Requirements for selective recruitment of Ets proteins and activation of mb‐1/Ig‐α gene transcription by Pax‐5 (BSAP) (original) (raw)

Journal Article

,

*To whom correspondence should be addressed. Tel: +1 303 398 1398; Fax: +1 303 398 1396; Email: hagmanj@njc.org

Search for other works by this author on:

,

Search for other works by this author on:

,

Search for other works by this author on:

,

Search for other works by this author on:

,

Search for other works by this author on:

,

Search for other works by this author on:

Search for other works by this author on:

Published:

01 October 2003

• PDF
• Split View
• • Article contents
  • Figures & tables
  • Video
  • Audio
  • Supplementary Data
• Cite

Cite

Holly Maier, Rachel Ostraat, Sarah Parenti, Daniel Fitzsimmons, Lawrence J. Abraham, Colin W. Garvie, James Hagman, Requirements for selective recruitment of Ets proteins and activation of mb‐1/Ig‐α gene transcription by Pax‐5 (BSAP), Nucleic Acids Research, Volume 31, Issue 19, 1 October 2003, Pages 5483–5489, https://doi.org/10.1093/nar/gkg785
Close

Navbar Search Filter Mobile Enter search term Search

Abstract

Pax‐5, a member of the paired domain family of transcription factors, is a key regulator of B lymphocyte‐specific transcription and differentiation. A major target of Pax‐5‐mediated activation is the mb‐1 gene, which encodes the essential transmembrane signaling protein Ig‐α. Pax‐5 recruits three members of the Ets family of transcription factors: Ets‐1, Fli‐1 and GABPα (with GABPβ1), to assemble ternary complexes on the mb‐1 promoter in vitro. Using the Pax‐5:Ets‐1:DNA crystal structure as a guide, we defined amino acid requirements for transcriptional activation of endogenous mb‐1 genes using a novel cell‐based assay. Mutations in the β‐hairpin/β‐turn of the DNA‐binding domain of Pax‐5 demonstrated its importance for DNA sequence recognition and activation of mb‐1 transcription. Mutations of amino acids contacting Ets‐1 in the crystal structure reduced or blocked mb‐1 promoter activation. One of these mutations, Q22A, resulted in greatly reduced mb‐1 gene transcript levels, concurrent with the loss of its ability to recruit Fli‐1 to bind the promoter in vitro. In contrast, the mutation had no effect on recruitment of the related Ets protein GABPα (with GABPβ1). These data further define requirements for Pax‐5 function in vivo and reveal the complexity of interactions required for cooperative partnerships between transcription factors.

Received July 22, 2003; Revised and Accepted August 20, 2003

INTRODUCTION

Combinatorial association of DNA binding proteins is an important mechanism for regulating the expression of tissue‐specific genes in differentiated cells. In many cases, cooperative interactions between two or more proteins are important for establishing higher order complexes with increased affinity for DNA. One example of this can be found in developing B cells, involving proteins from the Pax and Ets families of transcription factors.

Members of the Pax family of transcription factors are important for the control of tissue specific transcription during many types of cellular differentiation including brain, eye and lymphoid development (1), and are also implicated in oncogenesis (2–5). Pax proteins bind DNA via a highly conserved bipartite structure called the paired domain, which consists of a β‐hairpin/β‐turn and two helix–turn–helix motifs connected by a linker peptide. The paired domain binds DNA sequences comprised of two half‐sites separated by one turn of the DNA helix, with each half‐site being recognized by one of the two paired domain DNA binding motifs (6). An inherent flexibility of the paired domain enables Pax proteins to bind sequences comprised of very degenerate nucleotide sequences. Therefore, it has been predicted that regulatory mechanisms, including protein–protein interactions, function to enhance the specificity of Pax family members, including Pax‐5.

Ets proteins are a family of transcription factors that control a wide variety of cellular processes including cellular proliferation and differentiation. Ets proteins bind DNA via the ETS domain, a highly conserved winged helix motif shared by all members of the Ets family. Due to the conserved nature of the ETS domain, many family members have very similar patterns of sequence recognition. This creates a potential problem for selective activation of target genes when more than one Ets family member is expressed in a single cell. To compensate for this, the binding specificity and activity of many Ets family members is regulated through protein–protein interactions with other transcription factors (7–9).

Pax‐5 is necessary for B lymphocyte development and is also required for development of the mammalian midbrain segment (10,11). Moreover, it has been suggested that Pax‐5 mediates the commitment of cells to the B lineage, because B cell progenitors exhibit lineage plasticity in the absence of Pax‐5 (12). In part, the ability of Pax‐5 to promote B cell development is due to its role in activating the expression of B lineage specific genes, including mb‐1 (Ig‐α). We have previously shown that Pax‐5 recruits multiple members of the Ets transcription factor family including Fli‐1, Ets‐1 and GABPα (with GABPβ1) to bind a suboptimal Ets binding site (5′CCGGAG) within the early B‐cell specific mb‐1/Ig‐α promoter (between –86 and –64) in vitro (13–16). Several lines of evidence suggest that Pax‐5 and Ets proteins must work together to activate the promoter. First, binding of these proteins (Pax‐5 and Ets‐1) to DNA is highly cooperative (D.Fitzsimmons, in preparation). The absence of either partner results in greatly reduced or undetectable DNA binding, suggesting that ternary complex assembly is necessary for stable binding of these proteins to the mb‐1 promoter. Second, in vivo footprinting of the mb‐1 promoter detected coordinate occupation of Pax‐5 and Ets binding sites in early B cells (17). Third, mutation of either the Ets or Pax‐5 binding site in reporter assays similarly decreased mb‐1 promoter activity in B cells (13). Finally, the requirement of both components for transcriptional activation is strongly implied by studies of endogenous mb‐1 gene activation in retrovirally transduced plasmacytoma cells (16). In this system, transcriptional activation by Pax‐5 is dependent on an unmethylated Ets binding site. Methylation of this site blocks binding of Ets proteins in vitro, prevents ternary complex formation and inhibits transcriptional activation in reporter assays.

The interaction between Pax‐5 and Ets proteins occurs solely between each protein’s DNA binding domain. The crystal structure of the ternary complex of the paired DNA binding domain of Pax‐5 and the ETS domain of Ets‐1 bound to mb‐1 promoter DNA was solved recently (18). The paired domain of Pax‐5 is comprised of two domains—the N‐terminal domain (NTD) and the C‐terminal domain (CTD)—which are connected by a short linker. Both the NTD and CTD are comprised of three α‐helices and make sequence‐specific contacts to DNA through a helix–turn–helix motif. The NTD also contains a β‐hairpin at its N‐terminus that appears to make a number of contacts to the sugar‐phosphate backbone of DNA. The crystal structure revealed that residues in the β‐hairpin and one of the α‐helices in the NTD are responsible for cooperative interactions between Pax‐5 and select Ets family members.

In order to determine the relative importance of residues within the β‐hairpin of the NTD of Pax‐5 in a relatively physiological setting, we used a novel cell‐based system to analyze the effects of point mutations on the activation of endogenous mb‐1/Ig‐α transcription and recruitment of endogenous Ets proteins. Our results suggest that the β‐hairpin plays an essential role in stabilizing Pax‐5 on the mb‐1 promoter, enabling transcriptional activation. In contrast, mutations within the β‐hairpin did not cause a reduction in DNA binding to a different Pax‐5‐DNA binding site, highlighting the differential utilization of motifs within Pax‐5 for binding to distinct nucleotide sequences. Mutation of amino acids predicted to participate in protein–protein interactions resulted in drastic differences in the ability of Pax‐5 to recruit Ets partners. One of these mutations revealed that Pax‐5 differentially recruits two Ets family members, Fli‐1 and GABPα (with GABPβ1). These data lead us to conclude that the β‐hairpin of Pax‐5 is essential for mb‐1 promoter transcriptional activity and that Pax‐5 likely utilizes different amino acid residues to recruit different members of the Ets family to bind DNA. These differences may lead to the preferential recruitment of one Ets protein over another in vivo, leading to optimal transcriptional activity.

MATERIALS AND METHODS

Cell lines and cell culture

µM.10 cells, a clone of the J558LµM cell line (16) were cultured in RPMI 1640 medium containing 10% FBS (Gemini Bioproducts, Woodland, CA), 2 mM l‐glutamine, 50 µg/ml gentamicin, 1× HT Media Supplement (Sigma, St Louis, MO), 0.3 mg/ml xanthine (Sigma) and 1 µg/ml mycophenolic acid (Sigma). The pre‐B cell line 40E1 was cultured in Iscove’s Modified Dulbecco’s Medium (IMDM; Cellgro, Herndon, VA) containing 10% FBS (Gemini), 2 mM l‐glutamine (Invitrogen, Carlsbad, CA), 50 µg/ml gentamicin (Invitrogen), and 50 µM β‐mercaptoethanol (Sigma). The ΦNX retroviral packaging line was kindly provided by Phillipa Marrack (19) and cultured in IMDM containing 10% FBS, 2 mM l‐glutamine and 50 µg/ml gentamicin. All cell lines were grown at 37°C and 6% CO2.

Plasmid constructs

Retroviral vectors for expression of mutated Pax‐5 proteins were prepared as follows. Plasmid pPax‐5‐AgeWT was prepared by mutagenesis of the coding region of Pax‐5 to include an AgeI restriction site at codons 149–151, which changes the nucleotide sequence to CAACCGGTC without changing the encoded amino acid sequence. PCR mutagenesis (15) was performed using the plasmid Pax‐5.S1 (13) as template and primer pairs 5′CTCTTAACCGGTCCCAGCTTCCAGTCACAGCAT and 5′AGGAAGTCTCTGCCCGGAAGC. The amplified fragment was ligated into Pax‐5.S1 cut with BspE1 and Ecl136II, and filled in with Klenow. To produce full length Pax‐5 with paired domain mutations, the resulting plasmid was cut again with Ecl136II and AgeI, filled in with Klenow, and ligated with fragments amplified from previously mutated Pax‐5 sequences (15) using primers 5′CTCATCATGGATTTAGAGAAAAATTATCC and 5′TTG GTTGGGTGGCTGCTGTACT. To generate retroviral vectors, mutated Pax‐5 sequences were excised with Ecl136II and SalI, filled in with Klenow and ligated into the XhoI site of the MSCV2.2‐IRES‐GFPα retroviral vector provided by Dr P. Marrack (20). All plasmids were confirmed by DNA sequencing. The plasmid pCL‐Eco (21) was a kind gift from Dr I. Verma.

Retrovirus production and transduction

Retrovirus production and transduction of µM.10 cells were performed as previously described (16). Transduced cells were analyzed for surface mIgM expression on day 3 post‐transduction. Cells were also sorted for GFP expression using identical gates based on GFP expression and collected for further analysis (RNA and protein).

Flow cytometry and western blotting

For flow cytometry, biotin‐conjugated anti‐IgM was purchased from Caltag Laboratories (Burlingame, CA). Streptavidin‐conjugated allophycocyanin (APC) was purchased from BD Pharmingen (San Diego, CA). Dead cells were excluded by staining with propidium iodide. Flow cytometry was performed on a FACScalibur™ machine (Becton Dickinson, San Diego, CA). For western blotting, rabbit anti‐human Pax‐5 antibody was purchased from Geneka (Montreal, Canada). HRP‐conjugated F(ab)′2 donkey anti‐rabbit IgG was purchased from Amersham Biosciences Corp. (Piscataway, NJ). Whole cell lysates of equivalent cell numbers were lysed in 2× Laemmli Buffer (Bio‐Rad, Hercules, CA). Equivalent amounts of whole cell lysates or nuclear extracts (12 µg) were separated on 10% SDS– PAGE gels. Western blotting was performed as previously described (16).

Preparation of RNA and real‐time RT–PCR

Recovery of RNA from transduced cells and real‐time RT–PCR analysis of β_‐actin_ and mb‐1 transcripts was performed as previously described (16).

Recombinant proteins and electrophoretic mobility shift assays (EMSA)

Recombinant Pax‐5 (1–149) was described previously (15). Recombinant Ets‐1 (331–440) and GABPα/β1 complex was kindly provided by Dr Cynthia Wolberger (Johns Hopkins University). The plasmid for in vitro translation of the Fli‐1 ETS domain was described previously (13). Annealing, labeling of DNA probes and electrophoretic mobility shift assays were performed as previously described (14). Probe sequences were as follows: mb‐1: 5′TCGAAGGGCCACTGGAGCCCATCTCCGGCACGGC and 5′TCGAGCCGTGCCGGAGATGGGCTCCAGTGGCCCT; _I_ϵ: 5′TCGAGCTGA GGGCACTGAGGCAGAGCGGCCCCTA and 5′TCGATAGGGGCCGCTCTGCCTCAGTGCCCTCAGC; M17 promoter Sp1 binding site: 5′TCGAGGCACCTCCTCTTTCTG ACT and 5′TCGAGTCAGAAAGAGGAGGTGCCT. Nuclear extracts were prepared as previously described (16), except Buffer A contained 10 mM HEPES (pH 7.9), 10 mM KCl, 1.5 mM MgCl2, 1 mM DTT plus protease inhibitors (1.5 µM PMSF, 10 µg/ml leupeptin, 10 µg/ml aprotinin) and Buffer C contained 20 mM HEPES (pH 7.9), 0.4 M NaCl, 1.5 mM MgCl2, 1 mM DTT, 0.5% NP40 plus protease inhibitors. For supershifting, anti‐Ets‐2 and anti‐Fli‐1 antisera was purchased from Santa Cruz Biotechnologies, Inc. (Santa Cruz, CA). Anti‐GABPβ1 antisera was provided by Tom Brown (Pfizer Central Research, Groton, CT). Anti‐Pax‐5 antisera is described above.

RESULTS

Transcriptional activation of endogenous mb‐1/Ig‐α genes by Pax‐5 mutants

To assess the relative contribution of amino acids in the NTD of Pax‐5 to the formation of Pax‐5:Ets:DNA ternary complexes, we made use of a recently developed cell‐based system to analyze the ability of Pax‐5 mutant proteins to recruit endogenous Ets partner proteins and activate endogenous mb‐1/Ig‐α genes (16). Briefly, µM.10 is a plasmacytoma cell line expressing three of the four B‐cell receptor (BCR) components necessary for display of IgM on the plasma membrane (mIgM) of B cells: µ Ig heavy chain, λ Ig light chain and B29/Ig‐β. In the absence of the fourth component, mb‐1/Ig‐α, mIgM remains in the cytoplasm. Importantly, µM.10 cells do not express endogenous Pax‐5. Upon induction of mb‐1/Ig‐α transcription by retrovirally expressed Pax‐5 proteins, expression of mIgM can be detected on the cell surface by flow cytometry, providing an indirect measure of mb‐1/Ig‐α transcriptional activation in individual cells. To express wild‐type and mutant Pax‐5 proteins, we generated recombinant mouse stem cell viruses (MSCV) that transcriptionally link expression of enhanced green fluorescent protein (GFP) to expression of Pax‐5 via an internal ribosomal entry sequence [IRES; (20)]. Therefore, flow cytometry enables levels of mIgM expression to be analyzed with respect to the relative expression of Pax‐5 as indicated by GFP fluorescent intensity. Moreover, expression of GFP allows for cell sorting and recovery of relatively homogeneous populations of transduced cells for biochemical analysis.

µM.10 cells were transduced with retroviruses encoding control (cGFP), or full length wild‐type or mutated Pax‐5 proteins, then analyzed for mIgM expression by flow cytometry (Fig. 2A). GFP+ cells were also sorted for analysis of mb‐1 transcripts using real time PCR (Fig. 2B), and western blot analysis of Pax‐5 expression, which confirmed similar expression of Pax‐5 proteins in all transduced cell populations (Fig. 2C). While the cGFP‐expressing cells did not express mIgM, transduction with wild‐type Pax‐5 retroviruses induced mIgM expression on 45.7 ± 0.64% of GFP+ cells. Gly19 is ordered in the crystal structure and forms part of the β‐hairpin (Fig. 1). We predicted that mutation to alanine would have no effect on the hydrogen bonds it can make or in the ϕ–ψ conformation it can adopt. Indeed, mutation of Gly19 to arginine (G19R) did not affect mIgM expression (46.1 ± 0.59%) or mb‐1 transcript levels. Asn21 is important both for folding of the β‐hairpin/β‐turn structure and for DNA binding. Asn21 contacts the mainchain oxygen of Arg59 between helices 2 and 3 of the amino‐terminal subdomain (NTD) and hydrogen bonds to two phosphates of the DNA backbone (Fig. 1). Mutation of Asn21 to alanine (N21A) completely abolished the ability of Pax‐5 to activate mb‐1/Ig‐α transcription, suggesting that the β‐hairpin/β‐turn motif is indeed required for DNA binding in cells. A similar loss of function was observed with mutation of Leu23 to alanine or glycine. Leu23 mediates intermolecular contacts with Ets‐1 via van der Waals interactions with Lys399 of Ets‐1, which assembles an intramolecular salt bridge with Asp398. Leu23 also makes two other potential contacts within Pax‐5: a van der Waals contact with Val60, and a weak interaction with Arg59 (Fig. 1). Pax‐5 L23A produced some IgM+ cells (1.3 ± 0.11%), while L23G resulted in no IgM+ cells. A compound mutation with both Q22A and L23A produced the same phenotype as L23A alone. These data suggest that Leu23 does indeed have an important role for the structure and/or function of the β‐hairpin/β‐turn. A complete lack of activation was observed with mutations F27A or G30S, which are predicted to disrupt the β‐hairpin and β‐turn, respectively. The phenyl ring of Phe27 anchors the antiparallel β‐sheet structure of the β‐hairpin through stacking interactions with a ribose group of the DNA. Gly30 makes contacts with the minor groove of DNA, facilitates formation of the β‐turn, and is equivalent to an amino acid in Pax‐1 that is mutated in the undulated mouse (22). Arg56 forms an intermolecular salt bridge with Asp398 of Ets‐1, which was shown previously to be essential for recruitment by Pax‐5 [Fig. 1; (13)], and contacts the phosphodiester backbone of mb‐1 promoter DNA. Mutation of Arg56 to alanine also eliminated activation of mb‐1, suggesting it plays an important role in Pax‐5 DNA binding and/or protein–protein interactions in cells.

When the three amino acids suspected of participating in recruitment of Ets partners (Gln22, Leu23 and Arg56) were mutated, only Q22A maintained the ability to activate mb‐1 transcription, albeit at significantly reduced levels (Fig. 2A and B). Gln22 has been previously identified as playing a key role in Pax‐5:Ets interactions (14). Gln22 forms hydrogen bonds with Gln336 and Tyr395 of Ets‐1 (Fig. 1). These interactions stabilize changes in the conformation of Tyr395 of Ets‐1, which rotates approximately 90° in order for Ets‐1 to bind stably to the suboptimal binding site of the mb‐1 promoter. Gln22 also contacts the phosphodiester backbone. We previously showed that mutation of Pax‐5 Gln22 to alanine reduces recruitment of Ets‐1 by 75–80% in vitro, but has little or no effect on binding of Pax‐5 by itself to various sites (14). Here, we show that the Q22A mutation significantly reduces the ability of Pax‐5 to activate transcription of endogenous mb‐1 genes. In the µM.10 system, Pax‐5 Q22A activates mIgM expression in only 29.6 ± 2.1% of GFP+ cells (Fig. 2A), a 35.2% reduction compared to wild‐type. In addition, the mean fluorescence intensity (MFI) of the mIgM+ population is reduced compared to that of wild‐type Pax‐5 (34.1 ± 1.4% versus 53.8 ± 1.2%). Real‐time PCR analysis revealed that the transcriptional effect of this mutation is a 64% reduction in mb‐1 transcripts (Fig. 2B), reflecting both the reduced number of mIgM+ cells and the lower MFI.

DNA binding by Pax‐5 mutants

To determine whether the phenotypes observed in transduced µM.10 cells were due to inhibition of DNA binding or an inability to recruit Ets partners, nuclear extracts were prepared from transduced cells sorted for equivalent levels of GFP expression and expanded in culture for 1 week prior to preparation of nuclear extracts. Figure 3A shows binding by proteins in the nuclear extracts to two Pax‐5 binding sites: the mb‐1 ternary complex site and the high affinity binding site of the Ig heavy chain locus Iϵ promoter (23). As expected, both wild‐type Pax‐5 and Pax‐5 G19R bound equivalently to the mb‐1 probe and recruited Ets partners to assemble ternary complexes. Mutations in residues predicted to make contacts with the DNA backbone, including N21A, F27A and G30S, all greatly decreased Pax‐5 binding to the mb‐1 probe, corresponding with the lack of detectable mb‐1 transcripts. Ternary complexes were likewise reduced. Mutations in Leu23 and Arg56 are predicted to disrupt contacts with Ets partner proteins, but also affect intramolecular interactions (made by Leu23; see previous section) or contacts with DNA (Arg56). These properties result in the mutants’ significantly decreased abilities to bind the mb‐1 probe, in addition to reduced ternary complex assembly. Interestingly, the ability of these mutated Pax‐5 proteins to bind the Iϵ probe was not greatly affected, suggesting different amino acid requirements for recognition of this site. Surprisingly, the Q22A mutation eliminated the faster migrating ternary complexes, while slower migrating ternary complexes were enhanced slightly (by 2‐fold). The residual ability of Pax‐5 Q22A to activate mb‐1/Ig‐α transcription is likely due to activities of the slow migrating complexes. Identification of proteins within these ternary complexes is discussed below.

Identification of Ets proteins in µM.10 ternary complexes

Nuclear proteins from cells transduced with wild‐type Pax‐5 retroviruses assembled distinct ternary complexes when incubated with the mb‐1 promoter probe. When transduced with Pax‐5 Q22A, the faster migrating ternary complexes disappeared, while slower migrating complexes were slightly increased. To identify components of these complexes, a supershifting experiment was performed using antibodies against Pax‐5, the Ets proteins Fli‐1 and Ets‐2, and the partner of GABPα, GABPβ1 (supershifting antibodies recognizing GABPα are not available). The experiment revealed that Pax‐5 and Fli‐1 are present together in the faster migrating complexes (Fig. 4). In support of the detection of Fli‐1 together with Pax‐5, we previously detected these complexes in supershifting assays using extracts of cell lines representing pre‐B and immature B cells (13). Fli‐1 contains homologs of each of the amino acids contacted in Ets‐1 by Pax‐5, thus the Q22A mutation that inhibits Ets‐1 binding by Pax‐5 should have similar effects on recruiting Fli‐1. Reactivity with anti‐GABPβ1 antiserum shows that the slower migrating complexes include Pax‐5 and GABPα/β1. Previous analysis with anti‐Ets‐1 antiserum suggested that the slower migrating complexes include Ets‐1 (16). However, this antiserum cross‐reacts with GABPα, and Ets‐1 is undetectable in µM.10 cells by western blotting (H.Maier, data not shown). Together with previous data, these results suggest that the Q22A mutation has a profound effect on recruitment of some (Ets‐1 and Fli‐1) but not all Ets proteins recruited to the mb‐1 promoter by Pax‐5.

Differential recruitment of recombinant Fli‐1 and Ets‐1 versus GABPα/β by Pax‐5 in vitro

We previously showed that Ets proteins including Fli‐1, Ets‐1 and GABPα are recruited by Pax‐5 to bind the mb‐1 promoter in vitro. The molecular basis for this recruitment includes a network of amino acids that are largely shared between these proteins. However, our data suggest differential requirements for Gln22 of Pax‐5 for recruitment of Fli‐1 versus the GABPα/β1 complex. To further address this issue, we performed EMSA to test the relative abilities of recombinant wild‐type Pax‐5 or Pax‐5 Q22A paired domains to recruit ETS domains of Fli‐1, Ets‐1, or GABPα in vitro. The ETS domain of GABPα was tested together with an amino terminal fragment of GABPβ1 (24). These proteins were previously shown to associate stably with each other in the absence of DNA, and the β subunit is necessary for high affinity DNA binding to DNA by the GABPα/β complex in vitro (25–27). DNA binding by Pax‐5 wild‐type and Q22A proteins, which bind the promoter with similar affinities (D.Fitzsimmons, manuscript in preparation), was adjusted to similar levels. Addition of recombinant ETS domains each resulted in the detection of similar levels of ternary complexes with wild‐type Pax‐5 (Fig. 5). In contrast, the Q22A mutation greatly reduced levels of ternary complexes including ETS domains of Fli‐1 (2.5% of wild‐type) or Ets‐1 (7.7% of wild‐type), but had only a very small effect on recruitment of GABPα/β1 (71% of wild type). We conclude that differences in the recruitment of Ets proteins by Pax‐5 are intrinsic to their respective DNA‐binding domains. Moreover, the difference between recruitment of Fli‐1 and Ets‐1 versus GABPα/β defines an additional level of complexity in interactions between Pax‐5 and its Ets partners.

DISCUSSION

Protein–protein interactions between transcription factors are an important mechanism for transcriptional regulation of tissue and developmental stage‐specific genes. We have previously shown that one such interaction, between the paired domain of Pax‐5 and Ets family members, is important for transcriptional activation of the early B‐cell specific mb‐1/Ig‐α promoter (13,14). However, previous studies relied on transient transfection assays to determine the relative importance of the Pax‐5 and Ets binding sites and the use of in vitro binding assays to define important amino acids leading to ternary complex formation. The development of a novel cell‐based system to analyze endogenous gene expression in the context of chromatin (16) allowed us to assay structural requirements for Pax‐5 DNA binding and interactions with Ets proteins in a more physiological manner. The crystal structure of the Pax‐5:Ets‐1:DNA ternary complex was used as a guide to select residues that are involved in DNA‐binding and/or interaction with Ets partner proteins. Mutated full‐length Pax‐5 proteins were retrovirally expressed in a specialized plasmacytoma cell line that allows us to assess transcriptional activation of endogenous mb‐1/Ig‐α genes via flow cytometry. This system provides an additional advantage in that it relies on recruitment of endogenously expressed Ets partner proteins by Pax‐5. Therefore, this system allows for the physiological analysis of the importance of Pax‐5 amino acid contributions to DNA binding and recruitment of Ets proteins and the activation of mb‐1/Ig‐α transcription.

Previous studies have revealed differential requirements for the NTD and CTD of the paired domain of Pax‐5 at different binding sites (6,14,15,28). For example, in the absence of the CTD, Pax‐5 is still able to bind certain nucleotide sequences, such as the site identified within the Ig heavy chain switch region γ2a promoter. Both NTD and CTD regions are required for binding sites within the CD19, Iϵ and mb‐1 promoters, but amino acid sequence requirements are different for each of these sites. For example, Gly30 is essential for binding the mb‐1 and CD19 promoters, but not the Iϵ site. Amino acids of the β‐hairpin, including Asn21 and Phe27, are required for binding mb‐1, but not the CD19 or Iϵ sites. Together, these data suggest differential utilization of motifs within the paired domain, and of individual amino acids, for DNA recognition.

Amino acids of the β‐hairpin/β‐turn motif are highly conserved among members of the Pax family in humans, mice and more distantly related organisms including Caenorhabditis elegans and jellyfish (15). Interestingly, mutations in Pax proteins resulting in disease often reside within this motif (22,29,30). Our previous studies defined the importance of the paired domain β‐hairpin/β‐turn for Pax‐5 DNA binding and Pax‐5:Ets ternary complex formation in vitro (14,15), but we were unable to assess effects of these mutations on Pax‐5 function in a more physiological context. In the present study, we confirmed previous results and showed that many mutations in the β‐hairpin/β‐turn reduce Pax‐5 binding to the mb‐1 promoter and prevent activation of mb‐1 transcription, but do not significantly affect binding to the Iϵ Pax‐5 binding site. Thus, the β‐hairpin/β‐turn is an essential motif required for paired domain function in B cells, and likely, other cell types.

Gln22 of Pax‐5 was previously identified as being particularly important for interactions with Ets‐1, since mutation of Gln22 to an alanine resulted in a 75–80% reduction in Ets‐1 recruitment in in vitro binding assays (14). When introduced into the µM.10 cell system, Pax‐5 Q22A resulted in a 64% reduction in mb‐1 transcript levels. Surprisingly, in vitro DNA binding analysis of nuclear extracts from sorted Pax‐5 Q22A‐expressing cells revealed that formation of only one of two types of Pax‐5:Ets ternary complexes was inhibited by the Q22A mutation (Fig. 3A). Further examination revealed that complexes reduced by the Q22A mutation contain Fli‐1, while the unaffected complexes contain GABPα/GABPβ1 (Fig. 4). These data are only partially explained by our previous structural studies of complexes containing Pax‐5 and Ets‐1. Fli‐1 includes residues corresponding to Gln336, Tyr395, Glu398 and Lys399 of Ets‐1. Therefore, we expect Fli‐1 to assemble nearly identical networks of protein–protein contacts with Pax‐5. In contrast, at the position of Lys399 in Ets‐1, GABPα contains a glycine, which may indirectly impair assembly of the intermolecular salt bridge with Arg56 of Pax‐5. In addition, although GABPα includes a glutamine similar to Gln336 of Ets‐1, this amino acid is complexed with the GABPβ1 protein (24), and therefore may not be available for interaction with Pax‐5.

The question arises as to the identity of the Ets protein(s) that regulate mb‐1 transcription together with Pax‐5 in vivo. Previous studies showed that Pax‐5 assembles ternary complexes in vitro with Fli‐1, Ets‐1 and GABPα/β in pre‐B cell nuclear extracts (13). However, in this study, evidence was obtained for activation of the mb‐1 promoter by Pax‐5 with Fli‐1, GABPα/β, or both Ets proteins in µM.10 cells (which do not express detectable Ets‐1; H.Maier, data not shown). Unfortunately, it is not technically feasible at this time to definitively determine which of the three potential Ets partners functions together with Pax‐5 in vivo. This is due both to the lack of highly specific antibody reagents suitable for chromatin immunoprecipitation (ChIP) of Ets proteins, and to the presence of an additional Ets binding site immediately downstream of the Pax‐5:Ets ternary complex binding site in the mb‐1 promoter. In EMSA assays with nuclear extracts or recombinant Ets proteins, binding to the second site is independent of Pax‐5 binding in vitro. Thus, although ChIP assays may be able to determine whether an Ets protein(s) binds the promoter in vivo, they would be unable to determine which of the two adjacent sites is bound by the immunoprecipitated protein. Our studies suggest that Pax‐5 may be able to activate transcription with more than one kind of Ets protein. This is not entirely surprising in light of previous data showing that the ability to interact with Pax proteins is a highly conserved property of Ets proteins from species as distantly related as humans and C.elegans (15). Further studies are necessary to understand how Pax‐5 and different Ets proteins assemble ternary complexes, and how well the various complexes are able to activate transcription of the mb‐1 promoter. In conclusion, the difference in the molecular mechanisms for recruitment of Ets‐1 and Fli‐1 versus GABPα/β1 by Pax‐5 increases the complexity of how Pax‐5 assembles complexes with Ets proteins, a paradigm for partnerships between factors that are members of extensive families of proteins with similar properties.

ACKNOWLEDGEMENTS

H.M. is supported by National Institutes of Health training grant T32 AI 07405. S.P. was supported by National Cancer Institute grant 5R25‐CA049981‐15. J.H. was supported by Public Health Service grants from the National Institutes of Health (R01 AI37574, R01 AI056322 and P01 AI22295), by a grant from the Rocky Mountain Chapter of the Arthritis Foundation and by generous gifts from the Monfort Family Foundation and Milheim Foundation.

Figure 1. Location of residues in the Pax‐5:Ets‐1:DNA ternary complex discussed in the text. An overview of the structure is represented on the left and an expanded view of the β‐hairpin and the Pax:Ets interface is shown on the right. The ETS domain of Ets‐1 is shown in magenta and the paired domain of Pax‐5 in blue. Residues are labeled with a color denoting whether they come from the ETS domain (magenta) or the paired domain (blue). Hydrogen bonds are represented by green dashed lines and van der Waals interactions are represented by yellow dashed lines. Asn29 and Arg31 were not analyzed in this study but are shown for reference.

Figure 2. Activation of endogenous mb‐1/Ig‐α gene expression by wild‐type and mutant Pax‐5 proteins. (A) Representative flow cytometric analysis of mIgM expression in control (cGFP) and Pax‐5‐transduced (GFP+) µM.10 plasmacytoma cells (16). mIgM expression indicates activation of mb‐1/Ig‐α transcription by Pax‐5. Numbers refer to the percentage of GFP+ cells that are mIgM+ and are the average of three experiments. (B) Real‐time RT–PCR of mRNA from sorted transduced cells (GFP+) for mb‐1 transcripts. Results are the average of three experiments. (C) Representative western blot of sorted GFP+ whole cell lysates (each lane contains protein from equivalent cell numbers) showing equivalent expression of Pax‐5 proteins.

Figure 3. Effects of Pax‐5 mutations on DNA binding and recruitment of endogenous Ets proteins. (A) Gel shift assays using nuclear extracts prepared from sorted cGFP and Pax‐5‐transduced cells incubated with mb‐1 and Iϵ promoter sequences. The M17 promoter Sp1 binding site is a loading control for protein concentration and intactness of proteins. (B) Pax‐5 protein expression in nuclear extracts. A representative western blot of nuclear extract proteins (12 µg) was immunoblotted with Pax‐5 specific antibodies to show relative levels of expression.

Figure 4. Identification of Ets proteins in ternary complexes. mb‐1 probe mobility shift assays using nuclear extracts from wild‐type Pax‐5‐transduced µM.10 plasmacytoma cells (16) were performed as in Figure 3 with the inclusion of antisera specific for Pax‐5, Fli‐1, Ets‐2 or the GABPα partner, GABPβ, as indicated. The asterisk indicates degraded Pax‐5 protein.

Figure 5. Ternary complex assembly by Pax‐5 wild‐type or Q22A paired domains with ETS domains of Fli‐1, Ets‐1, or GABPα/β1. EMSA was performed as previously described (13) using recombinant wild‐type or Q22A Pax‐5 (1–149) proteins, alone or together with ETS domains of Fli‐1 (residues 279–366), Ets‐1 (residues 331–440), or GABPα (residues 320–429). Recombinant GABPα ETS domain was added in equimolar concentrations with GABPβ1 (5–157). After non‐denaturing polyacrylamide gel electrophoresis, DNA binding was quantified using a Molecular Dynamics Typhoon PhosphorImager system.

References

Dahl,E., Koseki,H. and Balling,R. (

1997

) Pax genes and organogenesis.

Bioessays

,

19

,

755

–765.

Chi,N. and Epstein,J.A. (

2002

) Getting your Pax straight: Pax proteins in development and disease.

Trends Genet.

,

18

,

41

–47.

Bennicelli,J.L., Fredericks,W.J., Wilson,R.B., Rauscher,F.J.,III and Barr,F.G. (

1995

) Wild type PAX3 protein and the PAX3‐FKHR fusion protein of alveolar rhabdomyosarcoma contain potent, structurally distinct transcriptional activation domains.

Oncogene

,

11

,

119

–130.

Adshead,J.M., Ogden,C.W., Penny,M.A., Stuart,E.T. and Kessling,A.M. (

1999

) The expression of PAX5 in human transitional cell carcinoma of the bladder: relationship with de‐differentiation.

BJU Int.

,

83

,

1039

–1044.

Iida,S., Rao,P.H., Ueda,R., Chaganti,R.S. and Dalla‐Favera,R. (

1999

) Chromosomal rearrangement of the PAX‐5 locus in lymphoplasmacytic lymphoma with t(9;14)(p13;q32).

Leuk. Lymphoma

,

34

,

25

–33.

Czerny,T., Schaffner,G. and Busslinger,M. (

1993

) DNA sequence recognition by Pax proteins: bipartite structure of the paired domain and its binding site.

Genes Dev.

,

7

,

2048

–2061.

Oikawa,T. and Yamada,T. (

2003

) Molecular biology of the Ets family of transcription factors.

Gene

,

303

,

11

–34.

Sharrocks,A.D. (

2001

) The ETS‐domain transcription factor family.

Nat. Rev.

,

2

,

827

–837.

Verger,A. and Duterque‐Coquillaud,M. (

2002

) When Ets transcription factors meet their partners.

Bioessays

,

24

,

362

–370.

Adams,B., Dorfler,P., Aguzzi,A., Kozmik,Z., Urbanek,P., Maurer‐Fogy,I. and Busslinger,M. (

1992

) Pax‐5 encodes the transcription factor BSAP and is expressed in B lymphocytes, the developing CNS and adult testis.

Genes Dev.

,

6

,

1589

–1607.

Urbanek,P., Wang,Z.Q., Fetka,I., Wagner,E.F. and Busslinger,M. (

1994

) Complete block of early B cell differentiation and altered patterning of the posterior midbrain in mice lacking Pax5/BSAP.

Cell

,

79

,

901

–912.

Nutt,S.L., Heavey,B., Rolink,A.G. and Busslinger,M. (

1999

) Commitment to the B‐lymphoid lineage depends on the transcription factor Pax5.

Nature

,

401

,

556

–562.

Fitzsimmons,D., Hodsdon,W., Wheat,W., Maira,S.M., Wasylyk,B. and Hagman,J. (

1996

) Pax‐5 (BSAP) recruits Ets proto‐oncogene family proteins to form functional ternary complexes on a B‐cell‐specific promoter.

Genes Dev.

,

10

,

2198

–2211.

Wheat,W., Fitzsimmons,D., Lennox,H., Krautkramer,S.R., Gentile,L.N., McIntosh,L.P. and Hagman,J. (

1999

) The highly conserved beta‐hairpin of the paired DNA‐binding domain is required for assembly of Pax‐Ets ternary complexes.

Mol. Cell. Biol.

,

19

,

2231

–2241.

Fitzsimmons,D., Lutz,R., Wheat,W., Chamberlin,H.M. and Hagman,J. (

2001

) Highly conserved amino acids in Pax and Ets proteins are required for DNA binding and ternary complex assembly.

Nucleic Acids Res.

,

29

,

4154

–4165.

Maier,H., Colbert,J., Fitzsimmons,D., Clark,D.R. and Hagman,J. (

2003

) Activation of the early B‐cell‐specific mb‐1 (Ig‐alpha) gene by Pax‐5 is dependent on an unmethylated Ets binding site.

Mol. Cell. Biol.

,

23

,

1946

–1960.

Sigvardsson,M., Clark,D.R., Fitzsimmons,D., Doyle,M., Akerblad,P., Breslin,T., Bilke,S., Li,R., Yeamans,C., Zhang,G. et al. (

2002

) Early B‐cell factor, E2A and Pax‐5 cooperate to activate the early B cell‐specific mb‐1 promoter.

Mol. Cell. Biol.

,

22

,

8539

–8551.

Garvie,C.W., Hagman,J. and Wolberger,C. (

2001

) Structural studies of Ets‐1/Pax5 complex formation on DNA.

Mol. Cell

,

8

,

1267

–1276.

Swift,S., Lorens,J.B., Achacoso,P. and Nolan,G.P. (

1999

) Rapid production of retroviruses for efficient gene delivery to mammalian cells using 293T cell‐based systems.

Curr. Protoc. Immunol.

, Unit 10.28. Suppl. 31.

Hawley,R.G., Lieu,F.H., Fong,A.Z. and Hawley,T.S. (

1994

) Versatile retroviral vectors for potential use in gene therapy.

Gene Ther.

,

1

,

136

–138.

Naviaux,R.K., Constanzi,E., Haas,M. and Verma,I.M. (

1996

) The pCL vector system: rapid production of helper‐free, high‐titer, recombinant retroviruses.

J. Virol.

,

70

,

5701

–5705.

Balling,R., Deutsch,U. and Gruss,P. (

1988

) undulated, a mutation affecting the development of the mouse skeleton, has a point mutation in the paired box of Pax‐1.

Cell

,

55

,

531

–535.

Rothman,P., Si,S., Gorham,B., Glincher,L., Alt,F. and Boothby,M. (

1991

) Identification of a conserved lipopolysaccharide‐plus‐interleukin‐responsive element located at the promoter of the germ‐line ϵ transcripts.

Mol. Cell. Biol.

,

11

,

5551

–5561.

Batchelor,A.H., Piper,D.E., de la Brousse,F.C., McKnight,S.L. and Wolberger,C. (

1998

) The structure of GABPα/β: an ETS domain‐ankyrin repeat heterodimer bound to DNA.

Science

,

279

,

1037

–1041.

Suzuki,F., Goto,M., Sawa,C., Ito,S., Watanabe,H., Sawada,J. and Handa,H. (

1998

) Functional interactions of transcription factor human GA‐binding protein subunits.

J. Biol. Chem.

,

273

,

29302

–29308.

Thompson,C.C., Brown,T.A. and McKnight,S.L. (

1991

) Convergence of ets‐ and notch‐related structural motifs in a heteromeric DNA binding complex.

Science

,

253

,

762

–768.

Chinenov,Y., Henzl,M. and Martin,M.E. (

2000

) The alpha and beta subunits of the GA‐binding protein form a stable heterodimer in solution.

J. Biol. Chem.

,

275

,

7749

–7756.

Xu,H.E., Rould,M.A., Xu,W., Epstein,J.A., Maas,R.L. and Pabo,C.O. (

1999

) Crystal structure of the human Pax6 paired domain‐DNA complex reveals specific roles for the linker region and carboxy‐terminal subdomain in DNA binding.

Genes Dev.

,

13

,

1263

–1275.

Tassabehji,M., Read,A.P., Newton,V.E., Patton,M., Gruss,P., Harris,R. and Strachan,T. (

1993

) Mutations in the PAX3 gene causing Waardenburg syndrome type I and type 2.

Nature Genet.

,

3

,

26

–30.

Vogan,K.J., Epstein,D.J., Trasler,D.G. and Gros,P. (

1993

) The splotch‐delayed (Sp d) mouse mutant carries a point mutation within the paired box of the pax3 gene.

Genomics

,

17

,

364

–369.

I agree to the terms and conditions. You must accept the terms and conditions.

Submit a comment

Name

Affiliations

Comment title

Comment

You have entered an invalid code

Thank you for submitting a comment on this article. Your comment will be reviewed and published at the journal's discretion. Please check for further notifications by email.

Citations

Views

Altmetric

Metrics

Total Views 717

464 Pageviews

253 PDF Downloads

Since 3/1/2017

Citations

34 Web of Science

×

Email alerts

Citing articles via

• Latest

• Most Read

• Most Cited

More from Oxford Academic