The dynamin-related GTPase Opa1 controls mitochondrial fusion and inner membrane cristae structure. In this study, we showed that Opa1 deficiency in pancreatic beta cells causes defects in glucose-stimulated ATP production. We also found that Opa1 is important for maintenance of ETC complex IV in beta cells. In the absence of Opa1, complex IV levels were decreased, and consequently its activity was reduced in RIP2-Opa1KO islets. Decreased complex IV activity accounts for defects in glucose-stimulated ATP production, oxygen consumption, calcium dynamics, and insulin secretion. Because the basal levels of ATP remained unaltered in the absence of Opa1, the reduced complex IV levels are apparently sufficient to support basal ATP levels, but the extra “load” of secretion in response to higher glucose concentrations likely overtaxes the Opa1-null mitochondria. Therefore we suggest that the effect of mitochondrial morphology on glucose-stimulated ATP production is mediated by the mitochondria's inability to meet the increased demand of secretion due to decreased level and activity of complex IV. Consistent with our findings, a previous study showed that knockdown of Opa1 in cultured cells causes a reduction in oxygen consumption (Chen et al., 2005
In addition to pancreatic beta cells, RIP2-Cre is also known to express in the hypothalamus. We measured animal behaviors and physiology because we sought to assess the effect of Opa1 deletion on the hypothalamus, which controls food intake and therefore fat content in the body. We observed no differences in food intake, fat content, and locomotive activity between control and RIP2-Opa1KO mice. Therefore the effect of Opa1 deficiency on the hypothalamus is negligible. The data also suggest that reduced insulin secretion and hyperglycemia did not affect general animal physiology in RIP2-Opa1KO mice.
It would be interesting to speculate whether the inability to respond to high glucose and reduction in beta cell mass occur in parallel during the lifespan of RIP2-Opa1KO mice. We found that young RIP2-Opa1KO mice have a normal beta cell mass but exhibit significantly higher blood glucose levels (control, 130 ± 3 mg/dl; and RIP2-Opa1KO, 173 ± 10 mg/dl; n ≥ 27), suggesting that the stimulus–secretion coupling defect may arise first. During postnatal growth, beta cell mass did not increase due to defects in cell proliferation, further contributing to hyperglycemia in adult RIP2-Opa1KO mice.
How the loss of Opa1 affects complexes IV and I remains to be determined. One possible hypothesis is that a reduction in complexes IV and I may result from reduced levels of mtDNA. Because mtDNA encodes 13 subunits of ETC complexes (7 subunits of complex I, 1 subunit of complex III, 3 subunits of complex IV, and 2 subunits of complex V), decreases in its copy number could lower ETC complex levels. Consistent with this idea, the loss of mitofusins, which are required for mitochondrial outer membrane fusion, has been shown to cause a decrease of mtDNA levels and impaired respiration in skeletal muscle (Chen et al., 2010
). However, we found similar copy numbers of mtDNA in control and Opa1-deficient beta cells. In addition, levels of other complexes, such as complexes III and V, which contain subunits encoded by mtDNA, were unaffected in RIP2-Opa1KO islets. Furthermore, subunits IV and Vb of complex IV, which are encoded by nuclear DNA, also decreased. Therefore it is unlikely that the reduced levels of complexes IV and I simply result from decreased mtDNA levels.
Second, fragmentation of mitochondria due to fusion defects may produce an uneven distribution of mtDNA. In this hypothesis, total amounts of mtDNA in cells remain unchanged, but mtDNA amounts per individual mitochondria vary. Some mitochondria have more mtDNA, whereas others contain no or reduced mtDNA copies. Because ETC subunits are encoded by both nuclear DNA and mtDNA, improper ratios of ETC subunits will be generated and assembly will be altered, likely affecting oxidative phosphorylation (Chen et al., 2010
). However, we found that ETC complexes are assembled normally in the absence of Opa1. In addition, this hypothesis is unable to explain our observation that complexes IV and I were specifically affected. A third hypothesis is that the partial release of cytochrome c
from the inner membrane cristae destabilizes complex IV. Supporting this idea, cytochrome c
binds to complex IV and is required for the stability and activity of this complex (Vempati et al., 2009
). Because the closed morphology of cristae junctions maintains cytochrome c
inside the inner membrane cristae (Frezza et al., 2006
), opening cristae junctions could relocate a fraction of cytochrome c
to the intermembrane space in Opa1-null mitochondria, compromising its interactions with complex IV and supercomplex containing complexes IV and I. It would be important to test these hypotheses and further understand the role of Opa1 in mitochondrial structure and function in future studies.
Previous reports have suggested that beta cell mitochondrial dysfunction is involved in type 2 diabetes (Lamson and Plaza, 2002
Lowell and Shulman, 2005
Prentki and Nolan, 2006
). For example, the maintenance of normal mtDNA is critical for normal beta cell function, as loss of mtDNA leads to hyperglycemia in mice carrying beta cell–specific deletion of Tfam, which is a mitochondrial transcriptional factor required for mtDNA maintenance (Silva et al., 2000
). Because mitochondrial fusion is thought to be important for mtDNA maintenance, we initially expected RIP2-Opa1KO mice to lose mtDNA. On the contrary, Opa1 loss resulted in a reduction in the quantity and activity of complex IV without mtDNA deficiency. Therefore our experiments provide a new model for mitochondrial diabetes. In addition, our findings also show differences in the effect of Opa1 loss in tissues and cultured cells. To better understand physiological roles of Opa1, it would be important to examine its function in vivo as well as in vitro. We speculate that the observed differences between islets and cultured cells may result from varying demand on glycolysis for ATP production. Moreover, we propose that changes in Opa1 levels may associate directly with the pathogenesis and progression of this disease. In support of this idea, a recent study showed that islet Opa1 levels decrease before the onset of type 2 diabetes in ob/ob mouse models (Keller et al., 2008
). Furthermore, a Goto–Kakizaki rat model for type 2 diabetes exhibits fragmented mitochondria with altered cristae morphology that is different from the normal tubular shape observed in beta cells (Mizukami et al., 2008
). Such changes in mitochondrial morphology may be a consequence of defects in Opa1. This study clearly demonstrates the physiological importance of Opa1 in mammals and provides a mechanism by which mitochondrial structure and dynamics are linked to metabolism.