Roles of the Ig κ Light Chain Intronic and 3′ Enhancers in Igk Somatic Hypermutation1

Somatic hypermutation (SHM) of the rearranged Ig genes is required for the affinity maturation of Abs. SHM is almost exclusively targeted to the rearranged Ig loci, but the mechanism of this gene-specific targeting remains unclear. The Ig κ L chain locus contains multiple enhancers, including the MAR/intronic (iEκ) and 3′ enhancers (3′Eκ). Previous transgenic studies indicate that both κ enhancers are individually necessary for SHM of Igk. In contrast, later studies of Ag-selected Vκ genes in 3′Eκ−/− mice found no absolute requirement for 3′Eκ in κ SHM. To address the roles of the two κ enhancers in SHM in a physiological context, we analyzed SHM of the endogenous Igk in mice with a targeted deletion of either iEκ or 3′Eκ in Peyer’s patch germinal center B cells. Our findings indicate that, although 3′Eκ is quantitatively important for SHM of Igk, iEκ is not required for κ SHM. In addition, a reduction of κ mRNA levels is also detected in activated 3′Eκ−/− B cells. These findings suggest that iEκ and 3′Eκ play distinct roles in regulating Igk transcription and SHM.

B cells diversify the repertoire of their Ag receptor genes through multiple and distinct processes. For example, somatic joining of the V, D, and J gene segments, mediated by the RAGs, occurs during early B cell development (1). In addition, somatic hypermutation (SHM) 4 of the rearranged V(D)J sequence, mediated by the activation-induced deaminase gene (AID), occurs in activated B cells in response to antigenic challenge (2). The combinatory effects of SHM and the selection for B cells expressing Ig with a higher affinity for Ag leads to affinity maturation (3). SHM is a tightly regulated process such that mutations are targeted almost exclusively to the rearranged Ig loci, with rare but significant mistargeting to oncogenes (3). However, the mechanism that targets SHM to rearranged Ig loci remains unclear.
Accumulating evidence has identified a connection between SHM and transcription (4). For example, the expression levels of transgenes strongly correlate with the levels of mutation (5)(6)(7). In addition, hypermutation of the Ig genes begins just downstream of the promoter (8,9) and ends well before the C region (10). This ϳ2-kb window from the promoter for SHM is observed regardless of which J gene segment is rearranged (8,11). Consistent with the notion that the distance to the promoter determines the mutability of a region, the insertion of a promoter upstream of the C region in a transgenic Igk construct caused SHM within the C region at levels similar to the V region (12).
Three enhancers have been identified within the Igk locus: one within the intron between J 5 and C (called the intronic enhancer, or iE ) (13), another 9 kb 3Ј of C (the 3Ј enhancer, or 3ЈE ) (14), and a third 8 kb downstream of 3ЈE (the downstream enhancer, or Ed) (15). The deletion of either iE or 3ЈE through gene targeting leads to a quantitative decrease in Igk rearrangement (16,17), and a near-complete block in Igk rearrangement is observed with the loss of both enhancers, indicating that these two enhancers are collectively essential for Igk rearrangement (18).
However, the roles of these enhancers in regulating SHM have been more controversial. Transgenic studies indicated that both iE and 3ЈE were necessary for the SHM of Igk transgenes (19,20). In this context, deletion of iE or 3ЈE resulted in a dramatic reduction of SHM of the V region within Igk transgenes to levels equivalent to that caused by PCR error (20). However, analysis of mice bearing a deletion of the endogenous 3Ј enhancer leads to a different conclusion. When immunized with the hapten 2-phenyl-5-oxazolone (ph-Ox), 3ЈE Ϫ/Ϫ mice underwent wild-type (Wt) levels of SHM of their V regions, demonstrating that 3ЈE is not absolutely required for SHM of Igk (21). In addition, the amount of anti-ph-Ox Abs present in the sera of the immunized 3ЈE Ϫ/Ϫ mice was comparable with that of Wt mice. However, in these mice, the SHM of the endogenously rearranged V Ox1/J 5 exon was analyzed after repeated immunization with ph-Ox. Because B cells expressing an Ig with higher avidity to the hapten are selected for survival during Ag-driven affinity maturation, it is possible that the functions of 3ЈE in SHM might be masked by such stringent selection. In this context, mice with reduced capacity for SHM might still produce B cells with hypermutated Ig . In this regard, it was noted that the mutation frequency of the strong hot spot was potentially decreased in 3ЈE -deleted loci, leading to the suggestion that the Ag-driven B cell activation and/or somatic mutation might be less efficient in 3ЈE -deleted cells (21).
In addition to reduced SHM, Igk transgenes lacking 3ЈE also have dramatically reduced expression (20). Because transcription is closely correlated with the efficiency of SHM, it was hypothesized that the defect in SHM of the 3ЈE -deleted Igk transgene was caused by the reduced transcription. In 3ЈE Ϫ/Ϫ mice, surface expression of Igk on mature resting B cells is reduced. However, Igk transcription appeared normal in 3ЈE Ϫ/Ϫ B cells activated by LPS (17). Because stimulation of B cells by LPS involves many pathways but not the ones involved in signaling through the Ag receptor or CD40, it remains to be tested whether 3ЈE is important for transcription during an immune response. iE plays a quantitatively more important role than 3ЈE in activating Igk rearrangement (18). However, in contrast to 3ЈE Ϫ/Ϫ

B cells, expression in iE
Ϫ/Ϫ B cells is normal (16). Although transgenic studies have indicated a critical role of iE in SHM of the transgenic Igk (20), the physiological roles of iE in SHM of Igk remain to be determined. To address these issues, we used a protocol to analyze SHM of Igk in germinal center (GC) B cells during a polyclonal immune response. In this context, Peyer's patch (PP) B cells are exposed to continuous antigenic challenge in the gut lumen, and hence undergo robust SHM (22). We also examined splenic GC B cells following immunization with the Ag Sheep RBC (SRBC), which induces a polyclonal response (23)(24)(25). Additionally, only mutations within the 5Ј end of the J C intron, which have no bearing on the specificity of the Ab and therefore will not be selected by Ag, were analyzed. Our findings indicate a quantitatively important role for 3ЈE , but not iE , in SHM.

Generation of iE
Ϫ/Ϫ mice iE and associated MAR of the endogenous allele was replaced with the LoxP-flanked PGK-Neo r gene in ES cells exactly as previously described (16). The heterozygous mutant ES cells were used to generate chimeric mice that transmitted the mutant allele into germline. To excise the PGK-Neo r gene from the targeted allele, the germline heterozygous mutant mice with the PGK-Neo r gene inserted in the targeted allele were bred with CMV-Cre transgenic mice that express the Cre enzyme in zygotes. The resulting heterozygous mutant mice with the PGK-Neo r gene deleted from the targeted allele were intercrossed to generate homozygous iE Ϫ/Ϫ mice.

Immunizations and splenic and PP GC B cell sorting
PP were harvested from iE Ϫ/Ϫ , 3ЈE Ϫ/Ϫ , and Wt littermates. Cells were stained with anti-mouse B220 or CD19 conjugated to PE (BD Pharmingen) and anti-mouse GL7 conjugated to FITC (BD Pharmingen). Live GL7 Ϫ and GL7 ϩ B cells were sorted using a FACSVantage cell sorter (BD Biosciences). Alternatively, PP cells were stained with anti-mouse Ig conjugated to PE (BD Pharmingen) and anti-mouse GL7 Abs, and then Ig ϩ GL7 ϩ cells were sorted. 3ЈE Ϫ/Ϫ and Wt littermates of 5 mo of age were immunized with 2 ϫ 10 8 SRBC (catalog no. R3378; Sigma-Aldrich). At day 7, mice were given a booster of 2 ϫ 10 8 SRBC. At day 14, spleens and lymph nodes were collected from these mice and purified into singlecell suspensions, which were stained with anti-mouse B220 (BD Pharmingen) Abs conjugated to PE (BD Pharmingen) and peanut agglutinin (PNA)-FITC (Vector Laboratories). Live (PI Ϫ ) B220 ϩ PNA high GC B cells were sorted.

Analysis of SHM
Genomic DNA from sorted cells was purified, and the J C intron was amplified by PCR using KOD hot-start polymerase (Calbiochem) according to the manufacturer's protocol, with the following primers: V D.bam, 5Ј-ccggatccGGCTGCAGSTTCAGTGGCAGTGGRTCWGGRAC-3Ј; K1.eco, 5Ј-ggcgaattcGTGACAAATTTTAGAATAAGAGTCACACCTC-3Ј.
Sequences in lowercase were added to introduce restriction sites for cloning. The PCR conditions were 95°C for 2 min, followed by 32-35 cycles of 94°C for 20 s, 57°C for 30 s, and 68°C for 1 min, concluding with a final extension at 68°C for 10 min. Purified PCR products were cloned into pBluescript and sequenced. Clones were sequenced using the T3 primer. Sequences were aligned using ClustalX software and scored for mutations. The unpaired t test was used. A value of p Ͻ 0.05 was considered significant.

Real-time quantitative PCR analysis
Unstimulated and stimulated ϩ B cells from 3ЈE Ϫ/Ϫ and Wt littermates were harvested, and their RNA was purified, converted to cDNA, and analyzed by real-time PCR as described previously (26). Primers for C and C are as described (26).
The University of California, San Diego, Animal Subjects Committee approved all experiments that involved mice.

Normal Igk expression in iE
Ϫ/Ϫ B cells To examine whether Igk expression in activated B cells requires iE , surface expression on B cells derived from the PP of Wt and iE Ϫ/Ϫ littermates was analyzed by flow cytometry. The B cell activation in PP was similar between Wt and iE Ϫ/Ϫ mice (Fig.  1A). In addition, surface Ig levels were normal in unactivated (GL7 Ϫ ) as well as activated iE Ϫ/Ϫ B cells (GL7 ϩ ) when compared with those in Wt controls (Fig. 1, B and C). Because gene rearrangement is compromised in iE Ϫ/Ϫ B cells, as published previously (16), knockout mice exhibit a much higher proportion of Ig-negative (Ig-positive) cells. Thus, consistent with the conclusion from the transgenic studies (20), iE is not required for expression in activated B cells.
To determine the role of iE in SHM, GC ϩ B cells ( ϩ B220 ϩ GL7 ϩ ) were sorted from the pooled PP of three 4-moold Wt and iE Ϫ/Ϫ littermates. To amplify only the rearranged Igk allele from genomic DNA purified from the sorted B cells, we used the degenerate V primer, which binds to ϳ90% of V gene segments (27), and a primer ϳ600 bp downstream of J 5 ( Fig. 2A). PCR products were cloned into pBluescript and sequenced. Because mutations within the intronic region 3Ј of J 5 are not selected by Ag, we analyzed the mutation frequency in the 500-bp intronic sequence. The mutation rate was 12.6 ϫ 10 Ϫ3 mutations/bp for Wt B cells and 15.8 ϫ 10 Ϫ3 for the iE Ϫ/Ϫ B cells ( Fig. 2B and Table I). Although SHM appeared slightly elevated in iE Ϫ/Ϫ B cells, this was due to the lower number of sequences analyzed with zero mutations, because the mutation frequency on sequences containing at least one mutation was equivalent between iE Ϫ/Ϫ and Wt (Table I). The mutational spectra were also similar between Wt and iE Ϫ/Ϫ B cells (Fig. 2C), resulting in similar percentages of transitions and transversions ( Table I). The mutations identified from both groups showed the hallmarks of mutations generated by SHM, because they were RGYW/WRCY hot spot targeted and biased toward transitions (Table I). Thus, SHM is normal in iE Ϫ/Ϫ mice.

Reduced mRNA levels in activated 3ЈE
Ϫ/Ϫ B cells Because deletion of 3ЈE leads to a reduction in surface expression in unactivated B cells (17), we examined how expression is affected in activated 3ЈE Ϫ/Ϫ B cells. Previous studies indicated that 3ЈE was not necessary for expression in B cells activated by LPS (17). However, because LPS activates B cells independently of the BCR, we analyzed mRNA levels in Wt and 3ЈE Ϫ/Ϫ B cells activated in vitro by multiple stimuli using real-time PCR. Consistent with earlier findings, there was a ϳ3-fold decrease in mRNA levels in resting 3ЈE Ϫ/Ϫ B cells when compared with that in resting Wt B cells (17) (Fig. 3A). In addition, similar levels of expression were detected between 3ЈE Ϫ/Ϫ and Wt B cells stimulated with LPS (Fig. 3A). However, when stimulated with anti-IgM and anti-CD40 Abs that mimic B cell activation during an immune response, a ϳ3-fold reduction in mRNA levels was detected in 3ЈE Ϫ/Ϫ B cells when compared with that in Wt B cells (Fig. 3A). Because the lower surface expression of IgM in 3ЈE Ϫ/Ϫ B cells might affect the signaling through the BCR, we treated B cells with PMA and ionomycin, which directly activate signaling pathways downstream of the BCR independent of BCR engagement. mRNA levels were 5-fold lower in 3ЈE Ϫ/Ϫ B cells than in Wt B cells after stimulation with PMA/ionomycin (Fig. 3A). The differences in mRNA levels also resulted in analogous differences in surface expression in the activated B cells (Fig. 3B). To further ensure that the reduced transcription in activated 3ЈE Ϫ/Ϫ B cells is not due to impaired activation, we compared the activation efficiency between Wt and 3ЈE Ϫ/Ϫ B cells under each of the stimulation conditions. The percentage of activated B cells, which express the early activation marker CD69 and are actively dividing, were similar between 3ЈE Ϫ/Ϫ and Wt B cells after various stimulations (Fig. 3B).
To examine mRNA levels in activated B cells in vivo, we sorted ϩ GC B cells from the PP of five pairs of Wt and 3ЈE Ϫ/Ϫ littermates. The mRNA levels of were ϳ2-fold reduced in 3ЈE Ϫ/Ϫ GC B cells than in Wt GC B cells (Fig. 4A). However, when the total PP B cells were analyzed by flow cytometry, surface expression in 3ЈE Ϫ/Ϫ GC B cells was only slightly lower than Wt (Fig. 4B). This is caused in part by the reduction of surface IgM in activated GC B cells (B220 ϩ PNA high ) compared with non-GC B cells (28,29) (Fig. 4C), likely as a result of a reduction in C mRNA levels in the B220 ϩ PNA high GC B cells (Fig. 4D). The lower mRNA levels are not due to a defect in B cell activation, because the proportion of activated B cells (Fig. 4E), or specifically activated ϩ B cells (Fig. 4C), is similar between the PP of 3ЈE Ϫ/Ϫ and Wt mice. When compared with that in 3ЈE Ϫ/Ϫ B cells stimulated by PMA and ionomycin in vitro, the reduction of mRNA levels in 3ЈE Ϫ/Ϫ GC B cells vs Wt controls is less dramatic, likely due to that the constant stimuli of PP B cells include LPS present on the bacterial flora.

Quantitative decrease in SHM in 3ЈE
Ϫ/Ϫ GC B cells Although 3ЈE was essential for the SHM of an Igk transgene, analysis of Ag-driven SHM of Igk in 3ЈE Ϫ/Ϫ mice indicated that 3ЈE is not required for SHM (20,21). However, it remains possible that 3ЈE could still have an influence on overall efficiency of SHM. Additionally, given the correlation between SHM and transcription efficiency, the reduction in mRNA levels we observed in 3ЈE Ϫ/Ϫ GC B cells suggests the possibility of a  defect in SHM in these cells. To address this possibility, we re-examined SHM in 3ЈE Ϫ/Ϫ mice by analyzing mutations within the J C intron in GC B cells after immunization with SRBCs. Five-month-old Wt and 3ЈE Ϫ/Ϫ littermates were immunized with 2 ϫ 10 8 SRBCs, and then given a booster injection at day 7. At day 14, spleen cells were purified, stained with an Ab for the B cell marker B220, and with PNA, which stains brightly for GC B cells, and GC B cells (B220 ϩ , PNA high ) were sorted. Genomic DNA from these GC B cells was purified, and V J 5 rearrangements were amplified by PCR, cloned, and sequenced. In Wt GC B cells, we detected an average of 3.7 mutations per 500 bp of sequence in the samples (7.3 ϫ 10 Ϫ3 mutations per base) (Table II and Fig.  5A). In contrast, only 1.3 mutations per 500 bp were detected in sequences derived from 3ЈE Ϫ/Ϫ GC B cells (2.6 ϫ 10 Ϫ3 per base), an ϳ3-fold reduction. Furthermore, the number of unmutated sequences was ϳ2.5-fold greater in 3ЈE Ϫ/Ϫ GC B cells (51%) compared with Wt (18%) (Fig. 5A). This did not fully account for the SHM defect, because the average number of mutations on sequences containing at least one mutation was still 2-fold reduced in 3ЈE Ϫ/Ϫ GC B cells, compared with Wt (Table II). The mutation spectrum was similar between Wt and 3ЈE Ϫ/Ϫ splenic GC B cells (Fig. 5B), and displayed equivalent percentages of hot-spot mutations and transitions (Table II).
To confirm the defect in Igk SHM observed in 3ЈE Ϫ/Ϫ splenic GC B cells, we also examined SHM in GC B cells from PP of 6-mo-old unimmunized mice. GC B cells (CD19 ϩ GL7 ϩ ) were purified from the PP, and the J C intronic region was sequenced (Fig. 5C). In Wt GC B cells, the mutation frequency of the analyzed region was 16 ϫ 10 Ϫ3 (Table II)

. However, in 3ЈE
Ϫ/Ϫ GC B cells, the mutation frequency was reduced 2.5-fold. The same proportion of sequences contained mutations in both groups, but the sequences from 3ЈE Ϫ/Ϫ GC B cells had a lower mutation load than the Wt. Wt and 3ЈE Ϫ/Ϫ B cells show a similar spectrum of mutations (data not shown). Additionally, the mutations analyzed from both groups are targeted to RGYW/WRCY hotspots and transition biased (Table II). The variation in the mutational frequency Ϫ/Ϫ B cells vs those in stimulated Wt B cells. ϩ B cells were purified by MACS and activated by a variety of stimuli. Untreated (UT) B cells were harvested immediately after sorting, and stimulated B cells were harvested 2 days after stimulation. Quantitative real-time PCR using primers specific for the and C regions was used to analyze the gene expression of and , respectively. All samples were normalized to C expression. B, surface expression profile in stimulated Wt (solid gray) and 3ЈE Ϫ/Ϫ (dotted black) B cells. Only cells within the gates displayed in Fig. 3C are shown. C, Analysis of the activation efficiency in 3ЈE Ϫ/Ϫ and Wt B cells by flow cytometry. Samples from each type of stimulation were stained with anti-B220, -CD69, and -Abs. CD69 expression (y-axis) and cell size (x-axis) of live, B220 ϩ lymphocytes are shown. The percentage of CD69 ϩ , blasting Wt (left column) and 3ЈE Ϫ/Ϫ (right column) B cells are listed inside the corresponding gate. IONO, Ionomycin.

FIGURE 4. expression in Wt and 3ЈE
Ϫ/Ϫ PP B cells. A, Analysis of mRNA levels in ϩ Wt (Ⅺ) and 3ЈE Ϫ/Ϫ (f) PP GC (GL7 ϩ ) and non-GC (GL7 Ϫ ) B cells. mRNA levels were normalized to CD19 mRNA levels. mRNA levels of all samples are shown relative to Wt non-GC sample. B, surface expression profile of non-GC (left) and GC (right) B cells of Wt (solid gray) and 3ЈE Ϫ/Ϫ (dotted black) PP. C, Comparison of surface expression in PP B cells in 3ЈE Ϫ/Ϫ and Wt mice. Total B220 ϩ PP cells were plotted for surface expression and intensity of PNA staining. D, Analysis of C mRNA levels in B220 ϩ PNA high GC B cells. Samples were normalized to CD19 mRNA levels and are shown relative to Wt non-GC. E, Proportions of non-GC (B220 ϩ PNA low ) and GC (B220 ϩ PNA high ) in Wt (left) and 3ЈE Ϫ/Ϫ (right) PP. Gates and percentages of GC and non-GC B cells are shown.
in the Wt GC B cells between different experiments was likely due to the different ages of mice analyzed, because older mice accumulate more mutations in Igk of GC B cells.
Although the sorted GC cells in the previous two experiments included ϩ B cells, which are a significant population in 3ЈE Ϫ/Ϫ mice, we would not expect that these ϩ B cells contribute to the SHM data, because the rearranged Igk loci are mostly deleted through RS rearrangement in 3ЈE Ϫ/Ϫ ϩ B cells (17). However, to rule out the possibility that a larger proportion of ϩ B cells in sorted GC B cells could lead to skewed data, we repeated the analysis of sorted ϩ GC cells ( ϩ GL7 ϩ ) from the PP of 6-mo-old Wt and 3ЈE Ϫ/Ϫ littermates. In Wt and 3ЈE Ϫ/Ϫ GC ϩ B cells, the mutation frequencies of the analyzed region were 22 ϫ 10 Ϫ3 and 10 ϫ 10 Ϫ3 , respectively, a 2.2-fold difference (Fig. 5D, Table II). Therefore, 3ЈE is quantitatively important for SHM.

Discussion
Recruitment of the SHM machinery specifically to the rearranged Ig loci is critical for Ab maturation as well as the maintenance of genetic stability via the prevention of hypermutation in other parts of the genome. cis-elements within the Ig loci have been thought to play critical roles in this recruitment process. Transgenic studies implicated essential roles for both iE and 3ЈE in driving SHM (20). In this study, we report the first analysis of SHM of the endogenous Igk allele in the absence of iE . In contrast with the findings of the transgenic studies (20), we have observed no apparent defect in SHM in iE Ϫ/Ϫ mice. Given the very small size of transgenes compared with the endogenous 3.2-Mb Igk locus, the discrepancy is likely due to the presence of other cis-elements that play redundant roles in activating SHM. One potential candidate is the recently discovered downstream enhancer (Ed), which was not contained in the transgenes (15).
Previous studies of the functions of 3ЈE in SHM have reached conflicting conclusions regarding its importance. Transgenic studies indicated that 3ЈE is critical for the expression and SHM of Igk in randomly integrated transgenes in the presence as well as absence of stringent Ag selection (20,30). However, analyses of 3ЈE Ϫ/Ϫ mice showed that 3ЈE is not important for transcription following LPS-induced activation (17) and also that 3ЈE is not required for Ag-driven SHM (21). However, these authors noted a potentially reduced mutation frequency and proposed that 3ЈE might be involved in promoting the optimal efficiency of SHM. By analyzing SHM in the J C intron following immunization with the polyclonal response-inducing Ag SRBC, or in unimmunized PP, we confirm that the 3ЈE is indeed not required for SHM. However, our findings clearly demonstrate that 3ЈE is quantitatively important for SHM.
Because of the correlation between transcription and SHM, we analyzed mRNA levels in 3ЈE Ϫ/Ϫ B cells activated by a number of stimuli. Consistent with earlier observations (17), normal mRNA levels are observed in 3ЈE Ϫ/Ϫ B cells activated by LPS. However, expression appears to be impaired in 3ЈE Ϫ/Ϫ B cells Ϫ/Ϫ mice. when signaling pathways downstream of the BCR and CD40 are activated. We therefore conclude that 3ЈE , similar to its function in developing and mature B cells, is an important mediator of transcription in activated B cells. The surface expression of Wt and 3ЈE Ϫ/Ϫ GC B cells is similar, possibly due to reduced IgH transcription in GC B cells. Therefore, the modest defect in transcription, but not surface expression, observed in 3ЈE Ϫ/Ϫ GC B cells could explain the reduced mutation of Igk in 3ЈE Ϫ/Ϫ mice. Our findings that neither 3ЈE nor iE is critical for SHM suggest that these two enhancers might play redundant roles in activating SHM. A similar scenario has been observed in the regulation of Igk rearrangement by the two enhancers (18). Furthermore, mice bearing a deletion of the core region of the IgH E enhancer show no defect in SHM of IgH (31). E can drive a low level of SHM in transgenic constructs (32-34), but not others (35,36), implicating a role for multiple elements in the regulation of SHM at the IgH locus.
Transcription factors binding to the enhancer elements might be involved in SHM recruitment. In this context, two accidentally introduced E2A binding sites, in addition to those already present in iE and 3ЈE , were found to increase SHM in a transgene with no apparent impact on the transcription of the transgene (25). As both iE and 3ЈE contain E2A binding sites, these enhancers might activate SHM via the recruitment of E2A. iE 's two E2A binding sites have been shown to be important for Igk rearrangement (26). Therefore, similar to their redundant roles in the regulation of rearrangement, the intronic and 3Ј enhancers might play redundant roles in activating SHM (18). Targeted replacement of J with a rearranged V J driven by a constitutive promoter in the context of an enhancerless Igk allele might reveal the redundant roles of the two Igk enhancers in targeting SHM to Igk loci.