacII to reveal specific patterns for all three Pms2 genotypes. mutations was described previously in T cells of Pms2 / mice (23). (C) q-PCR analysis of total RNA from MEFs and B cells, using primers located EK/EK −/− Mutations at A:T could have been generated by the error-prone in exon 2. White bars, WT; gray bars, Pms2 ; black bars, Pms2 .(D) DNA polymerase η (24). Interestingly, an interaction between Western blot analysis of protein extracts from splenic tissue. Relative PMS2 MutLα and Pol η has been reported recently (25), and it has been protein levels (normalized to β tubulin expression and compared with WT) were 21% in Pms2EK/EK mice and 65% in Pms2+/− mice. suggested that this interaction could facilitate error-free DNA replication by repairing the mismatches introduced by Pol η. The − − increase in mutations at A:T seen in Pms2EK/EK and Pms2 / mice is consistent with this notion. PMS1. Similarly, MLH3 protein levels were not affected in any of EK/EK Reduced Survival and Cancer Phenotype in Pms2 Mutant Mice. the Pms2 genotypes (Fig. 1D). − − Similar to Pms2 / mice (n =28),Pms2EK/EK mice (n = 39) had Defective MMR in Pms2EK/EK Cells. To determine the effect of the significantly lower survival compared with age-matched Pms2+/EK PMS2E702K mutation on MMR, we tested extracts from Pms2EK/EK, mice (n = 27) and WT mice (n = 28) (P < 0.001) (Fig. 2F). This − − Pms2+/EK, Pms2 / , and WT MEF cells for their ability to promote difference in survival was due primarily to increased predisposition repair of heteroduplexes containing a G-T mismatch and a single- to lymphomas (90%; 26/29 mice), mostly B cell lymphomas (83%; stranded nick located either 5′ or 3′ to the mismatched base. No 20/24 mice) (Table S2), although other tumors also occurred, in- − − MMR activity was detected in Pms2EK/EK or Pms2 / extracts, cluding sarcoma (5/29 mice), hepatic adenoma (2/29 mice), hem- angioma (1/29 mice), intestinal adenocarcinoma (1/29 mice), and whereas robust repair was seen in WT cells (Fig. 2A). These results − − uterine adenocarcinoma (1/29 mice). Compared with Mlh1 / demonstrate that the endonuclease function of PMS2 is essential mice, which have a median survival time of 7 mo (13, 26), the re- − − for MMR in cell extracts. In contrast to previous biochemical duced survival of Pms2EK/EK and Pms2 / mice is more moderate α ′ studies showing that MutL is not essential for 5 -directed MMR in (12 and 13 mo, respectively). This suggests the existence of a partial fi a system using puri ed protein (6, 7), our results indicate that the MMR defect in these mice in vivo, possibly due to either functional endonuclease activity is essential in both 5′-and3′-directed MMR redundancy with other MutL complexes, such as MLH1-MLH3, or in whole cell extracts. an abundance of preexisting DNA nicks that occur during normal It has been reported that low levels of human WT PMS2, but not DNA metabolic processes, such as replication. − − high levels of the E705K mutant protein, can restore MMR in The B cell lymphoma phenotype in Pms2EK/EK and Pms2 / − − − − Pms2 / mouse cells (9). To further confirm that the repair de- mice differs from that in Msh2 / (T cell lymphoma) (27) and − − ficiency in Pms2EK/EK extracts was caused by loss of endonuclease Mlh1 / (B and T cell lymphoma) (26) mice, and indicates that activity and not by the reduced level of mutant PMS2E702K protein, PMS2 endonuclease activity has an important tumor-suppressor we added decreasing amounts of WT human MutLα protein to function in the B cell lineage. The similarities in the survival and − − EK/EK −/− Pms2 / cell extracts. Significant MMR activity was still detected tumor spectra of Pms2 and Pms2 mice suggest that en- when WT human MutLα levels were reduced to levels comparable donuclease activity is the major function of PMS2 required for to the MutLα levels observed in Pms2EK/EK cells (Table S1). The prevention of genome instability and tumorigenesis. human and mouse PMS2 proteins share a high homology in their Pms2EK/EK Mutant Mice Show a Decrease in ex Vivo CSR. To examine amino acid sequence, and the observation that the WT human whether inactivation of the endonuclease function of PMS2 affects fi −/− − − protein can ef ciently reconstitute MMR in Pms2 mouse cells CSR, we purified splenic B cells from Pms2EK/EK, Pms2 / , and further demonstrates that PMS2 functions are conserved. Thus, our WT mice and stimulated them ex vivo to induce switching from − − results strongly suggest that the repair-deficient phenotype of IgM to IgG3, IgG1, or IgG2b. Pms2 / mice showed a ∼40–70% Pms2EK/EK mice is specifically caused by mutation of the endo- reduction in switching efficiencies compared with WT mice (P < nuclease domain. 0.001) (Fig. 3A). This result is consistent with previous reports on

2of6 | www.pnas.org/cgi/doi/10.1073/pnas.1008589107 van Oers et al. Downloaded by guest on September 30, 2021 GENETICS

Fig. 2. MMR is impaired and leads to genomic instability in Pms2EK/EK mice. (A) MMR in whole cell extracts of MEFs. Repair was quantiﬁed from three independent experiments. (B) MSI analysis. The percent instability was determined by the number of unstable alleles divided by the total number of alleles scored. (C) cII reporter gene mutation frequencies in spleen, liver, and small intestine of 10-wk-old WT, Pms2+/EK, Pms2EK/EK, and Pms2−/− mice. Data represent mean ± SD. (D) cII mutation spectra in splenic tissue. A total of 77 mutations were sequenced for WT mice, 33 for Pms2+/EK, 304 for Pms2EK/EK, and 267 for − − Pms2 / mice. Absolute mutation frequencies are shown. Similar data were obtained for liver and small intestine. (E) cII base substitution spectra. (F) The − − Pms2EK/EK tumor-free survival curve was signiﬁcantly different from both Pms2+/EK and WT curves (P < 0.001). Survival was similar in Pms2 / mice and mutant mice (P = 0.13).

− − truncated forms of PMS2 (16) and on Pms2 / mice (14, 15), which has been identified in truncated forms of human PMS2 (16). In showed significant deficiencies in CSR to all studied isotypes. contrast, although we detected a tendency toward a preferential Compared with WT mice, Pms2EK/EK mice also exhibited a ∼40– use of blunt junctions in Pms2EK/EK mice compared with WT mice 60% reduction in switching (P < 0.001), but this decrease was not (Table 1), we found no significant differences in the relative fre- − − significantly different from that observed in Pms2 / cells (P ≥ quencies of the different types of junctions (i.e., blunt junctions, 0.12). The defect in CSR was due to Pms2EK/EK samples accu- microhomologies, or insertions) or in the average length of mulating fewer cells that had switched at each cell division, not to microhomology (2.54 ± 0.53 vs. 2.6 ± 0.44; P = 0.47) between the a change in the doubling time of the stimulated B cells (Fig. 3B)or Pms2EK/EK and the WT mice (Table 1). These findings indicate a decrease in the frequency of B220+PNAhi germinal center B cells that the presence of the PMS2 protein, but not its endonuclease in the spleen (Fig. 3C). These findings suggest that PMS2 endo- activity, affects the microhomology-mediated end-joining mech- nuclease activity, as opposed to other protein functions, does not anisms of CSR, suggesting multiple roles for PMS2 during CSR. merely delay the switching process, but rather is the major con- −/− EK/EK tributor to PMS2 function during CSR. Somatic Hypermutation Is Not Impaired in Pms2 or Pms2 Mice. Although the present study and previous work have shown that − − EK/EK Pms2 / and Pms2 Mice Show Differences in Microhomology- PMS2 plays a role in postreplicative MMR and in the repair of Mediated Mechanisms at Sμ–Sγ3 Junctions. It has been suggested double-stranded DNA breaks in Ig switching regions, deficiency of that in the absence of the MLH1–PMS2 complex, an alternative the protein has a modest or no effect during SHM of the Ig locus end-joining pathway based on long microhomologies at the junc- (16–19). Considering the possibility that other factors (e.g., MLH3 tion might provide stability to the synaptic structure during CSR or apurinic/apyrimidinic endonuclease) might efficiently com- − − (14, 28). Indeed, our Pms2 / mice had Sμ–Sγ3 junctional DNA pensate for the absence of PMS2 and fulfill the endonucleolytic segments with extended microhomology (≥5 nts) and a longer requirements of V-region hypermutation, we investigated SHM in − − average length compared with WT mice (5.07 ± 1.4 vs. 2.6 ± 0.44; our mutant mice. Pms2EK/EK, Pms2 / and WT mice were immu- P = 0.04) (Table 1). Abnormal formation of switch junctions also nized with NP30-CGG, and splenic B cells were analyzed for the

van Oers et al. PNAS Early Edition | 3of6 Downloaded by guest on September 30, 2021 error-prone polymerase Pol η (29, 30), in which mutations at A sites and WA motifs exceed mutations at T sites and TW motifs. Furthermore, the overall mutation frequencies of the hotspots for AID (WRC/GYW) and Pol η (WA/TW), which are major hall- marks of the SHM process, were similar in WT, null, and mutant mice (P ≥ 0.17). The lack of effect of the mutant PMS2E702K protein on the frequency and characteristics of A:T mutations in the variable region gene, even in the presence of increased A:T mutations in the cII reporter gene, conﬁrms that PMS2 does not play a signiﬁcant role in SHM.

− − PMS2 Endonuclease Is Not Required for Spermatogenesis. Pms2 / male mice are sterile (10), indicating that PMS2 plays an important role in meiotic processes. Strikingly, in the present study, Pms2EK/EK male and female mice were fertile, indicating normal meiotic progression. Pms2EK/EK testes contained normal seminiferous epithelium with mature spermatozoa in the lumen, indicating completion of spermatogenesis (Fig. 4A). In contrast, the − − seminiferous tubules of Pms2 / mice were severely depleted in spermatogenic cells, and mature sperm cells were abnormal, with misshaped heads and truncated or coiled flagella (Fig. 4B, Bottom − − Right). Testis size in Pms2 / mice was about 70% of that in WT littermates, whereas testis size in Pms2EK/EK mice was similar to that in WT littermates (Fig. 4B, Left and Top Right). Progression through meiotic prophase I appeared to be normal in all three genotypes, however (Fig. 4C). These results indicate that the endonuclease function of PMS2 does not play an essential role in Fig. 3. Reduced class switching in Pms2EK/EK mice. (A) Pms2EK/EK (n = 5) and − − Pms2 / (n = 3) mice assayed in three independent experiments showed meiosis, and that the presence of the protein, whether in a scaf- a similar decrease in the relative switching to IgG3, IgG1, and IgG2b com- folding function or another unknown function, is essential for pared with their WT littermates (n = 5). The control switching was defined as spermatogenesis. PMS2 functions in the response to DNA damage 100%, and the data shown represent mean ± SD. (B) Similar proliferation of in mitotic cells (31), and it is possible that the protein also con- splenic B cells from WT, Pms2EK/EK, and Pms2−/− mice was evidenced by 5,6- tributes to checkpoint-signaling pathways during premeiotic S- carboxyfluorescein diacetate succinimidyl ester (CFSE) dilution histograms of phase and spermatogenesis. cells stimulated with IL-4 + LPS over 3 d. (C) FACS analysis of isolated splenic B cells showing similar proportions of germinal center B220+PNAhigh cells in − − Conclusion WT, Pms2EK/EK, and Pms2 / mice. Our analysis of Pms2EK mutant mice reveals essential roles of the PMS2 endonuclease activity in mutation avoidance and tumor suppression in mammalian tissues. It also demonstrates that the pattern of somatic mutations in the rearranged V 186.2 gene. H PMS2EK mutation acts as a separation-of-function mutation and Based on unique mutations and correcting for base composition of exerts distinct effects on different biological processes, including the VH186.2 sequence, we found no significant differences in the − − B cell antibody diversification and spermatogenesis. Although overall mutation frequency among WT, Pms2 / , and Pms2EK/EK ≥ both CSR and SHM are initiated by the same AID-dependent dU: mice (P 0.41) (Table 2). The frequencies of the different types of G mismatch, PMS2 endonucleolytic or scaffolding functions are ≥ substitutions (P 0.11) and of transversions or transitions were not essential for SHM, but significantly contribute to CSR. In fact, ≥ −/− EK/EK similar (P 0.35) in the Pms2 and Pms2 mice and not although PMS2 endonuclease activity is essential for MutLα- fi signi cantly different from those in their WT littermates. This dependent processing of CSR intermediates, the PMS2 protein indicates that neither the presence of the protein nor the endo- clearly provides additional functions in the process. Although it is nuclease activity of PMS2 is essential for the accumulation of possible that the differences between the WT and Pms2EK/EK mutations during SHM. Consistent with this, we found no differ- mutant mice are due in part to the decreased abundance of PMS2, ence among genotypes in the frequency of dinucleotide sub- the presence of longer microhomologies at CSR junctions in − − stitutions (P ≥ 0.84) or a decreased ratio of mutations in A as Pms2 / mice, but not in Pms2EK/EK mice, suggests that there is − − opposed to T (P ≥ 0.45). In fact, Pms2 / and Pms2EK/EK mice sufficient PMS2 protein in Pms2EK/EK mice to provide scaffolding retained the strand bias signature previously attributed to the and to preclude the need for long microhomologies during CSR.

Table 1. Sμ–Sγ3 junctions in stimulated Pms2+/+, Pms2−/−, and Pms2EK/EK splenic B cells Insertions Blunt Microhomology

% Number of Mouse sequences % P*%P* ≥1bp ≥2bp ≥5bp ≥8bp ≥10 bp P* Mean length ± SEM P†

Pms2+/+ 22 18 14 68 45 9 0 0 2.6 ± 0.44 Pms2−/− 20 20 0.89 15 0.9 65 55 25 15 15 0.83 5.07 ± 1.4 0.04 Pms2EK/EK 20 15 0.79 30 0.2 55 40 5 0 0 0.38 2.54 ± 0.53 0.47

*Pearson´s χ2 test. †Unpaired Student´s t test.

4of6 | www.pnas.org/cgi/doi/10.1073/pnas.1008589107 van Oers et al. Downloaded by guest on September 30, 2021 Table 2. Unique mutation frequency analysis of the VH186.2 region

VH186.2 region (273 bp)

Pms2+/+ (n =4) Pms2EK/EK (n =5) Pms2−/− (n =3)

Sequences analyzed 71 90 60 Mutated sequences 58 (81.7%) 77 (85.6%) 56 (93.4%) Unique mutations (total)* 152 (346) 176 (470) 117 (244) † − − − Overall mutation frequency 4.6 × 10 2 4.3 × 10 2 4.8 × 10 2 Dinucleotide substitutions* 15 18 10 Dinucleotide mutation frequency‡ 1.5 × 10−3 1.5 × 10−3 1.4 × 10−3 − − − G/C mutation frequency 5.7 × 10 2 5.4 × 10 2 5.8 × 10 2 − − − A/T mutation frequency 3.4 × 10 2 3.1 × 10 2 3.6 × 10 2 A>T strand bias ratio 2.5 2.4 3.3 Transversions frequency 2.7 × 10−2 2.6 × 10−2 2.7 × 10−2 − − − Transitions frequency 8.4 × 10 2 7.7 × 10 2 8.9 × 10 2 − − − AID hotspot§ WRC/ 6.4 × 10 2 8.2 × 10 2 9.6 × 10 2 /GYW 19.2 × 10−2 16.4 × 10−2 17.2 × 10−2 Pol η hotspot§ WA/ 9.4 × 10−2 7.8 × 10−2 11.6 × 10−2 − − − /TW 2.4 × 10 2 2.2 × 10 2 1.6 × 10 2

*Observed number of unique mutations: within each mouse, identical mutations were counted once. †Based on the theoretical maximum number of unique mutations: number of sites × number of mice in the category × 3 (because three possible substitutions can occur at each site). ‡ Based on the theoretical maximum number of unique mutated dinucleotides: number of dinucleotide sites × number of mice in the category × 9 (because nine possible combinations of dinucleotide substitutions can occur). §The underlined site of the motif is scored. W = A/T, R = A/G, Y = C/T, S = G/C. GENETICS The importance of PMS2 as a scaffolding protein, even at reduced currently unknown, and their identification will provide important cellular levels, is further highlighted by the finding that Pms2EK/EK insight into the mechanisms that control meiotic progression and/ male mice have no obvious defects in spermatogenesis. The or sperm differentiation. identity of the meiotic partners of PMS2 other than MLH1 is Materials and Methods Generation and Analysis of Pms2E702K Mice. Pms2E702K mice were created by a knock-in gene-targeting strategy (Fig. S1). All protocols involving animals were approved by the Albert Einstein College of Medicine’s Animal Care and Use Committee in accordance with the U.S. Public Health Service’s Animal Welfare Policy. RT-PCR using primers located in exon 12 (5′-CGATGTTTG- CAGAGATGGAG-3′) and exon 14 (5′-TAGCCCTTTCAGTGACTGGAG-3′)of Pms2 was done on total RNA isolated from MEF cell lines. Relative quanti- fication of Pms2 transcripts was performed on cDNA from resting splenic B cells and MEFs by real-time PCR using primers 5′-AACCGAAGGCGTGAG- TACAG-3′ and 5′-CACAGCGGTGCTTAAACTGA-3′. For Western blot analysis, equal amounts of total cellular protein from MEF and tissue extracts were separated on 3–8% NuPAGE gels (Invitrogen). Proteins were detected with anti-PMS2 (clone A16-4; BD PharMingen), anti-MLH1 (Bethyl Labora- tories), or anti-MLH3 (a gift from Paula E. Cohen, Cornell University, Ithaca, NY). Bands were quantified using GeneTools software (Syngene).

Mismatch Repair Assay. MEFs were cultured in DMEM high-glucose medium supplemented with 10% FBS (HyClone). Whole cell extracts were prepared as described previously (32), as were 5′ G-T heteroduplex and 3′ G-T heteroduplex DNAs (33, 34). MMR reactions using whole cell extracts were carried out following the method of Kadyrov et al. (32). Recombinant human MutLα − − used to complement the whole cell extracts of Pms2 / and Pms2EK/EK MEFs was isolated as described previously (35).

Microsatellite Instability Analysis. Tail DNA from six WT, six Pms2+/EK, six Pms2EK/EK, and six Pms2−/− mice was pooled and diluted to approximately Fig. 4. Testis morphology in male mice. (A) H&E staining of testis sections from − − one molecule per reaction. The microsatellite loci were amplified and ana- 10-wk-old WT, Pms2EK/EK,andPms2 / mice. Note the regions of degeneration in − − lyzed as described previously (36). the testis of Pms2 / males. In WT and Pms2EK/EK mice, the center portion of the tubule contains fully elongated spermatids with flagella extending into the − − cII lumen (asterisks), whereas these features are absent in Pms2 / mice. (Scale bar: Mutation Rate Analysis. Mutations in genomic DNA from the spleen, liver, EK/EK 50 μm.) (B) Comparison of testis size and analysis of spermatozoa in male mice of and small intestine of Pms2 mice crossed to Big Blue transgenic mice (37) − − λ different Pms2 genotypes. The 10-wk-old Pms2 / mice, but not the Pms2EK/EK were detected using the Select-cII Mutation Detection System for Big Blue mice, had significantly lower testis weight compared with WT controls. Sperm Rodents (Stratagene). Mutation frequency was defined as the ratio of mutant cells isolated from the cauda epididymis from Pms2−/− animals showed abnor- plaques to the total number of plaques screened. The mutation spectrum for mal morphology, whereas sperm cells in Pms2EK/EK and WT mice had a normal base substitutions was analyzed using the SHMTool Web server (38). appearance. (C) Examples of pachytene chromosome configurations after − − SYCP3 staining in WT, Pms2EK/EK,andPms2 / mice, indicating normal pro- Analysis of Tumors and Survival. Mice were observed until they became gression through prophase I. morbid or moribund. Tumors were removed and fixed in 10% buffered

van Oers et al. PNAS Early Edition | 5of6 Downloaded by guest on September 30, 2021 formalin. After being embedded in paraffin, sections were stained with H&E, Analysis of Meiotic Prophase I. Chromosome spreads were prepared as de- anti-B220 (BD PharMingen), or anti-CD3 (Lab Vision). Statistical analysis of scribed previously (40) with modifications. After cells were dispersed on tumor incidence was done using SPSS version 16.0 (SPSS). The Kaplan-Meier slides for 1 h, the slides were washed three times for 3 min each in 0.4% method was used to compare curves for survival, with significance evaluated Kodak Photo-Flo 200/water and air-dried. Then the slides were washed in by two-sided log rank statistics. 0.4% Kodak Photo-Flo 200/PBS and 0.1% Triton X-100/PBS for 3 min each, blocked for 3 min in blocking solution (3% BSA, 0.05% Triton X-100, 10% Ex Vivo CSR and Switch Junction Analysis. Splenic B cells were stimulated with goat serum in PBS), and incubated in an antibody against synaptonemal 50 μg/mL of LPS (Sigma-Aldrich), to induce switching to IgG3; LPS plus 50 ng/mL complex protein 3 (SYCP3, a gift from Paula E. Cohen, Cornell University, of rIL-4 (R&D Systems), to induce switching to IgG1; or LPS plus 30 μg/mL of Ithaca, NY) for 1 h at 37 °C in a humid chamber. The slides were then dextran sulfate (DS) (Sigma-Aldrich) to induce switching to IgG2b. After 4 d in washed, blocked, and incubated in a secondary antibody conjugated to fluorescein (Jackson Immunochemicals) for 30 min at 37 °C. Images were culture, surface IgM and IgG were stained and analyzed by FACS. PCR amplifi- captured with an Olympus BX61 upright microscope with a CoolSNAP HQ cation from genomic DNA, sequencing, and analysis of unique junctional Sμ–Sγ3 camera (Photometrics) and IPLab acquisition software (BD). regions were performed as described previously (39).

ACKNOWLEDGMENTS. We thank B. Jin for providing excellent technical Somatic Hypermutation Analysis. Eight-week-old mice were immunized i.p. assistance, Y. Zhang and R. Chahwan for providing additional technical with (4-hydroxy-3-nitrophenyl)acetyl (NP)30-CGG (BioSearch Technologies) in support, and P. Cohen (Cornell University, Ithaca, NY) for generously alum (Pierce) and boosted 4 wk later. TRIzol extraction (Invitrogen) of RNA was providing the MLH3 and SYCP3 antibodies. This work was supported by prepared 7 d after the boost from splenic B cells. Then high-ﬁdelity RT-PCR and a Rubicon Fellowship from the Netherlands Organization for Scientiﬁc Re-

nested PCR were performed to amplify VH186.2 joined to the IgG1 constant search (to J.M.M.v.O.), Postdoctoral Fellowship EX-2006-0732 from the Span- region, as described previously (39). Analysis of unique mutations was done ish Ministry of Education and Science (to S.R.), and National Institutes of − Health Grants GM45190 (to P.M.), CA72649 and CA102705 (to M.D.S.), and with SHMTool (38). The experimental error rate was estimated 2.2 × 10 4 CA76329 and CA93484 (to W.E.). M.D.S. is supported by the Harry Eagle γ ﬁ mutations/base by analysis of the C 1 segment. A signi cant bias of mutations Chair provided by the National Women’s Division of the Albert Einstein from A exceeding mutations from T in the WT data was used for quality as- College of Medicine. P.M. is an Investigator at the Howard Hughes sessment of contaminating PCR hybrids, as suggested recently (30). Medical Institute.

1. Iyer RR, Pluciennik A, Burdett V, Modrich PL (2006) DNA mismatch repair: Functions 22. Hegan DC, et al. (2006) Differing patterns of genetic instability in mice deficient in the and mechanisms. Chem Rev 106:302–323. mismatch repair genes Pms2, Mlh1, Msh2, Msh3 and Msh6. Carcinogenesis 27:2402– 2. Kunkel TA, Erie DA (2005) DNA mismatch repair. Annu Rev Biochem 74:681–710. 2408. 3. Marsischky GT, Kolodner RD (1999) Biochemical characterization of the interaction 23. Shao C, et al. (2002) Loss of heterozygosity and point mutation at Aprt locus in T cells − − between the Saccharomyces cerevisiae MSH2-MSH6 complex and mispaired bases in and fibroblasts of Pms2 / mice. Oncogene 21:2840–2845. DNA. J Biol Chem 274:26668–26682. 24. Kunkel TA, Pavlov YI, Bebenek K (2003) Functions of human DNA polymerases eta, 4. Kadyrov FA, Dzantiev L, Constantin N, Modrich P (2006) Endonucleolytic function of kappa and iota suggested by their properties, including fidelity with undamaged MutLalpha in human mismatch repair. Cell 126:297–308. DNA templates. DNA Repair (Amst) 2:135–149. 5. Kadyrov FA, et al. (2007) Saccharomyces cerevisiae MutLa is a mismatch repair 25. Kanao R, Hanaoka F, Masutani C (2009) A novel interaction between human DNA – endonuclease. J Biol Chem 282:37181 37190. polymerase eta and MutLalpha. Biochem Biophys Res Commun 389:40–45. ′ 6. Genschel J, Modrich P (2003) Mechanism of 5 -directed excision in human mismatch 26. Edelmann W, et al. (1999) Tumorigenesis in Mlh1 and Mlh1/Apc1638N mutant mice. – repair. Mol Cell 12:1077 1086. Cancer Res 59:1301–1307. 7. Genschel J, Modrich P (2009) Functions of MutLalpha, replication protein A (RPA), and 27. Lowsky R, et al. (1997) Defects of the mismatch repair gene MSH2 are implicated in ′ – HMGB1 in 5 -directed mismatch repair. J Biol Chem 284:21536 21544. the development of murine and human lymphoblastic lymphomas and are associated 8. Deschênes SM, et al. (2007) The E705K mutation in hPMS2 exerts recessive, not with the aberrant expression of rhombotin-2 (Lmo-2) and Tal-1 (SCL). Blood 89: – dominant, effects on mismatch repair. Cancer Lett 249:148 156. 2276–2282. 9. Erdeniz N, Nguyen M, Deschênes SM, Liskay RM (2007) Mutations affecting a putative 28. Schrader CE, Vardo J, Stavnezer J (2003) Mlh1 can function in antibody class switch MutLalpha endonuclease motif impact multiple mismatch repair functions. DNA recombination independently of Msh2. J Exp Med 197:1377–1383. Repair (Amst) 6:1463–1470. 29. Mayorov VI, Rogozin IB, Adkison LR, Gearhart PJ (2005) DNA polymerase eta 10. Baker SM, et al. (1995) Male mice defective in the DNA mismatch repair gene PMS2 contributes to strand bias of mutations of A versus T in immunoglobulin genes. exhibit abnormal chromosome synapsis in meiosis. Cell 82:309–319. J Immunol 174:7781–7786. 11. Andrew SE, et al. (2000) Mutagenesis in PMS2- and MSH2-deficient mice indicates 30. Steele EJ (2009) Mechanism of somatic hypermutation: Critical analysis of strand differential protection from transversions and frameshifts. Carcinogenesis 21:1291– biased mutation signatures at A:T and G:C base pairs. Mol Immunol 46:305–320. 1295. 31. Gong JG, et al. (1999) The tyrosine kinase c-Abl regulates p73 in apoptotic response to 12. Yao X, et al. (1999) Different mutator phenotypes in Mlh1- versus Pms2-deficient cisplatin-induced DNA damage. Nature 399:806–809. mice. Proc Natl Acad Sci USA 96:6850–6855. 32. Kadyrov FA, et al. (2009) A possible mechanism for exonuclease 1 independent 13. Prolla TA, et al. (1998) Tumour susceptibility and spontaneous mutation in mice eukaryotic mismatch repair. Proc Natl Acad Sci USA 106:8495–8500. deficient in Mlh1, Pms1 and Pms2 DNA mismatch repair. Nat Genet 18:276–279. 33. Dzantiev L, et al. (2004) A defined human system that supports bidirectional 14. Ehrenstein MR, Rada C, Jones AM, Milstein C, Neuberger MS (2001) Switch junction – sequences in PMS2-deficient mice reveal a microhomology-mediated mechanism of Ig mismatch-provoked excision. Mol Cell 15:31 41. 34. Fang WH, Modrich P (1993) Human strand-specific mismatch repair occurs by class switch recombination. Proc Natl Acad Sci USA 98:14553–14558. 15. Schrader CE, Edelmann W, Kucherlapati R, Stavnezer J (1999) Reduced isotype a bidirectional mechanism similar to that of the bacterial reaction. J Biol Chem 268: – switching in splenic B cells from mice deficient in mismatch repair enzymes. J Exp Med 11838 11844. 190:323–330. 35. Blackwell LJ, Wang S, Modrich P (2001) DNA chain length dependence of formation 16. Péron S, et al. (2008) Human PMS2 deficiency is associated with impaired immuno- and dynamics of hMutSalpha.hMutLalpha.heteroduplex complexes. J Biol Chem 276: – globulin class switch recombination. J Exp Med 205:2465–2472. 33233 33240. → 17. Cascalho M, Wong J, Steinberg C, Wabl M (1998) Mismatch repair co-opted by hyper- 36. Wong E, et al. (2002) Mbd4 inactivation increases C T transition mutations and mutation. Science 279:1207–1210. promotes gastrointestinal tumor formation. Proc Natl Acad Sci USA 99:14937–14942. 18. Di Noia JM, Neuberger MS (2007) Molecular mechanisms of antibody somatic hyper- 37. Dycaico MJ, et al. (1994) The use of shuttle vectors for mutation analysis in transgenic mutation. Annu Rev Biochem 76:1–22. mice and rats. Mutat Res 307:461–478. 19. Saribasak H, Rajagopal D, Maul RW, Gearhart PJ (2009) Hijacked DNA repair proteins 38. MacCarthy T, Roa S, Scharff MD, Bergman A (2008) SHMTool: A webserver for com- and unchained DNA polymerases. Philos Trans R Soc Lond 364:605–611. parative analysis of somatic hypermutation datasets. DNA Repair 8:137–141. 20. Chen PC, et al. (2005) Contributions by MutL homologues Mlh3 and Pms2 to DNA 39. Roa S, et al. (2008) Ubiquitylated PCNA plays a role in somatic hypermutation and mismatch repair and tumor suppression in the mouse. Cancer Res 65:8662–8670. class-switch recombination and is required for meiotic progression. Proc Natl Acad Sci 21. Wei K, et al. (2003) Inactivation of Exonuclease 1 in mice results in DNA mismatch USA 105:16248–16253. repair defects, increased cancer susceptibility, and male and female sterility. Genes 40. Kolas NK, et al. (2005) Localization of MMR proteins on meiotic chromosomes in mice Dev 17:603–614. indicates distinct functions during prophase I. J Cell Biol 171:447–458.

6of6 | www.pnas.org/cgi/doi/10.1073/pnas.1008589107 van Oers et al. Downloaded by guest on September 30, 2021