Ure(k)a! Sirtuins regulate mitochondria

Abstract

Increasing evidence suggests that multiple metabolic pathways are regulated by sirtuin-dependent protein deacetylation in the mitochondria. In this issue, Nakagawa et al. (2009) show that an uncharacterized sirtuin, SIRT5, deacetylates and activates a mitochondrial enzyme, carbamoyl phosphate synthetase 1, which mediates the first step in the urea cycle.

Mitochondria are key players in metabolism, energy maintenance, and apoptosis. Disruption of mitochondrial pathways can lead to metabolic disease, oxidative damage, and cancer. Therefore, elucidating how these pathways are regulated has therapeutic implications. There is compelling evidence that reversible acetylation of mitochondrial proteins is a key mechanism of metabolic regulation. In this issue of Cell, Nakagawa et al. (2009) describe the regulation of the first step of the urea cycle by a previously uncharacterized mitochondrial sirtuin, SIRT5, through deacetylation of the mitochondrial enzyme carbamoyl phosphate synthetase 1 (CPS1). A related study by Yu et al. (2009) describes the regulated acetylation of ornithine carbamoyltransferase (OTC), which catalyzes the second step of the urea cycle. These studies reveal that reversible protein acetylation is a major mechanism for regulating the urea cycle in mitochondria.

Conserved from bacteria to humans, sirtuins catalyze NAD⁺-dependent protein deacetylation. With some sirtuins, the ability to promote protein ADP-ribosylation has been reported (reviewed in Smith et al., 2008). In humans, there are seven sirtuins (SIRT1-SIRT7), three of which are localized to mitochondria (SIRT3, SIRT4 and SIRT5). SIRT3 deacetylates and activates mitochondrial acetyl-CoA synthetase 2 (ACS2) (Hallows et al., 2006; Schwer et al., 2006). ACS2 is regulated by reversible acetylation of lysine-635 (Hallows et al., 2006) (Figure 1). SIRT3 is a highly active mitochondrial deacetylase (Lombard et al., 2007) that deacetylates and activates complex I of the mitochondrial electron transport chain (Ahn et al., 2008) as well as isocitrate dehydrogenase and glutamate dehydrogenase (GDH) (Schlicker et al., 2008). Meanwhile, SIRT4 inactivates GDH through ADP-ribosylation (Haigis et al., 2006) (Figure 1). The functional link between reversible acetylation and ADP-ribosylation on GDH activity has not been determined. Previously, only cytochrome c was identified as a possible substrate for SIRT5 deacetylation (Schlicker et al., 2008) (Figure 1), even though SIRT5 is localized to the mitochondrial matrix and cytochrome c is associated with the inner membrane facing the intermembrane space. A recent mass spectrometry study revealed the diversity of acetylated proteins (more than 250) that may exist within mitochondria (Kim et al., 2006), although there are likely to be many more.

Regulation of mitochondrial enzymes by sirtuins.

In the mitochondrial matrix, SIRT3 deacetylates and activates acetyl CoA synthetase 2 (ACS2) leading to an increase in acetyl-CoA production from acetate. Deacetylation of glutamate dehydrogenase (GDH) increases formation of α-ketoglutarate, and deacetylation of complex I (electron transport chain) increases oxidative phosphorylation. SIRT4 ADP-ribosylates and inhibits GDH, which inhibits insulin secretion by pancreatic β-cells. SIRT5 deacetylates and activates CPS1, increasing flux through the urea cycle. Citrulline biosynthesis is also regulated by OTC acetylation through an as yet uncharacterized mechanism.

In their new study, Nakagawa et al. (2009) identify CPS1 as the first in vivo substrate regulated by the NAD⁺-dependent deacetylation activity of SIRT5. Using a polyclonal antibody to detect endogenous SIRT5 in mouse tissue, Nakagawa and co-workers first verified that SIRT5 is localized to the mitochondrial matrix, and then identified potential mitochondrial substrates using a SIRT5 affinity column. Analysis of the column eluate by SDS-PAGE showed a band at ~150 kDa, which the authors identified as CPS1 by mass spectrometry. To confirm this interaction, CPS1 and SIRT5 were co-immunoprecipitated from mouse liver samples. It is noteworthy that other unidentified protein bands were enriched from the SIRT5 affinity column. These may also represent SIRT5 substrates.

Next, Nakagawa et al. examined the activities of CPS1 and OTC. OTC catalyzes the second step of the urea cycle, converting ornithine and carbamoyl phosphate to citrulline (Figure 1). Interestingly, CPS1 activity in the liver mitochondria of mice lacking SIRT5 (but not SIRT3 or SIRT4) was ~30% lower than that of wild-type animals, though CPS1 levels remained unchanged. There was no apparent change in OTC activity in mice lacking SIRT3, SIRT4, or SIRT5 compared to wild-type mice. Furthermore, CPS1 activity from mice lacking SIRT5 was increased in vitro using recombinant SIRT5, and this increase correlated with deacetylation of CPS1. However, the identity and stoichiometry of the lysine(s) involved were not resolved. An important future experiment will be to determine how acetylation of specific lysine(s) controls CPS1 activity. For example, does stoichiometric acetylation of CPS1 on a particular lysine residue completely inactivate CPS1, similar to the acetylation ‘on/off’ switch of ACS2 (Hallows et al., 2006)? Or does acetylation affect the conformation of CPS1 or interactions with other proteins or small-molecule regulators?

In a related study, Yu et al. (2009) describe the specific acetylation of OTC on lysine-88, a residue that is mutated in patients with OTC deficiency, an X-linked genetic disorder causing death of newborns or hyperammonemia in adults. Acetylation of OTC decreased its affinity for carbamoyl phosphate and the maximum rate of catalysis. Lysine-88 acetylation could be modulated by adding extracellular glucose or amino acids. However, the protein acetyltransferase or the deacetylase responsible was not determined (Figure 1). Although the Nakagawa et al. study does not implicate SIRT3, SIRT4, or SIRT5 in the control of OTC activity, such a role might be revealed under different experimental conditions.

Because CPS1 catalyzes the first step of the urea cycle, Nakagawa et al. (2009) determined the physiological role of CPS1 regulation in primary hepatocytes from mouse liver. Under starvation conditions, which increase urea production, mouse hepatocytes lacking SIRT5 displayed decreased viability compared to wild-type cells. Furthermore, after 48 hours of starvation, CPS1 deacetylation and activation in liver mitochondria was increased in wild-type but not SIRT5-deficient animals. To directly correlate these results with the urea cycle, the authors measured a two-fold increase in blood ammonia in SIRT5-deficient compared to wild-type mice, indicating disruption of the urea cycle.

Nakagawa et al. hypothesized that SIRT5 might be activated under stress conditions, such as fasting, through an increase in NAD⁺ concentration. In line with previous reports (Yang et al., 2007), they found that under fasting conditions, NAD⁺ concentrations doubled in the mitochondria and that cellular Nampt (an NAD⁺ biosynthetic enzyme) increased 2-3 fold. However, the authors did not observe a corresponding increase in mitochondrial Nampt levels as previously observed in rats (Yang et al., 2007); mouse Nampt may not translocate to mitochondria. Nakagawa et al. suggest that increased cytoplasmic Nampt increases nicotinamide mononucleotide (NMN), which might traverse the membrane into the mitochondrial matrix leading to increased NAD⁺. Further studies are needed to demonstrate that the kinetic properties of the three enzymes – Nampt, Nmnat, and SIRT5 – would support this model.

During caloric restriction of wild-type mice, CPS1 activity increased though SIRT5 and CPS1 protein levels remained unchanged (Nakagawa et al., 2009). There was a decrease in CPS1 acetylation and a ~50% increase in mitochondrial NAD⁺, suggesting that an increase in SIRT5 activity may be responsible for the decrease in CPS1 acetylation and subsequent enhanced CPS1 activity.

The Nakagawa et al. study bolsters emerging evidence that sirtuin-mediated reversible acetylation is a major regulator of mitochondrial proteins. Perhaps most astonishing, the recent work of Nakagawa et al. and Yu et al. demonstrates that acetylation modulates the activity of three inextricably linked metabolic enzymes: GDH, CPS1 and OTC (Figure 1). Many intriguing questions remain. How are these mitochondrial proteins acetylated? Are they acetylated in the cytoplasm before import into mitochondria or are there as yet undiscovered mitochondrial protein acetyltransferases? Also, elucidating the specificity of mitochondrial sirtuins for a growing list of acetylated protein substrates may reveal how the different sirtuins regulate diverse mitochondrial processes. Regulation of the urea cycle through reversible acetylation provides additional support for the emerging theme that reversible protein acetylation is a major mechanism for regulating metabolic processes.