24 August 2007. Nearly 5 years ago, the first report appeared linking the G72/G30 locus on chromosome 13q to schizophrenia (Chumakov et al., 2002). That first paper also reported that a protein encoded by the G72 gene bound to and activated the enzyme D-amino acid oxidase (DAO) in vitro, earning the protein the name “D-amino acid oxidase activator” (DAOA) in some quarters.

A new study from Joseph Gogos and Maria Karayiorgou at Columbia University, New York, outlines a different scenario of how the G72 gene could contribute to schizophrenia. The results, published earlier this month in the online edition of Molecular Psychiatry, suggest that LG72, the protein encoded by the longest splice variant derived from G72, is found only in mitochondria, and not in peroxisomes where it might interact with DAO.

The researchers found that overexpression of LG72 causes mitochondrial fragmentation in several cell lines. In immature primary neurons, enforced expression of G72 caused increased dendritic arborization. The investigators did not replicate the previously reported association of LG72 with DAO, leading them to suggest that the protein may instead play a previously unrecognized role in neurons to modulate mitochondrial function.

A case of right suspect, wrong weapon?
DAO is one of a number of enzymes that oxidizes serine, an agonist of NMDA-type glutamate receptors. The findings of Chumakov and colleagues raised the possibility that changes in G72 expression could affect serine levels to lower NMDA receptor activity, an idea that fit nicely with the hypothesis that schizophrenia is a disease of glutamate hypofunction. In further support of the G72/serine connection, the original paper showed a genetic interaction between the G72 locus and DAO, the gene for DAO, on chromosome 12q24.

Since then, follow-up studies (see SRF related news story, as well as the meta analysis of Detera-Wadleigh and McMahon, 2006) have generally upheld the association between the G72/G30 locus and both schizophrenia and bipolar disorder; the results with DAO have been less consistent (for example, see Corvin et al., 2007 and Shinkai et al., 2007). In addition, the exact gene and variant responsible for the effect, as well as its protein product and function, have remained obscure (for review, see Abou Jamra et al., 2006).

To establish the subcellular localization of the G72 protein, first author Mirna Kvajo overexpressed G72 in COS7 cells and in rat hippocampal neuron cells. The LG72 protein has no homology to known proteins, but has high α helix content, suggesting it is membrane-bound. Both immunofluorescent staining of intact cells and Western blotting of cell fractions revealed LG72 in the mitochondrial compartment. Deletion of the first 25 N-terminal amino acids abolished mitochondrial localization of the protein. Endogenous LG72 was not detected by immunostaining in any of seven cell lines tested, with the exception of rare staining in HeLa cells, where mitochondrial localization was confirmed. The researchers also note that a protein band consistent with the presence of endogenous LG72 was detected by Western blotting of membrane-enriched fractions of human amygdala lysates.

The presence of overexpressed LG72 affected mitochondrial morphology, the investigators found. The protein was distributed in mitochondrial vesicles, and the mitochondria themselves appeared fragmented. The changes were not associated with induction of apoptosis, or loss of mitochondrial membrane potential in cell lines, suggesting that G72 might regulate normal mitochondrial morphology, involving the fission and fusion of organelles. In primary rat hippocampal neurons, expression of LG72 resulted in mitochondrial fragmentation and an increase in dendritic branching. The dendrites contained more mitochondria after LG72 expression, which the authors speculate may have enabled the higher degree of branching.

To look for association of LG72 and DAO in intact cells, the investigators overexpressed both proteins in COS cells. As reported, DAO localized to peroxisomes. There was no indication of colocalization with LG72, which was found only in mitochondria, nor did the two proteins appear physically associated by coimmunoprecipitation experiments. In addition, they found no increase in endogenous DAO activity in another cell line after LG72 overexpression. “Taken together with the lack of detectable subcellular colocalization and physical interaction, this result makes it unlikely that G72 can directly and robustly regulate DAO activity in the tested mammalian cell lines,” they conclude. The authors point out that the original study identified the G72/DAO protein-level interaction using a yeast two-hybrid screen and in vitro enzyme assays, and did not confirm the results in cells. Thus, it is possible that the interaction does not occur under physiological conditions.

“If our negative findings are confirmed in further studies, the ‘NMDA-hypofunction’ rationale behind the hypothesis that G72 is a susceptibility gene for schizophrenia would be weakened,” the authors write. “Nevertheless, the issue of whether or not the existing association studies support a role of variants in the vicinity of the G72 gene in schizophrenia and bipolar disorder susceptibility should be judged independently of the underlying functional rationale, especially because this gene was identified initially via a nonhypothesis-driven positional cloning approach.”

In other words, the genetic evidence itself warrants an intensive study of G72 function. If that function turns out to reside in mitochondria, it would fit a different but equally interesting theory of neurological disease. That is, of course, the hypothesis of mitochondrial dysfunction, which has been invoked in neurodegenerative diseases, and more recently, in schizophrenia and bipolar disorder (Ben-Shachar, 2002; Stork and Renshaw, 2005; Konradi et al., 2004)—Pat McCaffrey.

Reference:
Kvajo M, Dhilla A, Swor DE, Karayiorgou M, Gogos JA. Evidence implicating the candidate schizophrenia/bipolar disorder susceptibility gene G72 in mitochondrial function. Mol Psychiatry. 2007 Aug 7; [Epub ahead of print] Abstract