The B cell plays an important role in the humoral adaptive immune response, producing and secreting high affinity antibodies able to cause the killing of extracellular microbes by neutralising pathogens or toxins and opsonising pathogens to enhance phagocytosis or activate complement. Therefore, functions to prevent spread of intracellular pathogens (Janeways et al 2005). Hence, to avoid recognition pathogens continually alter the pattern of their surface coat proteins. (Delker et al 2009). Therefore, B cells have developed methods of generating an antibody repertoire able to respond to foreign antigen.
The first immunoglobulins expressed by mature B cells following development in the bone marrow is IgM and IgD. However, the B cell has the ability to produce antibodies of seven other isotypes (IgA1, IgA2, IgE, IgG1, IgG2, IgG3 and IgG4) with varying effector functions. All isotypes have similar structural features consisting of two heavy and two light chains, each comprising of a variable and constant region. The N terminal variable (V) region of each chain forms the antigen binding site and thus determines antigen specificity, whereas the C terminal heavy chain constant (C) region determines antibody effector function and varies between isotypes (Janeways et al 2005). Therefore, in order to gain a fully functional humoral immune response other secondary immunoglobulin classes need to be expressed. This is achieved by a process called class switch recombination (CSR).
Figure 1: Structure of a typical immunoglobulin. The N terminal variable region determines antigen binding specificity while the C terminal constant region determines antibody effector functions. Although image represents the typical antibody structure slight differences exist between isotypes, for instance IgE has an extra constant domain in its constant region. Janeways (2005) Immunobiology, Garland Science, New York
CSR is an antigen dependent deletional recombination event which occurs in germinal centres (but not exclusively) (Toellner et al 1996), formed when proliferating B cells enter primary follicles of peripheral lymphoid organs following antigen and T cell stimulation (Janeways et al 2005). The process operates by swapping the originally expressed heavy chain constant domain (CH ) exons Cµ and Cδ (which specify IgM and IgD, respectively) for one of the other downstream exons (5'-Cγ3, Cγ1, Cα1, Cγ2, Cγ4, Cε, C α2 -3') (Chaudhuri and Alt 2004). The selected CH exon is then joined to the previously rearranged V region exon produced during V(D)J recombination, therefore allowing the production of antibodies of varying isotypes that still maintain the same antigen specificity of the parental IgM antibody (Chaudhuri and Alt 2004).
The process of CSR requires key intronic DNA sequences called switch (S) regions located about 1-5kb upstream of each CH exon apart from that of Cδ (IgD) (Selsing 2004). S regions contain extensively repetitive GC rich sequences ranging between 1-12kb in length (Chaudhuri and Alt 2004), with very G rich non-template strands. CSR involves a recombination event between two S regions where intervening DNA including the originally expressed Cµ and Cδ exons are excised as a loop (REF)(FIGURE). S regions are important to CSR, as in vivo deletion of Sγ1 almost completely abolished CSR to IgG1 (0.4% of wild type) (Shinkura et al 2003).
Figure 2: Schematic of IgH locus prior to CSR, Selsing E (2006) Ig class switching: targeting the recombinational mechanism, Current Opinion in immunology 18, 1-6
CSR involves a recombination
event between two S regions, with the intervening
sequences, including Cμ, being deleted. As a result,
the VDJ exon is juxtaposed with a different downstream
CH gene5,6
Prior to CSR, transcription through the donor and recipient S regions, ending at the downstream associated CH exon is necessary. This germline transcription (GLT) is initiated from an inducible promoter (I-promoter) located upstream of the I-exon found immediately upstream of each S region (Chaudhuri and Alt 2004) (FIGURE). Resulting RNA is spliced to generate a non-coding sterile transcripts. The CH exon now expressed is determined by the binding of specific cytokines to the I-promoter, for example IL-4 stimulates the production of ε and γ transcripts and hence promotes class switching to IgE and IgG respectively (Nambu et al 2003).
in which transcription initiates from cytokine-inducible I-promoters (P) upstream
of the I-exon (I) and proceeds through S regions and the CH exons.
Figure 3: Mechanism of class switch recombination. The figure displays the mechanism by which CSR occurs from IgM to IgA. In the figure GLTs of the µ, δ and α are produced allowing access of the DNA to AID. Double stranded breaks are produced in the µ and α switch regions and the two S regions are joined while the intervening DNA is removed as a loop.
CSR is initiated by an enzyme called activation induced cytidine deaminase (AID) which was identified 11 years ago as a B cell specific gene upregulated following CSR stimulation during a mouse B cell lymphoma cell line PCR subtractive screen (Muramatsu et al 1999). Its requirement in CSR is clearly noted from experiments in murine AID-/- splenic B cells where its deficiency prevented CSR, even though antigen and cytokine stimulation was received (Muramatsu et al 2000). Furthermore, AID gene mutations result in human hyper-IgM type II (as unable switch from IgM expression) (Revy et al 2000) and its artificial expression using an artificial construct could initiate CSR in mouse fibroblast (Okazaki et al 2002). AID is also required for the initiation of another antibody diversification event called somatic hypermutation where antibody binding affinity is increased by the introduction of point mutations into V regions of the immunoglobulin gene loci (Chaudhuri and Alt 2004).
AID is a DNA editing deaminase which initiates CSR by converting deoxycytidine (dC) in S regions to deoxyuridine (dU) residues (Chaudhuri and Alt 2004). During CSR AID is thought to deaminate ssDNA substrates exposed via R loop formation following transcription. R loops form during transcription when nascent RNA forms a DNA-RNA hybrid with the template strand, leaving the ssDNA G rich non-template strand to become a substrate for AID. R loop formation at S regions has been observed in vitro and in vivo (Shinkura et al 2003, Yu et al 2003).
AID's action results in dU residues being abnormally present in DNA and U-G mismatch are subsequently processed into double stranded DNA break (DSB) intermediates by DNA repair pathways. The base excision pathway (BER) causes the removal of S region dU residues by the action of the enzyme uracil-DNA glycosylase (UNG), forming abasic sites (Chaudhuri and Alt 2004). UNG is an important class switching component as UNG-/- mice retain only 5% of the CSR ability of wild type mice (Stavnezer et al 2008). Newly created abasic sites are subsequently nicked by the enzyme apurinic/apyrimidinic endonuclease 1 (APE-1) leading to the formation of staggered ssDNA break (SSB) (Stavnezer et al 2008). Although, R loop formation suggests the non-template strand will be a preferential target for AID's action, several in vivo studies have shown that both the template and non-template strand are equally targeted, even at sites of R loop formation (Xue et al 2006). Thus close SSBs on neighbouring DNA strands are thought to generate staggered double stranded DNA breaks (DSBs). Faraway staggered SSBs are suggested to be processed into DSBs by mismatch repair (Stavnezer et al 2008). DSBs are generated at donor and recipient S region allow intervening DNA to be removed and S regions to be joined by the cell's non-homologous end joining machinery (Chaudhuri and Alt 2004).
Figure 4: Mechanism of dsDNA break formation by the BER pathway. Adapted from Stavnezer et al (2008) Mechanism and Regulation of Class Switch Recombination, Annual Review of immunology 26, 261-292
However, although much of the mechanism of CSR has been defined, its regulation is much less understood. One of the most intriguing questions yet to be answered is how AID is specifically recruited and targeted. It's known that AID mutations are limited to the Ig locus and only a few non-Ig genes at significantly lower rates, such as Bcl-6 and some proto-oncogenes in B cell lymphomas that express AID (Liu et al 2008, Nagaoka et al 2010). However, even at the IgH locus, AID is specifically targeted to S regions and excluded from CH exons and non-CSR required regions (Selsing 2006). Although physiological specificity exists, AID is clearly a very dangerous enzyme able to deaminate cytosines at many genomic sites (Liu et al 2008) and in vitro studies have shown the enzyme's ability to mutate any ssDNA substrate (Delker et al 2009). Furthermore, inappropriate AID targeting can lead to chromosomal translocations, such as IgH-Myc generating potential tumorigenesis when AID mediated DSBs are incorrectly repaired and fuse with abnormal genomic locations (Casellas et al 2009). Hence, it is clear accurate AID targeting is essential and important in limiting its genotoxic potential. However, no defined AID recognition sequence has been found and the S regions have no consensus sequences (Nagaoka et al 2010), so how is AID targeted?
Several suggestions have been made in terms of how AID may be targeted. Firstly, AID is known to deaminate dC residues in S regions most frequently within WRC (W=A/T, R=purine, C= cytidine, Y=C/T) motifs (Xue et al 2006). This is consistent with findings in Xenopus S region where DSBs were most regularly observed at sites where the motif was abundant (Chaudhuri and Alt 2004). This suggests that these sequences may have some sort of role in targeting AID, as it is also a SHM hotspot (Chaudhuri and Alt 2004). However, this motif is not S region specific and is found on many actively transcribed genes which are not AID targets. Therefore, a higher level of targeting must exist to determine which of these AID targeting motifs are accessible.
Another suggested AID recruitment mechanism is germ line transcription. In CSR stimulated splenic B cells, AID was found co-immunoprecipitated with RNA polymerase II, demonstrating that the two proteins interact (Nambu et al 2003). Furthermore, experiments at donor and recipient S regions in DNA repair deficient UNG-/- MHS2-/- mice showed AID mutations begin about 150 nucleotides downstream of the I-exons, continue through the S region, and decline at the 3' end as the associated CH exons is approached (Xue et al 2006). This indicates that AID mutations correlate with the start of germline transcription and supports a series of previous studies, including one where upstream promoter movement caused an equivalent upstream movement of AID mediated mutations (Delker et al 2009). Hence, it has been hypothesised that AID may actually be recruited to the transcriptional initiation complex (TIC) via binding to RNA polymerase and hence travel along the S region introducing mutations while transcription occurs (Nambu et al 2003, Stavnezer et al 2008). Thus the decline in AID mutations would result from the random unloading of AID from the transcription complex (Delker et al 2009), while the upstream promoter the TIC assembles at will be determined by specific cytokine signalling (Xue et al 2006). However, one caveat is that transcription alone would not be specific enough to recruit AID specifically to the S regions, so additional targeting methods must exist. The ssDNA binding protein replication protein A (RPA) has been implicated in maintaining ssDNA accessibility with the help of PKA phosphorylation (Rada 2009).
Figure 4: Proposed mechanism of transcription coupled mutagenesis. AID may be loaded onto the transcription elongation complex along with RNA polymerase II introducing mutations in the S regions as transcription occurs. Adapted from: Chaudhuri J & Alt F. W (2004) Class switch recombination: interplay of recombination, DNA deamination and DNA repair, Nature Reviews Immunology 4, 541-552
Chromatin histone regulation is another hypothesis proposed to recruit AID to S regions during CSR. Modification involves the covalent addition/removal of certain chemical groups to the N terminal tails of the four core histone proteins (H2A, H2B, H3, H4). These modifications are involved in maintaining eurochromatin and heterochromatin structure and regulating gene accessibility by weakening DNA:histone protein interactions or recruiting non-histone proteins with other regulatory functions (Kouzarides 2007). Histone modifications have been implicated in the regulation of other recombination processes, for example methylation of histone 3 at lysine 4 (H3K4me3) has been proposed to recruit RAG2 to the recognition signal sequences (RSS) during V(D)J recombination (Shimazaki et al 2009). Hence, it is possible that AID may be recruited to S regions by a similar mechanism in CSR.
A number of studies have noted H3K4me3 and H3K9me3 mark enrichment over S regions following CSR stimulation. The H3K4me3 is generally associated with active genes and its presence is usually used to distinguish between active and inactive genes (Schneider et al 2004). In tonsilar B cells prior to CSR stimulation, H3K4me3 was found over regulatory regions such as the 5' µ and 3' Cα enhancers and the I-promoters but not the intronic S regions. However, on stimulation the H3K4me3 was subsequently found over recipient S regions stimulated to undergo CSR (Chowdhury et al 2008). Similar results were seen in a mouse splenic B cell study, which further demonstrated that the donor Sµ regions maintained a high constitutive level of H3K4me3 which is consistent with its constitutive expression (Wang et al 2009). The pattern of H3K4me3 seems to correlate with histone acetylation, indicating that regulatory regions are accessible prior to CSR and recipient S regions only following stimulation (Wang et al 2009). As the H3K4me3 has been demonstrated in yeast to associate with a component of the NuA3 HAT complex (Wang et al 2009), this mark may play a role in regulating chromatin accessibility along with histone acetylation and may be required for AID recruitment.
In contrast to H3K4me3, H3K9me3 has historically been known as a repressive mark, functioning in constitutive and falculative heterochromatin formation, and active eurochromatin repression by recruiting members of the chromo-domain containing protein family, heterochromatin Protein 1 (HP1) (Kouzarides 2007). However, recently the association of this mark as being solely repressive has been challenged. In 2005 Vakoc et al found this modification over a number of actively transcribed genes, noting the H3K9me3 to increase and decrease over genes on induction of gene activation and repression respectively (Vakoc et al 2005). Additional studies supported this conclusion including a human cancer cell line study where some differentially expressed genes were expressed higher when the mark was present compared to its absence, once again indicating an activating role (Wiencke et al 2008).
Recently, the levels of the H3K9me3 mark were found to increase considerably at the Sγ and Sε regions and their corresponding CH exons following CSR stimulation to IgG and IgE (Chowdhury et al 2008). Another study found that the level of this modification at the recipient S region increased 24-48hrs after stimulation to class switch and was still retained 96hrs after stimulation (Kuang et al 2009). These findings aligned with a previous study showing that the H3K9 methyltransferase Suv39h1 stimulated CSR specifically to IgA, by possibly utilising its histone methyltransferase activity to inhibit the actions of a repressor specific to IgA class switching (Bradley et al 2006). Furthermore, Kuang et al (2009) demonstrated that AID-/- and wild type mice had the same H3K9me3 pattern indicating that this methylation occurs prior to AID cytidine deamination function and recombination (Kuang et al 2009). Hence, the mark may be laid down during CSR and assist AID recruitment.
MAKE AN IMAGE UP OF THE PROSED HISTONE MODEL!!
The fact that both H3K4me3 and H3K9me3 have been shown to increase over S regions, suggests they may play a role in recruiting AID. It is possible that the H3K4 and 9 methyltranseferase may exist in a complex with RNA polymerase II allowing the addition of these methyl groups to histone lysine residues, as the polymerase moves across the S region. Hence, the combination of the H3K9me3 and H3K4me3 marks over the S regions may create a unique positive sign on the genome that recruits AID. AID may possibly be found in the same complex with polymerase and subsequently recruited to sites where the marks are present and ssDNA exposed due to GLT.
Hence, to investigate the molecular mechanisms involved in recruiting and targeting AID we explored sites of the B cell genome where AID had been identified, by a previously performed ChIP-seq analysis using two algorithms: MACS and SICER. A cross linked ChIP (X-ChIP) assay using anti-AID antibody and real time PCR (qPCR) analysis was carried out to verify whether or not AID was actually bound at these identified sites. In addition, possible mechanisms used to target AID to these genomic sites were also investigated by performing X-ChIP against H3K4me3, H3K9me3 and RNA Polymerase II. Therefore, allowing the identification of whether the RNA polymerase II and histone modifications are always present at sites of AID action. All site analysis was performed on two isolated tonsilar B cell populations: resting B cells and those stimulated to undergo CSR by 48hr IL-4 and anti-CD40 antibody treatment.
