9 November 2006. Over the last few years, numerous schizophrenia-associated gene candidates have been proposed, raising hopes that some molecular handle on disease pathology may soon be within reach. One of the strongest gene candidates is disrupted in schizophrenia 1 (DISC1), and at the Society for Neuroscience annual meeting last month, some of the newest molecular and neuroscience data on DISC1 were on display.
DISC1 was discovered at the center of a chromosomal breakpoint that results in the translocation of a large segment of chromosome 1. The translocation has been linked to schizophrenia and other psychiatric diagnoses in a large Scottish family. Since that discovery, researchers have scrambled to figure out exactly what DISC1 does and how it may relate to psychopathology. In his presentation, Naoya Sawamura from Akira Sawa’s lab at Johns Hopkins University, Baltimore, Maryland, reported that DISC1 may be a repressor of cAMP response element (CRE)-regulated transcription.
Sawamura and colleagues previously showed that more DISC1 turns up in the nucleus in the brains of schizophrenia patients and those afflicted with substance addiction. To understand this redistribution, the researchers analyzed the DISC1 protein for sequences that might drive nuclear localization or export from the nucleus. Sawamura revealed that DISC1 contains a single nuclear localization signal and three potential nuclear export signals, only one of which appears to be functional.
Sawamura showed that DISC1 partially colocalizes with PML bodies, which may provide some clues as to what the protein is doing in the nucleus. PML bodies (named after the promyelocytic leukemia protein) are nuclear conglomerates that have been linked to various regulatory functions, including apoptosis, DNA repair, tumor suppression, gene regulation, and proteolysis. In addition to PML, the bodies contain many other proteins including Sp100, and they often colocalize with CRE binding protein (CREB) binding protein (CBP), which plays a central role in transcriptional regulation (LaMorte et al., 1998). The fact that DISC1 binds to PML bodies and also CREB2, otherwise known as activating transcription factor 4 (ATF4) (Morris et al., 2003), suggests that it may be playing some role in PML-mediated transcriptional regulation.
To investigate this theory, Sawamura and colleagues tested DISC1 in a CRE-dependent transcriptional assay. They found that in the presence of DISC1 ATF4-mediated repression of transcription was enhanced, suggesting that DISC1 acts as a co-repressor of genes. DISC1 has previously been linked to cAMP signaling: work from David Porteus’s lab showed that the protein interacts with phosphodiesterase 4B, an enzyme that degrades cAMP (see SRF related news story).
What drives localization of DISC1 to the nucleus is unclear, but Sawamura suggested that it might be ATF4. He reported that the two proteins interact via their leucine zippers, motifs essential for dimerization of many transcription factors.
…or disrupting microtubules?
Other potential binding partners for DISC1 include a family of proteins that have been linked to Bardet-Biedl syndrome (BBS), a multifaceted disorder that causes renal dysfunction, obesity, and neural and psychiatric symptoms. BBS can also cause polydactyly, or extra digits, suggesting the syndrome affects development. In his poster presentation, Atsushi Kamiya, also from the Sawa lab at Johns Hopkins, showed that DISC1 forms a complex with BBS proteins (there are eight of them) and PCM1 (pericentriolar material 1)—the PCM1 gene has also been linked to schizophrenia (see SRF related news story).
Kamiya found the BBS interaction by tagging DISC1 with the hemagluttinin (HA) antigen and then immunoprecipitating it from HEK293 cells. That brought along BBS1-8, with the BBS6 contribution being the highest. PCM1 turned up in a yeast two-hybrid screen. By using immunofluorescence, Kamiya found all three proteins localized to the centrosome, suggesting that they may conspire there to regulate microtubules, thereby impacting a variety of cellular processes, including synaptic growth, axonal transport, and cell migration.
By systematically deleting various parts of the protein, Kamiya found that DISC1 is probably the glue that holds the troika together. The N-terminal of DISC1 binds to PCM1, while the C-terminal end can bind to BBS1, 4, or 8 (the researchers have not yet tested the other BBS proteins because antibodies are unavailable, but they may also bind to the same region). In the case of the BBS proteins, it appears that their tetratricopeptide repeats (TPRs) contribute to the BBS-DISC1 interaction, though exactly how is not yet clear. These are degenerate stretches of 34 amino acids, and while deletion of the two N-terminal repeats of BBS 4 (there are 13 TPR motifs in total) abolished binding, when the last four were removed, binding was unaffected. Perhaps there are specific TPR sequences in the N-terminal that are essential for binding.
Regulating DISC1, regulating development
Most researchers in the field would undoubtedly want to know what regulates DISC1 and what DISC1 regulates. The answer to both is likely complex, but phosphorylation might be a good place to start addressing the former. Koko Ishizuka, also in Sawa’s group, has done just that, and reported in her poster presentation that the C-terminal of DISC1 is likely phosphorylated in vivo by cAMP-dependent protein kinase (PKA). Ishizuka used mass spectrometry to detect phosphorylated DISC1 fragments from cells treated with okadaic acid, which prevents dephosphorylation.
Two phosphorylated fragments turned up, one (amino acids 48-66) with three potential phosphorylation sites (tyrosine 50, serine 58, and tyrosine 63), the other with two (serines 713 and 715). Using an in vitro PKA assay, Ishizuka was able to show that serine 713 gets phosphorylated by the kinase—an S713A failed to incorporate phosphate.
The phosphorylation of DISC1 probably has real-life relevance, because when Ishizuka treated brain extracts with phosphatases, the electrophoretic mobility of the protein shifted, indicating that the protein is phosphorylated in vivo. Likewise, treatment of cell extract supernatants with okadaic acid also revealed a mobility shift, indicating that soluble DISC1 gets phosphorylated.
As for the regulatory role of DISC1 itself, while it may conspire with ATF4 and/or other nuclear factors to regulate transcription, another intriguing possibility is that it regulates localized translation, crucial for synaptic plasticity and axon growth. Daisuke Tsuboi, working in the lab of Kozo Kaibuchi at Nagoya University, Japan, reported in his poster that DISC1 binds to ribonucleoproteins, including Purα, hnRNP-U, and RACK1. Tsuboi and colleagues used affinity chromatography to identify these binding partners and then demonstrated that the mRNPs bound to DISC1 in vitro and colocalized with it in hippocampal dendrites. Furthermore, the researchers found that the interactions between DISC1 and these ribonucleoproteins were RNA-dependent, suggesting that DISC1 is a bona fide part of the mRNA binding apparatus. These interactions also suggest that DISC1 is involved in dendritic spine maturation.
Insights into the role of DISC1 are also coming in from in vitro cell studies and in vivo animal models. Mikhail Pletnikov and colleagues, also at Johns Hopkins, have tried both approaches. In his poster, Pletnikov showed that mutant DISC1 resulting from the chromosomal translocation found in the large Scottish pedigree inhibits neurite outgrowth. He used the suppressible tetracycline promoter system to switch DISC1 expression on and off in cells that were incubated with the neurotrophic nerve growth factor (NGF). In older cells (7 days), turning on DISC1 reduced the number of cells with neuritic processes by almost half.
How mutant DISC1 attenuated neuritic growth is unclear, but in support of the findings from Sawa’s lab, Pletnikov found that endogenous DISC1 ends up in the nucleus when mutant DISC1 is turned on. What’s more, Pletnikov found that ATF4 staining in the nucleus becomes more diffuse under these conditions, possibly reflecting an interaction between DISC1 and the transcription factor. Mutant DISC1 also reduced levels of LIS1, a DISC1 binding partner and causative factor in the neurodevelopmental disorder lissencephaly (see SRF related news story).
As for in vivo effects, Pletnikov and colleagues used the same tet-off system, coupled to the neuron-specific calmodulin kinase II promoter, to turn on and off mutant DISC1 in mouse forebrain. In his poster, Yangun Xu showed that mice expressing mutant DISC1 had no problem learning the Morris water maze, but they did have trouble remembering where the hidden platform was. The result suggests that they have impaired memory. Xu also reported that prepulse inhibition, as judged by the startle response at 74 and 78 dB, was significantly lower in the mice (they responded normally to louder sounds) and that they showed signs of anxiety, namely hyperactivity in the open field.
Interestingly, since there is growing support for the idea that schizophrenia is a developmental disorder, Xu reported that timing of the expression of mutant DISC1 is crucial for startle response effects. When the gene was turned on at birth, it had no effect on the startle response in the animals once they aged, but if the gene was turned on at conception, then impaired prepulse inhibition became evident in the adult animals. The authors conclude that “…expression of mutant DISC1 in the developing forebrain leads to neurobehavioral abnormalities similar to some features of schizophrenia.”
Stephanie Tankou from Sawa’s lab obtained similar results using mice expressing mutant DISC1 under control of the prion promoter. The prion gene is turned on very early in development (embryonic day 13) and in most places in the brain. Tankou reported in her poster that mice expressing the mutant DISC1 had some developmental problems, notably an increase in size of the lateral ventricle at 6 weeks. This returned to normal by 3 months, however. The number of animals tested was small, but if this finding can be corroborated, then it would show that the DISC1 mutant can have a dramatic effect on brain development. In fact, the Sawa lab has previously reported that silencing DISC1 in utero with RNAi had similar effects, suggesting that mutant DISC1 acts as a null phenotype. The symptoms elicited in the mice include increased startle response, open field anxiety, and poor motor coordination on the rotarod balance beam.
Neurodevelopment was also on the minds of Ken-Ichiro Kubo and colleagues at Keio University School of Medicine, Tokyo, Japan. Working in collaboration with the Sawa group, Kubo and colleagues from Kazunori Nakajima’s lab have shown perturbation of neuronal migration in DISC1 mutant mice (see SRF related news story). In his new work, Kubo has begun to look for the domain of DISC1 crucial to its activity by transferring several DISC1 constructs into developing embryos in utero. In his poster he showed that the C-terminal domain of DISC1, including binding sites to NUDEL and LIS1, is crucial for neuronal migration.
DISC1—The human connection
How do these various DISC1 interactions and effects relate to the human condition? Barbara Lipska and colleagues at the NIH, Bethesda, Maryland, previously showed that DISC1-interacting proteins, including NUDEL, LIS1, and Fez1, are downregulated in brain tissue from schizophrenia patients. Extending those studies, Shruti Mitkus reported by poster that there is no difference in expression levels of ATF4, ATF5, and NUDE between schizophrenia postmortem tissue and controls. But Mitkus did show that there were slight, though significant, increases in mRNA for CITRON (8 percent higher) and Kendrin (10 percent higher) in the dorsolateral prefrontal cortex (DLPFC) from patient samples. Patient hippocampal samples also had significantly lower expression of phosphodiesterase 4B (expression in the DLPFC was normal). The latter observation again tentatively links schizophrenia to cAMP signaling.
In her poster, Lipska addressed the issue of DISC1 in the human brain, namely, where does the protein get expressed? By using real-time quantitative RNA amplification techniques, Lipska found that the gene is broadly expressed in the normal brain with highest levels turning up in the hippocampus, amygdala, cerebellum, and the neocortex including the DLPFC and the entorhinal cortices. Lipska and colleagues used immunocytochemistry and in situ hybridization to examine protein expression at the cellular level and reported that DISC1 is found in neurons and astrocytes. The expression pattern turns out to be very similar among humans, primates, and mice. The researchers also looked at NUDEL and Fez1 with immunocytochemistry and found that, like DISC1, they are broadly, though not uniformly, distributed throughout the brain. Lipska’s group also reported that the subcellular distributions of DISC1 and Fez1 are similar, while the distribution of NUDEL is distinct from the other two proteins.—Tom Fagan.
[Editor’s note: We missed a few posters, and would welcome updates, from the authors or other attendees, on any other DISC1 related posters.]