Multiple cis-acting elements including the intronic enhancer and the 3′α enhancer (3′αE) regulate expression of the Ig heavy chain genes during B cell development. A 3′αE is composed of DNase I-hypersensitive sites, hs1,2, hs3a,b, and hs4, found 3′ of the murine Cα gene as well as 3′ of both human Cα genes, Cα1 and Cα2. Rabbits have 13 Cα genes, and we tested whether a 3′αE is associated with each of these genes. To identify 3′αE regions we developed a rabbit hs1,2 probe and used this to search for enhancer homologues of human hs1,2 in a genomic fosmid library. We identified a single hs1,2 fragment 8-kb downstream of Cα13, the presumed 3′-most Cα gene. We also identified and partially sequenced a new Cα gene, Cα14, located 6 kb upstream of Cα13. Genomic Southern blot analysis confirmed that the rabbit genome contains only one hs1,2 enhancer region. We tested the enhancer activity of the hs1,2 with the SV40, VH, and Iα promoters using the luciferase reporter gene in transient transfection assays and found that it significantly enhanced the activity of SV40 and VH promoters and slightly enhanced an Iα promoter. We conclude that the rabbit has a single hs1,2 enhancer that resides at the 3′ end of the IgH gene cluster and may constitute one of the cis-elements regulating the expression of IgH genes.
Immunoglobulin heavy chain genes are highly regulated during B cell development, and several cis-acting elements, including the intronic enhancer (Eμ)3 (1, 2, 3), and the 3′α enhancers (3′αE), are important for this regulation (4, 5, 6, 7, 8). The 3′αE regions are located downstream of the single Cα gene in mice and rats (4, 5, 6) and downstream of each Cα gene, Cα1 and Cα2, in humans (9, 10). The 3′αE contains four DNase I hypersensitivity sites (hs), denoted hs1,2, hs3a, hs3b, and hs4 (8, 11). Three of these hypersensitive sites, hs1,2, hs3a, and hs3b, are detected within DNA regions that have significant enhancer activity. The hs3a and 3b have a high degree of nucleotide sequence identity to each other, but not to hs1,2. All three enhancers appear to function late in B cell differentiation. The hs4, the most distal enhancer, does not share homology with hs1,2 and hs3a,b and appears to be active throughout B cell development (12, 13).
Each enhancer region includes a core segment containing a single octamer binding motif (ATTTGCAT) (9, 10, 14). The core region of hs1,2 is highly conserved among mouse, human, and rat (8). The nucleotide sequences flanking the core region, however, are less conserved and contain inverted repeats and GC-rich segments (4, 9, 11). Multiple binding sites for B cell-specific transcription factors were identified in the 3′αE, including Oct-1 and Oct-2 binding sites (6, 15, 16), as well as B cell lineage-specific activator protein and NF-κB binding sites (17, 18). The 3′αE has been shown to be important for both transcription and isotype switching. Murine, rat, and human 3′αE were shown in transfection assays to increase the transcription from various promoters (4, 5, 6). In vivo studies also showed that 3′αE increased the transcription of VH-dependent reporter genes in transgenic mice (19, 20, 21). Further, Alt and colleagues showed that disruption of the 3′αE by neo led to deficiencies in the secretion of some IgG isotypes, suggesting that 3′αE affects recombination during isotype switching (7). The 3′αE region was accessible to DNase I and was differentially methylated during B cell development (22, 23). Thus, 3′αE has been shown to influence the control of IgH genes and B cell development. The specific role of 3′αE, however, remains a mystery.
In the rabbit the IgH locus contains 13 nonallelic germline Cα genes, all of which are expressible (24, 25, 26). Within the rabbit IgH gene cluster, Eμ has been identified as an enhancer region (27, 28), but 3′αE has not been described. Because 3′αE is important for IgH gene expression and class switching, and because a 3′αE resides downstream of each of two Cα genes in the human genome, we investigated whether homologous regions are associated with each of the 13 rabbit Cα genes.
Materials and Methods
Cloning of the hs1,2 enhancer probes and Cα genes
A 70-bp 3′αE hs1,2 probe was PCR amplified from rabbit genomic DNA with primers taken from sequences of the core region of human and murine hs1,2: 5′-TATTTT/CTGGAAACAA/GCCT-3′ and 3′-T/ACA/GGGCACATGCAAAT-5′ (10). The hs1,2 sequence was determined using an ABI Prism 310 Genetic Analyzer and was >70% similar to mouse and human hs1,2. This region contained the core region with the octamer binding motif. Human and mouse hs4 probes (∼100-bp) were PCR amplified from genomic DNA with the following primers: human, 5′-TTTGGATCCAGGTATTTCTAAAAATGCT-3′ and 5′-TTGGATCCACCTCCCCCAATGCAAAT-3′; mouse, 5′-AAACATTTCTAAAAATGAT-3′ and 5′-CCCCAGCACCATGCAAAT-′. The identity of the probes was established by nucleotide sequence analysis. The mouse hs3b probe was obtained from a QM293Luc plasmid (gift from B. B. Birshtein, Albert Einstein College of Medicine, Bronx, NY). For PCR amplification of hs3 from rabbit genomic DNA we used as 5′ primer, 5′-ATCACCTGCAGAAA-3′, and as 3′ primers 5′-TGA/TTTCTCAGAACAG-3′ or 5′-AGGCA/CCACCTGGCC-3′.
A partial MboI genomic fosmid library from DNA of rabbit IgH E haplotype was prepared and screened with the rabbit 70-bp hs1,2 and 550-bp Cα probes as previously described (24, 29, 30). Fosmid clones that hybridized with these probes were purified, and the DNA was restriction mapped and analyzed by Southern blot with the Cα, 3′α hs1,2, hs3b (mouse), and hs4 (mouse and human) probes. Hybridization with the hs3b and hs4 probes was performed at low stringency in 6× hybridization solution (31) and washed in 2× wash buffer at 56oC. The nucleotide sequences of the identified Cα genes and 3′αE regions were determined.
pGL3-based luciferase reporter-gene constructs
Constructs containing both a promoter and the 900-bp XbaI/PstI putative 3′αE fragment were developed using the pGL3 basic vector with the luciferase reporter gene (Promega, Madison, WI). The plasmid constructs with either the putative rabbit 3′αE or one of several promoters were used as controls. The 900-bp XbaI/PstI fragment containing the putative hs1,2 was amplified by PCR from Fos15B with primers 5′-CGGGGATCCTCTAGAAGGA-3′ and 3′- TTTGGATCCAGGACCAGTGCTGAGTGC-5′ and inserted into the BamHI site of the pGL3 basic vector with and without a promoter in the 5′ to 3′ and 3′ to 5′ orientations relative to the reporter gene. The 700-bp EcoRI/BamHI fragment containing rabbit Eμ (27) was also inserted into the BamHI/SalI sites of pGL3 with and without a VH promoter.
The rabbit VH1a2 promoter fragment (240 bp) was amplified by PCR with primers 5′-TAACAAGCTTAAAAATTCATATGATCTGAATC-3′ and 3′-TCCTAAGCTTGGTGAGCGTCTGTGTTGA-5′ from the 2.3-kb HindIII fragment of cosmid clone cos 8.2 (32) containing VH1a2. This 240-bp fragment (−240 to −1 relative to the ATG start codon) containing the TATA box, heptamer, pyrimidine tract, and octamer was inserted into the pGL3 basic vector with and without 3′αE using HindIII sites in the 5′ to 3′ orientation relative to the reporter gene. For the Iα promoter we used the 180- and 450-bp fragments containing the rabbit Iα4 promoter region: −180 to −2 bp and −450 to −2 bp upstream of the putative ATG start site, respectively (33). Both Iα fragments contained TATA-less promoter regions and a putative TGF-β-responsive element. The 180- and 450-bp Iα promoter fragments were subcloned into SacI/XhoI sites of the pGL3 vector with and without the 900-bp fragment containing a putative enhancer.
After confirming the nucleotide sequences of all constructs, plasmid DNA was purified on CsCl gradients and used in transient transfection assays.
The rabbit B cell line 55D1, the plasmacytoma cell line 240E (27, 34), the murine A20 B cell line (American Type Culture Collection, Manassas, VA), and the rabbit CD4+ T cell line 484.3 (a gift from P. Medveczky, University of South Florida, Tampa, FL) were grown in RPMI 1640 supplemented with 10% FCS. The kidney cell line RK13 was maintained in DMEM with 10% FCS.
Transient transfection assays
The reporter constructs (20 μg) were transfected by electroporation into the various cell lines (20 × 106 logarithmically growing cells). Rabbit RK13 was transfected with Lipofectamine (Life Technologies, Gaithersburg, MD) according to the manufacturer’s protocol (Life Technologies). Transfections were performed three times with each construct, except for transfections with 484.3 T cells, which were transfected twice. Luciferase activity was measured in duplicate in each experiment, and values were normalized to Renilla luciferase activity according to the manufacturer’s protocol (Promega). To estimate enhancer activity, the luciferase activities of constructs containing both enhancer and promoter were compared with reporter gene activity of the promoter-only control plasmid and are reported as the fold increase in luciferase activity. The data from transient transfection assays were statistically analyzed by Student’s t test.
Cloning of 3′αE region(s)
To identify rabbit 3′αE regions, a genomic DNA fosmid library was screened with a Cα probe and with a rabbit 70-bp hs1,2 probe. We identified 12 fosmid clones that hybridized with the Cα probe; two of those fosmids also hybridized with hs1,2 probes. None of the fosmid clones hybridized with mouse and human hs4 and hs3b probes.
The DNA of the two clones, Fos15B and Fos 3D, that hybridized with the hs1,2 probe were overlapping and encompassed a total of 60 kb of DNA (Fig. 1). Hybridization with a Cα probe showed the DNA from Fos 3D contained two Cα genes, and nucleotide sequence analysis revealed that one of the Cα genes was identical with Cα13 (30). The other Cα gene has not been identified previously and is designated Cα14 (GenBank accession no. AF314407).
Because the hs3b and hs4 probes did not hybridize with the Cα-containing fosmid clones, we attempted to PCR-amplify hs3 and hs4 homologues from rabbit genomic DNA and Cα-containing fosmids using primers from the most conserved region of both mouse and human hs3a,b and hs4. We did not obtain a PCR product, and we conclude that the rabbit genome either does not contain hs3 and hs4 homologues, or if it does, they are highly divergent from mouse and human hs3 and hs4. Because we did not identify hs3- and hs4-like elements, we focused the remainder of our study on the hs1,2 putative enhancer.
We cloned and determined the nucleotide sequence from the 1.8-kb XbaI fragment located 8 kb 3′ of Cα13 that hybridized with the hs1,2 probe. From the nucleotide sequence of the 900-bp XbaI/PstI fragment (GenBank accession no. AF314408), we found a 123-bp segment that was 84% identical with the core region of human hs1,2 and 69% identical with that of mouse hs1,2 (Figs. 1 and 2). Significant homology was not found between rabbit putative hs1,2 and other enhancers (human and mouse hs3 and hs4). The region flanking hs1,2 was GC rich and contained single and dinucleotide repeats (GA) similar to those seen in human and mouse hs1,2 (8, 9). Based on sequence similarity to human and rodent hs1,2 we conclude that the 900-bp XbaI/PstI fragment contains a putative hs1,2 located 8 kb downstream of Cα13.
Southern blot analysis of genomic DNA with the hs1,2 probe
To determine whether the rabbit genome contains more than one hs1,2 homologue to the core region of human hs1,2, we performed Southern blot analysis of XbaI-digested rabbit appendix genomic DNA with the 900-bp XbaI/PstI fragment as a probe (35). We found a single 1.8-kb band that hybridized to this probe, a size corresponding to that in the fosmid clones (Fig. 3). Similarly, the probe hybridized to a single 2-kb BamHI fragment identical with the hs1,2-hybridizing fragment in the fosmid clones (data not shown). We conclude that the rabbit genome contains a single hs1,2 enhancer homologue of human hs1,2.
To confirm the germline configuration of DNA in Fos 15B and Fos 3D, Southern blot analysis of genomic DNA and Fos 15B digested with HindIII/EcoRI was conducted using the 900-bp fragment as the probe. A single 21-kb band was found both in genomic DNA and in DNA of Fos 15B (data not shown), indicating that DNA of Fos 15B was in germline configuration and did not contain insertions and/or deletions. We conclude that the 8-kb distance between Cα13 and hs1,2 is correct and that no other enhancers similar to hs1,2 are present within 40 kb downstream of Cα13.
Chromosomal organization of the Cα genes
The physical map of the previously identified 13 Cα genes showed several clusters of Cα genes that could not be linked within the IgH locus (24). To determine whether the 3′αE is located at the 3′ end of the IgH locus, as in mouse and human, we attempted to complete the physical map of the IgH region by identifying those regions of DNA missing from the original map.
A fosmid clone, Fos 6, from a previous study (36) contained a Cα gene downstream of Cα5. We analyzed the nucleotide sequence of this gene and identified it as Cα1, thus linking the Cα1-containing cluster to the upstream genes (Fig. 4). Ten additional Cα-hybridizing fosmid clones spanning about 165 kb of the Cα region were obtained from the fosmid library in this study and by restriction mapping and nucleotide sequence analysis were found to overlap each other as well as the clones identified in the previous map of the region. Fos 3B was found to have additional DNA downstream of Cα1 that contained the I exon associated with Cα12, thereby linking Cα12 within the locus (Fig. 4). We were not able to clone from this library the DNA that spans the region between Cα12 or Ca9 and the next downstream genes; however, we propose the genetic map shown in Fig. 4. According to this map, the hs1,2 resides at the 3′ end of the Cα gene cluster.
The 900-bp XbaI/PstI fragment containing the putative hs1,2 was tested for enhancer activity after cloning it into the luciferase reporter vector pGL3 with various promoters: SV40, rabbit VH1, and rabbit Iα4. We first established that these promoters were active in the reporter gene constructs following transient transfections of several B-lineage cell lines and a rabbit T cell line and measured luciferase activity. We determined the increase in luciferase activity over that for the pGL3 basic construct without any promoter (Fig. 5). The Iα promoters increased luciferase expression up to 27-fold in B cells and 20-fold in a T cell line, whereas the VH promoter increased expression by ∼4-fold. We conclude that both VH and Iα4 promoters are active in B and T cell lines.
To establish that the hs1,2 alone did not enhance expression of the reporter gene in the absence of a promoter, we transfected each cell line with constructs containing the hs1,2 enhancer in the absence of a promoter and showed that the expression of luciferase was similar to that found for the pGL3 basic vector without enhancer (data not shown). These data indicate that hs1,2 alone does not activate expression of the luciferase gene.
The hs1,2 was tested first with the SV40 promoter (Fig. 6 A), because this promoter is active with heterologous enhancers. The 900-bp fragment in the 5′ to 3′ orientation increased expression of the luciferase gene 2.5- to 4.5-fold in rabbit B cells (55D1), plasmacytoma cells (240E), kidney epithelial cells (RK13), and T cells (484.3); a 14-fold increase was found in the murine A20 B cell line. In the 3′ to 5′ orientation, the 900-bp hs1,2 fragment demonstrated somewhat lower enhancer activity in 240E and 55D1 cells; 484.3 and A20 cells were not tested in this assay. Analysis of these data by Student’s t test indicated that the 900-bp XbaI/PstI fragment had significant enhancer activity in the cell lines tested (p < 0.02).
Because the hs1,2 has been shown to enhance the activity of promoters located >100 kb from the hs1,2 (6, 8, 9), we tested whether rabbit hs1,2 can increase the in vitro transcription of a VH promoter, VH1, that lies >100 kb upstream of 3′α hs1,2 in the germline. In constructs with the VH promoter (Fig. 6 B), the hs1,2 increased luciferase expression ∼6-fold in 55D1 B cells, 3-fold in 240E plasmacytoma cells, and >20-fold in murine A20 B cells; the enhancement was independent of the orientation. By Student’s t test these differences are significant (p < 0.01), and we conclude that the 900-bp fragment has significant enhancer activity with the VH promoter in B cells.
We compared the enhancer activity of hs1,2 on the VH promoter with that of Eμ, which is considered a strong transcriptional enhancer, previously shown to enhance the activity of a VH promoter (3, 37). We found that Eμ enhanced VH promoter activity 12-fold in A20, ∼4-fold in 240E, and <2-fold in 55D1 (Fig. 6 B). Overall, the activity of hs1,2 was similar to that of Eμ in all cells tested, and we conclude that with the VH promoter, the rabbit hs1,2 has enhancer activity comparable to that of Eμ.
We also tested the hs1,2 with the Iα promoter, because Iα promoters are in close proximity to the 3′αE region and are potential targets for hs1,2 activity. We found that the hs1,2 had a slight effect (up to 4-fold) on a 180-bp region of the Iα promoter in B cells (55D1 and A20; Fig. 6 C). These data suggest that the 900-bp fragment has weak enhancer activity with the Iα4 promoter in the tested B cells.
To assess the tissue specificity of hs1,2, we transfected a rabbit kidney epithelial cell line, RK13, with the reporter constructs containing VH, Iα, and SV40 promoters (Fig. 6). We found a <2-fold increase in luciferase activity with the VH promoter regardless of the orientation (Fig. 6B) and almost no activity with Iα promoters (Fig. 6,C). However, the hs1,2 in both orientations enhanced activity of the SV40 promoter (2.5-fold) in both B cells and epithelial cells (Fig. 6,A). Similarly, a rabbit CD4+ T cell line transfected with the same set of constructs (Fig. 6) showed a 2-fold increase in luciferase expression with the SV40 promoter and a low level of activity with VH and Iα promoters. The data indicate that, depending on the promoter, hs1,2 can function in both B and non-B cells.
The regulation of Ig genes is generally thought to involve multiple cis-acting elements, one of which is the IgH 3′αE. Several studies have shown that the 3′αE is important for IgH gene expression and class switching (7, 38). Because each of the germline Cα genes in the human genome is associated with its own 3′αE, we thought that the rabbit genome with 13 Cα genes may have multiple Cα regions and hence multiple enhancers. Instead, the rabbit has a single hs1,2 homologue of the murine and human hs1,2 elements. We did not find a region homologous to the murine and human hs3a and 3b as well as hs4 regions. It is possible, however, that other enhancer elements that are not similar to hs3a,b and hs4 (12) are present in the germline.
While hs1,2 mapped 8 kb 3′ of Cα13 in our study, no hs1,2 was found downstream of the previously cloned Cα13 (24). Because the size of the HindIII restriction fragment (21 kb) on which both Cα13 and hs1,2 are found in each of two overlapping fosmid clones is the same as that found by a genomic Southern blot analysis, we are confident that hs1,2 resides 8 kb 3′of Cα13. In addition, the 8-kb XbaI fragment immediately 5′ of hs1,2 in Fos 15B and Fos 3D is identical with that found by genomic Southern blot analysis (data not shown). We suspect that in the previous study hs1,2 was not present in the Cα13-containing recombinant phage due to either a restriction polymorphism or deletion of DNA in the recombinant phage library.
Although we have not completed the physical map of the Cα chromosomal region, we think Cα13 is the 3′-most Cα gene. Burnet et al. (24) previously identified several clusters of Cα genes and by analyzing the DNA of 12 overlapping Cα-containing fosmid clones, we linked some of these clusters together (Fig. 4). We found that the order of the Cα genes 3′ of Cε is 5′-Cα4-Cα5-Cα1-Cα2-Cα3-Cα7-Cα10-Cα11-Cα12//Cα8-Cα9-//Cα14-Cα13-3′αE-3′.Although we could not directly link either the Cα8 and Cα9 gene cluster or the Cα14 and Cα13 gene cluster to the other Cα genes, we suggest that Cα13 is the 3′-most Cα gene because most of the Cα genes are generally separated by 10 kb, and no Cα genes are found within 30 kb 3′ of Cα13.
The organization and nucleotide sequence of rabbit hs1,2 are similar to those of hs1,2 in other species (8, 10). We found a 123-bp region containing an octamer motif that had significant homology to the human and mouse hs1,2 core regions and was flanked by segments with single nucleotide repeats and GA-rich repeats (8, 9). In fact, the rabbit hs1,2 contains nucleotide sequences similar to those in the human that serve as binding sites for Oct-1 and NF-κB, a set of transcription factors that regulates human hs1,2 (6, 14, 16, 17, 18). The NF-κB site in rabbit hs1,2 is located outside the core region, similar to that of mouse hs1,2. We think that this site is functional because a rabbit probe containing the NF-κB site of hs1,2 was shifted by nuclear extracts from anti-μ-activated primary rabbit B cells as tested by EMSA (unpublished observations). We did not, however, identify binding sites for B-cell lineage-specific activator protein within the 900-bp fragment, suggesting that in the rabbit, control of the IgH locus involves sets of transcription factors binding to hs1,2 different from those found in the mouse.
Functional assays showed that the rabbit hs1,2 enhances transcription from a viral promoter (SV40) and the promoter of rabbit VH1, and modestly from the Iα promoter region of rabbit Cα4. It is not clear which promoters are targets for hs1,2 in vivo, although in vitro 3′αE has been shown to act on several Ig promoters, including VH, Vκ, and Vλ (9, 10, 20). We also tested rabbit hs1,2 with an Iα promoter because these are located in close proximity to 3′αE. Rabbit hs1,2 weakly enhanced the Iα promoter. In contrast, Yanzhong et al. (39) recently showed that human hs1,2 strongly enhanced Iα1 and Iα2 promoters. While hs1,2 enhanced the activity of the rabbit 180-bp Iα promoter more than that of the 450-bp Iα promoter, we suggest that the overall low level of enhancement in our experiments may be because the hs1,2 enhancer and Iα promoter require a combination of transcription factors that is not expressed in the cell lines used in this study.
The rabbit hs1,2 enhanced the activity of the VH promoter in vitro at a level similar to or greater than that of Eμ, which is known to regulate Ig gene expression. The enhancement of VH and Iγ promoter activity by 3′αE has been shown in transgenic mice, suggesting that 3′αE can regulate promoters as far as 100 kb upstream (19, 20, 21). Further characterization of hs1,2 and its targets is required for determining how hs1,2 contributes to immune responses in vivo.
In mice and humans hs1,2 enhancer activity is B cell specific (5, 6, 8, 9, 10). In our experiments we found that while hs1,2 had a high level of activity in murine A20 B cells, the activity in rabbit B cells was considerably lower. Further, the activity in rabbit T and epithelial cells was nearly equal to that in rabbit B cells, raising the possibility that hs1,2 is not B cell specific. We think, however, that the hs1,2 enhancer probably is B cell specific, but that because the rabbit B cell lines were obtained from c-myc and c-myc/v-abl transgenic rabbits, in which expression of the oncogenes was controlled by Eμ (27, 34), the low level of activity may result from inhibition of hs1,2 by a high level expression of c-myc/v-abl. However, we cannot exclude that the difference in hs1,2 activity in mouse and rabbit cells is due to inherent differences between rabbit and mouse cell lines.
We undertook this study with the idea that a 3′αE would be located adjacent to each of the 13 Cα genes and that differences in the 3′αE associated with each gene might explain the differential expression of the Cα genes in various tissue. Our results demonstrate that only one hs1,2 resides in the germline. While additional functional analyses of hs1,2 will be required to elucidate the specific role of IgH 3′αE, we suggest that hs1,2 may interact with Iα promoter and regulate in part Cα gene expression in vivo. Studies to directly test this idea need to be performed.
This work was supported by National Institutes of Health Grant AI11234.
Abbreviations used in this paper: Eμ, intronic enhancer; 3′αE, 3′α enhancer; hs, hypersensitivity site.