The Partial Homeodomain of the Transcription Factor Pax-5 (BSAP) Is an Interaction Motif for the Retinoblastoma and TATA-Binding Proteins 1

Home , Meinrad Busslinger, Pax genes, Transcription factor

[CANCER RESEARCH (SUPPL,) 59, 1716s 1725s, April 1, 1999] The Partial Homeodomain of the Transcription Factor Pax-5 (BSAP) Is an Interaction Motif for the Retinoblastoma and TATA-binding Proteins 1

Dirk Eberhard and Meinrad Busslinger 2

Research Institute of Molecular Pathology, A-t030 Vienna, Austria

Abstract expression pattern, targeted inactivation of Pax-5 in the mouse germ- line revealed essential functions of this transcription factor in mid- Pax-5 codes for the transcription factor BSAP, which plays an impor- brain and B-cell development (7, 8). Interestingly, the human PAX-5 tant role in midbrain patterning, B cell development, and lymphoma gene is involved together with the immunoglobulin heavy chain locus formation. Pax-5 is known to control gene expression by recognizing its in a recurring t(9;14)(p13;q32) translocation associated with a subset target genes via the NH2-terminal paired domain and by regulating transcription through a COOH-terminal regulatory module consisting of of non-Hodgkin's lymphomas (9-11). Hence, PAX-5 can be activated activating and inhibitory sequences. The central region of Pax-5 contains by gain-of-function mutations to participate as an oncogene in tumor- a sequence with significant homology to the first a-helix of the paired-type igenesis. The recent genetic identification of Pax-5 target genes re- homeodomain. This partial homeodomain has been highly conserved vealed that Pax-5 controls their transcription either as an activator or throughout vertebrate evolution because it is found not only in Pax-5 but repressor depending on the specific regulatory sequence context (6, also in the related Pax-2 and Pax-8 members of the same Pax subfamily. 12). Structure-function analysis, furthermore, demonstrated that Pax-5 Here we report that the partial homeodomain binds the TATA-binding recognizes its target genes via the NH2-terminal paired domain and protein (TBP) and retinoblastoma (Rb) gene product. Both TBP and Rb controls u'anscription through a COOH-terminal regulatory module were shown by coimmunoprecipitation experiments to directly associate consisting of activating and inhibitory sequences (13). In addition, the with Pax-5 in vivo. The conserved core domain of TBP and the pocket central sequences containing the partial homeodomain were also region as well as COOH-terminal sequences of Rb are required for interaction with the partial homeodomain of Pax-5 in in vitro binding shown to contribute to the transcriptional activity of Pax-5 (13). assays. Furthermore, Pax-5 was specifically bound only by the underphos- Here we demonstrate by different protein binding assays that the phorylated form of Rb. These data indicate that Pax-5 is able to contact TBP and the Rb protein directly interact with the transcription factor the basal transcription machinery through the TBP-containing initiation Pax-5 in vivo and in vitro. Deletion analysis, furthermore, identified factor TFIID, and that its activity can be controlled by the cell cycle- the partial homeodomain of Pax-5 as an essential recognition motif for regulated association with Rb. both TBP and Rb. These data suggest therefore that the partial homeodomain controls the activity of Pax-5 by linking it either Introduction through TBP to the basal transcription machinery or through Rb to the The Pax gene family codes for transcription factors that play control of cell proliferation. important roles in embryonic development, cell differentiation, and Materials and Methods human disease. A hallmark of these developmental regulators is their conserved DNA-binding motif, the so-called paired domain. The Expression Constructs. The expression plasmids coding for lull-length mammalian genome contains nine Pax genes that can be grouped into Pax-5 (pKW2T-hBSAP) and Pax-5 (1-268) have been described previously four distinct classes based on their similarity in sequence and expres- (13). Pax5-APD was constructed by subcloning a HindlIi-BamHI fragment sion (reviewed in Ref. 1). Two of these subclasses contain, in addition from pKW-ABSAP-ER (12) into pKW2T (t3). The plasmid pKW2T-Pax5- AHD contains a 25-amino-acid deletion (amino acids 229-253) that was to the paired domain, also a homeodomain as a second DNA-binding generated in pKW2T-hBSAP via PCR-mediated mutagenesis by introducing region. A sequence motif, which is homologous only to the first an XhoI site at the deletion site without affecting amino acids 228 and 254. The a-helix of the homeodomain (2), has been identified in the subfamily FLAG epitope was added at the NH2 terminus of Pax-5 by replacing a 260-bp consisting of Pax-2, Pax-5, and Pax-8 (3). This partial homeodomain HindlII-BamHI fragment of pKW2T-hBSAP with a corresponding PCR prod- has no DNA-binding activity and yet is conserved in members of the uct generated with the oligonucleotide 5'-CCCAAGCTTACCATGGATTA- Pax-2/5/8 family from sea urchin to man (4, 5), which suggests that it CAAGGACGACGATGACAAGTTAGAGAAAAATTA-3' and a corre- constitutes a protein interaction motif. sponding downstream Pax-5 primer. The HindlII-EcoRI insert containing the The Pax-5 gene codes for the transcription factor BSAP, 3 which is full-length Pax-5 cDNA was subsequently recloned in the expression plasmid expressed in the developing midbrain, all of the lymphoid tissues, and pRK7 (3) to generate pRK7-FLAGhBSAP, which was used for transient adult testis of the mouse (reviewed in Ref. 6). Consistent with this transfection in COS-7 cells. The HindlII site in the above oligonucleotide was converted into a BamHI site, and the frill-length FLAG-tagged Pax-5 cDNA was assembled in the prokaryotic expression plasmid pET2a (14) by ligating a Received 9/18/98; accepted 2/1/99. 260-bp BamHI PCR fragment (5' end) together with a 900-bp BamHI-EcoRI 1 Contributed as part of the April 1, 1999 Supplement to Cancer Research, "General Motors Cancer Research Foundation Twentieth Annual Scientific Conference: Develop- cDNA fragment (3' end). The COOH-terminal Pax-5 deletion mutants con- mental Biology and Cancer." This work was supported in part by a Grant from the taining the FLAG epitope were generated by using the corresponding BamHI- Austrian industrial Research Promotion Fund. D. E. was the recipient of a fellowship from EcoRI cDNA fragments of the previously described deletion clones (13) in the the Deutsche Forschungsgemeinschaftand European Community. above ligation reaction. The expression plasmid pET-FLAGhPax5-AHD was 2 To whom requests for reprints should be addressed, at Research Institute of Molec- ular Pathology, Dr. Bohr-Gasse 7, A-1030 Vienna, Austria. Phone: 43-I-797-30-884; Fax: similarly generated by cloning the corresponding BamHI-EcoRI fragment from 43-1-798-71-53; E-mail: [email protected]. pKW2T-Pax5-AHD. The TBP expression constructs were described previ- 3 The abbreviations used are: BASP, B-cell-specific activator protein; TBP, TATA- ously (15), and the E1A (13S) expression vector was obtained by subcloning binding protein; Rb, retinoblastoma; GST, glutathione S-transferase; EtBr, ethidium the respective HindIII-BamHI fragment from pH/3APr-I-Neo-13S (16) into bromide; TFIID, transcription factor IID; EBC1 buffer, 150 mM NaC1, 50 mM Tris-C1 (pH 8.0), l rnM EDTA, 0.2% NP40, I mM DTT, 400 /.ZMNa3VO4, 10 mM NaF, 0.1 mg/ml pKW2T. The GST-Rb (379-928), GST-Rb (C706F), GST-Rb (3,21), and Pefabloc, 5 /xg/ml pepstatin, 5 /xg/ml leupeptin, 5 /xg/ml aprotinin, 2 /xg/ml antipain, 2 GST-Rb (379-792) fusion constructs have been described previously (17). /xg/ml chymostatin, and 2 mM benzamidine hydrochloride; NETN buffer, 20 mM Tris-Cl The GST-p107 (385-1068) plasmid was provided by S. Mittnacht (London). (pH 7.9), 100 mM NaC1, 1 mM EDTA, 0.5% NP40, and 1 mM DTT; buffer BCI00, 20 mM The human Rb expression plasmid was generated by cloning full-length Rb Tris-C1 (pH 8.0), 100 rnM KC1, 0.1 mM EDTA, 5 mM MgC12, 20% glycerol, 1 mM DTE, 0.1 mg/ml Pefabloc; buffer A, 50 mM Tris-C1 (pH 8.0), 400 mM NaC1, 1 rnN EDTA, 0.2% cDNA into the BamHI site of pcDNA3 (Invitrogen). The expression plasmid NP40, 0.5 mg/ml BSA, and 1 mM DTT. pECE-Ap34-HA has been described previously (18). 1716s

Antibodies. A polyclonal rabbit anti-hPax-5 antibody, which was directed protein A-Sepharose (Pharmacia) for 2 h at 4~ After washing with binding against the paired domain of human Pax-5 (amino acids 17-145; Ref. 3), was buffer, the immunoprecipitated proteins were resuspended in 2• SDS sample affinity-purified. The polyclonal rabbit anti-laminin serum was purchased from buffer, eluted from the beads by boiling, and separated by SDS-PAGE. Serotec Ltd. (Oxford, England) and the anti-FLAG M1 and M2 affinity gels Proteins were transferred to an Immobilon-P membrane (Millipore) and mon- from Eastman Kodak Co. (New Haven, CT). The mouse monoclonal anti-TBP itored for the presence of TBP using the monoclonal anti-TBP antibody 3G3 antibody, which recognizes the NH2-terminal region of hTBP, was described (19). Antibodies were revealed by an enhanced chemiluminescence (ECL) previously (19). The polyclonal anti-Rb antibody (C15), which is directed Western blot detection system (Amersham). against the COOH terminus of human Rb, was obtained from Santa Cruz Results Biotechnology Inc. (Santa Cruz, CA). Cell Lines. The monkey kidney cell line COS-7 and the human osteosar- The Transcription Factor Pax-5 (BSAP) Binds to TBP in Vivo. coma cell line U2-OS (20) were grown in high glucose DMEM supplemented Functional analyses (13) and evolutionary sequence comparisons (4, with 10% FCS. The human B cell line BJA-B expressing Pax-5 (21) was 5) revealed that the transcription factor Pax-5 (BSAP) uses different cultured in RPMI 1640 containing 10% FCS. Cell Transfection and Extract Preparation. COS-7 cells were grown on conserved domains to exert its function in gene regulation. Because 6-cm dishes to 80% confluency and then were transfected by Lipofectamine some of these domains may mediate binding of Pax-5 to components (Life Technologies) according to the manufacturer's instructions. Typical of the basal transcription machinery, we have screened for such transfection reactions contained 2 #g of total plasmid DNA and 12 /xl of interactions by a biochemical approach. To this end, a FLAG epitope Lipofectamine. After 48 h, whole-cell lysates were prepared by lysing the cells was inserted at the NH a terminus of Pax-5, whose transcriptional in 300 ~1 of EBC1 buffer for 30 min on ice. Cell debris was removed by activity was not altered by this modification (data not shown). Initial centrifugation at 14,000 rpm for 30 min at 4~ and the supernatant was used binding experiments indicated that at least one component of the basal for immunoprecipitations. U2-OS cell lysates were prepared as described for transcription machinery, the human TBP (23-25), is capable of bind- COS-7 cells. Nuclear extracts of BJA-B cells were prepared according to ing in vitro to the FLAG-tagged Pax-5 protein (see Fig. 2). To Dignam et al. (22). demonstrate that Pax-5 can also interact with TBP in vivo, we over- Purification of Recombinant Proteins from Bacteria. hPax-5 proteins tagged with the FLAG epitope (DYKDDDDK) at the NH 2 terminus were expressed human TBP and the FLAG-tagged Pax-5 protein by tran- expressed in the E. coli strain BL21-pLysS. GST and the GST-Rb and sient transfection in COS-7 cells (Fig. 1A). Extracts from transfected GST-pl07 fusion proteins were expressed in the E. coli strain DH5a. Over- cells were prepared, and the Pax-5 protein was precipitated with either night cultures of bacteria containing the respective expression plasmid were the monoclonal anti-FLAG M1 or M2 antibodies (Fig. 1A). The diluted 1:100 in LB-medium containing 1% glucose and grown at 37~ to an immunoprecipitates were subsequently analyzed by Western blotting absorbance of A6o0 nm of 0.6 before expression was induced by the addition of with a monoclonal antibody specific for the NH2-terminal sequence of 1 mM isopropyl-/3-D-thiogalactopyranoside for 2 h at 37~ Bacteria were TBP (19). A significant proportion of the TBP protein could be lysed in NETN buffer supplemented with protease inhibitors (0.1 mg/ml coprecipitated with Pax-5 (Fig. 1A, Lane 2), which was efficiently Pefabloc, 5 /xg/ml pepstatin, 5 /xg/ml leupeptin, 5 /xg/ml aprotinin, 2 /xg/ml depleted from the extract with the anti-FLAG M2 antibody (data not antipain, 2/xg/ml chymostatin, and 2 mM benzamidine hydrochloride) and 0.1 shown). The related M I antibody is known to bind to the anti-FLAG mg/ml lysozyme. After a 10-rain incubation on ice, the lysate was sonicated for 3 • 20 s. After removal of cellular debris by centrifugation, cell lysates epitope only in the presence of Ca 2+ ions (26). In the absence of containing FLAG-tagged proteins were incubated with the anti-FLAG M2 calcium, this control M I antibody was unable to precipitate Pax-5 affinity gel (Eastman Kodak Co.), and lysates containing GST-fusion proteins and, hence, TBP (Fig. 1A, Lane 3), which demonstrates that the were incubated with glutathione-Sepharose beads (Pharmacia) for 2 h at 4~ At the end of the binding reaction, the beads were extensively washed with t- NETN buffer containing 1% NP40. The anti-FLAG affinity gel was further t~ washed with buffer B [20 mM Tris-C1 (pH 8.0), 0.1 mM EDTA, 5 mM MgCI:, EtBr I~ L9 --A c~-FLAG ~ "g: 20% glycerol, and 700 mM KC1]. Beads were resuspended in 3 volumes of .,.., .,.., ~.1 ~ (~ TGME buffer [20 mM Tris-C1 (pH 8.0), 100 mM NaC1, 1 mM EDTA, 10% c~ t3 ._1 c::: M2 M1 M2 M2 ~ ~ glycerol, and 0.1% NP40] and stored at 4~ Protein-Protein Interaction Assays. For in vitro binding assays, beads IgH- IgH- coated with 2-5/xg of recombinant proteins were incubated for 2 h at 4~ with hTBP - 5/xl of 35S-proteins that were synthesized by a coupled in vitro transcription- translation system (TNT, Promega) in the presence of [35S]methionine. Bind- hTBP - 1 2 3 ing assays were performed either in buffer EBC 1 supplemented with 0.5 mg/ml IgL - BSA, in buffer BC100 containing 0.2% NP40, 0.5 mg/ml BSA or in buffer A. Where indicated, 100/xg/ml EtBr was included in the binding reaction. After 2 3 4 5 washing with 400 volumes of binding buffer, bound proteins were eluted from COS-7 BJA-B the beads by boiling in 2• SDS sample buffer and were applied to SDS- Fig. 1. In vivo interaction of Pax-5 and TBP. A, coimmunoprecipitation of Pax-5 and PAGE. 35S-proteins were detected by autoradiography. TBP from COS-7 cell extracts. Expression plasmids coding for human TBP and a Whole-cell lysates from COS-7 cells were mixed with the indicated anti- FLAG-tagged human Pax-5 protein were transiently transfected into COS-7 cells. Whole- bodies and incubated for 2 h at 4~ with constant rotation. If necessary, protein cell lysates prepared from these COS-7 cells were incubated, in the absence of Caa+ ions, A-Sepharose (Pharmacia) was added for 1 h to collect the immunocomplexes. with the anti-FLAG M2 (Lanes 2, 4, and 5) or MI (Lane 3) antibodies that were covalently cross-linked to agarose. Note that the calcium-dependent M1 antibody can only bind to the After stringent washing with binding buffer, the precipitated proteins were FLAG epitope in the presence of Ca 2+ (26). Where indicated, EtBr was included in the resuspended in 2 • SDS sample buffer, separated on SDS-PAGE, and analyzed binding reaction (33 ~g/ml EtBr in Lane 4; 100/xg/ml EtBr in Lane 5). The precipitated by Western blotting. proteins were separated by 10% SDS-PAGE, transferred to an lmmobilon-P membrane, For coprecipitation of endogenous proteins from BJA-B nuclear extract, 1 and probed with a monoclonal antibody specific for the NH2-terminal sequences of hTBP (19). Lane 1 contained 10% of the total input protein used for immunoprecipitation. The mg of nuclear extract was dialyzed against the buffer BC100. Dialyzed extracts positions of the hTBP and coeluted immunoglobulin heavy (IgH) and light (/gL) chain were supplemented with 0.2% NP40, a variety of protease inhibitors (0.1 proteins of the anti-FLAG antibodies are indicated to the left. B, association of endoge- mg/ml Pefabloc, 5 /xg/ml pepstatin, 5 /zg/ml leupeptin, 5 /xg/ml aprotinin, 2 nous TBP with Pax-5 in human B cells. A nuclear extract (1 rag) prepared from BJA-B /~g/ml antipain, 2/xg/ml chymostatin, and 2 mM benzamidine hydrochloride) cells was incubated with the Pax-5-specific anti-paired domain antibody (Lane 2) or a control anti-laminin antibody (Lane 3) in the presence of 100 /xg/ml EtBr. Antibody and 100/xg/ml EtBr before the affinity-purified polyclonal rabbit anti-hPax-5 complexes were precipitated with protein A-Sepharose and analyzed for the presence of antibody or the polyclonal anti-laminin antibody was added for 5 h at 4~ with TBP as described above. Lane 1 contained 5% (50 /xg) of the nuclear extract used for constant rotation. Immunocomplexes were captured by the addition of 20/xl of precipitation. 1717s

Downloaded from cancerres.aacrjournals.org on September 25, 2021. © 1999 American Association for Cancer Research. INTERACTING PROTEINS OF THE TRANSCRIPTION FACTOR PAX-5 presence of TBP in the immunoprecipitate depends on the specific TBP in BJA-B nuclear extracts suggests, therefore, that Pax-5 inter- interaction with Pax-5. The nuclear extracts used as a starting material acts in vivo with the TBP-containing multiprotein complex TFIID. usually contain some chromosomal DNA, and, therefore, it is con- Pax-5 Interacts with the Conserved Core Domain of TBP. The ceivable that the observed interaction may be mediated by simulta- domain of TBP that mediates the interaction with Pax-5 was next neous binding of both proteins to contaminating DNA. To rule out this identified by analyzing COOH-terminally truncated, in vitro synthe- possibility, the immunoprecipitations with the anti-FLAG M2 anti- sized TBP proteins in a Pax-5 interaction assay (Fig. 2). The associ- body were repeated in the presence of increasing amounts of EtBr, ation of the different mutant proteins with Pax-5 was internally which distinguishes between DNA-mediated and bona fide protein- controlled by the inclusion of full-length TBP in the binding reaction. protein interactions (27). The addition of EtBr to the binding reaction, Recombinant Pax-5 protein--immobilized on agarose beads--was however, did not influence the interaction between Pax-5 and TBP, able to efficiently bind full-length TBP as well as the truncation which confirms the specificity of protein complex formation (Fig. 1A, mutant lacking the last 68 amino acids (Fig. 2B, Lanes 7-9), whereas Lanes 4 and 5). Comparison of the TBP amount in the starting empty control beads failed to retain any TBP polypeptide (Fig. 2B, material (Lane 1; 1/10 of the input protein) and the immunoprecipitate Lanes 4-6). Interestingly, further deletion of the core domain in TBP (Lane 2) revealed that --10% of the TBP protein present in the (1-167) abolished all of the binding, which indicated that Pax-5 transfected cells was complexed with Pax-5. These data indicate, requires the core domain of TBP for stable complex formation (Fig. therefore, that Pax-5 can efficiently bind to the basal transcription 2B, Lane 8). The DNA-binding activity of TBP is known to depend on factor TBP under the in vivo conditions analyzed. the integrity of the entire core domain (31). The observation that the The high protein expression levels achieved in transfected COS-7 very COOH-terminal region of TBP can be deleted without affecting cells could, however, result in spurious protein-protein interactions. Pax-5 binding further demonstrates that the interaction between TBP Hence, we verified complex formation between endogenous Pax-5 and Pax-5 takes place in the absence of DNA binding. and TBP in the human B cell line BJA-B, which expresses both The core domain of TBP has been highly conserved in evolution as proteins at physiological concentrations (Fig. 1B). For this purpose, a it is 80% homologous between yeast and man (23, 28). As shown in polyclonal anti-Pax-5 antibody (3) was used to precipitate the endog- Fig. 2C, the TBP protein of yeast (32-34) is also able to form specific enous Pax-5 protein from nuclear extracts of BJA-B cells, and the complexes with Pax-5. Interestingly, the binding efficiency was sim- presence of TBP in the precipitate was monitored by immunoblotting. ilar for both the yeast and the human TBP proteins (compare Fig. 2, As shown in Fig. 1B (Lane 2), a substantial fraction of the endogenous B and C). We conclude, therefore, that the core domain sequences of TBP was coprecipitated with Pax-5 from BJA-B nuclear extracts. This TBP, which mediate the interaction with Pax-5, have been conserved coprecipitation was, however, not mediated by contaminating DNA, throughout evolution. because the anti-paired domain antibody used is known to completely The Partial Homeodomain of Pax-5 Participates in TBP Bind- block the DNA-binding activity of Pax-5 (3). Furthermore, TBP could ing. Structure-function analyses demonstrated previously (13) that not be precipitated with a control antibody directed against laminin several distinct domains of Pax-5 contribute to its transcriptional (Fig. 1B, Lane 3), thus confirming the specificity of the observed activity (see Fig. 3A). The COOH-terminal sequences of Pax-5 harbor TBP-Pax-5 interaction. Hence, we conclude that Pax-5 forms a stable a potent transactivation domain followed by an inhibitory element. In complex with TBP in BJA-B cells in vivo. It is well established that contrast, a weaker transactivation function was assigned to the internal the majority of the TBP protein is not present in free form in nuclear sequences of Pax-5 that encompass the conserved octapeptide motif extracts but instead is part of several multiprotein complexes includ- (3, 35) and a region with sequence similarity to the NH2-terminal part ing TFIID (reviewed in Refs. 28-30). The association of Pax-5 with of paired-type homeodomains (2, 3). To delineate the domain of Pax-5

A 55 Q 9.5 159 Core 339 Fig. 2. Pax-5 contacts the highly conserved core domain of TBP. A, schematic diagram of the human hTBP TBP protein. The relative positions of the Q-domain f t' t' Pax-5 and the evolutionary conserved core domain are in- Pstl ssp~ stu~ interaction dicated. Arrows denote the two direct repeat se- (98) (167) (271) i u quences of the core domain. COOH-terminallymm- i i i t cated TBP proteins were generated by in vitro i q 1-339 translation of a TBP expression vector that was lin- earized at codons 98, 167, and 271 with Pstl, Sspi, ; I 1-271 and Stul, respectively. The interaction of full-length and truncatedTBP proteins with Pax-5 is summarized 1-167 to the right. B, interaction of the core domain of 1-98 human TBP with Pax-5. In vitro translated 35S-TBPs were incubated in buffer EBC1 with the anti-FLAG M2 affinity gel, which was either directlyused (Con- trol; Lanes 4-6) or first coated with the bacterially expressed FLAG-tagged hPax-5 protein (Pax-5; Lanes 7-9). Protein complexes were collected by Input Control Pax-5 centrifugation and separated by SDS-PAGE followed B / / / / / C o', by fluorographyof the labeled TBP. Full-lengthTBP was added as an internal control protein to each binding reaction, and its position is indicated to the left. Lanes 1-3 contained 10% of the labeled input protein hTBP - (Input) used for the binding assay. C, evolutionary yTBP - conservation of the Pax-5 interactiondomain of TBP. In vitro translated 35S-TBP of yeast (yTBP) was as- 1 2 3 sayed for its ability to bind to immobilizedFLAG- tagged hPax-5 (Lane 1) or to the affinity matrix alone (Lane 2). The signal in Lane 3 corresponds to 10% of the total yTBP protein analyzed. 1 2 3 4 5 6 7 8 9 1718s

A TBP Pax-5 binding FLAG paired octa- homeodomain transactivation inhibitory Fig. 3. TBP interacts with the partial homeodo- tag domain peptide homology domain domain main of Pax-5. A, schematic diagram of Pax-5 de- 1-391 letion mutants. The structural organization of human Pax-5 is shown together with the extent of the various deletions. Numbers refer to the correspond- 1-358 ing amino acid positions of hPax-5 (BSAP; Ref. 3). A FLAG epitope was inserted at the NH 2 terminus of all of the hPax-5 proteins to facilitate purification 911 NNNNNNN'\NN\'N N N tl 1-268 and immobilization on the anti-FLAG M2 affinity gel. For generation of the different constructs, see 1-220 "Materials and Methods." B, analysis of purified hPax-5 proteins. Wild-type Pax-5 and the COOH- 4f terminal truncation mutants were expressed in E. 1-169 coIi and purified from the bacterial lysates by bind- ,NNNNx,'xNNNNx,N', I ing to the anti-FLAG M2 affinity matrix. The immobilized proteins were eluted and analyzed by AHD 12% SDS-PAGE followed by Coomassie blue staining, which indicated that all of the proteins were purified in similar quantities. The Control Lane contained untreated affinity beads. The size of marker proteins (given in kDa) is indicated to the left. The positions of coeluted immunoglobulin c heavy (lgH) and light (IgL) chain proteins are shown. C, the COOH-terminal Pax-5 sequences are kO dispensable for TBP binding. Similar amounts of the different immobilized proteins were assayed in 1oi - .~:~,~ hTBP buffer A for their ability to bind to in vitro synthe- 83 - : ;~ sized 35S-TBP (Lanes 3-7). Proteins were resolved by SDS-PAGE and detected by autoradiography. 1 2 3 4 5 6 7 Untreated M2 affinity matrix was analyzed in Con- trol Lane 2. The input in Lane 1 con'esponds to 10% of the total 35S-proteinused. D, TBP binds to O the partial homeodomain of Pax-5. The interaction of fnll-length Pax-5 and indicated deletion mutants with 35S-TBP was analyzed as described in C.

21 .... ; ~ 1 2 3 4 5

that mediates the interaction with TBP, we expressed full-length the same interaction domain in a variety of transcription factors (37, Pax-5 and a series of COOH-terminal truncation mutants (Fig. 3A) in 38, 45, 46). Given the association with TBP, we next investigated E. coli, affinity-purified these proteins (Fig. 3B), and subsequently whether Pax-5 is also able to interact with Rb. To this end, in vitro used them for in vitro binding assays with 3sS-TBP (Fig. 3C). In synthesized, radiolabeled Pax-5 protein was incubated with a bacte- agreement with the result of Fig. 2, TBP was bound by immobilized rially expressed GST-Rb (379-928) which consisted of GST linked to Pax-5 protein of full length (Lane 3) but not by untreated control the pocket region and COOH-terminal sequences of Rb (see Fig. 6.4). beads (Lane 2). Interestingly, the COOH-terminal sequences of Pax-5 As shown in Fig. 4A, Pax-5 was coprecipitated together with the could be eliminated up to amino acid 268 without affecting TBP GST-Rb fusion protein bound to glutathione-Sepharose beads. In binding (Lanes 4 and 5). Hence, the COOH-terminal transactivating contrast, neither GST nor the empty beads (control) were able to sequences as well as the inhibitory element of Pax-5 do not contribute interact with Pax-5 in this pull-down assay. Although significant, the to TBP binding, in contrast to other transcription factors that fre- binding affinity of Pax-5 was lower than that of the adenovirus E1A quently contact TBP via their transactivation domain (36-41). Dele- protein (Fig. 4A), which is known to efficiently associate with Rb tion of the partial homeodomain in Pax-5 (1-220) reduced, however, (47). In conclusion, Pax-5 and Rb are able to form a protein complex TBP binding to background level (Fig. 3C, Lane 6), which was aiso under these in vitro binding conditions. seen with a polypeptide (1-169) consisting only of the paired domain Deletion mutants of Pax-5 (Fig. 4D) were next analyzed in the GST (Lane 7). Interestingly, precise deletion of the homeodomain in the pull-down assay to identify the interaction domain involved in Rb full-length Pax-5 protein (AHD) was sufficient to prevent the inter- binding. The Pax5-APD and Pax-5 (1-268) proteins bound Rb as action with TBP (Fig. 3D, Lane 5). We conclude, therefore, that the efficiently as the fllll-length protein, indicating that both the DNA- partial homeodomain of Pax-5 and not its transactivation region is binding function of the paired domain and the COOH-terminal regu- involved in TBP binding. Hence, these data identify the partial ho- latory sequences of Pax-5 are dispensable for the interaction with Rb meodomain as a protein interaction motif which has been conserved in (Fig. 4B). In contrast, deletion of the partial homeodomain strongly different members of the Pax-2/5/8 family throughout vertebrate interfered with Rb binding (Fig. 4B). Hence, the homeodomain se- evolution (5). quences of Pax-5 constitute an interaction motif not only for the TBP The Partial Homeodomain of Pax-5 also Mediates Binding to but also for the Rb protein. the Rb Protein. The core domain of TBP is known to share signif- The proteins p 107 and p130 are structurally and functionally related icant sequence homology with the pocket domain of the Rb tumor to Rb but nevertheless differ from Rb in their association with distinct suppressor protein (37), which interacts with and, thus, controls a transcription factors (reviewed in Refs. 42 and 44). We, therefore, variety of transcription factors (reviewed in Refs. 42-44). Consistent assessed the specificity of the Pax-5 interaction with Rb family with this sequence similarity, TBP and Rb frequently associate with members, using the GST pull-down assay. Although bacterially ex- 1719s

Pax-5 - Pax-5 - Pax-5 -

Fig. 4. Pax-5 interacts with the Rb protein via its partial homeodomain. A, association of Pax-5 with Pax5-APD - the Rb protein. Bacterially expressed GST or EIA - E1A - GST-Rb (379-928) protein (2 /xg each), bound to glntathione-Sepharose, were incubated with in vitro translated 35S-Pax-5 or E1A (13S) protein in buffer 1 2 3 4 8 9 BCI00. After stringent washing of the Sepharose beads, the bound proteins were eluted by boiling, analyzed by SDS-PAGE, and detected by autora- Pax5-AHD - diography. Lane 1 contained 10% of the input protein. Untreated control beads were used in Lane 2. B, the homeodomain of Pax-5 is essential for complex formation with Rb. 35S-labeled full-length or mutant Pax-5 proteins were synthesized by in vitro Pax-5 - translation and then assayed for binding to GST-Rb (1-268) (379-928) as described in A. C, selective interaction of Pax-5 with members of the pocket domain protein family. 35S-labeled Pax-5 and E1A proteins 5 6 7 was analyzed for binding to GST-Rb (379-928) and GST-p107 (385-1068) as described in A. Sim- ilar amounts of both GST fusion proteins were used homeodomain in the binding assay, as shown by the equally effi- D homology cient complex formation of E1A with GST-Rb and GST-pl07. D, schematic diagram of the Pax-5 deletion mutants. The different domains and corre- Pax-5 sponding amino acid residues of Pax-5 are shown together with the extent of the different deletions. ~,x ~ Pax5-APD

Pax5-AHD

Pax-5(1-268)

pressed GST-pl07 and GST-Rb bound to the control E1A protein the less abundant underphosphorylated form of Rb, and control beads with equal efficiency, the p107 fusion protein was incapable of failed to bind any Rb protein (Fig. 5A). These data demonstrate, interacting with Pax-5 in contrast to Rb (Fig. 4C). These data indicate, therefore, that Pax-5 preferentially interacts with the active, under- therefore, that Pax-5 is able to selectively associate with Rb and thus phosphorylated form of Rb. to discriminate between different members of the Rb protein family in To verify this protein interaction in vivo, we expressed the FLAG- analogy to the E2F transcription factors (48-50). tagged Pax-5 protein together with wild-type Rb or a phosphorylation- Pax-5 Selectively Binds to the Underphosphorylated Form of incompetent Rb mutant (Ap34; Ref. 18) in transiently transfected Rb in Vivo. The activity of Rb is regulated by differential phospho- COS-7 cells. Pax-5-Rb protein complexes were precipitated with ryladon during the cell cycle (51-53). Rb is predominantly found in anti-FLAG antibodies from COS-7 cell extracts and detected by the active, underphosphorylated form in quiescent (Go) cells as well as Western blot analysis with an Rb-specific antibody. As shown in Fig. in the early G 1 phase of the cell cycle. At the transition between G1 5B, only the underphosphorylated form of the wild-type Rb protein and S phase, Rb is rapidly phosphorylated and maintains this hyper- was coprecipitated by the anti-FLAG M2 antibody (Lane 3), whereas phosphorylated, inactive configuration throughout the remainder of the calcium-dependent control M1 antibody (26) failed to precipitate the cell cycle. A variety of cellular transcription factors as well as Pax-5, and thus Rb, in the absence of Ca 2+ (Lane 2). Moreover, a many viral oncoproteins are known to preferentially interact with the significant increase of coprecipitated Rb protein was obtained with underphosphorylated form of Rb (reviewed in Refs. 42 and 44). To transfected COS-7 cells expressing the Rb mutant Ap34, (Lane 4), investigate the effect of Rb phosphorylation on Pax-5 binding, we which can no longer be phosphorylated by cyclin-dependent kinases tested the endogenous wild-type Rb isoforms of the human osteosar- (18). We conclude, therefore, that Pax-5 associates specifically in vivo coma cell line U2-OS (20) for their ability to interact with Pax-5. with the underphosphorylated Rb protein, suggesting that in prolifer- Extracts of U2-OS cells were incubated and subsequently precipitated ating cells this interaction is regulated by differential phosphorylation with recombinant Pax-5 protein immobilized on affinity beads. As of Rb. shown in Fig. 5A (Lane 1), cycling U2-OS cells predominantly Pax-5 Interacts with the Pocket Domain and COOH-Terminal contained the hyperphosphorylated, more slowly migrating form of Sequences of Rb. The sequences of Rb that participate in the Rb. However, the immobilized Pax-5 protein specifically precipitated interaction with Pax-5 were next delineated by analyzing a series 1720s

family play essential roles in development, organogenesis, and disease A (1), little is known about how these developmental regulators exert their function at the transcriptional level. Elucidation of their transcriptional role will depend not only on the identification of critical PO4 - Rt target genes but also on the characterization of partner proteins that Rt regulate the transcriptional activity of Pax proteins. In this study, we 1 2 3 have demonstrated that TBP and Rb directly interact with Pax-5 in vivo and in vitro, possibly coupling this regulator to the basal tran- U2-OS scription machinery and control of cell proliferation, respectively. The partial homeodomain of Pax-5 is essential for the binding of both TBP B o~-FLAG and Rb and, thus, constitutes a protein interaction motif. In agreement / / with this finding, the partial homeodomain is dispensable for DNA binding of Pax-5 (3) and is, furthermore, present in different members of the Pax-2/5/8 family from sea urchin to man (4, 5). Hence, this PO 4 - Rb protein interaction motif has been highly conserved in evolution RI~ despite the fact that it is homologous only to one third of the entire homeodomain of other Pax proteins (Fig. 7). The homeodomain 1 2 3 4 homology region encompasses helix I but excludes the sequences of the NH2-terminal arm, helix II, and the 'recognition' helix III (Fig. 7). COS-7 Other transcription factors are also known to use their homeodomain Fig. 5. Selective interaction of Pax-5 with the underphosphorylatedform of Rb. A, for interacting with TBP or Rb. Oct-l/2 (15) and Msx-1 (58) both formation of Rb-Pax-5 complexes in extracts from U2-OS osteosarcomacells. Whole-cell lysates (100/xg) were incubated with either the untreated anti-FLAG M2 affinity matrix associate with TBP through their homeodomain, although in the case (Lane 2) or the M2 affinity gel coupled to purified FLAG-tagged hPax-5 protein (Lane 3). of Msx- 1, this interaction appears to be mediated by the NH2-terminal Bound proteins were precipitated, separated by 6% SDS-PAGE, and analyzedby Western arm sequences that are absent in Pax-5. Recently, another member of blotting with a polyclonalanti-Rb antibodyrecognizing the COOH-terminal Rb sequences (C15; Santa Cruz BiotechnologyInc.). Lane 1 contained 10% (10/zg) of the total input the Pax family, Pax-3, has been shown to bind to Rb through helix I protein. The positions of the underphosphorylated and hyperphosphorylated (P04) Rb and II of its homeodomain (59), which suggests that this protein uses proteins are indicated to the left. B, in vivo interaction of Pax-5 and Rb proteins. Expression plasmids directing the synthesis of FLAG-tagged hPax-5 (Lanes 1-4) and of the same interaction motif identified here for Pax-5. It is, therefore, the wild-typeRb protein (Lanes 1-3) or the phosphorylation-incompetentRb mutant Ap34 likely that all of the Pax proteins associate with Rb and/or TBP via (Ref. 18; Lane 4) were transiently transfected into COS-7 cells. Cell lysates were helix I except for Pax-1 and Pax-9, which lack homeodomain se- immunoprecipitated with the anti-FLAG M2 antibody (Lanes 3 and 4) or the Ca2+- dependent M1 antibody (Ref. 26; Lane 2), as described in the legend to Fig. 1. The quences (1). Our mutational analysis of Pax-5 has ruled out a contri- immunoprecipitated proteins were analyzed by SDS-PAGE and Western blotting using bution of the NH2-terminal paired domain and COOH-terminal trans- the Rb-specific antibody C15. The positions of the differentially phosphorylated Rb activation region to the binding of TBP and Rb. These experiments proteins are indicated to the left. Lane I contained 10% of the input cell lysate. did not, however, address a possible involvement of the conserved octapeptide of Pax-5, which has indeed been implicated in Rb and TBP binding by recent deletion analysis (D. E., data not shown). of mutant Rb proteins (Fig. 6A) in the GST pull-down assay. Hence, the partial homeodomain of Pax-5 is necessary but not suffi- Mutant GST-Rb proteins were expressed in E. coli, purified on cient for the interaction with these two proteins. glutathione-Sepharose, and quantitated by SDS-PAGE, as shown The Pax-5-binding Domains of TBP and Rb. TBP and Rb fre- in Fig. 6B. These proteins were subsequently analyzed for their quently recognize the same or overlapping sequences in different potential to bind in vitro synthesized, radiolabeled Pax-5 or control transcription factors such as PU.1 (37), E2F-1 (38, 46), c-Myc (45) E1A protein. As shown in Fig. 6C (Lane 4), the naturally occurring and E1A (45). The association with the partial homeodomain of Pax-5 Rb mutation C706F (54) significantly reduced the interaction of Rb is yet another manifestation of the ability of TBP and Rb to contact the with Pax-5 in comparison with GST-Rb (379-928) containing same interaction motif. The reason for the similar binding specificity wild-type pocket domain and COOH-terminal sequences (Lane 3). resides in the significant sequence similarity that is shared between Deletion of the entire exon 21 sequences (55), which code for one the COOH-terminal core domain of TBP and the pocket region of Rb third of the subdomain B of the Rb pocket, resulted in complete (37). These homology regions of TBP and Rb are also essential for the loss of Pax-5 binding (Lane 5). Interestingly, truncation of the interaction with Pax-5. Different transcription factors can, however, COOH-terminal sequences in GST-Rb (379-792) also reduced be distinguished by their precise requirement for interacting se- Pax-5 binding to the background level seen with GST alone, quences in TBP and Rb. The entire core domain of TBP is necessary indicating that the pocket domain of Rb is not sufficient to mediate for the association of Oct-l/2 (15) and the Zta transactivator of the the interaction with Pax-5 (compare Lanes 2 and 6). In contrast, the EBV virus (60), whereas the COOH-terminal 68 amino acids are E1A protein bound with equal affinity to GST-Rb regardless of the dispensable for the binding of Pax-5, similar to the interaction with presence and absence of the COOH-terminal sequences (Fig. 6C, p53 (41), c-Rel (61), and E1A (36). Likewise, the pocket domain Lanes 9 and 12), whereas both mutations in the pocket domain (A/B) of the Rb protein is sufficient for binding of most cellular severely impaired the interaction of Rb with E1A (Lanes 10 and transcription factors and viral oncoproteins that contain the charac- 11) as described previously (54, 56, 57). Hence, the binding of teristic LxCxE sequence as their Rb interaction motif (reviewed in Pax-5 to Rb depends not only on an intact pocket domain but also Ref. 43). Consistent with the absence of this sequence element, Pax-5 on the COOH-terminal sequences of Rb in clear contrast to tran- additionally requires the COOH-terminal Rb sequences for efficient scription factors such as E1A. binding in analogy to the E2F transcription factors (62, 63). The consensus Rb recognition sequence of the E2F family members con- Discussion sists of 18 amino acids (64, 65) that, interestingly enough, shows some Identification of the Partial Homeodomain of Pax-5 as a Protein homology with the partial homeodomain of the Pax-2/5/8 proteins Interaction Motif. Although the transcription factors of the Pax (Fig. 7) and the corresponding sequences of Pax-3 and Pax-6 (59). 1721s

pocket domain I I Pax-5 Fig. 6. Binding of Pax-5 to the pocket domain Rbl ~\\\\\\\\~ ~',,\\\\\\\\~ I interaction and COOH-terminal sequences of Rb. A, schematic 1 393 573 645 772 928 diagram of the GST-Rb fusion proteins (17). The structural organization of Rb is shown together GST-Rb (379-928) mr~,.-~i ~,,,\\\\\\\\.~ [~.,,,\\~.\\\\"~ +++ with the relevant mutations: i.e., the Cys-to-Phe substitution at amino acid 706 (C706F), the deletion of exon 21 (AEx21), and the truncation of COOH-terminal sequences. The interaction of the GST-Rb (C706F) different GST-Rb proteins with Pax-5 is summa- 706 rized to the right, B, analysis of the purified GST-Rb proteins. Bacterially expressed GST and GST-Rb (AEx21) GST-Rb fusion proteins were purified on glutathi- 703 737 one-Sepharose and analyzed by SDS-PAGE followed by staining with Coomassie blue. Asterisks denote the recombinant proteins of correct size. GST-Rb (379-792) ~:~" &\\\\\\\\~ ~\\\\\\\\"~1 The migration of molecular mass standards (sizes 379 792 given in kDa) is shown in Lane M. C. GST pull- down assay. In vitro translated Pax-5 protein was incubated with 2/xg of GST (Lane 2) or GST-Rb GST-Rb / GST-Rb fusion proteins (Lanes 3-6) in buffer EBC1 sup- 8 / / C plemented with 0.1% NP40. After stringent washing, bound proteins were ehited by boiling, separated by SDS-PAGE, and analyzed for the presence ~,~ : ...... kD of Pax-5 by immunoblotting with a polyclonal anti- ~ ~ ~'~ ~( 101 Pax-5 antibody. In parallel, the same binding con- Pax-5 ditions were used for studying the interaction of in vitro translated ssS-E 1A protein (13S isoform) with the GST-Rb fusion proteins (Lanes 9-12). The 1 2 3 4 5 6 35S-E1A protein was detected by auto-radiography. ! ...... - 51 Lanes 1 and 7 contained 10% of the input Pax-5 or E1A protein, respectively. - EIA

7 8 9 10 11 12

Functional Significance of the Interaction of Pax-5 with TBP binding site in the -30 promoter region (66). Transcription of the and Rb. The recent characterization of Pax-5 target genes revealed CD19 gene is entirely lost in B-lymphocytes of Pax-5 (-/-)mice (8), that this transcription factor can act as an activator or repressor which suggests that Pax-5 functions to recruit the basal transcription depending on the specific regulatory sequence context (6, 12). One of machinery to the CD19 promoter. The identification of an interaction the identified target genes is CD19 which codes for a B-cell surface between Pax-5 and TBP now suggests a direct mechanistic link, protein and contains, instead of a TATA box, a high-affinity Pax-5- because TBP is an essential component of the basal transcription factor TFIID, which mediates transcription initiation (29, 30). Genetic and functional analyses of Pax-5 have thus far failed to 228 partial homeodomain 254 provide firm evidence that Rb is involved in controlling the activity of Pax-5 (AHD) KQMRGDL ...... deleted ...... EPIKPEQT Pax-5 this transcription factor. Hence, it is more difficult to speculate about

Pax-2 the functional significance of the Rb-Pax-5 interaction. Rb is known Pax-8 KHLRTDA ~ P Q SPSHTKGEQG to play a dual role at both the transcriptional and cellular level, as it

, [ I ' ' j J , can simultaneously repress the transcription of genes involved in cell Pax-3 cycle progression and activate the expression of genes promoting Pax-6 terminal differentiation (reviewed in Refs. 42-44). Rb exerts its t i ~ :: , ] J IiJ i repression function by different mechanisms. The interaction with Rb is known to sequester and inactivate transcription factors such as E2F E2F-I LDYHFG:LE 'G~IR ...... iD L F'D by masking their transactivation domain and, thus, preventing access E2F-2 DDYLWG~I~ E2F-3 GDYLLS~G of coactivators and basal transcription factors such as TFIID (46, 64). E2F-4 HDYIYN~D 9nlE~vc ...... iD, r~ In addition, Rb can actively repress the activity of several transcrip- E2F-5 DDYNFN~ ~.~.'~vc- 9 . - .... i ~Lr:~ i tion factors bound to the same promoter by preventing their interac- Consensus ....DY ir. E .i'~..- ..... iDLF~ tion with the transcription machinery (67). Moreover, the association of Rb with certain transcription factors results in the recruitment of Fig. 7. Sequence conservation of the partial homeodomain in members of the Pax-2/5/8 family. The first 40 amino acids of the homeodomain of Pax-3 (78) and Pax-6 (79) are histone deacetylase, which is able to silence entire control regions by compared with the respective sequences of the human Pax-2 (80), Pax-5 (BSAP; Ref. 3), promoting the formation of inactive chromatin (68-70). Rb also and Pax-8 (81) proteins. Gray overlay highlights those amino acids of the Pax-215/8 appears to repress the transactivation function of Pax-3, at least in proteins that are identical to the corresponding residues of the Pax-3 and/or Pax-6 sequences. Brackets denote helix I and II of the paired-type homeodomain as determined transiently transfected cells (59). By analogy, it is, therefore, conceiv- by X-ray crystallographic analysis (82). The first 9 amino acids of Pax-3 and Pax-6 able that Rb is involved in silencing the 3' enhancer of the immuno- constitute the "NH2-terminal arm" of the homeodomain. The extent of sequence deletion in the Pax-5 (AHD) protein is indicated together with the respective amino acid positions. globlin heavy chain locus as well as in repressing the lymphoid- The Rb-binding domain present in the COOH-terminal transactivation region of E2F specific PD-1 and J-chain genes, which are known to be under transcription factors (64, 65) shows some sequence homology with the paired-type negative control by Pax-5 (6, 12). In addition to its cell cycle function, homeodomain of Pax proteins, as shown by the sequence alignment according to Wiggan et al. (59). Dashes in the E2F sequence indicate a gap of 6 amino acids that was Rb also plays an important role in development, as Rb (-/-) embryos introduced for optimal alignment. exhibit defective differentiation in several tissues including the lens, 1722s

Downloaded from cancerres.aacrjournals.org on September 25, 2021. © 1999 American Association for Cancer Research. INTERACTING PROTEINS OF THE TRANSCRIPTION FACTOR PAX-5 nervous system, and hematopoietic precursor cells (71-73). Consist- 18. Hamel, P. A., Gill, R. M., Phillips, R. A., and Gallie, B. L. Transcriptional repression ent with its differentiation function, Rb associates and cooperates with of the E2-containing promoters EIIaE, c-rnyc, and RB1 by the product of the RB1 gene. Mol. Cell. Biol., 12: 3431-3438, 1992. lineage determination factors such as MyoD (74, 75) and members of 19. Brou, C., Chaudhary, S., Davidson, I., Lutz, Y., Wu, J., Egly, J-M., Tora, L., and the C/EBP family (76, 77). Interestingly, Rb is able to potentiate the Chambon, P. Distinct TFIID complexes mediate the effect of different transcriptional activators. EMBO J., 12: 489-499, 1993. DNA-binding and transactivation function of C/EBP proteins during 2o. Lee, W-H., Shew, J-Y., Hong, F. D., Sery, T. W., Donoso, L. A., Young, L-J., terminal adipocyte differentiation (76, 77). Consequently, it is also Bookstein, R., and Lee, E. Y-H. P. The retinoblastoma susceptibility gene encodes a possible that Rb positively regulates the activity of Pax-5 during nuclear phosphoprotein associated with DNA binding activity. Nature (Lond.), 329: B-cell and/or CNS development. Finally, it is important to note that 642-645, 1987. 21. Barberis, A., Widenhorn, K., Vitelli, L., and Busslinger, M. A novel B-cell lineage- Pax-5 is activated as an oncogene by a specific chromosomal trans- specific transcription factor present at early but not late stages of differentiation. location in a subset of non-Hodgkin's lymphomas (9-11). Hence, it Genes Dev., 4: 849-859, 1990. Dignam, J. D., Lebovitz, R. M., and Roeder, R. G. Accurate transcription initiation by will be interesting to see whether the deregulated expression of the 22. RNA polymerase lI in a soluble extract from isolated mammalian nuclei. Nucleic oncoprotein Pax-5 contributes to tumor formation by interfering with Acids Res., 11: 1475-1489, 1983. the normal function of the tumor suppressor protein Rb in analogy to 23. Hoffmann, A., Sinn, E., Yamamoto, T., Wang, J., Roy, A., Horikoshi, M., and Roeder, R. G. Highly conserved core domain and unique N terminus with presump- viral oncoproteins (42, 56, 57). tive regulatory motifs in a human TATA factor (TFIID). Nature (Lond.), 346: 387-390, 1990. Acknowledgments 24. Kao, C. C., Lieberman, P. M., Schmidt, M. C., Zhou, Q., Pei, R., and Berk, A. J. Cloning of a transcriptionally active human TATA binding factor. Science (Wash- ington DC), 248: 1646-1650, 1990. We thank L. Tora (Strasbourg) for providing the anti-TBP antibody, 25. Peterson, M. G., Tanese, N., Pugh, B. F., and Tjian, R. Functional domains and S. Mittnacht (London) for the GST-Rb and GST-p107 plasmids, T. Wirth upstream activation properties of a cloned human TATA binding protein. Science (Wtirzburg) for TBP expression vectors, P. Hamel (Toronto) for the pECE- (Washington DC), 248: 1625-1630, 1990. Ap34-HA plasmid, H. Stunnenberg (Nijmegen) for E1A expression vectors, 26. Prickett, K. S., Amberg, D. C., and Hoop, T. P. A calcium-dependent antibody for identification and purification of recombinant proteins. Biotechniques, 7: 580-589, and M. Cotten and P. Pfeffer for critical reading of the manuscript. t989. 27. Lai, J. S., and Herr, W. Ethidium bromide provides a simple tool for identifying References genuine DNA-independent protein associations. Proc. Natl. Acad. Sci. USA, 89: 5658-6962, 1992. 1. Mansouri, A., Hallonet, M., and Gruss, P. Pax genes and their roles in cell differen- 28. Heruandez, N. TBP, a universal eukaryotic transcription factor? Genes Dev., 7." tiation and development. Curr. Opin. Cell Biol., 8: 851-857, 1996. 1291-1308, 1993. 2. Krauss, S., Johansen, T., Korzh, V., and Fjose, A. Expression of the zebrafish paired 29. Goodrich, J. A., and Tjian, R. TBP-TAF complexes: selectivity factors for eukaryotic box gene pax[zf-b] during early neurogenesis. Development (Camb.), 113: 1193- transcription. Curr. Opin. Cell Biol., 6: 403-409, 1994. 1206, 1991. 30. Burley, S. K., and Roeder, R. G. Biochemistry and structural biology of transcription 3. Adams, B., Drrfler, P., Aguzzi, A., Kozmik, Z., Urbfinek, P., Maurer-Fogy, I., and factor IID (TFIID). Annu. Rev. Biochem, 65: 769-799, 1996. Busslinger, M. Pax-5 encodes the transcription factor BSAP and is expressed in B 31. Horikoshi, M., Yamamoto, T., Okhuma, Y., Weil, P. A., and Roeder, R. G. Analysis lymphocytes, the developing CNS, and adult testis. Genes Dev., 6:1589-1607, 1992. of structure-function relationship of yeast TATA box binding factor TFIID. Cell, 61: 4. Czerny, T., Bouchard, M., Kozmik, Z., and Busslinger, M. The characterization of 1171-1178, 1990. novel Pax genes of the sea urchin and Drosophila reveal an ancient evolutionary 32. Horikoshi, M., Wang, C. K., Fujii, H., Cromlish, J. A., Weil, P. A., and Roeder, R. G. origin of the Pax2~5~8 family. Mech. Dev., 67: 179-192, 1997. Cloning and structure of a yeast gene encoding a general transcription initiation factor 5. Pfeffer, P. L., Gerster, T., Lun, K., Brand, M., and Busslinger, M. Characterization of TFIID that binds to the TATA box. Nature (Lond.), 341: 299-303, 1989. three novel members of the zebrafish Pax2~5~8 family: dependency of Pax5 and Pax8 33. Schmidt, M. C., Kao, C. C., Pei, R., and Berk, A. J. Yeast TATA-box transcription expression on the Pax2.1 (noi) function. Development (Camb.), 125: 3063-3074, factor gene. Proc, Natl. Acad. Sci. USA, 86: 7785-7789, 1989. 1998. 34. Cavallini, B., Faus, I., Matthes, H., Chipoulet, J. M., Winsor, B., Egly, J.-M., and 6. Busslinger, M., and Nutt, S. L. Role of the transcription factor BSAP (Pax-5) in B-cell Chambon, P. Cloning of the gene encoding the yeast protein BTFIY, which can development. In: 1. G. Monroe and E. V. Rothenberg (eds.), Molecular Biology of substitute for the human TATA box-binding factor. Proc. Natl. Acad. Sci. USA, 86: B-Cell and T-Cell Development, pp. 83-110. Totowa, NJ: Humana Press Inc., 1998. 9803-9807, 1989. 7. Urbfinek, P., Wang, Z-Q., Fetka, I., Wagner, E. F., and Busslinger, M. Complete 35. Buni, M., Tromvoukis, Y., Bopp, D., Frigerio, G., and Noll, M. Conservation of the block of early B cell differentiation and altered patterning of the posterior midbrain paired domain in metazoans and its structure in three isolated human genes. EMBO in mice lacking Pax5/BSAP. Cell, 79: 901-912, 1994. J., 8: 1183-1190, 1989. 8. Nutt, S. L., Urb~nek, P., Rolink, A., and Busslinger, M. Essential functions of Pax5 36. Lee, W. S., Kao, C. C., Bryant, G. O., Liu, X., and Berk, A. J. Adenovirus EIA (BSAP) in pro-B cell development: difference between fetal and adult B lymphopoi- activation domain binds the basic repeat in the TATA box transcription factor. Cell, esis and reduced Vto-DJ recombination at the IgH locus. Genes Dev., 11: 476-491, 67: 365-376, 1991. 1997. 37. Hagemeier, C., Bannister, A. J., Cook, A., and Kouzarides, T. The activation domain 9. Busslinger, M., Klix, N., Pfeffer, P., Graninger, P. G., and Kozmik, Z. Deregulation of transcription factor PU. 1 binds the retinoblastoma (RB) protein and the transcrip- of PAX-5 by translocation of the E/x enhancer of the IgH locus adjacent to two tion factor TFIID in vitro: RB shows sequence similarity to TFIID and TFllB. Proc. alternative PAX-5 promoters in a diffuse large-cell lymphoma. Proc. Natl. Acad. Sci. Natl. Acad. Sci. USA, 90: 1580-1584, 1993. USA, 93: 6129-6134, 1996. 38. Hagemeier, C., Cook, A., and Kouzarides, T. The retinoblastoma protein binds E2F 10. Iida, S., Rao, P. H., Nallasivam, P., Hibshoosh, H., Butler, M., Louie, D. C., Dyomin, residues required for activation in vivo andTBP binding in vitro. Nucleic Acids Res., V., Ohno, H., Chaganti, R. S. K., and Dalla-Favera, R. The t(9;14)(p13;q32) chro- 21: 4998-5004, 1993. mosomal translocation associated with lymphoplasmacytoid lymphoma involves the 39. Xu, X., Prorock, C., Ishikawa, H., Maldonado, E., Ito, Y., and Grlinas, C. Functional PAX-5 gene. Blood, 88: 4110-4117, 1996. interaction of the v-Rel and c-Rel oncoproteins with the TATA-binding protein and 11. Morrison, A. M., JSger, U., Chott, A., Schebesta, M., Haas, O. A., and Busslinger, M. association with transcription factor liB. Mol. Cell. Biol., 13: 6733-6741, 1993. Deregulated PAXk5 transcription from a translocated IgH promoter in marignal zone 40. Metz, R., Bannister, A. J., Sntherland, J. A., Hagemeier, C., O'Rourke, E. C., Cook, lymphoma. Blood, 92: 3865-3878, 1998. A., Bravo, R., and Kouzarides, T. c-Fos-induced activation of a TATA-box-contain- 12. Nutt, S. L., Morrison, A. M., Drrfler, P., Rolink, A., and Busslinger, M. Identification ing promoter involves direct contact with TATA-box-binding protein. Mol. Cell. of BSAP (Pax-5) target genes in early B-cell development by loss- and gain-of- Biol., 14: 6021-6029, 1994. function experiments. EMBO J., 17: 2319-2333, 1998. 41. Liu, X., Miller, C. W., Koeffler, P. H., and Berk, A. J. The p53 activation domain 13. DSrfler, P., and Busslinger, M. C-terminal activating and inhibitory domains deter- binds the TATA box-binding polypeptide in Holo-TFI1D, and a neighboring p53 mine the transactivation potential of BSAP (Pax-5), Pax-2 and Pax-8. EMBO J., 15: domain inhibits transcription. Mol. Cell. Biol., I3: 3291-3300, 1993. 1971-1982, 1996. 42. Weinberg, R. A. The retinoblastoma protein and cell cycle control. Cell, 81: 323-330, 14. Studier, W., Rosenberg, A. H., Dunn, J. J., and Dubendorff, J. W. Use of T7 RNA 1995. polymerase to direct expression of cloned genes. Methods Enzymol., 185: 60-89, 43. Wang, J. Y. J. Retinoblastoma protein in growth suppression and death protection. 1990. Curr. Opin. Genet. Dev., 7: 39-45, 1997. 15. Zwilling, S., Annweiler, A., and Wil-th, T. The POU domains of the Octl and Oct2 44. Mulligan, G., and Jacks, T. The retinoblastoma gene family: cousins with overlapping transcription factors mediate specific interaction with TBP. Nucleic Acids Res., 22: interests. Trends Genet., 14: 223-229, 1998. 1655-1662, 1994. 45. Hoteboer, G., Timmers, H. T. M., Rustgi, A. K., Billaud, M., van 't Veer, L. J., and 16. Weigel, R. J., Devoto, S. H., and Nevin, J. R. Adenovirus 12S EIA gene represses Bemards, R. TATA-binding protein and the retinoblastoma gene product bind to differentiation of F9 teratocarcinoma cells. Proc. Natl. Acad. Sci. USA, 87." 9878- overlapping epitopes on c-Myc and adenovirus E1A protein. Proc. Natl. Acad. Sci. 9882, 1990. USA, 90: 8489-8493, 1993. 17. Kaelin, W. G. J., Pallas, D. C., DeCaprio, J. A., Kaye, F. J., and Livingston, D.M. 46. Pearson, A., and Greenblatt, J. Modular organization of the E2F1 activation domain Identification of cellular proteins that can interact specifically with the T/E1A-binding and its interaction with general transcription factors TBP and TFIIH. Oncogene, 15: region of the retinoblastoma gene product. Cell, 64: 521-532, 1991. 2643-2658, 1997. 1723s

47. Whyte, P., Buchkovich, K. J., Horowitz, J. M., Friend, S. H., Raybuck, M., Weinberg, 70: 337-350, 1992. R. A., and Harlow, E. Association between an oncogene and an anti-oncogene: the 65. Shah, B., Durfee, T., and Lee, W.-H. Disruption of RB/E2F-1 interaction by single adenovirus E1A proteins bind to the retinoblastoma gene product. Nature (Lond.), point mutations in E2F-1 enhances S-phase entry and apoptosis. Proc. Natl. Acad. Sci. 334: 124-129, 1988. USA, 93: 679-684, 1996. 48. Dyson, N., Dembski, M., Fattaey, A., Ngwa, C., Ewen, M., and Helin, K. Analysis of 66. Kozmik, Z., Wang, S., Drrfler, P., Adams, B., and Busslinger, M. The promoter of pl07-associated proteins: p107 associates with a form of E2f that differs from the CD19 gene is a target for the B-cell-specific transcription factor BSAP. Mol. Cell. pRB-associated E2F-1. J. Virol., 67: 7641-7647, 1993. Biol., 12: 2662-2672, 1992. 49. Beijersbergen, R. L., Kerkhoven, R. M., Zhu, L., Cat-lee, L., Voorhoeve, P. M., and 67. Weintraub, S. J., Chow, K. N. B., Luo, R. X., Zhang, S. H., He, S., and Dean, D. C. Bernards, R. E2F-4, a new member of the E2F gene family, has oncogenic activity Mechanism of active transcriptional repression by the retinoblastoma protein. Nature and associates with p107 in vivo. Genes Dev., 15: 2680-2690, 1994. (Lond.), 375: 812-815, 1995. 50. Ginsberg, D., Vairo, G., Chittenden, T., Xiao, Z-X., Xu, G., Wydner, K. L., DeCaprio, 68. Luo, R. X., Postigo, A. A., and Dean, D. C. Rb interacts with histone deacetylase to 1. A., Lawrence, J. B., and Livingston, D. M. E2F-4, a new member of the E2F repress transcription. Cell, 92: 463-473, 1998. transcription factor family, interacts with p107. Genes Dev., 8: 2665-2679, 1994. 69. Magnaghi-Jaulin, L., Groisman, R., Naguibneva, i., Robin, P., Lorain, S., Le Villain, J. P., 51. Mihara, K., Cat, X-R., Yen, A., Chandler, S., Driscoll, B., Murphree, A. L., T'Ang, A., Troalen, F., Trouche, D., and Harel-Bellan, A. Retinoblastoma protein represses tran- and Fung, Y-K. T. Cell cycle-dependent regulation of phosphorylation of the human scription by recruiting a histone deacetylase. Nature (Lond.), 391: 601-605, 1998. retinoblastoma gene product. Science (Washington DC), 246: 1300-1303, 1989. 70. Brehm, A., Miska, E. A., McCance, D. J., Reid, J. L., Bannister, A. J., and Kouzar- 52. Buchkovich, K., Duffy, L. A., and Harlow, E. The retinoblastoma protein is phos- ides, T. Retinoblastoma protein recruits histone deacetylase to repress transcription. phorylated during specific phases of the cell cycle. Cell, 58:1097-1105, 1989. Nature (Lond.), 391: 597-601, 1998. 53. Chen, P-L., Scully, P., Shew, J-Y., Wang, J. Y., and Lee, W-H. Phosphorylation of 71. Jacks, T., Fazeli, A., Schmitt, E. M., Bronson, R. T., Goodell, M. A., and Weinberg, the retinoblastoma gene product is modulated during the cell cycle and cellular R. A. Effects of an Rb mutation in the mouse. Nature (Lond.), 359: 295-300, 1992. differentiation. Cell, 58: 1193-1198, 1989. 72. Lee, E. Y-H. P., Chang, C-Y., Hu, N., Wang, Y-C. J., Lai, C-C., Herrup, K., Lee, 54. Kaye, F. J., Kratzke, R. A., Gerster, J. L., and Horowitz, J. M. A single amino acid W-H., and Bradley, A. Mice deficient for Rb are nonviable and show defects in substitution results in a retinoblastoma protein defective in phosphorylation and neurogenesis and haematopoiesis. Nature (Lond.), 359: 288-294, 1992. oncoprotein binding. Proc. Natl. Acad. Sci. USA, 87: 6922-6926, 1990. 73. Clarke, A. R., Maandag, E. R., van Roon, M., van der Lugt, N. M. T., van der Valk, 55. Bookstein, R., Shew, J-Y., Chen, P. L., Scully, P., and Lee, W-H. Suppression of M., Hooper, M. L., Berns, A., and te Riele, H. Requirement for a functional Rb-1 gene tumorigenicity of human prostate carcinoma cells by replacing a mutated RB gene. in murine development. Nature (Lond.), 359: 328-330, 1992. Science (Washington DC), 247: 712-715, 1990. 74. Gu, W., Schneider, J. W., Condorelli, G., Kaushal, S., Mahdavi, V., and Nadal- 56. Hu, Q., Dyson, N., and Harlow, E. The regions of the retinoblastoma protein needed Ginard, B. Interaction of myogenic factors and the retinoblastoma protein mediates for binding to adenovirus E1A or SV40 large T antigen are common sites for muscle cell commitment and differentiation. Cell, 72: 309-324, 1993. mutations. EMBO J., 9: 1147-1155, 1990. 75. Schneider, J. W., Gu, W., Zhu, L., Mahdavi, V., and Nadal-Ginard, B. Reversal of 57. Kaelin, W. G. J., Ewen, M. E., and Livingston, D. M. Definition of the minimal terminal differentiation mediated by p107 in Rb-/- muscle cells. Science (Wash- simian virus 40 large T antigen- and adenovirus E1A-binding domain in the retino- ington DC), 264: 1467-1471, 1994. blastoma gene product. Mol. Cell. Biol., 10: 3761-3769, 1990. 76. Chen, P-L., Riley, D. J., Chen-Kiang, S., and Lee, W-H. Retinoblastoma protein 58. Zhang, H., Catron, K. M., and Abate-Shen, C. A role of the Msx-1 homeodomain in directly interacts with and activates the transcription factor NF-IL6. Proc. Natl. Acad. transcriptional regulation: residues in the N-terminal arm mediate TATA binding Sci. USA, 93: 465-469, 1996. protein interaction and transcriptional repression. Proc. Natl. Acad. Sci. USA, 93: 77. Chen, P-L., Riley, D. J., Chen, Y., and Lee, W-H. Retinoblastoma protein positively 1764-1769, 1996. regulates terminal adipocyte differentiation through direct interaction with C/EBPs. 59. Wiggan, O., Taniguchi-Sidle, A., and Hamel, P. A. Interaction of the pRB-family proteins Genes Dev., 10: 2794-2804, 1996. with factors containing paired like homeodomains. Oncogene, 16: 227-236, 1998. 78. Goulding, M. D., Chalepakis, G., Deutsch, U., Erselius, ,l. R., and Gruss, P. Pax-3, a 60. Lieberman, P. M., and Berk, A. J. The Zta trans-activator protein stabilizes TFIID novel murine DNA binding protein expressed during early neurogenesis. EMBO J., association with promoter DNA by direct protein-protein interaction. Genes Dev., 5: 10: 1135-1147, 1991. 2441-2454, 1991. 79. Ton, C. C. T., Hirvonen, H., Miwa, H., Weil, M. M., Monaghan, P., Jordan, T., van 61. Kerr, L. D., Ransone, L. J., Wamsley, P., Schmitt, M. J., Boyer, T. G., Zhou, G., Berk, Heyningen, V., Hastie, N. D., Meijers-Heijboer, H., Drechsler, M., Royer-Pokora, B., A. J., and Verma, I. M. Association between proto-oncoprotein Rel and TATA- Collins, F., Swaroop, A., Strong, L. C., and Saunders, G. F. Positional cloning and binding protein mediates transcriptional activation of NF-KB. Nature (Lond.), 365: characterization of a paired box- and homeobox-containing gene from the aniridia 412-419, 1993. region. Cell, 67: 1059-1074, 1991. 62. Huang, S., Shin, E., Sheppard, K. A., Chokroverty, L., Shan, B., Quian, Y. W., Lee, 80. Eccles, M. R., Wallis, L. J., Fidler, A. E., Spurr, N. K., Goodfellow, P. J., and Reeve, E. Y., and Yee, A. S. The retinoblastoma protein region required for interaction with A. E. Expression of the PAX2 gene in human fetal kidney and Wilms' tumor. Cell the E2F transcription factor includes the T/EIA binding and carboxy-terminal se- Growth Diff., 3: 279-289, 1992. quences. DNA Cell Biol., 1l: 539-548, 1992. 81. Kozmik, Z., Kurzbauer, R., D6rfler, P., and Busslinger, M. Alternative splicing of 63. Hiebert, S. W., Chellappan, S. P., Horowitz, J. M., and Nevins, J. R. The interaction Pax-8 gene transcripts is developmentally regulated and generates isoforms with of RB with E2F coincides with an inhibition of the transcriptional activity of E2F. different transactivation properties. Mol. Cell. Biol., 13: 6024-6035, t993. Genes Dev., 6: 177-185, 1992. 82. Wilson, D. S., Guenther, B., Desplan, C., and Kuriyan, J. High resolution crystal 64. Helin, K., Lees, J. A., Vidal, M., Dyson, N., Halow, E., and Fattaey, A. A cDNA structure of a paired (Pax) class cooperative homeodomain dimer on DNA. Cell, 82: encoding a pRB-binding protein with properties of the transcription factor E2F. Cell, 709-719, 1995.

Discussion

The topic of Dr. Busslinger's presentation will be published shortly, and, hence, the reader is referred to the following manuscript for detailed information: Nutt, S. L., Vambrie, S., Steinlein, P,, Rolink, A., Kozmik, Z., Weith, A., and Busslinger, M. Independent regulation of the two Pax-5 alleles during B-cell development. Nature Genet., in press, 1999.

Dr. Phillip Sharp: Let me ask the first question, As far as I know, Dr. Sharp: But do you think that the allele-specific regulation the allele-specific regulation of Pax-5 is an unprecedented mechanism described by you might be more common? of gene regulation in mammalian cells. As you outlined in the last Dr. Busslinger: I am certain that this is a more general phenome- slide, this regulatory mechanism is distinguishable from imprinting by non. I would not be surprised, if, for example, the gene coding for the being reversible in terms of switching within the cell lineage. Could osteoblast-specific transcription factor Cbfa-1 (Osf-2) showed a sim- you comment on this? ilar monoallelic transcription pattern. However, I have to remind you Dr. Busslinger: Yes, as far as I can tell, this seems to be the first how difficult it is to demonstrate allele-specific regulation in geneti- example of this kind of transcriptional regulation. It is, however, cally unmanipulated cells, even if the two alleles can be distinguished important to realize that it is quite difficult to demonstrate this by a nucleotide polymorphism in the transcribed region. Due to the phenomenon for a given gene. We were extremely lucky to have reversibility of the phenomenon, it is impossible to grow a single cell identified a target gene whose expression is totally dependent on the into a colony to demonstrate allele-specific transcription by reverse transcription factor Pax-5. Such target genes are normally not avail- transcription-PCR analysis. By the time the colony has been ex- able for other transcription factors. Hence, mice heterozygous for panded, the expression from both alleles has been randomized. This is mutations in other transcription factor genes cannot be used for the main reason why all our analyses had to be performed at the analyses similar to those of the Pax-5 mutant mouse. single-cell level, 1724s

Speaker: You have partly answered my question already. I was just tissues where no haploinsufficient phenotype is seen. A similar situ- wondering if you felt that this regulation might apply to many genes ation may also hold true for other lineage determination genes. which are expressed at a low level. In other words, a low level of Dr. Nicholas Hastie: A couple of questions: First, as you men- transcription might always tend to be monoallelic by chance. tioned, Pax-5 does not necessarily show a haploinsufficient pheno- Dr. Busslinger: Let me rephrase your question. Essentially you type, while other Pax genes do. What do you know about monoallelic would like to know the mechanism underlying the allele-specific expression of other Pax genes? transcriptional regulation. At present, I think that we do not have to Dr. Busslinger: We did not perform any experiments with other invoke a grand novel mechanism to explain this phenomenon. Low- Pax genes. The groups of Dr. Howard Cedar and Dr. Andrew Chess level transcription may indeed mean that only one allele is selected by apparently have experimental data indicating that the Pax-3 and Pax-6 stochastic choice to be transcribed at an early point in development. genes are also subject to monoallelic transcriptional regulation. As I This allele then continues to be expressed, unless switching of ex- have never seen these data, I cannot further comment on this issue. pression to the second allele occurs. One could argue that the revers- Dr. Nicholas Hastie: The other question relates to the point made ibility of allele-specific transcription may be responsible for the earlier about whether many genes, which are expressed at a low level, haploinsufficient phenotypes of Pax genes. It is conceivable that the go through some stochastic program of transcriptional regulation. This switch of expression from the initially transcribed wild-type allele to observation may be relevant also for cancer, the actual topic of this the mutant allele may adversely affect the development of a tissue. A meeting. For instance, a tumor suppressor gene could be transcribed cell which undergoes such a switch past the lineage commitment stage from the inherited mutant allele during a window, where the wild-type may not be able to further participate in normal development. In the allele is not expressed. There may be considerable selective advantage case of Pax-5, we know from conditional gene inactivation experi- for such monoallelically transcribing cells to lose the second allele. ments that this transcription factor is indeed required from the begin- From this point of view, the phenomenon of allele-specific regulation ning to the end of B cell development. A similar situation may also could have broader significance. apply to other Pax genes. Dr. Busslinger: I agree that this is a possibility. Speaker." I am a bit surprised about your findings. Many genes are Dr. Suzanne Cory: I was fascinated by your talk and am coming known to be involved in lineage determination of cells, but only few back to the mechanism, following on from your answer to Jerry of them seem to be haploinsufficient. One could argue that there is an Adams. You have observed a progression from monoallelic to bial- evolutionary strategy to avoid haploinsufficiency. What are your thoughts on this issue? lelic to monoallelic Pax-5 expression during B cell development. Do Dr. Busslinger: Let me get the following point straight. Pax-5 is these transitions correlate with the concentration of Pax-5 during B haploinsufficient at the cellular level, as I presented it in my talk. cell development. In other words, does the level of Pax-5 start off low, However, Pax-5 does not cause a clinical syndrome, probably due to then increase and go back down low again? a peculiarity of the B-lymphoid system. B cells are either selected for Dr. Busslinger: This is a difficult question to answer, as it is not expansion or are deleted depending on whether they have undergone easy to distinguish a 2-fold difference in transcription factor concen- in-frame or out-of-frame rearrangements of immunoglobulin genes. In tration. Interestingly, however, the expression of the CD19 gene, such a system, a 2-fold difference in cell number will not result in a which is strictly dependent on Pax-5 function, shows an inverse phenotypic manifestation. However, this situation may be different for correlation relative to Pax-5. Immature B cells express less CD19 on lineage determination genes expressed in solid tissues. the surface compared to the early pro-B and mature B cells. In other Speaker: Well, the lymphoid system is not the only one where words, higher Pax-5 expression levels correlate with lower CD19 lineage decisions are taking place. So the issue still is that very few expression. In this context I would like to mention the elegant exper- genes involved in lineage determination seem to be haploinsufficient. iments performed in Nick Hastie's laboratory that clearly demon- Dr. Busslinger: Yes, but then again it is known even for the Pax strated a narrow dosage sensitivity of Pax-6. Moderate overexpression gene family that haploinsufficiency does not affect every tissue where of the Pax-6 gene due to extra copies on a wild-type background a given Pax gene is expressed. As I have shown in my presentation, resulted in as similar Small eye phenotype in the mouse as mutation the Pax-5 gene proceeds from a monallelic to a biallelic transcription of one of the two endogenous alleles. It is, therefore, a possibility that mode during B cell development. In analogy it is, therefore, possible the dosage sensitivity and allele-specific regulation of Pax genes may that other Pax genes may also be biallelically transcribed in some be related phenomena.

1725s

Downloaded from cancerres.aacrjournals.org on September 25, 2021. © 1999 American Association for Cancer Research. The Partial Homeodomain of the Transcription Factor Pax-5 (BSAP) Is an Interaction Motif for the Retinoblastoma and TATA-binding Proteins

Dirk Eberhard and Meinrad Busslinger

Cancer Res 1999;59:1716s-1725s.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/59/7_Supplement/1716s

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/59/7_Supplement/1716s. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.