¹Leukaemia Research Fund Centre at the Institute of Cancer Research, Chester Beatty Laboratories, 237 Fulham Road, London SW3 6JB, UK

²Cardiothoracic Surgery, National Heart and Lung Institute, Imperial College School of Medicine, London SW3 6LY, UK

³Institut de Recherches sur le Cancer de Lille, U124 INSERM, 59045 Lille, France

⁴Department of Human Anatomy, Oxford University, Oxford OX1 3QX, UK

^aAuthor for correspondence

^bCurrentt address: CNRS URA 1160, IBL, Institut Pasteur de Lille, 59019, Lille, France

BTB/POZ-domain C₂H₂ zinc(Zn)-finger proteins are encoded by a subfamily of genes related to the Drosophila gap gene krüppel. To date, two such proteins, PLZF and LAZ-3/BCL-6, have been implicated in oncogenesis. We have now identified a new member of this gene subfamily which encodes a 62 kDa Zn-finger protein, termed LRF, with a BTB/POZ domain highly similar to that of PLZF. Both human and mouse LRF genes, which localized to syntenic chromosomal regions (19p13.3 and 10B5.3, respectively), were widely expressed in adult tissues and cell lines. At approximately 9.5 - 10.0 days of embryonic development, the mouse LRF gene was expressed in the limb buds, pharyngeal arches, tail bud, placenta and neural tube. The LRF protein associated in vivo with LAZ-3/BCL-6, but not with PLZF to which it was more related. Although the LRF, or LAZ-3/BCL-6, BTB/POZ domain could readily homodimerize, no heterodimerization was detected in vivo between the LRF and LAZ-3/BCL-6 BTB/POZ domains and interaction between full length LRF and LAZ-3/BCL-6 required the presence of both the BTB/POZ domain and Zn-fingers in each partner protein. As expected from the above results, LRF and LAZ-3/BCL-6 also colocalized with each other in the nucleus. Taken together, our findings suggest that BTB/POZ-domain Zn-finger proteins may function as homo and heterodimeric complexes whose formation, and hence the resultant effect on transcription of their downstream target genes, is determined by the levels and expression domains of a given partner protein.

Transcription factor; development; limb; lymphoma; dimerization

Introduction

Differential control of gene expression plays a key role in complex cellular processes through which a given cell acquires specific characteristics (Karin, 1990). Studies of how genes are regulated at the level of transcription have led to identification of a large number of gene families which encode cell and/or promoter specific transcription factors as well as the components of the basal transcriptional machinery. One of the largest such gene families encodes proteins with one or more C₂H₂ Zinc(Zn)-finger motifs (El-Baradi and Pieler, 1991), also known as krüppel-like Zn-fingers, initially characterized in the Xenopus laevis RNA polymerase III transcription factor TFIIIA (Miller et al., 1985).

Recently, subfamilies of krüppel-like Zn-finger genes have been identified which contain divergent and highly conserved N-terminal motifs, the KRAB (krüppel-associated box) and BTB/POZ (for broad complex, tramtrack, and bric á brac/poxvirus and zinc finger) domains (Bardwell and Treisman, 1994; Constantinou-Deltas et al., 1992; Zollman et al., 1994), both of which appear to function in transcriptional repression (Chang et al., 1996; Deweindt et al., 1995; Friedman et al., 1996; Kim et al., 1996; Li et al., 1997; Moosmann et al., 1996; Witzgall et al., 1994). The BTB/POZ domain is an approximately 120 amino acid long, highly conserved, hydrophobic region and is present in two distinct classes of transcriptional regulators - the POK (BTB/POZ and krüppel) proteins and the Bach proteins which contain a C-terminal basic leucine zipper motif instead of a Zn-finger region (Oyake et al., 1996). The BTB/POZ domain is also present in some viral (Koonin et al., 1992) and cellular proteins (Xue and Cooley, 1993), which do not possess any obvious DNA binding motifs and whose function may not be associated with transcription. The BTB/POZ domain has been shown to be sufficient in mediating self-association of a number of BTB/POZ domain proteins (Bardwell and Treisman, 1994; Chen et al., 1995), including LAZ-3/BCL-6 (Dhordain et al., 1995) and PLZF (Dong et al., 1996), and in mediating transcriptional repression when fused to a heterologous DNA-binding region (Chang et al., 1996; Deweindt et al., 1995; Li et al., 1997; Seyfert et al., 1996). In addition, for a number of POK proteins, the BTB/POZ domain appears to confer specific, often speckled, nuclear localization pattern (Bardwell and Treisman, 1994; Dhordain et al., 1995; Dong et al., 1996).

Two POK proteins, PLZF (Promyelocytic Leukaemia Zinc Finger) and LAZ-3/BCL-6 (Lymphoma Associated Zn-finger-3/B-Cell Lymphoma-6), have been implicated in oncogenesis (Baron et al., 1993; Chen et al., 1993a,b; Kerckaert et al., 1993; Miki et al., 1994; Ye et al., 1993). The PLZF gene is translocated with the RAR locus in rare cases of all-trans-retinoic acid (RA) resistant (Licht et al., 1995) acute promyelocytic leukaemia (APL). As a result of this translocation oncogenic PLZF-RAR protein is expressed in leukaemic cells. In diffuse large cell lymphoma (DLCL), on the other hand, expression of the LAZ-3/BCL-6 gene is deregulated as a result of its translocations to immunoglobulin loci, or mutations in its regulatory regions (Bernardini et al., 1997; Migliazza et al., 1995; Ye et al., 1995). Both PLZF and LAZ-3/BCL-6 have recently been shown to function as transcriptional repressors through recruitment of nuclear receptor/Sin3/histone deacetylase co-repressor complexes (Dhordain et al., 1997; Guidez et al., 1998; He et al., 1998; Hong et al., 1997).

In addition to being associated with tumorgenesis, both PLZF and LAZ-3/BCL-6 have been implicated in normal haemopoietic development, regulating myelopoiesis and lymphopoiesis, respectively. PLZF is expressed in early myeloid progenitor cells and its expression is downregulated during granulocytic differentiation (Reid et al., 1995), suggesting a role in granulopoiesis. Consistently with these conclusions, ectopic expression of PLZF in myeloid progenitor cells inhibits their ability to undergo growth factor induced neutrophilic differentiation (Shaknovich et al., 1998). LAZ-3/BCL-6, on the other hand, is expressed mainly in germinal centre B-cells (Cattoretti et al., 1995; Onizuka et al., 1995) and mice lacking LAZ-3/BCL-6 do not possess germinal centres and have impaired lymphopoiesis (Dent et al., 1997; Ye et al., 1997).

Haemopoiesis is a very tightly regulated process. Appropriate expression and activity of various regulatory proteins is essential to ensure correct lineage commitment and differentiation of cells (Shivdasani and Orkin, 1996). Therefore, the activities of LAZ-3/BCL-6 and PLZF, which probably depend on formation of appropriate complexes through their BTB/POZ domains, may prove to be important targets during oncogenesis. For example, deregulated expression of LAZ-3/BCL-6 in DLCL, or expression of PLZF-RAR in APL cells, could disrupt a functional balance of regulatory networks with which LAZ-3/BCL-6, or PLZF, interact. To understand better the molecular mechanisms underlying the function of POK proteins in normal cellular processes and oncogenesis, as well as to understand their regulatory circuitry, we set out to identify other members of the POK gene family with emphasis on BTB/POZ domain sequences closely related to PLZF. We have cloned a cDNA encoding a novel POK protein, LRF (for Leukaemia/Lymphoma Related Factor), which interacts in vivo with the product of the LAZ-3/BCL-6 proto-oncogene, but not with PLZF, suggesting that it may be a target for the oncogenic activity of LAZ-3/BCL-6 in DLCL.

Results

Isolation of a novel POK protein, LRF

Exploiting the high degree of amino acid sequence identity within the BTB/POZ domain, we chose to use RT/PCR strategy to isolate partial cDNAs encoding previously uncharacterized proteins. Degenerate oligonucleotide primers were designed with a bias toward sequences which are more related to the PLZF BTB/POZ domain than to those of other POK proteins (see Materials and methods). Consequently, out of the total 700 clones screened, most positive clones corresponded to the PLZF sequences. Two clones contained identical sequences encoding a previously unidentified BTB/POZ domain which was very similar to that of PLZF, and to a lesser degree that of LAZ-3/BCL-6, particularly in those regions from which the PCR oligonucleotide primers 1, 2 and 3 were derived (see Figure 1b). Using the amplified cDNA as a ³²P-labelled probe, we then cloned a cDNA encoding the remainder of this POK protein, called LRF for Leukaemia/Lymphoma Related Factor. In parallel, its avian homologue was isolated through low stringency screening of a chicken heart cDNA library with the ³²P-labelled PLZF probe, further underlying the close homology between the LRF and PLZF BTB/POZ domain coding sequences.

The deduced amino acid sequence of the murine LRF protein and its conservation with the chicken sequence are shown in Figure 1a. Overall identity between these sequences is 65%, with the highest homologies lying within the BTB/POZ domain and the Zn-finger region (85% and 97% sequence identity, respectively). In common with other POK proteins, the region between the BTB/POZ domain and Zn-fingers (the P-Z region) was not well conserved, suggesting that it carries less structural and/or functional importance. Similarly, the most C-terminal sequences of the mouse and chicken LRF proteins were poorly conserved. The total number of amino acids for these two homologous proteins also differed, with mouse and chicken LRFs possessing 565 and 546 residues (M_r 61,757 and 61,318), respectively. Although LRF appears to possess four C₂H₂ Zn-finger motifs, it is worth noting that the 4th motif (underlined with a dotted line in Figure 1a) diverges from the classical krüppel-like Zn-finger in that the region between its C₂ and H₂ halves is seven instead of 12 amino acids. Recently, a murine gene called cKrox (Galera et al., 1994) and its human homologue hcKrox (Widom et al., 1997) have been described, which encode POK proteins with four Zn-finger motifs very similar to those of LRF (68%), including the fourth unusual Zn-finger, suggesting that cKrox and LRF proteins may recognise related DNA response elements.

Mouse and human LRF genes localize to syntenic chromosomal regions

Analysis of murine genomic clones revealed that, as for the PLZF protein, the LRF BTB/POZ domain is encoded by the same exon as the beginning of the Zn-finger region. Partial analysis of potential human LRF genomic clones showed that the exon/intron junctions for the exon encoding the BTB/POZ domain were conserved with corresponding regions of the murine LRF gene (data not shown) and a high degree of sequence homology (80% amino acid sequence identity) extended into the P-Z region (sequence between the BTB/POZ domain and Zn-fingers), confirming that they contained a direct human homologue of the mouse LRF gene. The total amino acid identity between the murine and human LRF sequences encoded by the above exon was 88%.

The mouse and human LRF genomic clones were used to determine the chromosomal localizations of the LRF genes by FISH. Mouse LRF was localized to a position which is approximately 52% of the distance from the heterochromatic-euchromatic boundary to the telomere of chromosome 10, an area corresponding to band B5.3 (Figure 2a). The human LRF genomic probe hybridized to a distal short arm of a group F chromosome (Figure 2b and B1). Cohybridization of LRF and chromosome 19q13.4 digoxigenin-dUTP labelled probes resulted in a specific labelling of both long and short arms of chromosome 19, and demonstrated that LRF localizes on the distal short arm (p13.3) of chromosome 19 (Figure 2b and B2). Of the metaphase cells hybridized with digoxigenin-dUTP labelled mouse and human LRF probes, 85% and 86% exhibited specific signals, respectively. The chromosomal localization of the mouse and human LRF genes is in agreement with linkage conservation between the mouse and human chromosomes. Numerous deletions, amplifications and translocations involving chromosome 19p13 have been reported in various haemopoietic malignancies (Heim and Mitelman, 1987; Mitelman, 1991). It remains to be seen if LRF, like the LAZ-3/BCL-6 and PLZF genes, may be involved in any of these translocations.

Expression patterns of various mouse and human LRF mRNA isoforms in different tissues and cell lines

Both mouse and human LRF genes were expressed widely with variable levels of expression in different samples. Among various murine tissues and cultured cells which were examined, two predominant LRF mRNA isoforms (labelled B and C in Figure 3a) were detected, although an additional transcript (A) of approximately 7 kb was present in some samples, for example in the heart (Figure 3a, lane 5). In the developing mouse, at 11, 15 and 17 d.p.c., the expression of LRF increased with gestation and the A mRNA isoform appeared to be the predominant transcript (Figure 3b). LRF expression was also detected by whole mount in situ hybridization in mouse embryos at approximately 9.5 to 10 days of gestation (Figure 3c, embryos 1 and 2,). Expression was detected in the first two pharyngeal arches, limb buds, the tail bud, placenta and the neural tube. In the forelimb bud, the highest level of staining was observed around a distal rim and gradually decreased into the more proximal tissues, suggesting that LRF expression increases along the proximo-distal axis of the developing forelimb bud. The hindlimb buds (Figure 3c, embryos 1 and 2, labelled h), which in a given embryo were at earlier stages of development, showed lower levels of LRF expression than the forelimb buds. Hybridization of approximately 9.5 - 10.0 d.p.c. embryos with a sense RNA LRF probe was used as a negative control; there was no staining other than that due to trapped probe in the otic vesicle (not shown).

Like its murine homologue, the human LRF gene was also widely expressed. Among the samples examined, the highest levels of LRF expression were detected in the breast cancer cell line T47D and adult skin (Figure 4, lanes 15 and 20). In nearly all samples, three transcripts were detected which corresponded in size to the mouse mRNA isoforms A, B and C. As in the murine tissues and cell lines, the expression of various LRF isoforms appeared to undergo tissue specific regulation. For example, the A form was not expressed in the liver (Figure 4b, lane 19). Furthermore, the B and C isoforms were down and up-regulated with RA treatment in K562 and HSG cells, respectively (Figure 4, lanes 4 - 5 and 11 - 12). This apparent regulation of LRF mRNA isoforms by RA was also seen in murine P19 cells (Figure 3, lanes 7 and 8), where the levels of the B isoform appear to be weakly up-regulated after RA treatment. It is not known at the present time whether the different LRF mRNA isoforms are generated through alternative splicing, differential promoter usage and/or alternative polyadenylation. Nevertheless, it is clear that expression of the LRF gene undergoes complex regulation resulting in expression of different transcripts in different tissues and at different stages of development.

LRF colocalises and interacts with the product of the LAZ-3/BCL-6 proto-oncogene in vivo

Given the overlapping pattern of expression with PLZF, particularly during development (Cook et al., 1995), and the high degree of similarity within their BTB/POZ domains (Figure 1b), we anticipated that LRF and PLZF may interact with each other. This was not the case, however, as we failed to detect interaction between these two proteins by coimmunoprecipitation and in the yeast two-hybrid assay (data not shown). Since the homology between the BTB/POZ domain sequences could also reflect the ability of two homologous proteins to interact with the same partner protein and since the LAZ-3/BCL-6 protein, at least in the yeast two-hybrid assay, appeared to readily interact with PLZF (data not shown), we attempted to test the possibility that LRF and LAZ-3/BCL-6 could complex with each other in vivo. LRF and LAZ-3/BCL-6 were readily coimmunoprecipitated from nuclear extracts of COS-1 cells which were transiently cotransfected with their respective expression vectors (Figure 5a), suggesting that at least in this system, the two proteins are part of the same complex. As predicted by the above results, the two proteins also colocalized in a specific punctate pattern in the nucleus (Figure 5b).

The in vivo association of full length LAZ-3/BCL-6 and LRF was corroborated using the yeast two-hybrid assay (Figure 6, lane 3). In order to characterize the regions within both proteins which were required for their interaction, various deletion mutants of LAZ-3/BCL-6 and LRF were generated and tested in the yeast two-hybrid assay for their abilities to interact with each other or with their wild type counterparts. Although either the LRF, or LAZ-3/BCL-6 BTB/POZ domain alone could readily self associate (Figure 6, lanes 1 and 2), no heterodimeric interaction was detected in vivo between the LRF and LAZ-3/BCL-6 BTB/POZ domains (Figure 6, lane 5). Deletion of the P-Z region (sequence between the BTB/POZ domain and Zn-fingers) in the LRF protein did not prevent its interaction with LAZ-3/BCL-6 (Figure 6, lane 4). In line with the above results, this LRF deletion mutant was unable to heterodimerize with just the BTB/POZ domain of LAZ-3/BCL-6 (Figure 6, lane 6). Furthermore, inability of LRF without the BTB/POZ domain to interact with the full length LAZ-3/BCL-6 (Figure 6, lane 7), or just its BTB/POZ domain (Figure 6, lane 8), indicated that this domain, although not sufficient, is nevertheless required for interaction between the two proteins. These results also excluded the possibility of interaction being mediated solely through the Zn-finger regions and supported a model in which LRF and LAZ-3/BCL-6 association requires contacts between both the Zn-fingers and the BTB/POZ domains of the two partner proteins. This conclusion was further corroborated by results demonstrating the lack of any interaction between the full length LRF and LAZ-3/BCL-6 without the Zn-finger region (Figure 6, lane 12), or just the LAZ-3/BCL-6 BTB/POZ domain (Figure 6, lane 13), as well as failure of the LRF BTB/POZ domain to interact with the full length LAZ-3/BCL-6 protein (Figure 6, lane 11). In addition, and in contrast to LAZ-3/BCL-6 (Figure 6, lane 10), the BTB/POZ domain and Zn-fingers of LRF could interact in the yeast two-hybrid assay (Figure 6, lane 9), possibly reflecting the ability of the LRF protein to engage in an intramolecular interaction in vivo.

Discussion

This study has identified a previously uncharacterized POK protein, LRF. Within the BTB/POZ domain, LRF is the closest homologue of the PLZF protein. Contrary to the in vitro results showing interaction between the isolated BTB/POZ domains of the tramtrak and GAGA proteins (Bardwell and Treisman, 1994), this study has shown that the heterodimerization between LAZ-3/BCL-6 and LRF, in vivo, requires both the BTB/POZ domain and the Zn-finger region. Although counterintuitive, we cannot exclude the possibility that a stable LAZ-3/BCL-6-LRF interaction requires simultaneous contacts between the BTB/POZ domain of LRF and Zn-fingers of LAZ-3/BCL-6 and vice versa, instead of BTB/POZ-BTB/POZ and Zn-finger-Zn-finger interactions. Nevertheless, as shown previously (Dhordain et al., 1995), the BTB/POZ domain is sufficient for in vivo homodimeric interactions. Furthermore, the speckled nuclear colocalization of LAZ-3/BCL-6 and LRF POK proteins suggests that they may exist in large, possibly heterogeneous, complexes whose formation may depend on the presence of two distinct interaction interfaces in each protein.

The degree of conservation of a given protein sequence motif through evolution can be taken as a direct reflection of its functional importance. This study has demonstrated that the BTB/POZ domain and, to a greater extent, the Zn-finger region of LRF are highly conserved motifs, both required for heterodimeric interactions. The breakdown of amino acid sequence conservation between the avian and murine LRF homologues within the P-Z region is observed for other members of this subfamily, including PLZF and LAZ-3/BCL-6 (Allman et al., 1996; Cook et al., 1995; Fukuda et al., 1995), and may prove to be their common feature. It was not surprising, therefore, that the P-Z region was not required for LAZ-3/BCL-6-LRF interactions in the yeast two-hybrid system. Nevertheless, this region may act as a hinge regulating and/or stabilising the potential intramolecular interaction within the LRF protein. An intramolecular interaction between the transactivation domain and Zn-finger region of EKLF has recently been proposed to regulate its protein interaction and DNA binding abilities (Chen and Bieker, 1996). Similarly, the Eves motif and DNA binding domain of c-Myb have been shown to engage in an inhibitory intramolecular interaction (Dash et al., 1996).

BTB/POZ domain proteins have been implicated in a number of developmental processes, both in vertebrates and invertebrate organisms. For example, PLZF (Hawe et al., 1996) as well as the Drosophila BTB/POZ domain proteins (abrupt and bric-à-brac) (Godt et al., 1993; Hu et al., 1995; Zollman et al., 1994) appear to be required for proper limb morphogenesis. Expression of LRF in sites of rapid cell growth in the embryo (pharyngeal arches, limb buds, tail bud, and possibly ventral neural tube) and in the placenta, suggests that it may play a role in stimulating cell proliferation. It remains to be determined how LRF interacts with signalling pathways that control cell proliferation during embryogenesis, e.g. Wnt, FGFs, RA and Sonic hedgehog (so far interactions between these pathways are best understood in limb development (Johnson and Tabin, 1997)).

Previous studies have indicated that, in analogy to LRF, the LAZ-3/BCL-6 mRNAs are expressed in many tissues (Allman et al., 1996; Cattoretti et al., 1995; Fukuda et al., 1995). Nevertheless, the LAZ-3/BCL-6 protein, at least within the lymphoid compartment, is restricted to germinal centre B cells and Burkitt's lymphoma cell lines with mature B cell characteristics (Cattoretti et al., 1995; Onizuka et al., 1995). Although we cannot yet evaluate expression of the endogenous LRF protein, the wide spread expression pattern of the LRF gene strongly suggests that patterns of LRF and LAZ-3/BCL-6 expression overlap, at least partially. This is corroborated by presence of LRF mRNAs in Burkitt's lymphoma cell lines including Daudi (data not shown and Figure 4a, lane 3). The wide spread expression pattern of LRF also suggests that it may exert its effects by interacting not only with LAZ-3/BCL-6, but also with a number of other spatially and temporally restricted POK proteins and/or other proteins. It is possible that the abilities of these proteins to form homo and heterodimeric complexes can widen the repertoire of their target genes and/or the ways in which they function as transcriptional regulators. The formation of LRF, or LAZ-3/BCL-6, homodimers and LRF-LAZ-3/BCL-6 heterodimers, and therefore, resultant transcriptional and physiological effects may depend on the relative strengths of various interactions and concentrations of the partner proteins. Although it is currently unclear as to what effect the LRF-LAZ-3/BCL-6 hetreodimer will have on transcription of specific genes, or whether it will even recognise the same target genes/response elements as LRF and/or LAZ-3/BCL-6, deregulated expression of LAZ-3/BCL-6 in lymphoma cells will inadvertantly disrupt the balance between different LRF containing protein complexes. Therefore, the function of the LRF protein must be considered as a potentially important target for the oncogenic activity of LAZ-3/BCL-6. Its role in lymphopoiesis and how it may be affected in DLCL remain to be established.

Materials and methods

Cloning of BTB/POZ domain sequences by degenerate RT/PCR

The relative positions of the degenerate primers used to identify cDNA sequences which encode the LRF BTB/POZ domain are indicated in Figure 1b. Primer 1, 5'-NRS NNN NAR NNN NSC NGT RTA; primer 2, 5'-GGN NNN YTS TGY GAY GTB GTB AT; primer 3, 5'-CNGT RTA NGC RWA YTC CAR VA, primer 4, 5'-TTY CMC RCC CAC CGS WSN GT; where N=A/G/C/T, V=A/G/C, B=T/G/C, H=A/C/T, D=A/T/G, Y=C/T/, S=C/G, R=G/A, W=A/T, K=G/T, M=A/C. Primer 1 was used, as previously described (Chen et al., 1992), to reverse transcribe (RT) 1 g of pooled polyadenylated (polyA⁺) RNA derived from 13.5 and 8.5 days post coitum (dpc) mouse embryos. The resultant cDNAs were amplified using primers 2 and 3, and 35 cycles of polymerase chain reaction (PCR) with denaturation, hybridization and extension steps at 95°C for 30 s, 55°C for 1 min 30 s and 72°C for 3 min, respectively. Final extension was followed by further incubation at 72°C for 30 min. All PCRs were carried out in a total volume of 100 l containing 2.5 l of RT reaction, 0.2 M dNTPs, 50 M KCl, 1.5 M MgCl₂, 20 g/ml bovine serum albumin, two units of Taq DNA polymerase and 800 ng of each primer. Amplified cDNAs were subcloned into the PCR^TM II TA cloning vector (Invitrogen), transferred to bacteria and colonies were screened by hybridization with ³²P-labelled oligonucleotide 4. Hybridization was carried out overnight at 45°C in 5´SSPE, 1´Denhardt's and 1 mg/ml sheared salmon sperm DNA. Hybridized filters were washed for 15 min at 45°C in several changes of 2´SSC/0.1% SDS and autoradiographed. As we have expected that most of the amplified cDNAs will correspond to PLZF sequences, after exposure the same filters were rehybridized with a PLZF BTB/POZ domain specific ³²P-labelled oligonucleotide probe (5'-ACCGAAACAGCCAGCACTAT) to eliminate PLZF containing clones from further analysis. Hybridization with the PLZF specific probe and subsequent washes were carried out as for oligonucleotide 4, except at 55°C.

Cloning of the chicken and mouse LRF cDNAs

The chicken LRF cDNA was cloned in parallel with the partial murine LRF BTB/POZ domain sequences by screening of an adult chicken heart gt10 cDNA library (Clontech) with a ³²P-labelled fragment of a cloned mouse PLZF cDNA, as described (Cook et al., 1995). A full length mouse LRF cDNA was cloned, using previously described conditions, by screening an adult mouse heart ZapII cDNA library (gift of Pierre Chambon, IGBMC, Illkirch, France) with a ³²P-labelled probe derived by RT/PCR

Chromosomal localisation of the mouse and human LRF genes using fluorescence in situ hybridization (FISH)

Partial sequences corresponding to the human homologue of the mouse LRF gene were identified among the expressed sequence tags (EST) present in the GenBank database (accession number N77854). PCR oligonucleotide primers derived from known murine and human LRF sequences were used by Genome Systems Inc. (St. Louis, MO, USA) to screen genomic DNA libraries constructed in the P1 bacteriophage. The isolated clones were confirmed by sequencing. Mouse and human P1 genomic clones were used to map the chromosomal localizations of both LRF genes by FISH (performed at Genome Systems Inc.). The DNAs were labelled by nick translation (Sambrook et al., 1989) with digoxigenin-dUTP, labelled probe was combined with sheared mouse or human DNA and hybridized to metaphase chromosomes derived from murine embryonic fibroblasts or PHA stimulated peripheral blood lymphocytes, respectively. Hybridization solution contained 50% formamide, 10% dextran sulphate and 2´SSC. Specific hybridization signals were obtained by incubating the hybridized slides in fluoresceinated anti-digoxigenin antibodies and then counterstaining with DAPI. Localization of the murine LRF gene was confirmed by co-hybridization with a probe specific for the centromeric region of mouse chromosome 10 (Shi et al., 1997). Total of 80 metaphase cells were analysed with 68 exhibiting specific labelling. For the human LRF gene, a genomic clone previously mapped to chromosome 19q13.4 (unpublished marker, Genome Systems Inc.) was used to confirm its localization. A total of 80 metaphase cells were analysed of which 69 exhibited specific labelling.

Northern analysis

Total and poly(A)⁺ RNAs were isolated from tissues and cell lines as previously described (Aviv and Leder, 1972; Chirgwin et al., 1979; Zelent, 1997). Northern blot analysis was performed following published procedures (Zelent et al., 1989). A fragment of mouse LRF cDNA encoding amino acids 1 - 177 (see Figure 1a) and a PCR derived human LRF DNA encoding a region corresponding to the amino acids 14 - 131 in Figure 1a, were ³²P-labelled by random priming (Sambrook et al., 1989) and used as probes. Hybridization and washing conditions were as described (Zelent et al., 1989). Blots were exposed for 7 days at -80°C using two intensifying screens and Kodak X-OMAT^TM film.

Human and mouse cell lines

Mouse: P19 (Jones-Villeneuve et al., 1982), emryonal carcinoma cells; Wehi3BD (Warner et al., 1969), myeloid cells; 32D (Valtieri et al., 1987), myeloid progenitors; 18.8, pre-B (Yancopoulos et al., 1984); Mel (Friend et al., 1965), erythroid precursors; A4 (Heyworth et al., 1990) and B6 (Greenberger et al., 1983), multipotential progenitors; LyD9 (Palacios et al., 1987), pro-B cells; J774 (Ralph and Nakoinz, 1977), macrophage; EL-4 (Ralph and Nakoinz, 1973), mature transformed T cells; ES, embryonic stem cells.

Human: Jurkat (Weiss, 1984), T cell leukaemia; Nalm-6 (Minowada et al., 1978), pre-B cell acute lymphoblastic leukaemia; Daudi (Klein et al., 1968), Burkitt's lymphoma; K562 (Lozzio and Lozzio, 1975), chronic myelogenous leukaemia; KG1 (Koeffler and Golde, 1978), acute myelogenous leukaemia; HL-60 (Collins et al., 1977), acute myeloid leukaemia; APL (Chen et al., 1992), acute promyelocytic lukaemia bone marrow cells; 293 (Graham et al., 1977), adenovirus transformed human embryo kidney cells; HSG (Shirasuna et al., 1981), salivary gland adenocarcinoma; T₂Cl₁₃ (Weima et al., 1988), teratocarcinoma cells; T47D (Keydar et al., 1979), breast carcinoma.

Whole mount in situ hybridization

In situ hybridization of 10 d.p.c. mouse embryos with antisense and sense digoxigenin-UTP labelled RNA probes, derived from the full length LRF cDNA by in vitro transcription, was performed using previously described procedures (Wilkinson, 1992). Probes were synthesized as described (Cook et al., 1995) and purified through microspin S400 HR columns (Pharmacia) to remove unincorporated digoxigenin-UTP.

Yeast and mammalian expression plasmids

Mammalian expression vectors for Flag-tagged murine LRF (LRF^FLAG) and the human LAZ-3/BCL-6 protein were constructed in the pSG5 plasmid (Green et al., 1988) using murine cDNA LRF clone encoding sequence shown in Figure 1a and human LAZ-3/BCL-6 cDNA (gift of Riccardo Dalla-Favera, Columbia University, NY, USA). Yeast expression vectors (Bartel et al., 1993; Durfee et al., 1993) for LRF BTB/POZ domain fused either to the GAL4 DNA binding or activation (ACT) domains were generated using a partial LRF cDNA clone encoding amino acids 1 - 177. Yeast expression vectors for full length LRF (amino acids 1 - 565), or LRF without the BTB/POZ domain (amino acids 136 - 565), fused in frame to Gal4 DBD and ACT domains were generated by in frame ligation of appropriate restriction enzyme fragments of the murine LRF cDNA. The GAL4 DNA binding and ACT domain fusions of LRF sequences lacking the P-Z region (sequence between the BTB/POZ domain and Zn-fingers, amino acids 136 - 374) were constructed using PCR amplified sequences encoding just BTB/POZ domain and Zn-fingers. All constructions were partially sequenced to confirm in frame ligation. PCR amplified regions were sequenced in their entirety to avoid potential PCR errors. Yeast expression vectors for either partial or full length LAZ-3/BCL-6 GAL4 DNA binding and ACT domain fusions were as previously described (Dhordain et al., 1995).

Coimmunoprecipitation and coimmunofluorescence

COS-1 (Gluzman, 1981) and CHO (Puck et al., 1958) cells were cultured in DMEM with 10% FCS at 37°C and 5% CO₂. At 50% confluence cells were transfected by lipofection using lipofectamine reagant (Gibco - BRL) according to the manufacturer's instructions. Transfections of COS-1 cells with 0.5 g of LAZ-3/BCL-6 and/or 0.5 g of LRF^FLAG mammalian expression vectors were carried out in 60 mm tissue culture plates. The total amount of transfected expression vector was kept constant at 1 g in all experiments. Approximately 18 h after transfection cells were labelled with ³⁵S-methionine/cysteine, lysed and protein complexes present in the cell lysates were coimmunoprecipitated with 2 g of anti-FLAG M2 antibody (Kodak) and resolved by SDS - PAGE as previously described (Koken et al., 1997). Electrophoresed proteins were transferred to a solid support and blotted with a rabbit polyclonal anti-LAZ-3/BCL-6 antibody (Dhordain et al., 1995), diluted at a 1 : 1000 in TBST (0.14 M NaCl, 2.7 mM KCl, 25 mM Tris-Cl, pH 7.5, 0.1% Tween-20). Proteins were detected with a horseradish peroxidase-conjugated donkey anti-rabbit secondary antibody and ECL reagents (Amersham).

CHO cells, grown on glass cover slips in 60 mm tissue culture plates, were transfected with 1 g of LAZ-3/BCL-6 and/or 1 g of LRF^FLAG mammalian expression vectors. Approximately 24 h post transfection, the cover slips with adhered CHO cells were transferred to multiwell plates and immunoflourescence was performed essentially as described (Reid et al., 1995). Stained cells were visualized and photographed as described before (Koken et al., 1997) using a BioRad MRC600 confocal imaging system. The polyclonal rabbit anti-LAZ-3/BCL-6 (Dhordain et al., 1995) and the monoclonal anti-FLAG M2 (Kodak) antibodies were used at 1 : 200 and 1 : 1000 dilutions in Ca²⁺ and Mg²⁺ free phosphate buffered saline, respectively.

Yeast two-hybrid assay

The Y190 strain of Saccharomyces cerevisiae (Harper et al., 1993) harbours an integrated lacZ reporter gene under the control of a Gal4 responsive promoter. Competent Y190 yeast cells were transformed, using the lithium acetate method (Gietz and Woods, 1994), with 1 g of each of the two yeast expression vectors pASI/pGBT9 (Durfee et al., 1993) and pACTII/pGAD424 (Bartel et al., 1993) containing either LAZ-3/BCL-6 or LRF cDNAs, partial or full length, fused in frame to the GAL4 DNA binding and ACT domains, respectively. Liquid and solid media, as well as conditions used for yeast growth were as described (Johnston, 1994). For the -galactosidase assay, individual colonies which grew on selective media were then streaked onto replica plates, lifted onto filter papers (Whatmann), permeabilized by immersion in liquid nitrogen for 10 s and stained for -galactosidase activity as described previously (Breeden and Nasmyth, 1985).

Note added in proof

Mouse and chicken cDNA sequences reported here have been deposited on GenBank database under accession numbers AF086830 and AF086831, respectively.

Acknowledgements

This research was supported by The Leukaemia Research Fund of Great Britain and the National Institutes of Health (Grant No. CA59936-01). JZ and NB were supported by a fellowship from the Samuel Waxman Cancer Research Foundation and British Heart Foundation Grant (No. F297). We thank L Wiedemann and PP Pandolfi for helpful comments and discussions. We are grateful to Drs P Chambon and R Dalla-Favera for the gifts of the mouse heart cDNA library and the human LAZ-3/BCL-6 cDNA, respectively.

Allman D, Jain A, Dent A, Maile RR, Selvaggi T, Kehry MR and Staudt LM. (1996). Blood 87, 5257-5268. MEDLINE

Aviv H and Leder P. (1972). Proc. Natl. Acad. Sci. USA 69, 1408-1412. MEDLINE

Bardwell VJ and Treisman R. (1994). Genes Dev. 8, 1664-1677. MEDLINE

Baron BW, Nucifora G, McCabe N, Espinosa RI, LeBeau MM and McKeithan TW. (1993). Proc. Natl. Acad. Sci. USA 90, 5262-5266. MEDLINE

Bartel PL, Chien CT, Sternglanz R and Fields S. (1993). Cellular Interactions in Development: A Practical Approach. Hartley, D. A. (ed.). Oxford University Press: Oxford, pp 153-179.

Bernardini F, Collyn-d'Hooghe M, Quief S, Bastard C, Leprince D and Kerckaert JP. (1997). Oncogene 14, 849-855. MEDLINE

Breeden L and Nasmyth K. (1985). Cold Spring Harbor Symp. Quant. Biol. 50, 643-650. MEDLINE

Cattoretti G, Chang CC, Cechova K, Zhang J, Ye BH, Falini B, Louie DC, Offit K, Chaganti R and Dalla-Favera R. (1995). Blood 86, 45-33. MEDLINE

Chang CC, Ye BH, Chaganti RSK and Dalla-Favera R. (1996). Proc. Natl. Acad. Sci. USA 93, 6947-6952. Article MEDLINE

Chen SJ, Chen Z, Chen A, Tong JH, Dong S, Wang ZY, Waxman S and Zelent A. (1992). Oncogene 7, 1223-1232. MEDLINE

Chen SJ, Zelent A, Tong JH, Yu HQ, Wang ZY, Derre J, Berger R, Waxman S and Chen Z. (1993a). J. Clin. Invest. 91, 2260-2267. MEDLINE

Chen W, Zollman S, Couderc J and Laski F. (1995). Mol. Cell. Biol. 15, 3424-3429. MEDLINE

Chen X and Bieker JJ. (1996). EMBO. J. 15, 5888-5896. MEDLINE

Chen Z, Brand NJ, Chen A, Chen SJ, Tong JH., Wang ZY, Waxman S and Zelent A. (1993b). EMBO. J. 12, 1161-1167. MEDLINE

Chirgwin JM, Przybyla AE, MacDonald J and Rutter WJ. (1979). Biochem. 18, 5294-5299.

Collins SJ, Gallo RC and Gallagher RE. (1977). Nature 270, 347-349. MEDLINE

Constantinou-Deltas CD, Gilbert J, Bartlett RJ, Herbstreith M. Roses AD and Lee JE. (1992). Genomics 12, 581-589. MEDLINE

Cook M, Gould A, Brand N, Davies J, Strutt P, Shaknovich R, Licht J, Waxman S, Chen Z, Gluecksohn-Waelsch S, Krumlauf R and Zelent A. (1995). Proc. Natl. Acad. Sci. USA 92, 2249-2253. MEDLINE

Dash AB, Orrico FC and Ness SA. (1996). Genes. Dev. 10, 1858-1869. MEDLINE

Dent AL, Shaffer AL, Yu X, Allman D and Staudt LM. (1997). Science 276, 589-592. Article MEDLINE

Deweindt C, Albagli O, Bernardin F, Dhordain P, Quief S, Lantoine D, Kerckaert JP and Leprince D. (1995). Cell Growth Diff. 6, 1495-1503. MEDLINE

Dhordain P, Albagli O, Ansieau S, Koken MHM, Deweindt C, Quief S, Lantoine D, Leutz A, Kerckaert JP and Leprince D. (1995). Oncogene 11, 2689-2697. MEDLINE

Dhordain P, Albagli O, Lin JN, Ansieau S, Quief S, Leutz A, Kerckaert JP, Evans RM and Leprince D. (1997). Proc. Natl. Acad. Sci. USA 94, 10762-10767. Article MEDLINE

Dong S, Zhu J, Reid A, Strutt P, Guidez F, Zhong HJ, Wang ZY, Licht J, Waxman S, Chomienne C, Zelent A and Chen SJ. (1996). Proc. Natl. Acad. Sci. USA 93, 3624-3629. Article MEDLINE

Durfee T, Becherer K, Chen PL, Yeh SH, Yang Y, Kilburn AE, Lee WH and Elledge SJ. (1993). Genes. Dev. 7, 555-569. MEDLINE

El-Baradi T and Pieler T. (1991). Mech. Dev. 35, 155-169. MEDLINE

Friedman JR, Fredericks WJ, Jensen DE, Speicher DW, Huang XP, Neilson EG and Rauscher FJ. (1996). Genes. Dev. 10, 2067-2078. MEDLINE

Friend C, Patuleia MC and de Harven E. (1965). Natl. Cancer Inst. Mono. 22, 505-522.

Fukuda T, Miki T, Yoshida T, Hatano M, Ohashi K, Hirosawa S and Tokuhisa T. (1995). Oncogene 11, 1657-1663. MEDLINE

Galera P, Musso M, Ducy P and Karsenty G. ( (1994). Proc. Natl. Acad. Sci. USA 91, 9372-9376. MEDLINE

Gietz RD and Woods RA. (1994). Molecular Genetics of Yeast: A Practical Approach: The Practical Approach Series. Johnston JR. (ed.). Oxford University Press: Oxford, pp. 121-134.

Gluzman Y. (1981). Cell 23, 175-182. MEDLINE

Godt D, Couderc JL, Cramton SE and Laski FA. (1993). Development 119, 799-812. MEDLINE

Graham FL, Smiley J, Russell WC and Nairn R. (1977). J. Gen. Virol. 36, 59-74. MEDLINE

Green S, Issemann I and Scheer E. (1988). Nucl. Acids. Res. 16, 369.

Greenberger JS, Sakakeeny MA, Humphries RK, Eaves CJ and Eckner RJ. (1983). Cell Biology 60, 2931-2935.

Guidez F, Ivins S, Zhu J, Soderstrom M, Waxman S and Zelent A. (1998). Blood 91, 2634-2642. MEDLINE

Harper JW, Adami GR, Wei N, Keyomarsi K and Elledge SJ. (1993). Cell 75, 805-816. MEDLINE

Hawe N, Soares V, Niswander L, Cattoretti G and Pandolfi PP. (1996). Blood 88, 1151 (supplement 1, abstract).

He LZ, Guidez F, Tribioli C, Peruzzi D, Ruthardt M, Zelent A and Pandolfi PP. (1998). Nat. Genet. 18, 126-135. MEDLINE

Heim S and Mitelman F. (1987). Cancer Cytogenetics. Alan R Liss, New York.,

Heyworth CM, Dexter TM, Kan O and Whetton AD. (1990). Growth Factors 2, 197-211. MEDLINE

Hong SH, David G, Wong CW, Dejean A and Privalsky ML. (1997). Proc. Natl. Acad. Sci. USA 94, 9028-9033. Article MEDLINE

Hu S, Fambrough D, Atashi JR, Goodman CS and Crews ST. (1995). Genes. Dev. 9, 2936-2948. MEDLINE

Johnson RL and Tabin C. (1997). Cell 90, 979-990. MEDLINE

Johnston JR. (ed.). (1994). Molecular Genetics of Yeast: A Practical Approach. Oxford University Press, Oxford.,

Jones-Villeneuve EMV, McBurney MW, Rogers KA and Kalnins VI. (1982). J. Cell. Biol. 94, 253-262. MEDLINE

Karin M. (1990). Curr. Opin. Cell. Biol. 2, 996-1002. MEDLINE

Kerckaert JP, Deweindt C, Tilly H, Quief S, Lecocq G and Bastard C. (1993). Nat. Genet. 5, 66-70. MEDLINE

Keydar I, Chen L, Karby S, Weiss FR, Delarea J, Radu M, Chaitcik S and Brenner HJ. (1979). Euro. J. Cancer 15, 659-670.

Kim SS, Chen YM, Oleary E, Witzgall R, Vidal M and Bonventre JV. (1996). Proc. Natl. Acad. Sci. USA 93, 15299-15304. MEDLINE

Klein E, Klein G, Nadkarni JS, Nadkarni JJ, Wigzell H and Clifford P. (1968). Cancer Res. 28, 1300-1310. MEDLINE

Koeffler HP and Golde DW. (1978). Science 200, 1153-1154. MEDLINE

Koken MHM, Reid A, Quignon F, Chelbi-Alix MK, Davies JM, Kabarowski JHS, Zhu J, Dong S, Chen SJ, Chen Z, Tan CC, Licht J, Waxman S, de The H and Zelent A. (1997). Proc. Natl. Acad. Sci. USA 94, 10255-10260. MEDLINE

Koonin EV, Senkevich TG and Chernos V.I. (1992). Trends Biochem. Sci. 17, 213-214. MEDLINE

Li JY, English MA, Ball HJ, Yeyati PL, Waxman S and Licht JD. (1997). J. Biol. Chem. 272, 22447-22455. Article MEDLINE

Licht JD, Chomienne C, Goy A, Chen A, Scott AA, Head DR, Michaux JL, DeBlasio A, Miller Jr W, Zelenetz AD, Willman CL, Chen Z, Chen SJ, Zelent A, Macintyre E, Vail A, Cortes J, Kantarjian H and Waxman S. (1995). Blood 85, 1083-1094. MEDLINE

Lozzio CG and Lozzio BB. (1975). Blood 45, 321-334. MEDLINE

Migliazza A, Martinotti S, Chen WY, Fusco C, Ye BH, Knowles DM, Offit K, Chaganti RSK and Dalla-Favera R. (1995). Proc. Natl. Acad. Sci. USA 92, 12520-12524. MEDLINE

Miki T, Kawamata N, Hirosawa S and Aoki N. (1994). Blood 83, 26-32. MEDLINE

Miller J, McLachlan AD and Klug A. (1985). EMBO. J. 4, 1609-1614. MEDLINE

Minowada J, Janossy G, Greaves M, Tsubota T, Srivasta B, Morikawa S and Tatsumi E. (1978). J. Natl. Cancer Inst. 60, 1269-1273. MEDLINE

Mitelman F. (1991). Catalog of Chromosome Aberrations in Cancer, Fourth edn. Wiley-Liss: New York.

Moosmann P, Georgiev O, LeDouarin B, Bourquin JP and Schaffner W. (1996). Nucleic Acids Res. 24, 4859-4867. Article MEDLINE

Onizuka T, Moriyama M, Yamochi T, Kuroda T, Kazama A, Kanazawa N, Sato K, Kato H and Mori S. (1995). Blood 86, 25-37.

Oyake T, Itoh K, Motohashi H, Hayashi N, Hoshino H, Nishizawa M, Yamamoto M and Igarashi K. (1996). Mol. Cell. Biol. 16, 6083-6095. MEDLINE

Palacios R, Karasuyama H and Rolink A. (1987). EMBO. J. 6, 3687-3693. MEDLINE

Puck TT, Cieciura SJ and Robinson A. (1958). J. Exp. Med. 108, 945-956.

Ralph P and Nakoinz I. (1973). J. Natl. Cancer Inst. 31, 883-890.

Ralph P and Nakoinz I. (1977). Cancer Res. 37, 546-550. MEDLINE

Reid A, Gould A, Brand N, Cook M, Strutt P, Li J, Licht J, Waxman S, Krumlauf R and Zelent A. (1995). Blood 86, 4544-4552. MEDLINE

Sambrook J, Fritsch EF and Maniatis T. (1989). Molecular Cloning: A Laboratory Manual, 2nd edn. Cold Spring Harbor Laboratory: Cold Spring Harbor, NY.,

Seyfert VL, Allman D, He Y and Staudt LM. (1996). Oncogene 12, 2331-2342. MEDLINE

Shaknovich R. PL Yeyati P, Ivins I, Melnick A, Lempert C, Waxman S, Zelent Z and Licht J. (1998). Mol. Cell. Biol. in press. .

Shi Y, Mohapatra G, Miller J, Hanahan D, Lander E, Gold P, Pinkel D and Gray J. (1997). Genomics 45, 42-47. Article MEDLINE

Shirasuna K, Sato M and Miyazaki T. (1981). Cancer 48, 745-752. MEDLINE

Shivdasani RA and Orkin SH. (1996). Blood 87, 4025-4039. MEDLINE

Valtieri M. Tweardy DJ, Caracciolo D, Johnson K, Mavilio F, Altmann S, Santoli D and Rovera G. (1987). J. Immun. 138, 3829-3835. MEDLINE

Warner NL, Moore MAS and Metcalf D. (1969). J. Natl. Cancer Inst. 43, 963-982. MEDLINE

Weima SM, van Rooijen MA, Mummery CL, Feijen A, Kruijer W, de Laat SW and van Zoelen EJJ. (1988). Differentiation 38, 203-210. MEDLINE

Weiss A. (1984). J. Immun. 133, 123-128. MEDLINE

Widom RL, Culic I, Lee JY and Korn JH. (1997). Gene. 198, 407-420. MEDLINE

Wilkinson DG. (1992). In Situ Hybridization; A Practical Approach. Wilkinson, DG. (ed.). IRL Press: Oxford, pp. 75-83.

Witzgall R, O'Leary E, Leaf A, Önaldi D and Bonventre JV. (1994). Proc. Natl. Acad. Sci. USA 91, 4514-4518. MEDLINE

Xue F and Cooley L. (1993). Cell 72, 681-693. MEDLINE

Yancopoulos GD, Desiderio SV, Paskind M, Kearney JF, Baltimore D and Alt FW. (1984). Nature 311, 727-733. MEDLINE

Ye BH, Cattoretti G, Shen Q, Zhang J, Hawe N, de Waard R, Leung C, Nouri-Shirazi M, Orazi A, Chaganti RSK, Rothman P, Stall AM, Pandolfi PP and Dalla-Favera. (1997). Nat. Genet. 16, 161-170. MEDLINE

Ye BH, Chaganti S, Chang CC, Niu H, Corradini P, Chaganti RSK and Dalla-Favera R. (1995). Embo. J. 14, 6209-6217. MEDLINE

Ye BH, Lista F, Lo Coco F, Knowles DM, Offit K, Chaganti RSK and Dalla-Favera R. (1993). Science 262, 747-750. MEDLINE

Zelent A. (1997). Methods in Molecular Biology, Vol. 89: Retinoid Protocols. Redfern C. (ed.). Humana Press Inc: Totowa, NJ, pp. 307-332.,

Zelent A, Krust A, Petkovich M, Kastner P and Chambon P. (1989). Nature 339, 714-717. MEDLINE

Zollman S, Godt D, Prive G, JL C and Laski F. (1994). Proc. Natl. Acad. Sci. USA 91, 10717-10721. MEDLINE

Figure 1 Deduced amino acid sequence of the LRF protein. (a) The murine LRF sequence, numbered on the left, and its conservation with the chicken protein. Amino acids which differ in the corresponding positions in the chicken homologue are indicated underneath the murine sequence. Gaps, introduced to obtain optimal alignment, are indicated with a hyphen. The BTB/POZ domain and Zn-finger sequences are underlined with double and single lines, respectively. The fourth Zn-finger, which diverges from the classical C₂-H₂ krüppel-like motif, is underlined with a dotted line. (b) Alignment of the LRF BTB/POZ domain with those of LAZ-3/BCL-6 and PLZF. Sequences, which are numbered on the left with respect to their methionine initiation codons, were aligned using Microgenie software (Beckmann). Black and grey backgrounds are used to indicate identical and conserved residues, respectively. Conserved amino acid substitutions are defined according to the scheme (A,S,T), (Q,N), (E,D), (F,Y), (H,K,R) and (I,L,M,V). Consensus sequence is shown at the top of the alignment. +/- represent positively/negatively charged amino acid. Arrows indicate the relative position of the degenerate oligonucleotide primers used in the cloning of partial LRF cDNAs by RT/PCR

Figure 2 Chromosomal localization of the mouse and human LRF genes. (a) Ideogram showing the location of the mouse LRF gene on chromosome 10 within a region corresponding to band B5.3. A1, FISH of metaphase chromosomes with the digoxigenin-dUTP labelled murine LRF P1 genomic clone. A2, cohybridization of metaphase chromosomes with the same probe as in A1 and a probe specific for the centromeric region of murine chromosome 10 (Shi et al., 1997), indicated with arrowheads and arrows, respectively. A1' and A2' represent maginfication of areas boxed in A1 and A2. (b) Ideogram showing the localization of the human LRF gene to chromosome 19p13.3. B1, FISH of metaphase chromosomes with the digoxigenin-dUTP labelled human LRF P1 genomic clone. Cohybridization of metaphase chromosomes with the above probe and a probe specific for the chromosomal 19q13.4 (unpublished marker, Genome Systems, Inc.), as indicated with arrowheads and arrows, respectively. B1' and B2' represent enlargements of areas boxed in B1 and B2

Figure 3 Expression of the murine LRF gene. (a) Expression of three mouse LRF (mLRF) transcripts (A, B and C) in selected adult tissues (lanes 1 - 6) and various cell lines (lanes 7 - 18). Where indicated, cells were treated for 24 h with 10^-6 M RA. Hybridization of the blot with a ³²P-labelled GAPDH cDNA probe (lower panel) was used to control for mRNA integrity. Either 3 g of poly(A)⁺ (lanes 1 - 12), or 10 g of total (lanes 13 - 22), RNA was loaded in each lane. Size markers (in kb) are indicated on the right of each panel. (b) Expression of two predominant mouse LRF transcripts (A and C) in embryos at 11 - 17 d.p.c. The 2 kb actin signal (lower panel) indicates RNA integrity. All lanes received 3 g of poly(A)⁺ RNA. (c) Whole mount in situ hybridization of two 9.5 - 10.0 d.p.c. mouse embryos. Staining in the otocyst (ot) is artefactual, due to probe trapping. LRF transcripts are present in the facial region, limb buds, tail bud (t), umbilical cord (c) and placenta (p); a low level of expression appears to be also present in the ventral neural tube (n). The specimen on the left (1) is at a slightly earlier developmental stage than that on the right (2). Hindlimb buds are half a day behind forelimb buds in their development at this stage. In the face, LRF expression is detected in the maxillary (mx) and mandibular (md) parts of the first pharyngeal arch and in the second arch (2). In the hindlimb bud (h), strong staining is present at the posterior margin in the specimen on the left (arrow); in the right hand specimen, hindlimb expression is more diffuse, but still biased to the posterior (top left) side of the bud. Forelimb expression is diffuse throughout the limb bud, but transcript levels are highest in the apical ectodermal ridge and underlying progress zone mesenchyme, with a slight bias towards the posterior aspect of the bud

Figure 4 Expression of the human LRF gene in different cell lines (b) and tissues (b). Where indicated, prior to RNA isolation cells were treated for 24 h with 10^-6 M RA. 3 g of poly(A)⁺ RNA were loaded in each lane. Three predominant human LRF transcripts (hLRF), equivalent in size to the murine transcripts A, B and C (see Figure 3) are indicated. The 2 kb actin signal (lower panel) reflects the RNA integrity and the size markers (in kb) are indicated on the right of each panel

Figure 5 LRF associates with the product of the LAZ-3/BCL-6 proto-oncogene in vivo. (a) Coimmunoprecipitation of epitope tagged LRF (LRF^FLAG) and LAZ-3/BCL-6, transiently coexpressed in COS-1 cells. ³⁵S-labelled proteins were coimmunoprecipitated with an anti-FLAG M2 (Kodak) antibody (lanes 1 - 4), resolved by SDS/PAGE and Western blotted with a polyclonal anti-LAZ-3/BCL-6 antiserum (lanes 5 - 8). Size markers in kDa are indicated to the right of the panel. (b) Coimmunofluorescence of LRF^FLAG and LAZ-3/BCL-6, transiently coexpressed in CHO cells. Cells were stained with mouse monoclonal anti-FLAG and rabbit polyclonal anti-LAZ-3/BCL-6 antibodies, counterstained with FITC and TRITC conjugated goat anti-mouse and anti-rabbit IgGs (Jackson ImmunoResearch Laboratories, Inc.), respectively, and analysed by laser scanning confocal microscopy at approximately 0.4 micron resolution. No crossreactivity was observed between the anti-LAZ-3/BCL-6 and anti-FLAG antibodies, or the secondary antibodies, when control cells transfected either with the LRF^FLAG or LAZ-3/BCL-6 expression vectors were used for coimmunofluorescence experiments

Figure 6 LRF and LAZ-3/BCL-6 interact with each other in the yeast two hybrid assay. -galactosidase filter assay showing homodimeric and heterodimeric interactions between the LAZ-3/BCL-6 and LRF proteins, and their indicated deletion mutants, when fused to the GAL4 activation (AD) and/or DNA binding (DBD) domains. LRF and LAZ-3/BCL-6 protein sequences are drawn in red and green colours, respectively, and different portions of each protein (BTB/POZ domain, Zn-finger motifs, and the intervening P-Z region) are as indicated. Interaction between indicated combination of GAL4 AD and DBD fusion proteins is reflected by appearance of a blue colour. Expression of indicated LAZ-3/BCL-6 or LRF GAL4 fusion proteins without a given partner protein did not induced -galactosidase expression and did not produce blue colour (data not shown)