Culture, differentiation, and transduction of mouse E12.5 pancreatic spheres: an in vitro model for the secondary transition of pancreas development.

During the secondary transition of rodent pancreatic development, mainly between E12.5 and E15.5 in mice, exocrine and endocrine populations differentiate from pancreatic progenitors. Here we describe an experimental system for its study in vitro. First, we show that spheres derived from dissociated E12.5 mouse pancreases differentiate within 7 days into most pancreatic exocrine and endocrine cell types, including beta cells. The proportion and spatial repartition of the different endocrine populations mirror those observed during normal development. Thus, dissociation and culture do not impair the developmental events affecting pancreatic progenitors during the secondary transition. Moreover, dissociated cells from mouse E12.5 pancreas were transduced with ecotropic MLV-based retroviral vectors or, though less efficiently, with a mixture of ALV(A)-based retroviral vectors and gesicles containing the TVA (Tumor Virus A) receptor. As an additional improvement, we also created a transgenic mouse line expressing TVA under the control of the 4.5 kB pdx1 promoter (pdx1-TVA). We demonstrate that pancreatic progenitors from dissociated pdx1-TVA pancreas can be specifically transduced by ALV(A)-based retroviral vectors. Using this model, we expressed an activated mutant of the YAP transcriptional co-activator in pancreatic progenitors. These experiments indicate that deregulated YAP activity reduces endocrine and exocrine differentiation in the resulting spheres, confirming and extending previously published data. Thus, our experimental model recapitulates in vitro the crucial developmental decisions arising at the secondary transition and provides a convenient tool to study their genetic control.

The RB gene family controls the maturation state of the EndoC-ßH2 human pancreatic ß-cells.

The functional maturation of human pancreatic ß-cells remains poorly understood. EndoC-ßH2 is a human ß-cell line with a reversible immortalized phenotype. Removal of the two oncogenes, SV40LT and hTERT introduced for its propagation, stops proliferation, triggers cell size increase and senescence, promotes mitochondrial activity and amplifies several ß-cell traits and functions. Overall, these events recapitulate several aspects of functional ß-cell maturation. We report here that selective depletion of SV40LT, but not of hTERT, is sufficient to revert EndoC-ßH2 immortalization. SV40LT inhibits the activity of the RB family members and of P53. In EndoC-ßH2 cells, the knock-down of RB itself, and, to a lesser extent, of its relative P130, precludes most events triggered by SV40LT depletion. In contrast, the knock-down of P53 does not prevent reversion of immortalization. Thus, an increase in RB and P130 activity, but not in P53 activity, is required for functional maturation of EndoC-ßH2 cells upon SV40LT-depletion. In addition, RB and/or P130 depletion in SV40LT-expressing EndoC-ßH2 cells decreases cell size, stimulates proliferation, and decreases the expression of key ß-cell genes. Thus, despite SV40LT expression, EndoC-ßH2 cells have a residual RB activity, which when suppressed reverts them to a more immature phenotype. These results show that the expression and activity levels of RB family members, especially RB itself, regulate the maturation state of EndoC-ßH2 cells.

The supply chain of human pancreatic ß cell lines.

Patients with type 1 or type 2 diabetes have an insufficiency in their functional ß cell mass. To advance diabetes treatment and to work toward a cure, a better understanding of how to protect the pancreatic ß cells against autoimmune or metabolic assaults (e.g., obesity, gestation) will be required. Over the past decades, ß cell protection has been extensively investigated in rodents both in vivo and in vitro using isolated islets or rodent ß cell lines. Transferring these rodent data to humans has long been challenging, at least partly for technical reasons: primary human islet preparations were scarce and functional human ß cell lines were lacking. In 2011, we described a robust protocol of targeted oncogenesis in human fetal pancreas and produced the first functional human ß cell line, and in subsequent years additional lines with specific traits. These cell lines are currently used by more than 150 academic and industrial laboratories worldwide. In this Review, we first explain how we developed the human ß cell lines and why we think we succeeded where others, despite major efforts, did not. Next, we discuss the use of such functional human ß cell lines and share some perspectives on their use to advance diabetes research.

New α- and SIN γ-retrovectors for safe transduction and specific transgene expression in pancreatic ß cell lines.

BACKGROUND: Viral vectors are invaluable tools to transfer genes and/or regulatory sequences into differentiated cells such as pancreatic cells. To date, several kinds of viral vectors have been used to transduce different pancreatic cell types, including insulin-producing ß cells. However, few studies have used vectors derived from « simple ¼ retroviruses, such as avian α- or mouse γ-retroviruses, despite their high experimental convenience. Moreover, such vectors were never designed to specifically target transgene expression into ß cells. RESULTS: We here describe two novel α- or SIN (Self-Inactivating) γ-retrovectors containing the RIP (Rat Insulin Promoter) as internal promoter. These two retrovectors are easily produced in standard BSL2 conditions, rapidly concentrated if needed, and harbor a large multiple cloning site. For the SIN γ-retrovector, either the VSV-G (pantropic) or the retroviral ecotropic (rodent specific) envelope was used. For the α-retrovector, we used the A type envelope, as its receptor, termed TVA, is only naturally present in avian cells and can efficiently be provided to mammalian ß cells through either exogenous expression upon cDNA transfer or gesicle-mediated delivery of the protein. As expected, the transgenes cloned into the two RIP-containing retrovectors displayed a strong preferential expression in ß over non-ß cells compared to transgenes cloned in their non-RIP (CMV- or LTR-) regulated counterparts. We further show that RIP activity of both retrovectors mirrored fluctuations affecting endogenous INSULIN gene expression in human ß cells. Finally, both α- and SIN γ-retrovectors were extremely poorly mobilized by the BXV1 xenotropic retrovirus, a common invader of human cells grown in immunodeficient mice, and, most notably, of human ß cell lines. CONCLUSION: Our novel α- and SIN γ-retrovectors are safe and convenient tools to stably and specifically express transgene(s) in mammalian ß cells. Moreover, they both reproduce some regulatory patterns affecting INSULIN gene expression. Thus, they provide a helpful tool to both study the genetic control of ß cell function and monitor changes in their differentiation status.

Modeling human pancreatic beta cell dedifferentiation.

OBJECTIVE: Dedifferentiation could explain reduced functional pancreatic ß-cell mass in type 2 diabetes (T2D). METHODS: Here we model human ß-cell dedifferentiation using growth factor stimulation in the human ß-cell line, EndoC-ßH1, and human pancreatic islets. RESULTS: Fibroblast growth factor 2 (FGF2) treatment reduced expression of ß-cell markers, (INS, MAFB, SLC2A2, SLC30A8, and GCK) and activated ectopic expression of MYC, HES1, SOX9, and NEUROG3. FGF2-induced dedifferentiation was time- and dose-dependent and reversible upon wash-out. Furthermore, FGF2 treatment induced expression of TNFRSF11B, a decoy receptor for RANKL and protected ß-cells against RANKL signaling. Finally, analyses of transcriptomic data revealed increased FGF2 expression in ductal, endothelial, and stellate cells in pancreas from T2D patients, whereas FGFR1, SOX,9 and HES1 expression increased in islets from T2D patients. CONCLUSIONS: We thus developed an FGF2-induced model of human ß-cell dedifferentiation, identified new markers of dedifferentiation, and found evidence for increased pancreatic FGF2, FGFR1, and ß-cell dedifferentiation in T2D.

Xenotropic retrovirus Bxv1 in human pancreatic ß cell lines.

It has been reported that endogenous retroviruses can contaminate human cell lines that have been passaged as xenotransplants in immunocompromised mice. We previously developed and described 2 human pancreatic ß cell lines (EndoC-ßH1 and EndoC-ßH2) that were generated in this way. Here, we have shown that B10 xenotropic virus 1 (Bxv1), a xenotropic endogenous murine leukemia virus (MuLV), is present in these 2 recently described cell lines. We determined that Bxv1 was also present in SCID mice that were used for in vivo propagation of EndoC-ßH1/2 cells, suggesting that contamination occurred during xenotransplantation. EndoC-ßH1/2 cells released Bxv1 particles that propagated to human 293T and Mus dunni cells. Mobilization assays demonstrated that Bxv1 transcomplements defective MuLV-based retrovectors. In contrast, common rodent ß cell lines, rat INS-1E and RIN-5F cells and mouse MIN6 and ßTC3 cells, displayed either no or extremely weak xenotropic helper activity toward MuLV-based retrovectors, although xenotropic retrovirus sequences and transcripts were detected in both mouse cell lines. Bxv1 propagation from EndoC-ßH1/2 to 293T cells occurred only under optimized conditions and was overall poorly efficient. Thus, although our data imply that MuLV-based retrovectors should be cautiously used in EndoC-ßH1/2 cells, our results indicate that an involuntary propagation of Bxv1 from these cells can be easily avoided with good laboratory practices.

[Punish or cherish: p53, metabolism and tumor suppression]. / Protéger et sévir : p53, métabolisme et suppression tumorale.

The p53 gene is essential for tumor suppression, but how it does so remains unclear. Upon genotoxic or oncogenic stresses, increased p53 activity induces transient cell cycle arrest, senescence or apoptosis, the three cornerstones of the so-called triumvirate. Accordingly, it has long been thought that p53 suppresses tumorigenesis by somehow counteracting cell proliferation or survival. However, several recently described genetically modified mice indicate that p53 can suppress tumorigenesis without triggering these three responses. Rather, as an important mechanism for tumor suppression, these mutant mice point to the ability of p53 to prevent the Warburg effect, that is to dampen glycolysis and foster mitochondrial respiration. Interestingly, these metabolic functions of p53 rely, in part, on its "unstressed" (basal) expression, a feature shared by its mechanistically linked anti-oxydant function. Together, these "conservative" activities of p53 may prevent tumor initiation by promoting and maintaining a normal oxidative metabolism and hence underly the "daily" tumor suppression by p53 in most cells. Conversely, destructive activities elicited by high p53 levels and leading to senescence or apoptosis provide a shield against partially or overtly transformed cells. This last situation, although relatively infrequent throughout life, is usual in experimental settings, which could explain the disproportionally high number of data implicating the triumvirate in tumor suppression by p53.

Distinct effects of the soluble versus membrane-bound forms of the notch ligand delta-4 on human CD34+CD38low cell expansion and differentiation.

Although Notch ligands are considered to activate signaling through direct cell-cell contact, the existence of soluble forms has been demonstrated. However, their roles remain controversial: soluble forms have been reported to mimic the biological activity of membrane-bound form, whereas other studies rather suggested an antagonistic activity toward their full-length counterparts. We previously observed that membrane-bound Delta4-expressing S17 stroma (mbD4/S17) reduced human CD34+CD38(low) cell proliferation and favored self-renewal. Here, we assessed the effects of a soluble form of Delta4 (solD4) by exposing CD34+CD38(low) cells to S17 feeders engineered to express solD4 (solD4/S17). In contrast to mbD4/S17, (a) solD4/S17 increased 10-fold cell production after 2 weeks, through enhanced cell proliferation, and (b) it did not preserve colony-forming cell and long-term culture-initiating cell potential of output CD34+ cells. mbD4 and solD4 appeared to also differ in their signaling. Indeed, mbD4, but not solD4, strongly activated both CSL (the nuclear mediator of Notch signaling) in Hela cells overexpressing Notch1 and transcription of some classic Notch target genes in CD34+CD38(low) cells. Furthermore, both biological effects and CSL activation elicited by mbD4 were strictly dependent upon the gamma-secretase complex, whereas solD4 enhanced cell expansion in a partially gamma-secretase-independent manner. Altogether, these results suggest that part of solD4 activity did not rely upon canonical Notch pathway.

Gene transfer to pre-hematopoietic and committed hematopoietic precursors in the early mouse yolk sac: a comparative study between in situ electroporation and retroviral transduction.

BACKGROUND: Hematopoietic development in vertebrate embryos results from the sequential contribution of two pools of precursors independently generated. While intra-embryonic precursors harbour the features of hematopoietic stem cells (HSC), precursors formed earlier in the yolk sac (YS) display limited differentiation and self-renewal potentials. The mechanisms leading to the generation of the precursors in both sites are still largely unknown, as are the molecular basis underlying their different potential. A possible approach to assess the role of candidate genes is to transfer or modulate their expression/activity in both sites. We thus designed and compared transduction protocols to target either native extra-embryonic precursors, or hematopoietic precursors. RESULTS: One transduction protocol involves transient modification of gene expression through in situ electroporation of the prospective blood islands, which allows the evolution of transfected mesodermal cells in their "normal" environment, upon organ culture. Following in situ electroporation of a GFP reporter construct into the YS cavity of embryos at post-streak (mesodermal/pre-hematopoietic precursors) or early somite (hematopoietic precursors) stages, high GFP expression levels as well as a good preservation of cell viability is observed in YS explants. Moreover, the erythro-myeloid progeny typical of the YS arises from GFP+ mesodermal cells or hematopoietic precursors, even if the number of targeted precursors is low. The second approach, based on retroviral transduction allows a very efficient transduction of large precursor numbers, but may only be used to target 8 dpc YS hematopoietic precursors. Again, transduced cells generate a progeny quantitatively and qualitatively similar to that of control YS. CONCLUSION: We thus provide two protocols whose combination may allow a thorough study of both early and late events of hematopoietic development in the murine YS. In situ electroporation constitutes the only possible gene transfer method to transduce mesodermal/pre-hematopoietic precursors and analyze the earliest steps of hematopoietic development. Both in situ electroporation and retroviral transduction may be used to target early hematopoietic precursors, but the latter appears more convenient if a large pool of stably transduced cells is required. We discuss the assets and limitation of both methods, which may be alternatively chosen depending on scientific constraints.

lyl-1 and tal-1/scl, two genes encoding closely related bHLH transcription factors, display highly overlapping expression patterns during cardiovascular and hematopoietic ontogeny.

The TAL-1/SCL and LYL-1 genes encode two closely related basic helix-loop-helix transcription factors involved in child T-acute lymphoblastic leukemia through chromosomal rearrangements and transcriptional deregulation. During ontogeny, Tal-1/SCL is required for hematopoietic cell generation, both in the yolk sac, where erythro-myeloid cells are first produced, then in the intra-embryonic compartment, where hematopoietic stem cells independently arise. We describe here the expression pattern of lyl-1 in mouse embryos from 7 to 14 days post coitus using in situ hybridization, as well as beta-Galactosidase (beta-Gal) expression in lyl-1-lacZ knock-in embryos, which express a C-terminally truncated Lyl-1 protein fused to the beta-Galactosidase (Lyl-1Delta/beta-Gal). In addition, we compare lyl-1 expression pattern with that of tal-1/scl. Similar to Tal-1/SCL, Lyl-1 mRNA expression occurs in the developing cardiovascular and hematopoietic systems. However, contrary to tal-1/scl, lyl-1 is not expressed in the developing nervous system. In lyl-1-lacZ knock-in heterozygous and homozygous embryos, beta-Gal expression completely correlates with Lyl-1 mRNA expression in the intra-embryonic compartment and is present: (1) in the developing hematopoietic system, precisely where hematopoietic stem cells emerge, and thereafter in the fetal liver; (2) in the developing vascular system; and (3) in the endocardium. In contrast, whereas Lyl-1 mRNA is expressed in yolk sac-derived endothelial and hematopoietic cells, Lyl-1Delta/beta-Gal is either absent or poorly expressed in these cell types, thus differing from Tal-1/SCL, which is highly expressed there at both mRNA and protein levels.

[Myc and cell competition in Drosophila]. / Myc et compétitions intercellulaires chez la drosophile.

Cell differentiation and organ shaping proceed not only upon instructive but also upon competitive cell-cell interactions. In the proliferating epithelium forming the larval Drosophila wing disc, cell competition contributes to the fidelity of the organogenesis. Several recent studies show how d-myc, encoding a bHLH/LZ transcription factor homologous to vertebrate Myc proteins, controls cell competition during wing development. In this model, any experiment leading to the confrontation of two populations differing by their levels of d-Myc expression, even in a two-fold ratio, gives rise to a competition characterized both by an overgrowth of the population having the highest level and an apoptotic elimination of the neighbouring << weakly >> expressing cells. As a consequence of the mutually compensating nature of these two processes, the final size of the wing remains unchanged. Importantly, lowering or elevating d-Myc expression to the same extent in all cells of the disc does not trigger competition. This indicates that competition is linked to a spatial heterogeneity in, and not to the absolute level of, d-Myc expression. Both vertebrate and Drosophila Myc proteins stimulate ribosome biogenesis, and genetic evidence in Drosophila suggests that this property underlies the strong competitive advantage imparted by its relatively high expression. Accordingly, it is proposed, although not proved, that the more the wing cells express d-Myc and amplify their protein synthesis apparatus, the more they bind, internalize, and transduce the vital and limiting growth factor Dpp, which in turn is presumed to increase d-Myc protein level. These findings suggest that wing organogenesis is a self-corrected process whereby d-Myc induction in overgrowing cells ensures the compensatory elimination of their neighbours. Moreover, they have important implications for the oncogenic role of vertebrate Myc proteins and possibly of related transcription factors.

The SCL relative LYL-1 is required for fetal and adult hematopoietic stem cell function and B-cell differentiation.

Hematopoietic stem cells (HSCs) arise, self-renew, or give rise to all hematopoietic lineages through the effects of transcription factors activated by signaling cascades. Lyl-1 encodes a transcription factor containing a basic helix-hoop-helix (bHLH) motif closely related to scl/tal, which controls numerous decisions in embryonic and adult hematopoiesis. We report here that Lyl-1 null mice are viable and display normal blood cell counts, except for a reduced number of B cells resulting from a partial block after the pro-B stage. Nevertheless, the deletion of Lyl-1 results in a diminution in the frequency of immature progenitors (Lin(-), CD34(-), sca-1(+), c-kit(+) [LSK], and LSK-side population [LSK-SP]) and in S(12) colony-forming unit (CFU-S(12)) and long-term culture-initiating cell (LTC-IC) content in embryonic day 14 fetal liver (E14 FL) and adult bone marrow (BM). More important, Lyl-1(-/-) E14 FL cells and BM are severely impaired in their competitive reconstituting abilities, especially with respect to B and T lineage reconstitution. Thus, ablation of Lyl-1 quantitatively and functionally affects HSCs, a cell population that transcribes Lyl-1 more actively than their differentiated progenies. Our results demonstrate for the first time that Lyl-1 functions are important for HSC properties and B-cell differentiation and that they are largely distinct from scl functions.

Development of a new bicistronic retroviral vector with strong IRES activity.

BACKGROUND: Internal Ribosome Entry Site (IRES)-based bicistronic vectors are important tools in today's cell biology. Among applications, the expression of two proteins under the control of a unique promoter permits the monitoring of expression of a protein whose biological function is being investigated through the observation of an easily detectable tracer, such as Green Fluorescent Protein (GFP). However, analysis of published results making use of bicistronic vectors indicates that the efficiency of the IRES-controlled expression can vary widely from one vector to another, despite their apparent identical IRES sequences. We investigated the molecular basis for these discrepancies. RESULTS: We observed up to a 10 fold difference in IRES-controlled expression from distinct bicistronic expression vectors harboring the same apparent IRES sequences. We show that the insertion of a HindIII site, in place of the initiating AUG codon of the wild type EMCV IRES, is responsible for the dramatic loss of expression from the second cistron, whereas expression from the first cistron remains unaffected. Thus, while the replacement of the authentic viral initiating AUG by a HindIII site results in the theoretical usage of the initiation codon of the HindIII-subcloned cDNA, the subsequent drop of expression dramatically diminishes the interest of the bicistronic structure. Indeed, insertion of the HindIII site has such a negative effect on IRES function that detection of the IRES-controlled product can be difficult, and sometimes even below the levels of detection. It is striking to observe that this deleterious modification is widely found in available IRES-containing vectors, including commercial ones, despite early reports in the literature stating the importance of the integrity of the initiation codon for optimal IRES function. CONCLUSION: From these observations, we engineered a new vector family, pPRIG, which respects the EMCV IRES structure, and permits easy cloning, tagging, sequencing, and expression of any cDNA in the first cistron, while keeping a high level of expression from its IRES-dependent second cistron (here encoding eGFP).

Microarray analysis of LIF/Stat3 transcriptional targets in embryonic stem cells.

Mouse embryonic stem (ES) cells can be propagated in vitro while retaining their properties of pluripotency and self-renewal under the continuous presence of leukemia inhibitor factor (LIF). An essential role has been attributed to subsequent activation of the Stat3 transcription factor in mediating LIF self-renewal response. To date, however, downstream target genes of Stat3 in ES cells are still unknown. To isolate these genes, we performed a microarray-based kinetic comparison of LIF-stimulated (undifferentiated) ES cells versus ES cells induced to differentiate by shutting down Stat3 activity through either LIF deprivation or, more specifically, expression of a Stat3 dominant-negative mutant. In each case, we chose the earliest time at which ES cells lose their self-renewal properties, as illustrated by a decrease in the number of embryoid bodies and blast cell colony formation as well as germ layer marker expression. Comparison of the two independent approaches revealed similarly regulated genes that are likely to be involved in the Stat3 effects on ES cell self-renewal. For instance, upregulation of growth factors such as the transforming growth factor-beta relative Lefty1 or transcriptional regulators such as Id1 and Id2 and down-regulation of the groucho-like protein Aes1 (grg5) were found. Promoter analysis of the aes1 gene revealed three functional Stat3 consensus sites, as shown by luciferase assays. Furthermore, chromatin immunoprecipitation experiment demonstrated that Stat3 is recruited to the promoter of aes1 in ES cells. These data demonstrated that the aes1 gene is a direct transcriptional target of Stat3 in ES cells.

FLI1 monoallelic expression combined with its hemizygous loss underlies Paris-Trousseau/Jacobsen thrombopenia.

Paris-Trousseau syndrome (PTS; also known as Jacobsen syndrome) is characterized by several congenital anomalies including a dysmegakaryopoiesis with two morphologically distinct populations of megakaryocytes (MKs). PTS patients harbor deletions on the long arm of chromosome 11, including the FLI1 gene, which encodes a transcription factor essential for megakaryopoiesis. We show here that lentivirus-mediated overexpression of FLI1 in patient CD34(+) cells restores the megakaryopoiesis in vitro, indicating that FLI1 hemizygous deletion contributes to the PTS hematopoietic defects. FISH analysis on pre-mRNA and single-cell RT-PCR revealed that FLI1 expression is mainly monoallelic in CD41(+)CD42(-) progenitors, while it is predominantly biallelic in the other stages of megakaryopoiesis. In PTS cells, the hemizygous deletion of FLI1 generates a subpopulation of CD41(+)CD42(-) cells completely lacking FLI1 transcription. We propose that the absence of FLI1 expression in these CD41(+)CD42(-) cells might prevent their differentiation, which could explain the segregation of the PTS MKs into two subpopulations: one normal and one composed of small immature MKs undergoing a massive lysis, presumably originating from either FLI1(+) or FLI1(-) CD41(+)CD42(-) cells, respectively. Thus, we point to the role of transient monoallelic expression of a gene essential for differentiation in the genesis of human haploinsufficiency-associated disease and suggest that such a mechanism may be involved in the pathogenesis of other congenital or acquired genetic diseases.

Point mutations in BCL6 DNA-binding domain reveal distinct roles for the six zinc fingers.

The B-cell lymphoma 6 (BCL6) gene encodes a transcriptional repressor containing six C-terminal Krüppel-like zinc fingers. The zinc finger (ZF) cluster is necessary and sufficient for interaction with both DNA and several proteins and for nuclear targeting. However, the functional specificity of the six ZFs in these cellular roles is unknown. To characterize this domain, we mutated individually each ZF of BCL6. Our results reveal that mutation of the two N-terminal ZFs does not impair cognate DNA-binding, cellular localization of the protein nor the transcriptional repression capacity of BCL6. By contrast, mutation of any of the remaining ZFs abolishes the binding of BCL6 to DNA in vitro and the transrepressive function of the protein in vivo. Finally, none of the six mutations affect the interaction between BCL6 and class II histone deacetylases. Thus our experiments demonstrate that BCL6 uses each of the four C-terminus ZFs for binding to a target sequence while the two amino terminal fingers are likely engaged in other unknown function(s).

Class II histone deacetylases are directly recruited by BCL6 transcriptional repressor.

BCL6 is a member of the POZ/zinc finger (POK) family involved in survival and/or differentiation of a number of cell types and in B cell lymphoma upon chromosomal alteration. Transcriptional repression by BCL6 is thought to be achieved in part by recruiting a repressor complex containing two class I histone deacetylases (HDACs). In this study we investigated whether BCL6 could also target members of class II HDACs. Our results indicate that three related class II deacetylases, HDAC4, HDAC5, and HDAC7 can associate with BCL6 in vivo and in vitro. Using electron microscopy, we found that endogenous BCL6 and class II HDACs partially co-localize in the nucleus. Overexpression experiments showed that BCL6 and HDAC4, -5, or -7 are intermingled onto common nuclear substructures and form stable complexes. A highly conserved domain in the N-terminal region of HDAC5 and HDAC7 as well as the zinc finger region of BCL6 were found necessary for the complex formation in vivo and in vitro. Moreover, our data point to the zinc finger region of BCL6 as a multifunctional domain which, beside its known capacity to bind DNA, is involved in the nuclear targeting of the protein and in the recruitment of the class II HDACs, and hence constitutes an autonomous repressor domain. Since PLZF, a BCL6 relative, could also interact with HDAC4, -5, and 7, we suggest that class II HDACs are largely involved in the control of the POK transcription factors activity.