GANP regulates recruitment of AID to immunoglobulin variable regions by modulating transcription and nucleosome occupancy

Abstract

Somatic hypermutation in B cells is initiated by activation-induced cytidine deaminase-catalyzed C→U deamination at immunoglobulin variable regions. Here we investigate the role of the germinal centre-associated nuclear protein (GANP) in enhancing the access of activation-induced cytidine deaminase (AID) to immunoglobulin variable regions. We show that the nuclear export factor GANP is involved in chromatin modification at rearranged immunoglobulin variable loci, and its activity requires a histone acetyltransferase domain. GANP interacts with the transcription stalling protein Spt5 and facilitates RNA Pol-II recruitment to immunoglobulin variable regions. Germinal centre B cells from ganp-transgenic mice showed a higher AID occupancy at the immunoglobulin variable region, whereas B cells from conditional ganp-knockout mice exhibit a lower AID accessibility. These findings suggest that GANP-mediated chromatin modification promotes transcription complex recruitment and positioning at immunoglobulin variable loci to favour AID targeting.

Figure 1. GANP interacts with histones and associates with chromatin.

(a) GANP-associated proteins in Ramos cells. 2DICAL analysis was used to identify about 100 proteins co-immunoprecipitated with GANP, but not with IgG control. The proteins are classified according to their biological functions. (b) Schematic representation of chromatin fractionation using MNase. (c) Western blot (WB) analysis of GANP in chromatin fractions prepared from Ramos cells. Whole-cell lysate (WCL) is shown for control. (d) WB analysis of 293 T cells transfected with a GFP-GANP expression vector. Soluble (Sol) and insoluble (Insol) chromatin fractions were isolated by digesting native chromatin with MNase. Each fraction was blotted with anti-GFP Ab. (e) 293 T cells were transfected with Control or GANP siRNA and harvested after 3 days. The insoluble chromatin fraction was prepared and blotted with anti-GANP (upper) and anti-RNA Pol-II (lower) antibodies. Data in (c–e) are from one of three sets of independent experiments with similar results.

Figure 2. The histone acetylation activity of GANP modulates chromatin assembly and nucleosome organization.

(a) The HAT domain of GANP (HAT_G) possesses histone acetylation activity. Recombinant GST-HAT_G or GST-alone protein (100 ng) was incubated with histone H3 or H1 (1 μg) in the presence of acetyl-coenzyme A (10 μM). Acetylated histones H1 and H3 were analysed by WB. (b) MNase sensitivity assay for Ramos cell transfectants. Expression of GFP-alone, GFP-GANP and GFP-ΔHAT_G constructs were confirmed by blotting with the whole-cell lysate (WCL; input). Released mono- and poly- nucleosomes following MNase treatment were separated by agarose gel electrophoresis and detected by ethidium bromide (Et-Br). GANP presence in the chromatin fraction was detected using anti-GFP Ab. Histone acetylation was analysed by WB using anti-AcK, anti-H3 and anti-H3K9ac antibodies. Gel pictures are representatives from one of three sets of independent experiments with similar results.

Figure 3. GANP effects on nucleosome occupancy and histone modification at the rearranged Ig VH4(DP63)-JH4 locus.

(a) Fifteen primer sets were designed to assay MNase resistance across the entire IgV-region (upper panel). Each bar shows the position of amplified region (100 bp). MNase resistance profile of VH4(DP63)-JH4 locus for GFP-GANP Ramos transfectants (middle panel) and GFP-ΔHAT_G transfectants (lower panel). Values are the normalized amounts of the MNase-digested products to those of undigested samples. (b) Effect of ActD on MNase resistance profile VH4(DP63)-JH4 IgV-locus. Ramos cells were treated for 2 h with ActD (100 ng) followed by washing with PBS and incubation in the complete culture medium for 24 h. (c) GANP association to regions q1–q5 of the rearranged VH4(DP63)-JH6 locus and its effect on histone modifications. Scale is 100 bp. (d) ChIP-qPCR analysis of GFP-positive cells transfected with either GFP-alone (white columns), GFP-GANP (orange columns), or GFP-ΔHAT_G (green columns) using anti-GFP, anti-AcK, anti-H3K9ac and anti-H3K27ac antibodies. IgG was used as a negative control. Data represent mean±s.d. calculated from three (a, d) or two (b) independent experiments. *P<0.05 (two-tailed unpaired t-test).

Figure 4. GANP enhances recruitment of the transcription complex, transcription stalling factor Spt5 and DSIF (Spt4–Spt5 complex) to the IgV-locus.

(a) Association of RNA Pol-II and Spt5 with GANP. Lysates from Ramos cells were IP using anti-GANP Ab. WB was carried out with anti-RNA Pol-II pSer2/5, anti-Spt5 or anti-GANP Ab. IgG was used as a negative control. Whole-cell lysate (WCL) is shown in the left. (b) ChIP-qPCR analysis of Ramos GFP-positive cells transfected with either GFP-alone (white columns), GFP-GANP (orange columns) or GFP-ΔHAT_G (green columns) for RNA Pol-II CTD phosphorylation at Ser5 and Ser2. (c) ChIP-qPCR assay for endogenous GANP, histone H1, H1K63ac, Spt5 and CTCF. Two different GANP antibodies, GANP (N1) and GANP (C1), anti-H1K63ac and anti-Spt5 were used to show preferred recruitment of GANP, H1K63ac and Spt5 to the q5 region. CTCF was found to associate preferentially at the q4 region. (d) ChIP-qPCR analysis of DSIF (Spt4 and Spt5). Data represent mean±s.d. calculated from three (b–d) or two (a) independent experiments. *P<0.05 (two-tailed unpaired t-test).

Figure 5. Effects of downregulation of endogenous GANP on nucleosome occupancy and chromatin modification at Ig VH4(DP63)-JH4 locus.

(a) Knockdown of GANP in Ramos B cells by fluorescent Cy3-labelled siRNA. GANP and RNA Pol-II expression in Cy3-positive sorted cells with control and GANP-targeted siRNA were examined by WB. (b) Analysis of nucleosome occupancy at the IgV-locus in control and GANP-knockdown Ramos B cells by the MNase resistance assay. (c) ChIP-qPCR analyses for pSer2, pSer5, Spt5 and modified histones H1 and H3 at q1 to q5 regions of the IgV-locus in GANP knockdown B cells. Data represent mean±s.d. calculated from two independent experiments. *P<0.05 (two-tailed unpaired t-test).

Figure 6. AID targeting and SHM mutation distribution at the rearranged IgV-locus.

(a) Mutation spectra of AID-bound IgV-region, H3K9ac-bound IgV-region and CTCF-bound IgV-region in GFP-alone and GFP-GANP Ramos cell transfectants. DNAs isolated from anti-AID, anti-H3K9Ac and anti-CTCF Ab IP samples were subjected to sequencing analysis. SHM mutation profiles of IgV DNA starting from the primer 5 (DP63-05S) were shown. (b) Pie charts showing the distribution of numbers of mutations per a sequenced clone for AID-bound IgV-region from GFP-GANP and GFP-alone transfectants. Data are from one of three independent experiments with similar results.

Figure 7. GANP enhances the accessibility of AID to IgV_H-region DNA in mouse GC B cells.

(a) RT–qPCR analysis of Ganp, Aicda and IgV_H transcripts in GC B cells from WT, Ganp^Tg, Ganp^F/F and CD19-Cre/Ganp^F/F (n=3 mice per genotype). (b) ChIP analysis of GC B cells of WT and Ganp^Tg. Input and ChIP samples were serially diluted 1:3 (wedge) and various genomic DNA sequences (left margin) were amplified with specific primers and were detected by ethidium bromide (Et-Br) staining. Anti-histone H3 (H3) Ab was used as a positive control. (c) ChIP analysis in GC B cells of the control Ganp^F/F and Ganp-deficient CD19-Cre/Ganp^F/F, as in b. (d) ChIP-qPCR analysis of GC B cells from WT, Ganp^Tg, Ganp^F/F and CD19-Cre/Ganp^F/F. ChIP was amplified by qPCR with specific primers. Data (a, d) represent mean±s.d. calculated from three independent experiments *P<0.05 and **P<0.01 (two-tailed unpaired t-test).

Figure 8. A model of GANP-mediated AID targeting for SHM at the IgV-loci.

GANP associates with histones H1 and H3, the IgV transcription–elongation complex, and the transcription stall factor Spt5 in Ramos B cells. The rearranged IgV-locus supports two stably assembled nucleosomes at site A and site B. GANP is involved in chromatin remodelling at the IgV-locus through its HAT domain and causes destabilization of nucleosome assembly at site B (1). Chromatin remodelled IgV provides increased access for both pSer5 and pSer2 forms of RNA Pol-II leading to an increase in IgV transcription (2). GANP facilitates early recruitment of the DSIF transcription stalling complex (Spt4–Spt5) to the IgV-coding region, causing frequent stalling of transcription bubbles, which provide ssDNA substrates for AID-targeting (3).