Sumoylation and proteasomal activity determine the transactivation properties of the mineralocorticoid receptor.

MR is a hormone-activated transcription factor that carries a strong synergy inhibitory function at its N-terminus. Using this region as bait in a yeast two-hybrid screening, we isolated major components of the sumoylation pathway, including the SUMO-1-conjugating enzyme Ubc9, and SUMO-1 itself. We found that MR interacts with both, Ubc9 and SUMO-1 in mammalian cells, and that the receptor is sumoylated at four acceptor sites which are clustered within its AF-1 domain. We observed that MR can be poly-ubiquitinated and that proteasome activity is essential for MR-activated transcription. Disruption of the SUMO-1 attachment sites abolished MR sumoylation but interfered with neither the poly-ubiquitination of the receptor nor its transactivation potential on MMTV. However, the hormone-activated mutant displayed enhanced synergistic potential on a compound promoter and delayed mobility in the nucleus. FRAP analysis further showed that proteasome inhibition immobilizes a subpopulation of unliganded MR receptors in the nucleus, a phenomenon that is significantly attenuated in the presence of aldosterone. Interestingly, the ability of the hormone to counteract the immobilizing effect of MG132 requires the sumoylation-competent form of MR. Moreover, increasing exogenously SUMO-1 cellular levels resulted in a selective, dose-dependent inhibition of the activity of the sumoylation-deficient MR. This effect was observed only on a synergy-competent promoter, revealing a mode for negative regulation of synergy that might involve sumoylation of factors different from MR. The data suggest that the overall transcriptional activity of MR can be modulated by its sumoylation potential as well as the sumoylation level of MR-interacting proteins, and requires the continuous function of the proteasome.

The manifold actions of the protein inhibitor of activated STAT proteins on the transcriptional activity of mineralocorticoid and glucocorticoid receptors in neural cells.

Corticosteroid actions in the brain are exerted via the mineralocorticoid receptor (MR) and glucocorticoid receptor (GR). These receptors share several structural and functional similarities but their activation in the brain triggers distinct biological actions, for instance on neuronal survival or the regulation of the hypothalamo-pituitary-adrenal axis. Like other hormone-activated receptors, the transcriptional properties of the MR and GR depend on their ability to recruit a variety of co-regulators, which modulate their activity on target promoters, in a specific manner. The N-terminal regions of the MR and GR share the smallest degree of sequence conservation, whereas they display opposite effects on the transactivation properties of these receptors; thus, they may provide surfaces suitable for receptorspecific interactions with co-regulatory proteins. Here, we employed a yeast two-hybrid system to identify molecules interacting with the N-terminal part of the MR (amino acids 170-433). This approach resulted in the isolation of representative cDNAs from all members of the protein inhibitor of activated STAT (PIAS) family of proteins as potential MR-interacting partners. In neural cells, PIAS3 exhibited a strong and specific interaction with MR, but not GR, as indicated by mammalian two-hybrid assays and co-immunoprecipitation experiments in vivo. The interaction with MR was enhanced in the presence of aldosterone, an MR agonist, and was found to occur through a conserved, serine- and acidic amino acid residue-rich domain of PIAS3. To compare the modulatory properties of PIAS proteins on MR and GR transcriptional activity in a neural environment, MMTV reporter gene assays were performed in the human neuroblastoma cell line SK-N-MC. This analysis revealed that PIAS3 can inhibit MR, but not GR, transactivation in response to their corresponding ligands. Further, it showed that PIAS1 and PIASxbeta, but not PIASy, could also inhibit MR-mediated transcription despite the lack of detected physical interaction with MR. Interestingly, PIASxbeta and PIASy dose-dependently co-activated GR, whereas PIAS1 impaired GR-induced transcription. Taken together the results reveal differential modulatory roles of the PIAS proteins on the transcriptional properties of MR and GR, thus providing new insights into the bifurcating actions of these two receptors in neural cells where they are frequently co-localized.

Subtle shifts in the ratio between pro- and antiapoptotic molecules after activation of corticosteroid receptors decide neuronal fate.

Glucocorticoid receptor (GR) activation induces apoptosis of granule cells in the hippocampus. In contrast, neuroprotection is seen after mineralocorticoid receptor (MR) activation. To date there is no in vivo evidence for direct interactions between corticosteroids and any of the key regulatory molecules of programmed cell death. In this report, we show that the opposing actions of MR and GR on neuronal survival result from their ability to differentially influence the expression of members of the bcl-2 gene family; specifically, in the rat hippocampus, activation of GR induces cell death by increasing the ratio of the proapoptotic molecule Bax relative to the antiapoptotic molecules Bcl-2 or Bcl-x(L); the opposite effect is observed after stimulation of MR. The same results were obtained in both young and aged animals; however, older subjects (which were more susceptible to GR-mediated apoptosis) tended to express the antiapoptotic genes more robustly. Using a loss-of-function mouse model, we corroborated the observations made in the rat, demonstrating Bax to be essential in the GR-mediated cell death-signaling cascade. In addition, we show that GR activation increases and MR activation decreases levels of the tumor suppressor protein p53 (a direct transcriptional regulator of bax and bcl-2 genes), thus providing new information on the early genetic events linking corticosteroid receptors with apoptosis in the nervous system.

Targeted disruption of the bcl-2 gene in mice exacerbates focal ischemic brain injury.

Neuronal death after brain ischemia is mainly due to necrosis but there is also evidence for involvement of apoptosis. To test the importance of apoptosis, we investigated the effect of targeted disruption of the apoptosis-suppressive gene bcl-2 on the severity of ischemic brain injury. Transient focal ischemia for 1 hour was induced by occlusion of the middle cerebral artery in homozygous (n=7) and heterozygous (n=6) bcl-2 knockout mice as well as in their wildtype littermates (n=5). Bcl-2 ablation did not influence cerebral blood flow but it significantly increased infarct size and neurological deficit score at 1 day after reperfusion in a gene-dose dependent manner. The exacerbation of tissue damage in the absence of Bcl-2 underscores the importance of apoptotic pathways for the manifestation of ischemic injury after transient vascular occlusion.

Alterations in cell death and cell cycle progression in the UV-irradiated epidermis of bcl-2-deficient mice.

The effect of bcl-2 gene ablation on epidermal cell death induced by UV-B irradiation was investigated in mice. Exposure of depilated back skin of bcl-2-/- mice to 0.5 J/cm2 UV-B caused a prolonged increase in the number of epidermal cells showing nuclear DNA fragmentation compared to wild-type littermates. Consistently, skin explants from bcl-2-deficient mice exhibited a higher number of sunburn cells per cm epidermis (16.6+/-2.1 vs 7.0+/-1.5) following exposure to 0.1 J/cm2 UV-B in vitro. Furthermore, UV irradiation failed to increase pre-melanosomes in skin explants from mutant animals, and primary menalocyte cultures derived from bcl-2 null mutants were highly susceptible to UV-induced cell death compared to cultures from wild-type littermates. An accelerated reappearance of proliferating cells, showing nuclear immunoreactivity for Ki-67 and c-Fos, was observed in the UV-irradiated epidermis of bcl-2-deficient mice. Taken together, these findings suggest that effects of UV radiation on epidermal cell death and cell cycle progression are influenced by survival-promoting Bcl-2.

Mitotic silencing of human rRNA synthesis: inactivation of the promoter selectivity factor SL1 by cdc2/cyclin B-mediated phosphorylation.

We have used a reconstituted cell-free transcription system to investigate the molecular basis of mitotic repression of RNA polymerase I (pol I) transcription. We demonstrate that SL1, the TBP-containing promoter-binding factor, is inactivated by cdc2/cyclin B-directed phosphorylation, and reactivated by dephosphorylation. Transcriptional inactivation in vitro is accompanied by phosphorylation of two subunits, e.g. TBP and hTAFI110. To distinguish whether transcriptional repression is due to phosphorylation of TBP, hTAFI110 or both, SL1 was purified from two HeLa cell lines that express either full-length or the core domain of TBP only. Both TBP-TAFI complexes exhibit similar activity and both are repressed at mitosis, indicating that the variable N-terminal domain which contains multiple target sites for cdc2/cyclin B phosphorylation is dispensable for mitotic repression. Protein-protein interaction studies reveal that mitotic phosphorylation impairs the interaction of SL1 with UBF. The results suggest that phosphorylation of SL1 is used as a molecular switch to prevent pre-initiation complex formation and to shut down rDNA transcription at mitosis.

The bcl-2 knockout mouse exhibits marked changes in osteoblast phenotype and collagen deposition in bone as well as a mild growth plate phenotype.

Histological examination of long bones from 1-day-old bcl-2 knockout and age-matched control mice revealed no obvious differences in length of bone, growth plate architecture or stage of endochondral ossification. In 35-day-old bcl-2 knockout mice that are growth retarded or 'dwarfed'. the proliferative zone of the growth plate appeared slightly thinner and the secondary centres of ossification less well developed than their age-matched wild-type controls. The most marked histological effects of bcl-2 ablation were on osteoblasts and bone. 35-day-old knockout mouse bones exhibited far greater numbers of osteoblasts than controls and the osteoblasts had a cuboidal phenotype in comparison with the normal flattened cell appearance. In addition, the collagen deposited by the osteoblasts in the bcl-2 knockout mouse bone was disorganized in comparison with control tissue and had a pseudo-woven appearance. The results suggest an important role for Bcl-2 in controlling osteoblast phenotype and bone deposition in vivo.

Origin of eukaryotic programmed cell death: a consequence of aerobic metabolism?

A marked feature of eukaryotic programmed cell death is an early drop in mitochondrial transmembrane potential. This results from the opening of permeability transition pores, which are composed of adenine nucleotide translocators and mitochondrial porins. The latter share striking similarities with bacterial porins, including down-regulation of their pore size by purine nucleotides), suggesting a common origin. The porins of some invasive bacteria play a crucial role during their accommodation inside the host cell and this coexistence resembles the endosymbiotic origin of mitochondria. The above observations suggest that early in eukaryotic evolution, former invaders may have used porin-type channels to enter their host and to induce its death when the levels of its cytoplasmic purine nucleotides were dropped. The appearance of adenosine nucleotide translocators in the primitive eukaryotes, which permitted usage of the oxidative metabolism of the invaders, provided the basis for the permeability transition phenomena, now linked to the apoptotic process. Bcl-2-type molecules, being able to modulate the permeability transition pores by interaction with adenosine nucleotide translocators, may have played an essential role in conferring a means of controlling apoptosis.

Introduction of the negative selection marker into replacement vectors by a single ligation step.

Gene targeting is a powerful method for introducing mutations into the genome of embryonic stem cells. The most widely used approach is the positive-negative selection method in which a gene encoding a negative selection marker is cloned into the replacement vector to obtain an enrichment of properly targeted clones. Here, we present an alternative means to introduce any given negative selection marker at the ends of a replacement vector using a single ligation step, thereby avoiding laborious cloning procedures. Our results demonstrate that this fast and simple method consistently provides a high level of enrichment of appropriately targeted clones.

Cell density increases Bcl-2 and Bcl-x expression in addition to survival of cultured cerebellar granule neurons.

The proto-oncogene bcl-2 and its family members, bcl-x and bax are recognized as major regulators of cell death and survival. Although Bcl-2 and Bcl-x are expressed in brain, little is known how they are regulated in neurons. Here we have studied the expression of bcl-2, bcl-xL and bax mRNA in rat cerebellar granule neurons cultured under conditions which influence neuron survival. Insulin-like growth factor-1 and brain-derived neurotrophic factor supported the survival of these neurons, but affected neither the expression of bcl-2, bcl-xL nor bax mRNA. In contrast, bcl-2 and bcl-xL mRNAs were up-regulated in cerebellar granule neurons plated at high density exhibiting an increased neuronal survival. Western blots showed that cell density also increased Bcl-2 protein level. However, conditioned medium from dense cultures did not affect the level of bcl-2 mRNA nor survival of the neurons. This suggests that cell density promotes survival and regulates Bcl-2 expression in cerebellar granule neurons through a signaling pathway different from known neurotrophic factors.

Inactivation of bcl-2 results in progressive degeneration of motoneurons, sympathetic and sensory neurons during early postnatal development.

Bcl-2 is a major regulator of programmed cell death, a critical process in shaping the developing nervous system. To assess whether Bcl-2 is involved in regulating neuronal survival and in mediating the neuroprotective action of neurotrophic factors, we generated Bcl-2-deficient mice. At birth, the number of facial motoneurons, sensory, and sympathetic neurons was not significantly changed, and axotomy-induced degeneration of facial motoneurons could still be prevented by brain-derived neurotrophic factor (BDNF) or ciliary neurotrophic factor (CNTF). Interestingly, substantial degeneration of motoneurons, sensory, and sympathetic neurons occurred after the physiological cell death period. Accordingly, Bcl-2 is not a permissive factor for the action of neurotrophic factors, and although it does not influence prenatal neuronal survival, it is crucial for the maintenance of specific populations of neurons during the early postnatal period.

Novel human glutamate dehydrogenase expressed in neural and testicular tissues and encoded by an X-linked intronless gene.

Glutamate dehydrogenase, an enzyme central to glutamate metabolism, is deficient in patients with heterogeneous neurological disorders characterized by multiple system atrophy. There is evidence for multiplicity of human glutamate dehydrogenase, which may account for the heterogeneity of the above disorders. However, only one mRNA that is encoded by an intron-containing gene (GLUD1) is presently known. Because blindness due to neuroretinal degeneration can occur in rare forms of multiple system atrophy, we searched for retina-specific GLUD mRNA(s) by screening a lambda gt10 library derived from human retina. A novel cDNA encoded by an X chromosome-linked intronless gene, designated GLUD2, was isolated and characterized. Reverse transcription-polymerase chain reaction analysis of human tissues revealed that the novel cDNA is expressed in human retina, testis, and, at lower levels, brain. In vitro translation of mRNAs derived from GLUD1 and GLUD2 genes generated proteins with distinct electrophoretic characteristics. The retinal cDNA was expressed in the baculovirus heterologous system, producing a protein capable of catalyzing the oxidative deamination of glutamate. The mobility of the expressed protein on SDS-polyacrylamide gel electrophoresis and its catalytic properties were very similar to those of the naturally occurring human brain glutamate dehydrogenases. The novel gene will be useful for understanding the biology of human neural and testicular tissues and in the study of X-linked neurodegenerative disorders.

Isolation of the cDNA and chromosomal localization of the gene (TAX1) encoding the human axonal glycoprotein TAG-1.

The transient axonal glycoprotein (TAG-1) is a cell adhesion molecule that promotes neurite outgrowth and belongs to the immunoglobulin superfamily. We have isolated cDNAs encoding TAX1, the human homologue of TAG-1. Human TAX1 shows a high degree of homology to rat TAX1 and less to its chick counterpart, axonin-1, with 91 and 75% identity at the amino acid level, respectively. The numbers of immunoglobulin (IgC2) domains and fibronectin repeats present in TAG-1 are conserved among the three species. The highest degree of conservation occurs in the second IgC2 domain (98% with the rat and 82% with the chick). The human homologue also contains a putative N-terminal signal sequence and a C-terminal hydrophobic sequence, suggestive of linkage to the cell membrane via phosphatidylinositol. In addition, the two mammalian TAG-1 proteins share the RGD tripeptide, a motif known to mediate recognition of fibronectin by integrins. In situ hybridization to human metaphase chromosomes maps the TAX1 gene encoding human TAG-1 to a single location on chromosome 1q32.

The human glutamate dehydrogenase gene family: gene organization and structural characterization.

Glutamate dehydrogenase is a mitochondrially located, key metabolic enzyme. In addition to its general metabolic role, GLUD is important in neurotransmission. Significant alterations in GLUD enzymatic activity have been associated with certain neurodegenerative human disorders. Although a single species of human GLUD cDNA molecule has been identified so far, both genomic DNA Southern and cytogenetic analyses have indicated the presence of a GLUD gene family. Screening of a human genomic lambda-phage library with the GLUD cDNA, led us to the isolation of several clones divided into five structurally distinct contigs. We have confirmed the presence of all GLUD-specific sequences in the human genome by detailed genomic Southern analysis. This study allowed the identification of the entire functional GLUD gene, named GLUD1. The GLUD1 gene is about 45 kb long and it is organized into 13 exons. Its nucleotide sequence, exon-intron boundaries, and transcription start sites were determined. Potential binding sites for various regulatory factors such as Sp1, AP-1, and AP-2 were recognized at the promoter region of the gene. The members of the other contigs showed an organization clearly different from GLUD1. Two distinct GLUD-specific gene loci, termed GLUDP2 and GLUDP3, possibly represent truncated pseudogenes. Their high degree of similarity to GLUD1 is limited to the region surrounding exons 2, 3, and 4. Finally, two additional GLUD-specific genomic sequences, termed GLUDP4 and GLUDP5, are structurally similar with the 3' part of the GLUD cDNA sequence. These loci probably represent truncated GLUD pseudogenes generated by retrotransposition. The data presented here suggest that all human GLUD pseudogenes have diverged recently in evolution.