The E3 ubiquitin ligase Mule acts through the ATM–p53 axis to maintain B lymphocyte homeostasis

We generated a conditional Mule allele (Mule^fl) in which exons 76 and 77 (encoding the HECT domain) were flanked by loxP sites (Fig. 1). Mule^fl/y male mice were created using Flpe-deleter mice (Rodríguez et al., 2000) and bred with DeleterCre females (Schwenk et al., 1995) to produce Mule^del/+ females. According to Mendelian principles, 50% of the male progeny from a cross between a WT male and a Mule^del/+ female should be Mule^del/y (Mule KO) mice. However, only Mule^+/y males were born alive, implying that Mule^del/y mice die before birth (unpublished data). Analysis of Mule^del/y embryos at embryonic day 9.5 (E9.5) and 10.5 revealed various developmental anomalies, including enlarged heart, reduced somites, and greatly decreased body size compared with controls (unpublished data). The mutant embryos were dead by E12.5, indicating that Mule is essential for mouse embryogenesis.

We chose to characterize the effects of Mule deletion in B lymphocytes in adult mice because these cells are known to accumulate p53 and succumb to apoptosis under conditions of genotoxic stress (Murray-Zmijewski et al., 2008). Mule^fl/y (male) and Mule^fl/fl (female) mice showed normal B lymphocyte development and subset distribution as determined by flow cytometry (unpublished data). To generate mutants in which Mule was selectively deleted in B cells (B cell–specific Mule KO [BMKO] mice), Mule^fl/fl(y) mice (male or female) were crossed to Mb1Cre mice in which Cre recombinase is controlled by the B cell–specific Mb1 promoter (Hobeika et al., 2006). Southern blotting of DNA from purified splenic B cells of BMKO mice confirmed efficient Cre-mediated disruption of Mule in B cells but not in thymocytes (Fig. 2 A). Immunoblotting demonstrated the absence of Mule protein in purified splenic BMKO B cells (Fig. 2 B, left). Flow cytometric analysis revealed that Mule was also deleted in a large fraction of BMKO pro–B cells and pre–B cells (Fig. 2 B, right).

The main stages of B cell development in the BM are: pro–B (B220^lowIgM⁻cKit⁺), pre–B (B220^lowIgM⁻CD25⁺), immature B (B220^lowIgM⁺), and recirculating mature B (B220^highIgM⁺). Although the pro–B cell compartment of BMKO BM was normal, the pre–B cell compartment in the mutant was significantly decreased compared with controls (Fig. 2 C). Indeed, the pre–B cell population in BMKO mice contained significantly more apoptotic (Annexin V⁺) cells than the control (Fig. 2 D). Recirculating B220^highIgM⁺ mature B cells were also reduced by threefold in the mutant BM compared with controls, and apoptotic cells were increased by 1.5-fold (unpublished data). Thus, B cell development is partially blocked in the absence of Mule, and pre–B and mature B cells in BMKO BM are severely reduced as a result of increased apoptosis. When splenic B cells were examined, the percentage of B220⁺IgM⁺ B cells was reduced from 52.3% in controls to 38.5% in the mutant (Fig. 2 E, top). BMKO spleen contained ∼50% fewer cells than control spleen (Fig. 2 F, left). In addition, the B1a cell population that is normally present in the peritoneal cavity of control mice was virtually absent in BMKO mice (Fig. 2, E, [bottom] and F [right]). Thus, Mule is required for B cell development in the BM, as well as the homeostasis of mature B cells in the spleen and B1a cells in the peritoneum.

To determine whether impaired proliferation contributed to the reduced B cell numbers in BMKO mice, we performed in vivo BrdU incorporation experiments and observed that the percentage of BrdU⁺ B cells among total splenic B cells of BMKO mice was only ∼50% of controls (Fig. 3 A, top left). To analyze BrdU incorporation in various B cell subsets, B220⁺IgM⁺ cells from control and BMKO splenocytes were gated and then sorted by flow cytometry into CD93⁺ transitional (TS), CD21⁺CD23^low marginal zone (MZ), and CD21⁺CD23⁺ follicular (FO) B cells. The majority of TS B cells were BrdU⁺ in both control and BMKO mice but the mutant showed significantly fewer of these cells (Fig. 3 A, bottom left). A more striking deficit was found in the mutant MZ compartment, where the percentage of BrdU⁺ cells was reduced to ∼50% of that in controls (Fig. 3 A, top right). Mule-deficient FO B cells exhibited the most severe defect in proliferation, with the number of BrdU⁺ cells reaching only ∼30% of the control value (Fig. 3 A, bottom right). [³H]-thymidine incorporation assays confirmed that BMKO B cells proliferated much less vigorously than control B cells in response to stimulation with anti-IgM or anti-IgM plus anti-CD40 (Fig. 3 B). Although the activation marker CD69 was up-regulated normally on stimulated BMKO B cells, levels of CD25 and CD86 up-regulation were decreased compared with controls (Fig. 3 C). Thus, Mule is required for splenic B cell proliferation and activation.

The impaired B cell development in BMKO mice and defective proliferation of BMKO B cells in response to anti-IgM treatment suggested that Mule might be involved in BCR-induced signal transduction, which is initiated by protein tyrosine kinases (Kurosaki and Hikida, 2009). However, BCR ligation in purified splenic BMKO B cells resulted in a normal tyrosine phosphorylation pattern (Fig. 3 D). Thus, Mule is unlikely to be involved in BCR-mediated proximal signaling. Consistent with this finding, Ca²⁺ flux in BMKO B cells treated with anti-IgM was also normal (Fig. 3 E). Further investigation is required to dissect how Mule is involved in BCR-induced proliferation and activation.

The involvement of Mule in B cell proliferation and activation prompted us to investigate serum Ig production and humoral immune responses in our BMKO mice. Apart from IgG2b, serum levels of IgM, IgG1, IgA, IgG2a, and IgG3 were reduced by 2.5-, 2-, 4.5-, 32-, and 37-fold, respectively, in nonimmunized BMKO mice compared with controls (Fig. 4 A). However, plasma cell numbers in the BM of mutants and controls were similar (unpublished data), indicating that Mule is involved in Ig production rather than plasma cell differentiation or maintenance.

To determine whether Mule-deficient B cells could mount normal humoral responses, we immunized control and BMKO mice with either the T-independent (Ti) antigen TNP-Ficoll or the T-dependent (Td) antigen alum-precipitated 4-hydroxy-3-nitrophenylacetyl–chicken γ-globulin (NP_(15)-CGG). When serum was analyzed at 2–3 wk after immunization, Ti-immunized BMKO mice failed to mount a TNP-specific IgM or IgG3 response, whereas Ti-immunized control mice showed normal IgM and IgG3 responses (Fig. 4 B). In contrast, although Td-immunized BMKO B cells produced significantly less NP-specific IgG1 compared with controls at 7 d after immunization, this antibody (Ab) titer was normal at 14 and 21 d after immunization (Fig. 4 C, left). Consistent with this finding, the frequency of GC B cells in BMKO spleens was comparable to controls at day 14 after Td immunization as determined by flow cytometry (Fig. 4 C, right).

To examine the role of Mule in humoral responses in a more physiological setting, we infected BMKO and control mice with vesicular stomatitis virus (VSV). WT mice infected with VSV mount strong humoral responses characterized by a Ti-neutralizing IgM response by days 2–4, followed by a Td IgG response by days 4–6 (Hangartner et al., 2006). In control mice, total VSV-neutralizing Abs (IgM + IgG) appeared by day 2 after infection, peaked on day 4, and remained high until day 12 (Fig. 4 D, left). In contrast, total anti-VSV Abs in BMKO mice were undetectable on day 2 and were only 1/8 of control titers on day 4 after infection. However, by day 6, levels of total anti-VSV Abs in the mutant had caught up to the control, indicating that Mule-deficient B cells can mount an IgM response but with delayed kinetics. When we specifically examined VSV-neutralizing IgG, control mice first exhibited these Abs on day 6 after infection, and their levels continually increased until day 12 after infection (Fig. 4 D, right). However, in BMKO mice, anti-VSV IgG levels were only 1/5 of those in the control on day 6 after infection and failed to achieve control levels even by day 12.

In vitro, B cells undergo class switch recombination (CSR) and differentiate into either IgG3⁺ plasmablasts in response to LPS or into IgG1⁺ plasmablasts in response to LPS plus IL-4 (Stavnezer et al., 2008). To determine whether Mule affects CSR, we cultured naive BMKO and control B cells with LPS, or LPS plus IL-4, and measured the percentages of B cells producing IgG3 or IgG1. In both cases, Mule-deficient B cells produced 50% less of the expected isotype than controls (Fig. 5 A). CSR requires that a minimum of two rounds of B cell division be completed to allow IgG production (Hodgkin et al., 1996). To assess whether the CSR defect in BMKO B cells was associated with impaired proliferation, we labeled BMKO and control B cells with CFSE, activated them with LPS plus IL-4, and followed their cell division by flow cytometry. Compared with control B cells, BMKO B cells underwent one cell division less (Fig. 5 B). Thus, Mule is required for proper B cell proliferation and CSR in vitro, as well as normal Ti and Td humoral responses in vivo.

We next examined the effect of Mule deletion on the function of p53, a known Mule substrate and a key proapoptotic molecule involved in maintaining homeostasis (Chen et al., 2005; Murray-Zmijewski et al., 2008). To aid in these analyses and examine additional cell types, we created Mule-deficient embryonic stem (ES) cell lines and E1A/Ras-transformed Mule-deficient MEFs via Cre-mediated deletion of Mule^fl/y (Peitz et al., 2002). Immunoblotting showed that, compared with controls, p53 protein was clearly elevated in both Mule-deficient ES cells and MEFs at steady state (Fig. 6 A). Although p53 protein was not detectable by immunoblotting in either BMKO or control B cells at steady state (Fig. 6 A), quantitative real-time PCR analysis showed that mRNA levels of its proapoptotic target genes Bax, Mdm-2, Noxa, p21, and Puma (Villunger et al., 2003; Green and Kroemer, 2009), including p53, were all significantly higher in resting Mule-deficient B cells than in control B cells (Fig. 6 B). These data are consistent with our earlier observation of increased apoptosis in B cells isolated from unstressed BMKO mice (Fig. 2 D). Thus, loss of Mule has profound effects on p53 transcriptional activity at steady state.

To assess whether the increased p53 transactivation in BMKO B cells accounted for their defects, we crossed our BMKO mice to floxed p53 (p53^fl/fl) mice (Jonkers et al., 2001) to generate compound mutants lacking both Mule and p53 specifically in B cells (B cell–specific double KO [BDKO] mice). Quantitative real-time PCR analysis showed that the abnormal elevations of Bax, Mdm-2, Noxa, and Puma mRNAs (but not p21 mRNA) in BMKO B cells were abrogated upon genetic ablation of p53 (Fig. 6 B). These results suggest that the increase in Bax, Mdm-2, Noxa, and Puma mRNAs in BMKO B cells is p53 dependent.

Flow cytometric analysis showed that both the block in BM B cell development at the pre–B cell stage and the defect in splenic B cell proliferation disappeared upon genetic ablation of p53 (Fig. 6 C). B cell proliferation in response to anti-IgM or anti-IgM plus anti-CD40 was also restored upon loss of p53 (Fig. 6 D, left), and total B cell numbers in BDKO spleens returned to normal levels (Fig. 6 D, right). Our studies therefore suggest that Mule controls p53 activity essential for B cell development, proliferation, and homeostasis under steady-state conditions.

We noted that our BDKO mice still showed significantly lower levels of serum IgM and IgG1 compared with control mice and B cell–specific p53 KO mice (Fig. 7 A). Consistent with the reduced serum IgG1 in BDKO mice, the impaired LPS-plus-IL-4–induced CSR to IgG1 observed in BMKO B cells was not rescued by genetic ablation of p53 (Fig. 7 B). Indeed, the percentage of class-switched IgG1-producing cells was reduced even further in BDKO B cells, suggesting that both Mule and p53 are indispensable for CSR to IgG1. In addition, the B1a cells missing from the peritoneal cavity of BMKO mice were not restored by loss of p53 (Fig. 7 C). Because B1a cells are a major producer of IgM (Herzenberg, 2000), this finding may account for the reduced serum IgM levels in BMKO and BDKO mice (Fig. 4 A and Fig. 7 A). Notably, deletion of p53 alone in B cells affected neither LPS-plus-IL-4–induced CSR nor B1a cell numbers (Fig. 7, A–C). Thus, the defective CSR, impaired serum Ig production, and lack of B1a cells in BMKO mice are not attributable to increased p53 transactivation.

To examine how Mule regulates p53 protein levels under stress conditions, we treated Mule-deficient MEFs and BMKO B cells with genotoxic agents known to elicit a p53-dependent response. Control MEFs treated with doxorubicin (Dox) exhibited up-regulation of Mule protein as early as 10 min after Dox treatment (Fig. 8 A). p53 protein gradually accumulated in Dox-treated control MEFs and became phosphorylated at Ser18 (Fig. 8 B, left). Consistent with p53 being a Mule substrate (Chen et al., 2005), high levels of p53 protein were present in Mule-deficient MEFs even at steady state (Fig. 8 B, right, top row). However, phosphorylation of this p53 in response to Dox treatment was impaired (Fig. 8 B, right, middle row). Because genotoxic stress induces p53 phosphorylation by activated ATM (Lavin, 2008), we next examined ATM phosphorylation in the absence of Mule. In Dox-treated control MEFs, total ATM protein accumulated with time (Fig. 8 C, top row) and became phosphorylated at Ser1987 by 2 h after stimulation (Fig. 8 C, middle row). However, consistent with their decreased phospho-p53, Dox-treated Mule-deficient MEFs showed much lower levels of phospho-ATM (Fig. 8 C). Another relevant substrate of phospho-ATM is Brca1 (Cortez et al., 1999). In response to Dox treatment, Brca1 accumulated in a time-dependent manner in control MEFs but not in Mule-deficient MEFs (Fig. 8 D). Intriguingly, low levels of phospho-ATM and high levels of Brca1 were present in Mule-deficient MEFs even without genotoxic stress (Fig. 8, C and D), whereas control MEFs showed no detectable phospho-ATM or Brca1 at steady state. To investigate whether there was a physical interaction between Mule and ATM that might account for our observations, protein lysates from control MEFs that were left untreated or treated with Dox for 2 h were subjected to immunoprecipitation with anti-Mule Ab followed by anti–phospho-ATM (S1987) immunoblotting. The results showed that Mule was able to interact with phospho-ATM (Ser1987) after Dox treatment (Fig. 8 E, left). The specificity of this interaction was confirmed in a parallel immunoprecipitation experiment using Mule-deficient MEFs (Fig. 8 E, right).

As we are aware that the interpretation of p53-related experiments in E1A transformed MEFs may be equivocal, we extended our investigations in primary B cells. We exposed control and BMKO splenic B cells to two different genotoxic agents. Consistent with our MEF analyses, Mule was up-regulated in control splenic B cells as soon as 10 min after exposure to γ-irradiation (unpublished data). Although our earlier data showed that p53 protein was undetectable in control splenic B cells at steady state (Fig. 6 A), p53 protein was both stabilized and efficiently phosphorylated at Ser18 in control B cells subjected to γ-irradiation for 2 h (Fig. 9 A, left). However, neither stabilization nor Ser18 phosphorylation of p53 (phospho-p53-Ser18) was observed in γ-irradiated BMKO B cells (Fig. 9 A, left). Similarly, control B cells (but not BMKO B cells) treated with etoposide accumulated phospho–p53-Ser18 (Fig. 9 A, right). In γ-irradiated control B cells, total ATM protein accumulated with time and became phosphorylated at Ser1987 by 60 min after stimulation (Fig. 9 B, left). In contrast, consistent with their decreased phospho-p53-Ser18, γ-irradiated BMKO B cells showed much lower levels of both total ATM and phospho-ATM (Fig. 9 B, right). As we had observed for MEFs (Fig. 8 C), spontaneous phosphorylation of ATM at Ser1987 occurred in BMKO B cells (but not in control B cells) even in the absence of genotoxic stress (Fig. 9 B, middle row). Again replicating our observations in stressed MEFs, Brca1 protein readily accumulated in γ-irradiated control B cells but not in BMKO B cells (Fig. 9 C). Thus, Mule is required for efficient phosphorylation of ATM and induction of some of its substrates.

Although Mcl-1 is a known substrate for Mule’s E3 ubiquitin ligase activity, only slightly more Mcl-1 protein was present in BMKO B cells than in control B cells at steady state (Fig. 9 D). To investigate whether stress-induced Mcl-1 degradation is mediated by Mule, we treated purified splenic control and BMKO B cells with etoposide and analyzed lysates of these cells by immunoblotting. Mcl-1 protein degradation was highly efficient in stressed control B cells but largely inhibited in stressed BMKO B cells (Fig. 9 D), indicating that the E3 ubiquitin ligase activity of Mule is critical for controlling Mcl-1 protein levels in response to genotoxic stress. Consistent with their impaired Mcl-1 degradation and decreased p53 accumulation and phosphorylation, BMKO B cells treated with etoposide for 4–6 h showed a two to threefold reduction in caspase-3 and -9 cleavage (Fig. 9 E and not depicted) and were consequently resistant to etoposide-induced apoptosis (Fig. 9 F). However, BMKO B cells were just as sensitive as control B cells to apoptosis induced by dexamethasone (Fig. 9 F), which is a p53-independent stimulus. Thus, Mule is required for Mcl-1 degradation and proper ATM-mediated p53 responses to genotoxic stress.

In summary, our results show that, in addition to targeting Mcl-1 and p53 for polyubiquitination and degradation (Chen et al., 2005; Zhong et al., 2005), Mule is required for the regulation of ATM and its substrates. Importantly, this regulation is required to maintain cellular homeostasis under both steady-state and genotoxic stress conditions.

Protein ubiquitination plays key roles in the transcriptional regulation and signal transduction that control the balance between cell survival and cell death (Liu et al., 2005a). Mule has been identified as an E3 ubiquitin ligase that polyubiquitinates p53 and targets it for proteasomal degradation (Chen et al., 2005). In this study, we present data from examinations of Mule-deficient B cells and MEFs which indicate that Mule is essential for cellular homeostasis under both resting and stress conditions.

We found that B cell development, proliferation, homeostasis, and immune responses were impaired in BMKO mice, and that Mule maintains B cell homeostasis at steady state by regulating p53 activity. Elevated p53 transactivation is largely responsible for the blockade of B cell development and disturbed B cell functions observed in unstressed BMKO mice, as indicated by the rescue of these phenotypes upon genetic ablation of p53. Given that Mule-deficient B cells expressed a basal low level of phospho–ATM-S1987 that was not observed in control cells, we speculate that this activated ATM may constitutively phosphorylate p53 and activate it sufficiently to at least partially account for the two to fourfold increase in p53 transactivation seen in the mutant cells. In addition to increasing apoptosis, this elevated p53 transactivation could trigger cell cycle arrest. This hypothesis is supported by our observation of decreased BrdU incorporation by BMKO splenic B cells compared with controls. We recognize that further investigation is needed to clarify the extent to which cell cycle arrest contributes to the B cell deficit caused by loss of Mule. However, the abrogation of the abnormal increase in Puma and Noxa mRNAs in BMKO B cells upon genetic ablation of p53 argues in favor of the interpretation that p53–mediated apoptosis is responsible for the B cell defects seen in BMKO mice. Genetic ablation of Puma in BMKO mice through genetic breeding should resolve this issue clearly. Finally, although BMKO B cells exhibited elevated p53 transactivation, total p53 protein levels remained below the limit of detection by immunoblot. Consequently, we cannot be certain that p53 protein levels were coordinately increased in these cells, although elevated p53 protein was detected in both Mule-deficient MEFs and ES cells under the same conditions. In any case, these data suggest that Mule may influence the transactivation and functions of p53 in a manner that is not exclusively dependent on p53 polyubiquitination and degradation.

Modulation of p53–Mdm2 interaction is essential for normal p53 activity because genetic ablation of p53 prevents the lethality of Mdm2 KO mice (Brooks and Gu, 2006). B cell–specific Mdm2 deletion prevents any B cell development in these animals (unpublished data), a more severe phenotype than that resulting from B cell–specific loss of Mule. Because both Mule and Mdm2 target p53 for polyubiquitination and degradation, it is reasonable to speculate that Mule might compensate for the loss of Mdm2 in Mdm2-deficient mice and vice versa. However, the embryonic lethality of both Mule-null and Mdm2-null mice, and the defective B cell development in B cell–specific Mule and B cell–specific Mdm2 KO mutants, suggest that Mule and Mdm2 have nonoverlapping functions in maintaining p53 levels. Thus, both Mdm2 and Mule are needed to maintain p53 protein at the appropriate low levels at steady state.

Although p53 ablation rescued the defective B cell development, proliferation, and homeostasis in BMKO mice, it failed to restore normal B1a cells, CSR, or serum Ig levels. We speculate that the latter two deficits may be related to the impaired ATM activation in BMKO B cells. ATM-deficient mice show defects in B cell CSR and serum Ig levels that partially mirror those in BMKO mice (Xu et al., 1996; Lumsden et al., 2004). Because ATM stabilizes the DNA double-strand breaks required for Ig gene assembly (Bredemeyer et al., 2006), it is possible that some of the effects of Mule deficiency on B cell development are mediated through ATM. In addition, given that siRNA knockdown of Mule leads to increased DNA repair in manner dependent on DNA polymerase β (Parsons et al., 2009), loss of Mule in B cells could quite conceivably lead to defects in RAG-mediated V(D)J recombination. However, BMKO and control mice showed comparable percentages of intracellular Ig μ⁺ cells among B220⁺CD43⁺ BM B cells (unpublished data). Consistent with this finding, the Ig κ-to-λ light chain ratio in newly formed splenic B cells is not altered in BMKO mice. These results suggested that neither RAG-mediated V(D)J recombination of the Ig H locus and its expression nor the λ light chain usage are affected by loss of Mule. Thus, Mule is unlikely to be required for events associated with DNA recombination in B cells.

Given that Mule has multiple substrates, it is not surprising that Mule has different regulatory effects on different cell types. In neurons, Mule controls differentiation and proliferation in vivo by destabilizing its substrate N-Myc (Zhao et al., 2008). In our study, we have shown that Mule regulates B cell development, proliferation, and homeostasis through the ATM–p53 axis. Although Mule was previously identified as an ATM substrate in screening studies, the biological implications of Mule–ATM interaction were not explored (Matsuoka et al., 2007; Mu et al., 2007). Our study has unearthed the intriguing possibility that Mule plays opposite roles in ATM phosphorylation depending on whether the cell is at steady state or has experienced DNA damage. We theorize that, under steady-state conditions, Mule is required to suppress ATM phosphorylation, thereby preventing excessive activation of p53 that could disrupt cellular homeostasis. However, during responses to genotoxic stress, Mule protein quickly accumulates and binds to phospho-ATM. The importance of this Mule–ATM interaction is highlighted by the impaired ATM accumulation and phosphorylation observed in both Mule-deficient B cells and MEFs treated with DNA-damaging agents. Without Mule, neither of the ATM substrates phospho-p53 nor Brca1 can be efficiently induced. Thus, Mule is directly or indirectly involved in the ATM-mediated response to DNA damage. In support of this hypothesis, we have found that the numbers of ATM foci and 53BP1 foci induced by DNA damage are greatly reduced in Mule-deficient MEFs compared with controls (unpublished data). Whether Mule itself participates in the DNA damage response through its interaction with ATM, or whether it is Mule’s ubiquitination of its known substrates (such as histones [Liu et al., 2005b], TopBP1 [Herold et al., 2008], and/or DNA polymerase [Parsons et al., 2009]) that is important, remains to be determined.

In conclusion, we have demonstrated crucial functions for Mule in regulating the ATM–p53 axis under both steady-state and genotoxic stress conditions. Disruption of Mule function in various cell types leads to developmental and functional deficits that implicate this E3 ubiquitin ligase as a key player in a new molecular network required for the maintenance of cellular homeostasis.

Mule floxed (Mule^fl/fl(y)) mice were generated using Cre/LoxP recombination and the gene targeting and screening strategies outlined in Fig. 1. The experimental procedures used for the culture, transfection, and selection of E14K ES cells (129/Ola) have been previously described (Hao and Rajewsky, 2001). Chimeric mice were crossed to Flpe-deleter mice (Rodríguez et al., 2000) to generate Mule^fl/fl(y) mice.

To generate BMKO mice (Mule deleted specifically in B cells), Mule^fl/fl(y) mice were bred with Mb1Cre mice (C57BL/6; Hobeika et al., 2006). The resulting Mule^fl/fl(y)Mb1Cre mice were backcrossed for 6–10 generations to C57BL/6. p53 was deleted in Mule^fl/fl(y)Mb1Cre mice by crossing to P53^fl/fl mice (Jonkers et al., 2001). The resulting Mule^fl/+(y)P53^fl/+ F1 mice, with or without Mb1Cre, were intercrossed to generate Mule^fl/fl(y)P53^fl/flMb1Cre mice (BDKO mice; mixed genetic background). Because Cre expression can be toxic to cells (Schmidt-Supprian and Rajewsky, 2007), we included Mule^fl/+Mb1Cre, P53^fl/+Mb1Cre and Mule^fl/+P53^fl/+Mb1Cre mice in all initial analyses of the corresponding mutants. These control mice were phenotypically indistinguishable from the control Mule^fl/fl(y) animals used in this study, as judged by cellularity and size of spleen. Mice used in experiments were ≥2 mo old unless otherwise specified. All animal experiments were approved by the University Health Network Animal Care Committee.

Cells were lysed in 0.5% NP-40 lysis buffer (20 mM Tris-HCl, pH 8, 137 mM NaCl, 10% glycerol, 0.5% NP-40, and 2.5 mM EDTA), or RIPA buffer (50 mM Tris-HCl, pH 8, 150 mM NaCl, 0.5% NP-40, 0.5% sodium deoxycholate, and 0.1% SDS), with protease inhibitor and phosphatase inhibitor (Roche) freshly added. Lysates were incubated on ice for 30 min and cleared by centrifugation. Protein concentrations in lysates were determined using the BCA protein assay (Thermo Fisher Scientific). Cell lysates (20–100 µg) were loaded onto 4–12% bis-Tris gels to detect proteins below 200 kD, or onto 3–8% Tris-acetate gels to detect Mule. Fractionated proteins were transferred to a nitrocellulose membrane by i-Blot according to the manufacturer’s instructions (Invitrogen). The membranes were immunoblotted using Abs recognizing the following: ATM (GeneTex); Ser1987-ATM (Rockland); Brca1 (Santa Cruz Biotechnology, Inc.); p53 (Santa Cruz Biotechnology, Inc.); Ser18-p53 (R&D Systems); and phospho-tyrosine (Millipore). Infrared dye-labeled secondary Abs (anti–rabbit Alexa Fluor 680 [Invitrogen] and anti–mouse IR800 [LI-COR Biosciences]) were visualized with an Odyssey scanner (LI-COR Biosciences).

For immunoprecipitations, cell lysates prepared from 10⁸ MEFs were incubated with anti-Mule Ab and protein G–Sepharose beads at 4°C overnight. Immunocomplexes were washed five times with lysis buffer, followed by standard immunoblotting analysis as described in the previous paragraph.

Single cell suspensions were prepared from spleen or BM preparations that had been treated to lyse red blood cells. Cells (1–2 × 10⁶) were preincubated with Fc block for 15 min at 4°C and immunostained with Abs recognizing the following: IgM (II/41), IgD (11-26c.2a), IgG1 (X56), IgG3 (R40-82), Ig κ light chain (187.1), Ig λ light chain (R26-46), CD5 (53–7.3), CD9 (KMC8), CD16/CD32 (2.4G2), CD21 (7G6), CD23 (B3B4), CD24 (M1/69), CD25 (PC61), CD38 (90), CD43 (S7), CD45R/B220 (RA3-6B2), CD69 (H1.2F3), CD86 (GL1), CD93 (AA4.1), or CD117/ cKit (2B8; all obtained from BD or eBioscience unless otherwise specified). A mouse mAb against bacterially expressed Mule (aa 2375–2491) was generated by ProMab and used to detect Mule protein in immunoblotting and intracellular flow cytometry experiments. This anti-Mule mAb recognizes a specific 480-kD band in mouse cell extracts, as confirmed in Mule-deficient MEFs. Fluorescein-labeled PNA (peanut agglutinin) was used to detect GCs (Vector Laboratories). Flow cytometry data were acquired using either a FACSCalibur (BD) or FACSCanto (BD) flow cytometer, and analyzed with either the CellQuest software (BD) or the FlowJo analysis program (Tree Star).

Intracellular staining of Mule was performed as described previously (Hao et al., 2008). In brief, BM cells were fixed with 1.6% paraformaldehyde and incubated for 30 min at RT. After one wash in PBS, ice-cold methanol (100%) was added dropwise and cells were incubated on ice for 30 min. Washed cells were incubated for 30 min with Abs recognizing B220 (RA3-6B2), CD43 (S7), IgM (II/4.1), or Mule. Immunostained cells were incubated with Alexa Fluor 488– or Alexa Fluor 647–conjugated rabbit anti–mouse IgG and analyzed by flow cytometry.

For intracellular Ig μ staining, BM cells were first surface stained with the following fluorescent Abs recognizing B220, IgM, CD43, or CD25. B220⁺IgM⁻CD43⁺CD25⁻ cells were sorted by flow cytometry, fixed, and permeabilized using the Cytofix/Cytoperm kit according to the manufacturer’s instructions (BD). Fixed cells were stained with FITC–anti-IgM (II/4.1) and analyzed by flow cytometry.

For BrdU staining, mice were supplied with BrdU-containing drinking water (1 mg/ml) for 3.5 d. Total splenic B cells were isolated and immunostained as described in the previous paragraph, and sorted into TS (B220⁺IgM⁺CD93⁺), MZ (B220⁺IgM⁺CD21⁺CD23^low), and FO (B220⁺IgM⁺CD21⁺CD23⁺) subpopulations by flow cytometry. BrdU incorporation in each subset was detected by flow cytometry using the BrdU-Flow kit according to the manufacturer’s instructions (BD).

Splenocytes were stained using FITC-conjugated anti-B220 Ab as described in the previous section. 1 × 10⁷ stained cells were incubated at 37°C for 45 min with 5 µg/ml Indo-1 AM (Invitrogen) in RPMI-1640 medium supplemented with 10% FCS. Indo-loaded cells were resuspended at 5 × 10⁶/ml and kept on ice until use. 500-µl aliquots of Indo-loaded cells were warmed to 37°C for 10 min and a baseline reading was acquired for 1 min. Ca²⁺ flux induction was triggered by the addition of 20 µg/ml goat anti–mouse IgM (Jackson ImmunoResearch Laboratories, Inc.) followed, 7 min later, by addition of 100 nM thapsigargin as a positive control. Data collection was continued for another 5 min. The Ca²⁺ response was measured based on the ratio of violet (Ca²⁺ bound) to blue (Ca²⁺ free) fluorescence as detected by flow cytometry (LSRII; BD).

Purified splenic B cells were seeded in triplicate in 96-well U-bottom plates at 10⁵ cells/200 µl in RPMI-1640 containing 10% FCS. Cells were left untreated, or stimulated with various doses of anti-IgM (Jackson ImmunoResearch Laboratories, Inc.), or with anti-IgM plus 1 µg/ml anti-CD40 (BD). At 16, 40, or 64 h after seeding with stimuli, [³H]-thymidine (1 µCi) was added to each well and cells were cultured for another 8 h to determine proliferation at 24, 48, or 72 h [³H]-thymidine uptake was assessed using a liquid scintillation β-counter (TopCount reader).

Mice were immunized i.p. with either 25 µg of the Ti antigen TNP-Ficoll in PBS, or 100 µg of the alum-precipitated Td antigen NP₍₁₅₎-CG (Biosearch Technologies). Assays were performed using 96-well ELISA plates coated with 20 µg/ml TNP-BSA or 10 µg/ml NP₍₁₅₎-BSA. Serum samples collected on day 0 (before immunization) and on days 7, 14, and 21 after immunization were serially diluted and applied to the ELISA plates. Bound Abs were detected by the addition of biotinylated goat anti–mouse IgM, IgG1, and IgG3 (Santa Cruz Biotechnology, Inc.). Biotin binding was revealed using streptavidin-HRP and its substrate. Absorbance at 450 nm was measured using a FlexStation Microplate Reader (Molecular Devices).

VSV infection was performed as previously described (Junt et al., 2002). In brief, mice were i.v. injected with 10⁵ pfu VSV-Indiana strain (VSV-IND, Mudd-Summers isolate). Virus was propagated on BHK-21 cells at an MOI of 0.01 and plaqued on Vero cells. During survival experiments, the health status of mice was checked twice daily. Upon appearance of clinical signs of VSV replication in the CNS, such as paralysis or scrubby pelt, mice were removed from the experiment and counted as dead.

For Ig analyses, serum samples collected on days 0, 2, 4, 6, 10, and 12 after infection were diluted 40-fold in MEM containing 2% FCS. For the detection of neutralizing IgM and IgG, various serum dilutions were incubated with 500 pfu VSV-Ind for 90 min followed by transfer to a single cell layer of Vero cells and incubation for 1 h. After addition of an overlay (1:1 MEM/2% methylcellulose), cells were incubated overnight before fixation and staining for 1 h with 0.5% crystal violet (Sigma-Aldrich). The dilution that resulted in a 50% reduction in virus plaques was taken as the neutralizing titer. For IgG titers, serum samples were incubated with an equal volume of 0.1 M β-ME in PBS for 1 h at RT before dilution. Before use in experiments, all mouse sera were heated at 56°C for 30 min to inactivate complement.

Purified splenic B cells were cultured for 3–5 d in RPMI-1640 containing 10% FCS plus either 20 µg/ml LPS (InVivo Gene) or 20 µg/ml LPS plus 100–200 U/ml IL-4 (PeproTech). The percentages of B cells that had undergone class-switching to production of IgG3 or IgG1 were determined by flow cytometry.

Purified splenic B cells were incubated with 1 µM CFSE (Invitrogen) in serum-free RPMI-1640 medium for 10 min at 37°C with gentle mixing. After two washes, the labeled cells were seeded at a density of 5 × 10⁶ cells/ml in RPMI-1640 containing 10% FCS. Labeled cells were stimulated with 20 µg/ml LPS plus 100–200 U/ml IL-4 for 3–5 d. The proliferation capacity of the stimulated B cells was measured by flow cytometry as denoted by a decrease in CFSE fluorescence intensity.

MEFs derived from E14.5 embryos arising from the crossing of Mule^fl/+ and Mule^fl/y mice were transformed with pLPC-E1A-IRES-Ras^V12 according to a standard protocol. Transformed Mule^fl/y MEFs were transduced with His-TAT-NLS-Cre (Peitz et al., 2002), followed by limiting dilution and expansion to screen for single clones that had undergone Cre-mediated deletion of the floxed Mule allele. The elimination of Mule protein in MEFs was confirmed by immunoblotting as described in Immunoblotting and immunoprecipitation.

Targeted ES cell clones were transiently transfected with a Cre-encoding plasmid to delete the distal, loxP-flanked exons 76 and 77 of the mouse Mule gene as well as the neomycin resistance gene. DNA samples from neomycin-sensitive clones were digested with StuI and hybridized to probe c, as depicted in Fig. 1. The appearance in these mutants of only a band of 6.9 kb confirmed the total deletion of the targeted Mule exons. The elimination of Mule protein in ES cells was confirmed by immunoblotting as described in Immunoblotting and immunoprecipitation.

Total RNA was extracted from purified splenic B cells using the RNeasy kit (QIAGEN) according to the manufacturer’s protocol. Extracted RNAs were reverse transcribed using random hexamers and the SuperScript III kit (Invitrogen) as per the manufacturer’s protocol. The resulting cDNAs were diluted 1:10 and served as templates for real-time PCR reactions using the Power SYBR-Green PCR Master Mix (Applied Biosystems). The optimization of the real-time PCR reaction was performed according to the manufacturer’s instructions but was scaled down to a 10-µl final reaction volume containing 500 nM forward primer and 500 nM reverse primer. The PCR reaction mixtures were preincubated at 50°C for 2 min, followed by denaturation at 95°C for 10 min, and 40 cycles of 95°C for 15 s and 60°C for 1 min, using the ABI 7700HT Fast Real-Time PCR System (Applied Biosystems). Data were analyzed using SDS software (Applied Biosystems).

Primer sequences were as follows: p21 forward, 5′-GGCCCGGAACATCTCAGG-3′; p21 reverse, 5′-AAATCTGTCAGGCTGGTCTGC-3′; Puma forward, 5′-TCACCCTGGAGGGTCATGTA-3′; Puma reverse, 5′-GCGGGTGTAGGCACCTAGT-3′; Noxa forward, 5′-ATCAAGGGCTGAAGGGATTT-3′; Noxa reverse, 5′-AGAGAGGCACCTGGGATATG-3′; Mdm2 forward, 5′-CCAGCTTCGGAACAAGAGAC-3′; Mdm2 reverse, 5′-GTCGTTTTGCGCTCCAAC-3′; Bax forward, 5′-TGGAGATGAACTGGACAGCA-3′; Bax reverse, 5′-GAAGTTGCCATCAGCAAACA-3′; p53 forward, 5′-ACTGCATGGACGATCTGTTG-3′; and p53 reverse, 5′-GTGACAGGGTCCTGTGCTG-3′. Each sample was assessed in triplicate. Relative mRNA levels were normalized to the housekeeping gene 18S or the actin mRNA and calculated using the comparative threshold cycle method (2−ΔΔ^Ct).

For culture under stress conditions, purified splenic B cells were incubated at 37°C in RPMI-1640 containing 10% FCS in the presence or absence of etoposide (Sigma-Aldrich) or dexamethasone (Sigma-Aldrich) at the concentrations indicated in the figures. Alternatively, cells were subjected to 5 Gy γ-irradiation for the times indicated in the figures. MEFs were cultured at 37°C in DME containing 10% FCS in the presence or absence of 1 µM Dox (Sigma-Aldrich) for the times indicated in the figures. Proteins were analyzed by immunoblotting as described in Immunoblotting and immunoprecipitation. For apoptosis assays, purified splenic B cells were cultured for 20–24 h with the concentrations of etoposide indicated in the Figures. Cells were harvested and stained with Annexin V plus propidium iodide and analyzed by flow cytometry using a standard protocol.

The Student’s t test with tail 1 and type 3 was used for statistical analyses. Values are expressed as the mean ± SD. For statistical significance: *, P < 0.05; **, P < 0.005; ***, P < 0.0005.

We thank A. Berns for floxed p53 mice; the Princess Margaret Hospital/Ontario Cancer Institute Flow Cytometry Facility for cell sorting; the Campbell Family Institute for Cancer Research PCR Genotyping Facility for genotyping; and M. Saunders for scientific editing.

This work was supported by grants to T.W. Mak and Z. Hao by the Canadian Institutes of Health Research. This research was also funded in part by the Ontario Ministry of Health and Long Term Care. The views expressed do not necessarily reflect those of the Ontario Ministry of Health and Long Term Care.

The authors declare that they have no competing financial interests.