physiological activation of the igh 3′ enhancer in b lineage cells is not blocked by pax-5

Eur. J. Immunol. 1996.26: 2499-2507 IgH 3' enhancer activity in B lineage cells is not blocked by Pax-5 2499

Tove Andersson', Markus F. Neurath', Patrick A. Grant' and Sven Pettersson'

Department of Biosciences at Novum, Karolinska Institute, Huddinge, Sweden Laboratory of Immunology, I. Medical Clinic, University of Mainz, Mainz, Germany

Physiological activation of the IgH 3' enhancer in B lineage cells is not blocked by Pax4

The mouse 3' enhancer contains a high-affinity binding site for the paired box protein Pax-5. Here, we demonstrate by genomic footprinting that the rat 3' enhancer contains a low-affinity binding site for Pax-5, which is occupied in activated splenic B cells. Thus, binding of Pax-5 to the IgH 3' enhancer appears to be evolutionarily conserved in rodents. Analysis of Pax-5 expression in primary B cells demonstrates that Pax-5 remains expressed after 4 days of lipopoly- saccharide (LPS) induction, but is down-regulated in 5-day stimulated cells. Similarly, the expression of Pax-5 is down-regulated in vivo in activated large splenocytes, in contrast to small resting cells. Multimerization of the high- affinity Pax-5 binding site linked to a heterologous reporter gene demonstrates that Pax-5 can function as a transcriptional activator. In contrast, Pax-5 overex- pressed in cell lines represses both the mouse and the rat 3' enhancer. Surprins- ingly, cross-linking of the IgM receptor in BAL-17 cells containing a stably integrated 3' enhancer-dependent p globin reporter gene demonstrates that induction of 3' enhancer activity is not blocked by Pax-5. Moreover, stimulation of 3' enhancer p globin-transgenic splenocytes demonstrate that Pax-5 cannot repress activation of the 3' enhancer upon LPS induction or CD40 receptor stimulation. Hence, activation of the IgH 3' enhancer occurs independently of changes in Pax-5 gene expression. This indicates that previous studies conducted in vitro may be an oversimplification of the function of Pax5 and 3' enhancer activity.

1 Introduction

B lymphocytes secrete large quantities of antibodies when challenged with an antigen. To differentiate into plasma cells, a large portion of the B-lineage cells are subject to two consecutive recombination events: V(D)J recombination and heavy chain class switch recombination. Whereas the intragenic enhancer (Ep) controls V(D)J recombination, the subsequent class switch recombinant, at least in part, is controlled by the IgH 3' enhancer [l]. This 3' enhancer is located downstream of the C a gene in rodents [2-41 and forms, together with additional enhancer elements identified in this region [5, 61, a putative locus control region (Fig. 1).

Studies using mouse enhancer-dependent p globin- transgenic mice demonstrated that 3' enhancer activity is confined to the activated cell pool in vivo [7]. This is further supported by changes in the methylation pattern: this region is hypermethylated in pre-B and Tcell lines, but demethylated in plasma cells [S]. In addition, changes in chromatin structure correlate with the appearance of four

[I 155241

Received March 7, 1996; in revised form July 9, 1996; accepted July 23, 1996.

Correspondence: Sven Pettersson, Department of Biosciences at Novum, Karolinska Institute, S-141 57 Huddinge, Sweden Fax: +46-87 7455 38; e-mail: [email protected]

Abbreviations: CD40L: CD40 Iigand IgH. Immunoglobulin heavy chain Ep Heavy chain intragenic enhancer HLH: Helix-loop helix

Key words B lymphocyte differentiation / 3' enhancer / Immu- noglobulin gene expression / Pax-5 / Transcription factors

tissue-specific and cell stage-specific DNase I-hyper- sensitive sites in the region 3' of the locus, two of which map to the IgH 3' enhancer [5 , 81. These changes in the chromatin structure appear to be guided by extracellular signals. Ligand receptor-dependent stimulation via LPS, or the IgM or CD40 receptors, can transactivate the IgH 3' enhancer ([7, 91 and Grant et al., submitted). Targeted deletion of the IgH 3' enhancer in the mouse germ-line resulted in a 3' E-/- mouse whose B lymphocytes were deficient in class switching to certain C, genes, suggesting that this enhancer is linked to the machinery that controls class switch recombination [l]. Another possible role for the 3' enhancer is suggested by the results of studies using transgenic animals where an Ig reporter gene, potentiated by the mouse 3' enhancer, displays increased levels of Ig gene expression compared to Ig constructs potentiated by the Ep enhancer [lo]. Targetted deletion of the 3' enhancer in an IgG2a-producing cell line lacking the Ev enhancer abrogates y2a expression, further supporting the role for this enhancer in the regulation of IgH transcription [Ill.

We demonstrated recently that IgH 3' enhancer activity can be repressed in cell lines by the dominant negative helix-loop-helix (HLH) protein Id3 [12]. This protein is expressed in pre-B and B cell lines, but down-regulated in plasma cell lines [12]. Like Id3, the expression of the Pax-5 gene is inversely correlated to 3' enhancer activity in cell lines. The Pax-5 gene is a member of a vertebrate multi- gene family encoding proteins sharing the DNA-binding paired-box domain with a group of regulatory and tissue- specific genes in Drosophila [13, 141. The Pax-5 protein, also known as BSAP [ 151, NF-HB [ 161 and Sa-BP [ 171, has been shown to act as a transcriptional activator of the CD19 promoter [18] and the IgH IE promoter [19]. In Bcell lines, Pax-5 is expressed in cell types derived from pro-, pre- and mature B cells, but not in plasma cells [20].

0 VCH Verlagsgesellschaft mbH, D-69451 Weinheim, 1996 0014-2980/96/1010-2499$10.00 + .25/0

2500 T. Anderson et al. Eur. J. Immunol. 1996.26: 2499-2507

Furthermore, Pax-5-/- mice display no mature B lymphocytes, indicating the importance of PaxJ in early B lymphocyte development [21].

Previous studies have reported a high-affinity and a low- affinity binding site for Pax-5 in the mouse IgH 3’ enhancer [22, 231. However, while the high-affinity site is occupied in vivo, the occupancy of the low-affinity site is unclear [24]. Although binding sites for some transcription factors are conserved between the mouse and the rat enhancer, the high-affinity site for Pax-5 is deleted in the rat 3’ enhancer [2-41 (Fig. 1). However, the rat 3‘ enhancer contains a binding site homologous to the mouse low- affinity site. This observation raises two questions: is Pax- 5-mediated regulation of the enhancer conserved in rodents, and at what stage in the developmental program of B lineage cells does Pax-5 exert its function on enhancer activity?

By genomic footprinting experiments, we demonstrate that the rat Pax-5 binding site is occupied in vivo, suggesting that the Pax-5 protein can regulate both the mouse and the rat 3’ enhancer by binding two different Pax-5 binding sites. Analysis of a multimerized high-affinity Pax-5 binding site from the mouse enhancer linked to a heterologous reporter gene reveals that the Pax-5 protein can also function as an activator on this site. When the rat low-affinity and the mouse high-affinity Pax-5 binding site is positioned within its natural context, Pax-5 represses both rat and mouse enhancer activity in cell lines. Surprisingly, the use of splenic B cells from mouse 3‘ enhancer-dependent p globin reporter-transgenic mice, [7], or a BAL-17 cell line which contains a stably integrated 3’ enhancer- dependent reporter gene [19], demonstrates that physiological activation of the 3‘ enhancer is not blocked by Pax-5. The implication of Pax-5 function and 3’ enhancer activity in late B cell development in vivo is discussed.

2 Materials and methods

2.1 Plasmid constructs

The generation of the mouse and rat 3‘ Epp128 [2, 31 and the pp34SV3 [25] vector has been described. A 72-base pair oligonucleotide (top strand: 5’-CTAGAGTTG- AGCCACCCATCCTTGCCCGTTGAGCCACCCATC- CTTGCCCGTTGAGCCACCCATCCTTGCCCT-3’),

containing three copies of the mouse 3’ enhancer Pax-5 high-affinity binding site was cloned into the Xba I site of pp34SV3. This construct is referred to as pp3XP. Expres- sion plasmids containing the human Pax-S cDNA in sense and antisense configuration were kind gifts from Dr. M. Busslinger (Institute for Molecular Pathology, Vienna, Austria), the expression vector encoding the human basic HLH protein SEF2-1B from Dr. T. Grundstrom (Depart- ment of Applied Molecular Biology, University of Umel, Umel, Sweden) and the expression vector encoding the human Ets protein Elf-1 from Dr. C. B. Thompson (Howard Hughes Medical Institute, University of Chi- cago, Chicago, IL).

2.2 Cell lines and transfection

The mouse cell lines MPCll (plasmacytoma) and BAL-17 (B cell lymphoma) were used. The monkey kidney cell line COS was used as a representative of a non-lymphoid cell line. The origin of BAL-17 was as described [26] and the origin of MPC-11 and COS is described in the American Type Culture Collection (ATCC, Rockville, MD) catalog. MPCll and COS cells were grown in DMEM, while BAL-17 cells were grown in RPMI 1640, supplemented with 10 % FCS, penicillin, streptomycin (Gibco Grand Island, NY) and 50 pM 2-mercaptoethanol (Sigma, St. Louis, MO). MPCll and COS cells were transfected with 20 pg plasmid DNA by calcium phosphate co-precipitation [27]. Pax-5, SEF2-1B and Elf-1 expression vectors were co-transfected where indicated. The amount of expression plasmid was held constant by the addition of irrelevant plasmid. Cells were harvested 30-36 h after transfection and RNA was extracted from the cells.

A BAL-17 clone containing a stably integrated mouse 3’ enhancer-dependent p globin reporter gene [9] was cross-linked, at a density of 1 x lo7 cells/ml, with 100 pg affinity-isolated goat anti-mouse IgM (p chain specific) antibody (Sigma) per 1 x lo7 cells.

2.3 Primary cell preparation and culture conditions

A single-cell suspension of mouse or rat spleen was made as described [28]. Primary cells were washed twice in PBS. Erythrocytes were lysed using Gey’s solution. Resting and cycling lymphocyte populations were isolated by Percoll

Figure 1. Outline of the mouse IgH locus and the regulatory elements in the mouse and rat IgH 3’ enhancers. The variable (V), diversity (D), joining (J) and constant (C) region exons are shown as shaded boxes. S indicates the identified switch regions in the IgH locus. The Ep and IgH 3’ enhancers are represented as circles. At the 3’ end of the IgH locus, the positions of four hypersensitivity sites, two of which map to the 3’ enhancer (1-4), are indicated. The mouse IgH 3’ enhancer contains multiple binding sites, including the two Pax-5 binding sites flanking the NF-aP binding site. Many binding sites are conserved in the rat enhancer, but specifically note the deletion of the high-affinity binding site for Pax-5 in the rat enhancer (here indicated by *).

Eur. J. Immunol. 1996.26: 2499-2507 IgH 3' enhancer activity in B lineage cells is not blocked by Pax-5 2501

density gradient separation (Pharmacia, Uppsala, Swe- den) as described [7]. Where indicated, B cells were posi- tively selected by staining with anti-mouse B220 magnetic microbeads (Miltenyi Biotech, Bergisch Gladbach, Ger- many) and separated on a MACS column according to manufacturers instructions. Splenic cells were stimulated, at a cell concentration of 5 x lo6, with 50 pglml lipopoly- saccharide (LPS, Escherichia coli, serotype 0111 : B4; Sigma) in RPMI medium supplemented with 15% FCS, penicillin, streptomycin (Gibco) and 20 pM 2-mercaptoethanol (Sigma). RNA was then extracted from the cells. Cycling blast cells were isolated from dead cells by Ficoll-Hypaque separation (Pharmacia) prior to RNA extraction where indicated.

For stimulation with CD40 ligand (CD40L), 5 X lo6 cells per ml were seeded together with 5 X lo5 irradiated J558L hybridoma cells secreting a mouse (m)CD40L-CD8a fusion protein [29], a kind gift from Dr. P. Lane (Basel Institute for Immunology, Basel, Switzerland). For stimulation of splenic cells prior to nuclear extract preparation, trimeric CD40L was used.

2.4 RNA extraction and analysis

Total cytoplasmic RNA from cell lines and primary cells was extracted as described [2, 71. Ribonuclease protection analysis was performed as described [2]. The 70-kDa cognate heat-shock protein (Hsp-70) and Cp riboprobes used in these studies have been described [30].

2.5 Flow cytometry

Flow cytometry was carried out by staining cells with FITC-labeled anti-mouse CD23 antibody (PharMingen, San Francisco, CA). Flow cytometry data were obtained on a FACScan using Cellquest aquisition and analytic soft- ware (Becton Dickinson, Mountain View, CA).

2.6 Gel mobility shift assay

The Pax-5 oligonucleotide used in the protein-binding studies was synthesized on a Gene synthesizer 380B (Applied Biosystems, Fostor City, CA) and contains the mouse 3' enhancer Pax-5 high-affinity binding site and flanking sequences. The top strand of the annealed oligonucleotide is: 5'-agctTGTTGAGCCACCCATCCTTGCC- CATCTCCTGTCa-3'. Nuclear extract preparation and gel mobility shift assays were performed essentially as described [31]. The COP8 mouse fibroblast cell line extracts were kind gifts from Dr. M. Busslinger.

2.7 Dimethyl sulfate (DMS)-piperidine treatment of

Cells grown in suspension were washed twice with PBS, then resuspended in RPMI 1640 at lo8 cells/ml. Incubation with 0.1% dimethyl sufate (DMS) was performed for 1 min at room temperature, followed by incubation in cell lysis buffer (1 mM Tris-HC1 pH 7.5,400 mM NaCl, 2 mM EDTA, 0.2% SDS, 0.2 mg/ml proteinase K) at 37°C. After extraction of DNA with phenolkhloroform, the

DNA in vivo and in vitro

methylated DNA was precipitated with ethanol and the strand scission reaction was performed in 1 M piperidine at 90°C. The DNA was then reprecipitated and ligation- mediated (LM)-PCR was performed as described below. For DMS treatment of genomic DNA in v i m , the DNA was extracted by standard methods and digested overnight with restriction enzymes that do not cut the target sequences. The DNA was then incubated with 0.1 YO DMS for 30 s at room temperature and then treated as described above.

2.8 Ligation-mediated polymerase chain reaction (LM-

Footprinting in vivo by LM-PCR was carried out as described [24]. In brief, primer annealing was performed with 0.5 pmol primer 1 for 2 pg genomic DNA. For primer extension, Vent polymerase (New England Biolabs, Beverly, MA) was used. Linker ligation was performed overnight at 16 "C and exponential PCR amplification was done with primer 2 and the linker primer for 16-22 cycles (94°C for 1 min, Tm + 1°C for 2 min, 76°C for 10 min), followed by phenol/chloroform exctraction of samples, ethanol precipitation and analysis on 8 YO denaturing urea/ polyacrylamide gels. Primer sequences for LM-PCR for the rat Pax-5 site (non-coding strand) are (1) 5'-TC-

CTGCAGCAGGTTCAC-3' and (3) 5'-TTCCCCCA- CCTGCAGCAGGTTCACC-3 '. Labeled and unlabeled primers for LM-PCR were purified by polyacrylamide gel electrophoresis. The correctness of sequence ladders obtained by LM-PCR was verified by sequence analysis of cloned DNA using the plasmid pCRII-w3'e containing the core sequence of the rat IgH 3' enhancer [2, 241.

PCR)

GACTCA'MCTCTGAGCCA-3 ' , (2) 5'-ATTCCCCCAC-

3 Results

3.1 The rat 3' enhancer contains a Pax-5 binding site which is protected in unstimulated and LPS-activated B lymphocytes

Transient transfection studies suggested a role for the Pax-5 protein in regulation of the mouse 3' enhancer activity via binding of Pax-5 to the high-affinity binding site in the enhancer [22, 231. Functional analysis suggested that the mouse low-affinity binding site for Pax-5 regulates 3' enhancer activity, although at a lower level compared to the high-affinity site [23]. The rat enhancer contains a putative Pax-5 binding site in the same position as, and almost identical to, the low-affinity site for Pax-5 in the mouse enhancer [2, 221 (Fig. 2A). These two sites have a 59 YO (mouse) and 65 YO (rat) similarity to the consensus site for the Pax-2, Pax-5 and Pax-8 protein subclass [32]. Sequence alignment suggest that these sites belong to class I Pax-5 recognition sequences. Both binding sites have, in their 5' half, a good match to the class I consensus. In addition, both sequences contain a CA at positions 4 and 5, which is invariant in all class I sequences (Fig. 2A) [14].

To examine whether the rat Pax-5 binding site was occupied in vivo, genomic footprint analysis was carried out. As can be seen, a footprint is observed in crude

2502 T. Anderson et al.

A Pax-2/5/8 consensus:

Mouse low-affinity:

Rat Pax-5:

A C T AC G N G C ~ G T O A C

(3

(3 (3 (3

O(3

2

t b 0

Pax5 2 02 0 0 0 0 k a

DMS FOOTPRINT I NG non-coding strand

Figure 2. Occupancy in vivo of the rat 3' enhancer Pax-5 binding site. (A) Sequence comparison of the rat Pax-5 binding site and the mouse low-affinity binding site with the consensus recognition sequence for the paired domain of the Pax-2, Pax-5 and Pax-8 sub- family [32]. Boxes indicate residues which match the consensus. (B) DMS footprint analysis in vivo at the rat Pax-5 binding site in splenic B cells. B220+ MACS-isolated splenic B cells were stimulated with LPS for 48 h. After treatment with DMS in vivo, DNA from T cells (B220- fraction) and unstimulatedstimulated B cells was extracted and subjected to LM-PCR footprinting in vivo as described in Sect. 2.7. Protected guanine residues at the rat Pax-5 binding site on the noncoding strand are indicated with circles.

splenic Bcells and Bcells stimulated with LPS in vitro (Fig. 2B). Subsequent gel retardation analysis demonstrated that the Pax-5 protein can specifically bind to an oligonucleotide containing the rat Pax-5 binding site (data not shown). Thus, Pax-5 may exert its effector function in vivo via two separate Pax-5 binding sites in the mouse and the rat 3' enhancer.

3.2 The Pax4 protein can function as a transactivator of the high-affinity Pax-5 binding site from the 3' enhancer

The Pax-5 protein has been demonstrated to be a transcriptional activator of the CD19 promoter [18] and the IgH IE promoter [19]. To test whether Pax-5 could function

Eur. J. Immunol. 1996.26: 2499-2507

as a transcriptional activator of the 3' enhancer, three copies containing the mouse high-affinity Pax-5 binding site were linked to a p globin gene promoter (Fig. 3). This minimal promoter construct (pp3XP) was transiently transfected, with or without a Pax-5 expression vector, into MPCll and COS cells. In the absence of Pax-5, no transcriptional activity of the pp3XP minimal promoter was observed (Fig. 4A and B and in J558L cells, data not shown). However, a strong dose-dependent induction of reporter gene activity was detected when the Pax-5 expression vector was co-transfected (Fig. 4A and B). In contrast, co-transfection of Pax-5 antisense, or expression vectors for two other transcription factors, the bHLH protein SEF2-1B and the Ets protein Elf-1 (constructs shown in Fig. 3), did not induce transcription from this promoter.

3.3 Pax-5-dependent repression of both the mouse and the rat IgH 3' enhancer is observed in cell lines

Recent studies of cell lines suggested that the Pax-5 protein can repress the mouse IgH 3' enhancer [22, 231. To test whether a rat 3' enhancer-dependent human p globin reporter construct (shown in Fig. 3) was subject to Pax-5 repression, as demonstrated for the mouse enhancer, this construct was transfected into MPCll cells together with a Pax-5 cDNA expression vector. As can be seen, the rat enhancer, as well as the mouse enhancer, strongly poten- tiate transcription of the reporter gene (Fig. 5A and B). However, a fourfold reduction of both mouse and rat 3' enhancer-dependent p globin transcription was observed in the presence of a Pax-5 sense expression vector, but remained unaffected by an antisense Pax-5 expression vector.

3.4 Pax-5 cannot repress ligandreceptor-dependent transactivation of the IgH 3' enhancer

To determine whether Pax-5-mediated repression of the 3' enhancer could be observed in vivo, resting splenic cells from the p globin-transgenic mice [7] were stimulated with LPS or CD40L. After 48 h, RNA was prepared and expression levels of the p globin reporter gene versus Pax-5 expression were determined by ribonuclease protection. While strong induction of enhancer-dependent p globin activity is observed upon LPS and CD40L stimulation, the Pax-5 expression levels remained virtually unaffected in comparison to the levels observed in resting cells (Fig. 6A and B). The CD40L-induced activation was specific, since addition of specific antibodies blocked this activity (data not shown). To determine the percentage of cells responding to CD40L stimulation, CD23 expression levels were determined. This molecule is a low-affinity receptor for the Fc of IgE (FCERII), which is up-regulated on Bcells following CD40 ligation [33, 341. Flow cyto- metric analysis demonstrated a strong increase in CD23 expression, suggesting that a large majority of the resting cells respond upon stimulation with CD40 (Fig. 6C). Acti- vation of the 3' enhancer cannot be explained by the absence of protein expression, since Pax-5 is present in a DNA-binding form as shown by gel retardation experiments of nuclear extracts from LPS- or CD40L-activated primary Bcells (Fig. 6D, lanes 4-7), in agreement with previous results [35].

Eur. J. Immunol. 1996.26: 2499-2507 IgH 3’ enhancer activity in B lineage cells is not blocked by Pax-5 2503

To avoid any possible clonotypic difference within the resting Bcell pool, we took advantage of a clone of the mouse B lymphoma cell line BAL-17 which contains a stably integrated mouse 3‘ enhancer-dependent p globin reporter construct [9]. This cell line models primary B lymphocyte responses in a number of aspects, including signal- ing from the IgM receptor and can, upon IgM cross- linking, transactivate the IgH 3’ enhancer [9]. Following incubation of the subclone BAL-17-D3 with goat anti- mouse IgM (p-chain specific) antibodies, RNA was iso-

A

&l

a a 0 > 0 P Q

R 0

Q P

!4

.I Y - Y

$

3 0 k

Plasmid:

-r

M P C l l

-r

I 2 3 4

B

Figure3. Map of reporter constructs and expression plasmids. All reporter gene constructs contains a pUC backbone. The mouse and rat 3’Epp128 plasmids contain the human fl globin gene with its endogenous promoter potentiated by the mouse or the rat 3’ enhancer. The pp3XP reporter gene is driven by three copies of the mouse Pax-5 high-affinity binding site, potentiated by an SV40 enhancer downstream of the fl globin gene. CMV29-31 contains the mouse cDNA encoding the basic HLH protein SEF2-1B, driven by the hCMV promoter and cloned into the pBR322-derived vector pML2. pcDNA Elf-1 contains the human cDNA encoding the Ets protein Elf-1, potentiated by the hCMV promoter and cloned into pBR322-derived pcDNAIneo.

lated 48 and 120 h after stimulation. Ribonuclease protection analysis of p globin and endogenous Pax-5 expression revealed that, while p globin transcription gradually increases with incubation time, the levels of Pax-5 RNA and protein remain constant (Fig. 6D, lane 8-10, and 6E). These data, in combination with the data obtained from primary Bcells, strongly suggest that the Pax-5 protein cannot block ligandreceptor-mediated transactivation of the enhancer.

cos

-I 7

1 2 3 4 5 6

p g: - , 4 8 1 2 ,, 1 2 ,, 1 2 , - Pax-5 P a x - 5 SEF2-1B Plasmid: - Pax-5 sense pax-5 Elf-1 sense a n t i -

sense antisense

Figure 4. The Pax-5 protein can function as a transcriptional activator. (A) Histogram showing Pax-5-inducible activation of a multimerized Pax-5 binding site, as determined by ribonuclease protection analysis. The pfl3XP fl globin reporter construct was transiently transfected into MPCll plasmacytoma cells alone (lane 1) or together with 6 pg expression vector containing the Pax-5 cDNA in sense (lane 2) or anti-sense (lane 3) configuration. Co-transfection of the same amount of expression vector encoding the protein SEE-1B (CMV29-31) is shown in lane 4. An Hsp-70 riboprobe was used as a reference in the assay and the /3 globin values from three independent experiments were normalized to this reference by densitometric quantification. (B) Same as in (A) except that the transfection was carried out in COS cells. The amount of Pax-5 sense expression vector was titrated from 4 to 12 pg (lanes 2-4) and 12 pg Elf-1 expression vector was transfected as a control (lane 6).


30 -

20 -

10 -

O +

A M P C l l B M P C l l 150 1 4ol

1 1

1 7 7 4 -I -r 1

1 2 3 4 Plasmid: - - Pax-5 Pax-5 - Pax-5 Pax-5 sense antisense, Plasmid: - -1

pS128 rat 3'Epl3128 sense antisense, - 1 pl3128 mouse 3'Epl3128

Figure 5. The Pax-5 protein represses mouse and rat IgH 3' enhancer activity. (A) Histogram showing transactivation of the enhancer- dependent p globin reporter construct 3'Epp128 containing the mouse 3' enhancer following transient transfection into MPCll cells. Lane 2 shows the increase in 3'Epp128 expression compared to expression of the empty vector pp128 (lane 1). Co-transfection of 8 pg human Pax-5 expression vector in sense configuration is shown in lane 3 (75 % reduction of p globin expression) and anti-sense configuration in lane 4. Expression of p globin and endogenous Hsp-70 (reference) was analyzed in a ribonuclease protection assay. p globin values from three independent experiments were quantified by densitometry and normalized to the reference values. (B) Same analysis as in (A), but with 3'Epp128 containing the rat 3' enhancer instead. Co-transfection of the Pax-5 sense expression vector is shown in lane 3 (85 % reduction in p globin expression) or the anti-sense vector in lane 4.

3.5 Paxd gene expression in B lineage cells is subject to down-regulation both in vivo and in vitro

To determine the expression pattern of Pax-5 in vivu, cytoplasmic RNA was prepared from Percoll-fractionated small resting cells and large in vivo activated cells. Whereas small cells displayed high Pax-5 expression, a considerable decrease in the Pax-5 level (75 %) was observed in large cells activated in vivo (Fig. 7). The cell pool activated in vivo is, however, quite heterogenous with respect to stages in Bcell development. To correlate sys- tematically Pax-5 gene expression with the previously reported pattern of 3' enhancer activation [7], splenic lymphocytes were stimulated in vitro with LPS. At different time points, cytoplasmic RNA was prepared and subjected to ribonuclease protection. As can be seen, the expression of Pax-5 remains unaffected after 96 h of stimulation. However, a significant reduction of Pax-5 expression is observed after 120 h (Fig. 7). This reduction correlates well with the timepoint when a majority of Bcells activated in v i m differentiate into the secretory cell stage ([36] and references therein). To exclude the possibility that the observed down-regulation of Pax-5 was due to cell death, the 96-h and 120-h cell pools were subjected to Ficoll-Hypaque separation prior to RNA preparation to exclude dead cells. Subsequent analysis of these activated, viable, cell populations revealed similar results to the nonseparated LPS-activated cells (Fig. 7). Furthermore, to underline that the LPS-activated cell pools contained Bcells, we determined the expression of C p by a Cp- specific riboprobe. As can be seen, increased levels of Cp expression are observed in the activated cells (Fig. 7).

4 Discussion

Here, we demonstrate that the Pax-5 binding site in the rat 3' enhancer is occupied in activated B cells, suggesting that Pax-5-mediated regulation of the enhancer may be evolutionarily conserved in rodents. Kinetic studies of Pax-5 expression in primary B cells demonstrate that Pax-5 is subject to down-regulation at later stages of Bcell development. Transient transfections of a minimal promoter construct containing the mouse Pax-5 high-affinity binding site demonstrates that Pax-5 can function as a transcriptional transactivator. This result is in agreement with studies showing that Pax-5, like Pax-1, Pax-3, and Pax-6, can activate transcription through a multimerized high-affinity Pax binding site from the CD19 promoter [32]. Surpris- ingly, physiological ligandheceptor-dependent activation of the 3' enhancer in primary B lineage cells demonstrates that 3' enhancer activation cannot be blocked by Pax-5. Similarly, 3' enhancer activation by cross-linking of the IgM receptor in BAL-17 cells cannot be blocked by Pax-5- mediated repression.

Our observation that transactivation of the 3' enhancer cannot be blocked by the Pax-5 protein, either in cell lines or in primary splenic B lymphocytes, raises some concern over the model proposed by two groups over the last years [22, 231. We are not suggesting that Pax-5-mediated repression does not exist, but rather that Pax-5-mediated regulation of 3' enhancer activity is clearly more complicated than simply an off-on control of enhancer activity. More- over, our data strongly indicate that Pax-5 is an important physiological regulator of enhancer activity, since the binding site for this protein is evolutionarily conserved. Given

Eur. J. Immunol. 1996.26: 2499-2.507 IgH 3' enhancer activity in B lineage cells is not blocked by Pax-5 250.5

Figure 6. Analysis of endogenous Pax-5 gene expression and 3' enhancer activation following stimulation of splenic cells with LPS, CD40L or cross-linking of the surface IgM receptor. (A) Percoll purified high-density resting splenic cells from transgenic mice carrying a mouse IgH 3' enhancer-human p globin construct were induced with LPS for 48 h. RNA was prepared and analyzed by ribonuclease protection. The positions of correctly initiated 0 globin, Pax-5 and Hsp-70 (reference) protected transcripts are indicated. RNA prepared from noninduced resting splenic cells was analyzed as a negative control. (B) Same as (A), but the B cells were co-cultured with JS58L hybridoma cells secreting a mouse CD40L-CD8a fusion protein. (C) CD23 expression is up-regulated in CD40L-activated splenic B cells. Flow cytometry of small resting splenic cells before (top) and after (bottom) 48 h of CD40 stimulation, as described in (B). Heavy lines show CD23 expression and light lines show negative control staining. The percentage of CD23-expressing cells in the CD40-stimulated population was 90 % compared to 47 % in small, resting cells. However, both the negative and the positive staining cells in the small cell population show an increased CD23 staining following CD40 stimulation. (D) Gel mobility shift assay showing nuclear extracts prepared from small, resting splenocytes (lane 4) or BAL-17 cells incubated with a radiolabeled oligonucleotide spann- ing the mouse high-affinity Pax-5 binding site. Nuclear extracts were also prepared following stimulation of resting splenocytes with LPS or a CD40L trimer (lane 5-7) and BAL-17 cells (lane 8) with anti-IgM (p-chain specific) antibodies (lanes 9, lo), as indicated. In lane 2, control COP8 extracts are shown, and in lane 3, extracts from COP8 cells transfected with a Pax-5 expression vector. Positions of free probe, nonspecific binding (*) and of the Pax-5 complex are indicated. (E) A clone of BAL-17 cells, BAL-17-D3, stably transfected with the mouse 3' enhancer-dependent p globin reporter construct, was cultured in normal medium with and without anti-IgM (pchain specific) antibodies for 48 and 120 h. RNA was prepared and analyzed by ribonuclease protection. As a negative control, RNA was extracted from noninduced cells. Positions of correctly initiated transcripts are indicated.


Figure 7. Kinetics of Pax-5 and Cp expression in LPS-stimulated splenic B cells and expression of Pax-5 in large cells activated in vivo. Splenocytes were cultured with LPS and at the indicated time points, RNA was prepared and analyzed by ribonuclease protection. At 96 and 120 h, the cycling blast cell population was isolated by Ficoll-Hypaque separation prior to RNA preparation. The last two lanes show ribonuclease protection of RNAprepared from Percoll-purified high-density resting (small), and low-density cycling (large) splenocytes, respectively. Positions of correctly initiated Pax-5, Cp and Hsp-70 (reference) protected transcripts are indicated. A radiolabelled Msp I-digested PBR322 plasmid was used as a size marker.

that this enhancer element has been linked to class-switch recombination, it is equally possible that Pax-5 may be directly involved in the machinery directing class-switch recombination. It will be of considerable interest to learn whether Pax-5 mutant pro-B cells can undergo class switch recombination, since pro-B cells from RAG-2 -/- mice can be triggered to undergo class-switch recombination to IgE by stimulation with CD40L and IL-4 (Dr. E Melchers, personal communication).

The 3' enhancer has also been implicated in the regulation of immunoglobulin expression in transgenic animals [ 101. Targetted deletion of the 3' enhancer in an IgG2a- secreting cell line which lacks the Ep enhancer resulted in abrogation of IgH gene expression, which supports a role for the 3' enhancer in the regulation of IgH expression in this cell line [ll]. Moreover, plasma cells display increased levels of immunoglobulin production, but the precise mechanism(s) behind this up-regulation has not yet been resolved. Our studies demonstrate that Pax-5 gene expression is down-regulated in antibody-secreting cells. If Pax-5 repression of the 3' enhancer occurs in vivo, it may act on B-lineage cells before they become secretory cells. While our data demonstrate that Pax-5 does not inhibit activation of the 3' enhancer, they are insufficient to reveal whether additional enhancer activity would be achieved in the absence of Pax-5. Our data presented here suggest that Pax-5 does not interfere with triggering of enhancer activity. Instead, Pax-5 may regulate the levels of IgH expression by modulating 3' enhancer activity in activated Bcells, but not in plasma cells. Recent experiments demonstrated that cross-linking of the OX40 receptor on CD40L-stimulated B cells results in elevated steady-state

levels of immunoglobulin mRNA, and a concomitant down-regulation of Pax-5 gene expression and loss of Pax-5 binding to the 3' enhancer [37]. Thus, Pax-5 may, at least in part, repress 3' enhancer activity.

Recent experiments have shown that blocking the binding of Pax-5 to the high-affinity site with a triple helix-forming oligonucleotide leads to recruitment of a factor, NF-aP, to a site near the Pax-5 site [24]. Interestingly, this site is identical to the ets-like pA (hereafter denoted aP) site, which contributes to the activity of the tissue-specific domain B in the enhancer [31]. The a P site is conserved in the same position in both the mouse and the rat enhancer [31], which suggests that a target for Pax-5-mediated regulation may be conserved in rodent evolution. The NF-aP protein is present in both B and plasma cells and is likely to belong to the ets-family of transcription factors, since a pan-ets antibody could supershift the NF-aP complex [24]. In addition, blocking of the Pax-5 high-affinity binding site in vivo in the 3' enhancer in IgG and IgA cell lines, and to a lesser extent in IgM cell lines, resulted in increased accumulation of immunoglobulin mRNA from endogenous heavy-chain genes [24]. It is conceivable, therefore, that the NF-aP factor is coupled to the Pax-5-regulated enhancer activity. Once the cell reaches the secretory stage, Pax-5 expression is down-regulated and NF-aP can bind to its cognate site, thereby derepressing the Pax-5- dependent 3' enhancer activity. Such a mechanism is com- patible with recently published data [24, 371.

We thank Annika Samuelsson for flow cytometry analysis and Dr. J . Gamble for critical reading of this manuscript. We also gratefully acknowledge Dr. M. Busslinger, Dr. Z Grundstrom and Dr. C. Thompson for plasmids and Dr. M. Busslinger for helpful discus- sions. This work was supported by Cancerfonden and the Swedzkh Medical Research Council.

5 References

1 Cogne, M., Lansford, R., Bottaro, A., Zhang, J., Gonnan, J.,

2

3

4

5 6

7

8

9

10

11

12

13

Yo&, E, Cheng, H. L. and Alt, F. W., Celi1994. 77: 737. Pettersson, S., Cook, G. P., Briiggeman, M., Williams, G. T. and Neuberger, M. S., Nature 1990. 344: 165. Dariavach, P., Williams, G. T., Campbell, K., Pettersson, S. and Neuberger, M. S., Eur. J . Immunol. 1991. 21: 1499. Lieberson, R., Giannini, S. L., Birshtein, B. K. and Eckhardt, L. A., Nucleic Acids Res. 1991. 19: 933. Madisen, L. and Groudine, M., Genes Dev. 1994. 8: 2212. Michaelson, J. S., Giannini, S. L. and Birshtein, B. K., Nucleic Acids Res. 1995. 23: 975. Arulampalam, V., Grant, P. A., Samuelsson, A., Lendahl, U. and Pettersson, S., Eur. J . Immunol. 1994.24: 1671. Giannini, S. L., Singh, M., Calvo, C. F., Ding, G. and Birsh- tein, B. K., J. Immunol. 1993. 150: 1772. Grant, P. A., Thompson, C. B. and Pettersson, S., EMBO J . 1995. 14: 4501. Arulampalam, V., Furebring, C., Samuelsson, A., Lendahl, U., Lundkvist, I., Borrebaek, C. and Pettersson, S., Int. Immunol. 1996, in press. Lieberson, R., Ong, J. , Shi, X. and Eckhardt, L., EMBO J . 1995. 14: 6229. Meyer, K. B., Skogberg, M., Margenfeld, C., Ireland, J. and Pettersson, S., Eur. J . Immunol. 1995. 25: 1nO. Walther, C., Guenet, J. L., Simon, D., Deutsch, U., Jostes, B., Goulding, M. D., Plachov, D., Balling, R. and Gruss, P., Genomics 1991. I I : 424.

Eur. J. Immunol. 1996.26: 2499-2507 IgH 3’ enhancer activity in B lineage cells is not blocked by Pax-5 2507

14 Czerny, T., Schaffner, G. and Busslinger, M., Genes Dev. 1993. 7: 2048.

15 Adams, B., Dorfler, P., Aguzzi, A., Kozmik, Z., Urbanek, P., Maurer-Fogy, I. and Busslinger, M., Genes Dev. 1992.6: 1589.

16 Liao, F., Giannini, S. L. and Birshtein, B. K., J. Immunol. 1992.148: 2909.

17 Waters, S. H., Saikh, K. U. and Stavnezer, J., Mol. Cell. Biol. 1989. 9: 5594.

18 Kozmik, Z., Wang, S., Dorfler, P., Adams, B. and Busslinger, M., Mol. Cell. Biol. 1992. 12: 2662.

19 Liao, F., Birshtein, B. K., Busslinger, M. and Rothman, P., J. Immunol. 1994. 152: 2904.

20 Barberis, A., Widenhorn, K., Vitelli, L. and Busslinger, M., Genes Dev. 1990. 4: 849.

21 Urbanek, P., Wang, Z. Q., Fetka, I., Wagner, E. E and Buss- linger, M., Cell 1994. 79: 901.

22 Singh, M. and Birshtein, B. K., Mol. Cell. Biol. 1993.13: 3611. 23 Neurath, M. E, Strober, W. and Wakatsuki, Y., J. Immunol.

24 Neurath, M. E, Max, E. E. and Strober, W., Proc. Natl. Acad.

25 Cook, G . P. and Neuberger, M. S., NucleicAcids Res. 1990.18:

1994. 153: 730.

Sci. USA 1995. 92: 5336.

3565.

26 Mizuguchi, J. O., Tsang, W., Morrison, S. L., Beaven, M. A.

27 Graham, E L. and van der Eb, A. J., Krology 1973. 52: 456. 28 Pettersson, S., Sharpe, M. J., Gilmore, D. R., Surani, M. A.

29 Lane, P., Brocker, T., Hubele, S., Padovan, E., Lanzavecchia,

30 Mason, J. O., Williams, G. T. and Neuberger, M. S., Genes

31 Grant, P. A., Arulampalam, V., Ahrlund-Richter, L. and Pet-

32 Czerny, T. and Busslinger, M., Mol. Cell. Biol. 1995.15: 2858. 33 Maliszewski, C. R., Grabstein, K., Fanslow, W. C., Armitage,

R., Spriggs, M. K. and Sato,T. A., Eur. J. Immunol. 1993.23: 1044.

34 Wortis, H. H., Teutsch, M., Higer, M., Zheng, J. and Parker, D. C., Proc. Natl. Acad. Sci. USA 1995. 92: 3348.

35 Wakatsuki, Y., Neurath, M. E, Max, E. E. and Strober, W., J. Exp. Med. 1994. 179: 1099.

36 Pettersson, S., Pobor, G. and Coutinho, A., Eur. J . Immunol. 1982. 12: 653.

37 Stuber, E., Neurath, M., Calderhead, D., Fell, H. P. and Stro- ber, W., Immunity 1995.2: 507.

and Paul, W. E., J. Immunol. 1985. 137: 2162.

and Neuberger, M. S., Int. Immunol. 1989. 1: 509.

A. and McConnell, E, J. Exp. Med. 1993. 177: 1209.

Dev. 1988.2: 1003.

tersson, s., Nucleic Acids Res. 1992. 20: 4401.

physiological activation of the igh 3′ enhancer in b lineage cells is not blocked by pax-5

Documents