NF-kappa B binds to the immunoglobulin S gamma 3 region in vivo during class switch recombination

Abstract

Ig class switch recombination (CSR) is dependent upon the expression of activation-induced deaminase and targeted to specific isotypes by germ-line transcript expression and isotype-specific factors. NF-kappaB plays critical roles in multiple aspects of B cell biology and has been implicated in the mechanism of CSR by in vitro binding assays and altered S/S junctions derived from NF-kappaB p50-deficient mice. However, the pleiotropic contributions of NF-kappaB to gene expression in B cells has made discerning a direct role for NF-kappaB in CSR difficult. We now observe that binding of NF-kappaB components p50 and p65 is detected on Sgamma3 in vivo following lipopolysaccharide (LPS) activation and repressed by LPS + IL-4, suggesting a direct role for this factor in CSR. In vivo footprinting confirms occupancy of a previously defined NF-kappaB recognition site in Sgamma3 with the same temporal kinetics as found in the chromatin immunoprecipitation analysis. Binding of NF-kappaB components p50 and p65 was also detected on Sgamma1 following B cell activation. H3 histone hyper acetylation at Sgamma1 is strongly correlated with NF-kappaB binding, suggesting that NF-kappaB mediates chromatin remodeling in the Sgamma3 and Sgamma1 region.

Figure 1

NF-κB p50 homo- and heterodimers are expressed in the nuclei of resting and LPS-stimulated splenic B cells. (A) Gelshift assays were performed with nuclear extracts from splenic B cells (3.5 µg) and 70Z/3 cells (5 µg) stimulated with LPS for 44 and 72 h, respectively. Nuclear extracts were incubated with the B2 oligo and the κB oligo probes, as indicated. (B) Nuclear extracts derived from unstimulated splenic B cells or those activated with LPS for 21, 44, 68 and 92 h were analyzed in gelshift assays using the B2 oligo or the κB oligo probes, as indicated.

Figure 2

ChIP analysis indicates that NF-κB p50 and p65 associate with Sγ3 and Sγ1 regions. GLT (A) and post-switch transcripts (B) were analyzed by semi-quantitative RT-PCR using cDNA derived from splenic B cells that were unstimulated (Un) or activated with LPS or LPS + IL–4 for 48 h or 5 days, respectively. GLT and Aicda (A) and post-switch (B) PCR products were harvested after 33 cycles (lanes 1, 4, 7), 31 cycles (lanes 2, 5, 8), and 29 cycles (lanes 3, 6, 9). Gapdh PCR products were harvested after 30 cycles (lanes 1, 4, 7), 28 cycles (lanes 2, 5, 8), 26 cycles (lanes 3, 6, 9). (C) Schematic diagrams (not to scale) showing the γ3 and γ1 loci in the I exons, and S regions are shown. The positions of primer sets are indicated by the arrows. Open circles represent NF-κB sites. (D–F) ChIP analyses were carried out on splenic B cells that were either unstimulated (Un), or activated with LPS or LPS + IL–4 for 48 h by quantitative real time PCR and the relative enrichment was normalized to the input DNA and the nonspecific control antibody ChIP. (D) DNA were analyzed by resolving radioactively labeled PCR products by PAGE and representative gels are shown. ChIP samples were immunoprecipitated by αp50, αp65, or Ig control (Ig) and input control (In) are shown. PCR products were harvested after 35 cycles. (E) ChIP samples were analyzed by quantitative real-time PCR in duplicate and averaged. The results from five samples derived from two independent experiments are shown and the error bars represent the SEM for these samples. All ChIP PCR values were normalized against input DNA in the same sample. Pair-wise comparisons of ChIP results were analyzed using the Student's t-test and the p values were calculated. The p values indicated by one or two stars (*) were p ≤0.05 and p ≤0.005, respectively, whereas a bracket without a star indicates p = 0.07. (F) ChIP analyses using the αp50 antisera were carried out on splenic B cells that were activated with LPS for 0, 22 and 68 h. The error bars represent the SEM for these samples.

Figure 3

LM-PCR analysis of Sγ3 DNA in unstimulated and LPS activated splenic B cells indicates a dynamic occupancy of SNIP- and SNAP-binding sites. (A) A partial restriction map of the Sγ3 region of the IgH locus [47] is shown. The relative positions of the primers used for first strand synthesis, amplification, and labeling are indicated flanking the repetitive switch sequence. Restriction sites are abbreviated as B, BglII; H, HindIII; K, KpnI; and S, SacI. B cells were treated DMS for 1 min at 37°C. The methylated DNA were cleaved and the fragments were amplified and labeled according to the LM-PCR protocol [48]. The amplified fragments were resolved on a 4% denaturing gel and analyzed with a phosphorimager using ImageQuant software (Molecular Dynamics). (B, C) LM-PCR was performed on the non-coding (B) and coding (C) strands of Sγ3 DNA prepared from BALB/c nu/nu splenocytes that were unstimulated or LPS activated for 22 and 68 h. DNA methylated in vitro serves as a reference for methylation in the absence of bound protein. Numbers indicate the positions relative to the 3′ end of the labeling primers, where position 1 for the coding strand is equivalent to nucleotide 2574 and position 1 for the non-coding strand is equivalent to nucleotide 505 of the germ-line Sγ3 sequence MUSIGHANA. Boxes indicate positions of the SNIP- and SNAP-binding motifs. Residues that are strongly (>50%; ●) or moderately (30–50%; ○) protected from methylation or are hypermethylated (∗) in vivo as determined by densitometry are shown. Quantitation of band intensities was performed using ImageQuant software (Molecular Dynamics). Band intensities associated with the coding strand residues were normalized to the average intensity of residues 229–227. (D) Histograms comparing the coding strand signal intensities of ratios from in vivo and in vitro methylated DNA derived from unstimulated (upper panel) and LPS-activated 22 h (middle panel) and 68 h (lower panel) stimulated cells are shown. The light gray histogram bars indicate residues associated with the SNIP and SNAP motifs and residue numbers are indicated. (E) A summary of the methylation patterns for each time point are shown above the sequence of Sγ3 repeats 38 and 39.

Figure 4

Histone H3 hyper Ac is induced by LPS in the area spanning the Iγ1–Cγ1 locus. Anti-H3Ac histone antisera were used in ChIP assays carried out on nuclei derived from splenic B cells. Cells were activated with LPS or with LPS + IL–4 for 48 h or were unstimulated. A representative experiment is shown using three independent ChIP samples. All quantitative PCR values were normalized against input DNA to obtain the Ac histone index. Averaged histone Ac indices and SD downstream of the Sμ region (Sμ-D) and Sγ1 (Sγ1-D) are shown.