Obesity is a major global public health problem, and understanding its pathogenesis is critical for identifying a cure. In this study, a gene knockout strategy was used in post‐neonatal mice to delete synoviolin (Syvn)1/Hrd1/Der3, an ER‐resident E3 ubiquitin ligase with known roles in homeostasis maintenance. Syvn1 deficiency resulted in weight loss and lower accumulation of white adipose tissue in otherwise wild‐type animals as well as in genetically obese (ob/ob and db/db) and adipose tissue‐specific knockout mice as compared to control animals. SYVN1 interacted with and ubiquitinated the thermogenic coactivator peroxisome proliferator‐activated receptor coactivator (PGC)‐1β, and Syvn1 mutants showed upregulation of PGC‐1β target genes and increase in mitochondrion number, respiration, and basal energy expenditure in adipose tissue relative to control animals. Moreover, the selective SYVN1 inhibitor LS‐102 abolished the negative regulation of PGC‐1β by SYVN1 and prevented weight gain in mice. Thus, SYVN1 is a novel post‐translational regulator of PGC‐1β and a potential therapeutic target in obesity treatment.
Selective inhibition or knock‐out of ER‐resident E3 ubiquitin ligase SYVN1 increases PGC‐1β activity and basal energy expenditure, preventing weight gain in a mouse model for obesity.
Loss of Synoviolin (Syvn1) decreases body weight and white adipose tissue accumulation in mice.
Syvn1 deficiency induces mitochondrial activation and increases energy expenditure.
SYVN1 ubiquitinates peroxisome proliferator‐activated receptor coactivator (PGC)‐1β.
SYVN1 negatively regulates PGC‐1β function and target gene expression.
Selective inhibition of SYVN1 improves the phenotype of genetically induced obesity.
Obesity is characterized by the excessive accumulation of adipose tissue, as well as increased risk of diabetes, hypertension, cardiovascular diseases, and depression, and is an enormous economic and social burden (Wickelgren, 1998). White adipose tissue (WAT) functions as an energy reservoir and as an endocrine organ, and various inflammatory mediators and cytokines are overexpressed in the WAT of obese individuals, which leads to chronic inflammation (Hotamisligil, 2006). Metabolism in adipocytes has been extensively studied (Rosen & MacDougald, 2006; Lefterova & Lazar, 2009); however, there are presently no effective and safe pharmacological options for obesity prevention and treatment. A better understanding of the underlying mechanisms is therefore necessary to develop suitable treatments for obesity and associated metabolic diseases.
The mitochondrion is a key organelle in cellular energy control and has been implicated in obesity (Bournat & Brown, 2010). Peroxisome proliferator‐activated receptors (PPARs) (Viana Abranches et al, 2011) and their coactivators (PPAR coactivator (PGC)‐1α, PGC‐1β, and PGC‐1‐related coactivator) (Puigserver et al, 1998; Andersson & Scarpulla, 2001; Lin, 2001; Kressler et al, 2002) play important roles in mitochondrion biogenesis and energy metabolism, including H2O2‐based respiration and β‐oxidation of fatty acids. Gene knockout studies have suggested functional differences between PGC‐1α and PGC‐1β, for instance in terms of lethality (Scarpulla, 2008). Interestingly, PGC‐1β overexpression results in an increase in mitochondrion number and respiratory function in cultured cells (St‐Pierre et al, 2003); furthermore, PGC‐1β transgenic mice show high energy expenditure and resistance to obesity (Kamei et al, 2003). In addition to the mitochondrion, the endoplasmic reticulum (ER) is also thought to play an important role in obesity (Cnop et al, 2012). For example, mutations or deficiency in genes that function in the unfolded protein response produces insulin resistance in a mouse model (Cnop et al, 2012; Wang & Kaufman, 2012). Nonetheless, the role of the ER‐associated protein degradation (ERAD) pathway in obesity remains obscure.
Synoviolin (Syvn1), a mammalian homolog of Hrd1p/Der3p, is an E3 ubiquitin (Ub) ligase that was identified from the cDNA of rheumatoid synovial cells (Amano et al, 2003) and plays important roles in the ERAD pathway (Yamasaki et al, 2005). In a series of studies, we have demonstrated the importance of SYVN1 expression in arthritis (Amano et al, 2003) and fibrosis (Hasegawa et al, 2010). Inflammation plays a critical role in regulating metabolic status, and Syvn1 is a key target for inflammatory cytokines such as tumor necrosis factor α (TNFα), interleukin (IL)‐1, and IL‐17 (Gao et al, 2006; Toh et al, 2006, 2010).
In the present study, conditional Syvn1 knockout mice were generated to clarify the role of Syvn1 in obesity. The results highlight a novel function for SYVN1 in the control of body weight and mitochondrial biogenesis through negative regulation of PGC‐1β.
Generation of Syvn1‐deficient mice
To study the post‐neonatal function of SYVN1, tamoxifen (Tam)‐inducible Syvn1 knockout mice were generated that carry homozygous floxed‐Syvn1 alleles and a Cre‐estrogen receptor (ER) transgene (Hayashi & McMahon, 2002) (Fig 1A). Efficient recombination was confirmed in Syvn1 conditional knockout (CAG‐Cre‐ER; Syvn1flox/flox) and control (Syvn1flox/flox and Syvn1flox/+) mice following Tam administration by PCR (genome), real‐time PCR (mRNA), and Western blotting (protein) (Fig 1B–D). A significant reduction in body weight was observed in CAG‐Cre‐ER;Syvn1flox/flox mice as early as 1 week after Tam administration until it was approximately 80% that of control mice (Fig 1E and F). A slight decrease in weight was also observed in Syvn1 heterozygous CAG‐Cre‐ER;Syvn1flox/+ mice (Supplementary Fig S1). Three experiments were used to determine whether the weight loss was due to low food intake, absorption‐related malnutrition, or both. First, the daily food intake was measured; however, differences were almost not found in terms of food consumption between CAG‐Cre‐ER;Syvn1flox/flox and control mice at day 1 and 7, and higher food consumption was observed in CAG‐Cre‐ER;Syvn1flox/flox mice than in control mice at day 14 (Fig 1G). Second, macroscopic and microscopic examinations showed comparable amounts of food residue in the gut and no histological abnormalities in the intestine of CAG‐Cre‐ER;Syvn1flox/flox mice compared to controls. Third, serum biochemical tests showed no significant differences in the levels of several biomarkers of nutrition and liver and kidney function between the two groups (Supplementary Table S1). However, a detailed anatomical analysis showed a marked reduction of WAT in CAG‐Cre‐ER;Syvn1flox/flox mice relative to controls (Fig 1H–J).
Loss of Syvn1 in ob/ob mice and db/db mice causes body weight loss
Two well‐established mouse models of obesity (ob/ob and db/db) in which the leptin signal is constitutively inactivated (Tartaglia et al, 1995) were used to determine whether Syvn1 deficiency is associated with a reduction in body weight at the level of the central nervous system under conditions of constitutive food intake. The expression level of SYVN1 was higher in ob/ob and db/db than in ob/+ and db/+ mice (Fig 2A). Moreover, Tam administration resulted in a significant loss of body weight in CAG‐Cre‐ER;Syvn1flox/flox;ob/ob and CAG‐Cre‐ER;Syvn1flox/flox;db/db compound mutants (Fig 2B and C). An anatomical dissection revealed a reduction in fat mass in CAG‐Cre‐ER;Syvn1flox/flox;ob/ob and CAG‐Cre‐ER;Syvn1flox/flox;db/db mice compared to Syvn1flox/flox;ob/ob and Syvn1flox/flox;db/db mice, respectively (Fig 2D and E, and Supplementary Fig S2). No differences in food intake were noted across groups (Fig 2F and G). Taken together, these results indicate that SYVN1 directly controls body weight at the level of peripheral energy expenditure, and not at the level of the central nervous system.
WAT is the source of body weight loss resulting from Syvn1 deletion
To directly test the effect of Syvn1 knockout on peripheral energy expenditure in WAT, adipose‐specific Syvn1 knockout mice were generated by crossing Syvn1flox/flox mice with the adiponectin (Adipoq)‐Cre line (Herman et al, 2012; Kleiner et al, 2012) to obtain Adipoq‐Cre;Syvn1flox/flox mice. Cre‐mediated Syvn1 deletion in WAT was confirmed by PCR (Fig 3A) and Western blotting (Fig 3B). The body weight of Adipoq‐Cre;Syvn1flox/flox mice was approximately 90% that of controls (Fig 3C and D), an effect that was associated with a reduction in WAT (Fig 3E and F). These results indicate that SYVN1 directly targets WAT to control body weight.
SYVN1 directly interacts with PGC‐1β
To understand the molecular mechanism, we first used microarray analysis to establish the expression of genes in WAT. Genes involved in β oxidation and mitochondria biogenesis (Supplementary Fig S3) were increased in CAG‐Cre‐ER;Syvn1flox/flox mice compared with Syvn1flox/flox mice. Several studies have indicated that thermogenic nuclear receptors PPARα and β and their coactivators of the PGC‐1 family transcriptionally regulate peripheral energy expenditure via β oxidation and mitochondria biogenesis (Lowell & Spiegelman, 2000; Scarpulla, 2006); therefore, the interaction of SYVN1 with these factors was tested in vitro. Glutathione‐S‐transferase (GST)‐tagged SYVN1 lacking the transmembrane domain (GST‐SYVN1ΔTM) bound HA‐tagged PGC1β but not PPARα or γ or PGC‐1α (Fig 4A). In a GST pull‐down assay, PGC‐1β bound the SyU domain (amino acids, aa 236–270) of SYVN1 (Fig 4B), which is highly conserved from Caenorhabditis elegans to humans, but not in yeast SYVN1 orthologs (Supplementary Fig S4A). In addition, an R266A/R267A double mutation in the SyU domain decreased this interaction (Fig 4C), but had no effect on the E3 ligase activity of SYVN1 (Supplementary Fig S4B). The GST pull‐down assay mapped the SYVN1‐binding domain of PGC‐1β to aa 195–367 containing an LXXLL motif of middle portion (Supplementary Fig S4C). To verify the interaction in cellulo, HA‐PGC‐1β and FLAG‐tagged SYVN1 (SYVN1/FLAG) were co‐transfected into HEK 293T cells. HA‐PGC‐1β co‐immunoprecipitated with SYVN1/FLAG but not the control FLAG vector (Fig 4D). To further investigate the interaction between SYVN1 and PGC‐1β, whole‐cell lysates of HEK 293 cells, in which SYVN1 and PGC‐1β were expressed, were precipitated with anti‐SYVN1 antibody or a control non‐immune mouse immunoglobulin (Ig)G and probed with an antibody against PGC‐1β in an immunoblotting assay. Endogenous PGC‐1β was detected in the precipitate with anti‐SYVN1 but not with IgG (Fig 4E). These results clearly indicate that SYVN1 interacts in vivo with PGC‐1β under normal physiological conditions.
Since SYVN1 is an ER‐resident protein and PGC‐1β translocates into the nucleus, their subcellular localization was investigated by immunofluorescence staining in transiently transfected HEK 293T cells. HA‐PGC‐1β was mainly detected in the nucleus (Fig 4F), as previously reported (Kelly et al, 2009). However, when coexpressed with SYVN1/FLAG, the two proteins predominantly colocalized in the perinuclear region, but were not observed in the nucleus (Fig 4F). In contrast, the coexpression of HA‐PGC‐1β with SYVN1R266A, R267A/FLAG, or SYVN1ΔSyU/FLAG, which do not interact with PGC‐1β (Fig 4B and C), resulted in the nuclear localization of HA‐PGC‐1β (Fig 4F, Supplementary Fig S4D). These results indicate that SYVN1 traps PGC‐1β in the perinuclear region.
PGC‐1β is a novel substrate of SYVN1
An in vitro assay was carried out to determine whether PGC‐1β is a substrate of SYVN1, which is an E3 Ub ligase (Amano et al, 2003). Polyubiquitinated PGC‐1β was detected in the presence of ATP, HA‐Ub, E1, E2, and SYVN1 (Fig 4G). The ubiquitination of PGC‐1β was also examined in vivo. FLAG‐tagged Ub and HA‐PGC‐1β were coexpressed with wild‐type (WT) or mutant SYVN1 (3S)—which lacks Ub ligase activity (Amano et al, 2003)—in HEK 293T cells. Ubiquitinated HA‐PGC‐1β was observed in WT SYVN1‐ but not SYVN1 3S‐expressing cells (Fig 4H). Collectively, these results suggest that PGC‐1β is a SYVN1 substrate.
Since ubiquitinated proteins are degraded by the proteasome, experiments were performed to verify whether the level of PGC‐1β protein is regulated by SYVN1. PGC‐1β protein level was markedly elevated in the WAT of CAG‐Cre‐ER;Syvn1flox/flox (Fig 4I) and Adipoq‐Cre;Syvn1flox/flox (Supplementary Fig S4E) mice; however, the transcript level of Ppargc1b (encoding PGC‐1β) was unaltered (Supplementary Fig S4F). PGC‐1β protein level was 1.4‐fold higher in Tam‐treated skin fibroblasts from CAG‐Cre‐ER;Syvn1flox/flox mice than in vehicle‐treated cells (Fig 4J). A similar observation was made by knocking down Syvn1 expression using small interfering RNA (siRNA), which resulted in a near‐complete disappearance of SYVN1 expression in HEK 293 cells: PGC‐1β protein expression in Syvn1 siRNA‐treated cells was 2.5‐fold higher than in controls (Fig 4K). Finally, to investigate the role of SYVN1 in PGC‐1β degradation, the proteasome inhibitor MG‐132 was applied to cultured skin fibroblasts. Similar to Tam treatment, MG‐132 upregulated PGC‐1β protein levels in vehicle‐treated cells by 1.6‐fold, but produced no additional effects on Tam‐treated skin fibroblasts (Fig 4J). To confirm whether the SYVN1–PGC‐1β interaction is critical for SYVN1‐mediated PGC‐1β degradation, the half‐life of PGC‐1β was measured in mouse embryonic fibroblasts (MEFs) coexpressing HA‐PGC‐1β and WT SYVN1 or SYVN1ΔSyU (Supplementary Fig S4G). WT SYVN1 overexpression significantly shortened the half‐life of HA‐PGC‐1β, whereas SYVN1ΔSyU had no effect (Fig 4L). These results indicate that PGC‐1β protein expression is negatively regulated by SYVN1 at the post‐transcriptional level and strongly suggest that SYVN1 is a major E3 ligase for PGC‐1β in cells.
Negative regulation of PGC‐1β by SYVN1
PGC‐1β functions as a coactivator of several transcription factors including PPARα and is implicated in various biological processes such as mitochondrial biogenesis, β oxidation, and body weight control (Scarpulla, 2008). To investigate the role of SYVN1 in the regulation of the coactivator and mitochondrial biogenesis functions of PGC‐1β, the activity of PPAR luciferase (PPRE X3‐TK‐luc) (Kim et al, 1998)—which contains three PPAR binding sites and is specifically regulated by PPAR and its coactivators—was measured. As previously reported (Lin et al, 2003), reporter activity was induced by treatment with Wy‐14643, a PPARα agonist. Syvn1 but not control siRNA transfection significantly enhanced reporter activity under Wy‐14643 induction (Fig 5A). Furthermore, the co‐transfection of PGC‐1β and Syvn1 siRNA activated reporter activity. These results indicate that loss of Syvn1 enhances PPARα‐mediated transcription in a PGC‐1β‐dependent manner. Conversely, SYVN1 overexpression inhibited the coactivator function of PGC‐1β (Fig 5B), but this effect was decreased for the SYVN1R266A, R267A mutant (Fig 5C).
The role of SYVN1 in the regulation of mitochondrial biogenesis was also examined. Syvn1 siRNA‐treated cells contained large numbers of mitochondria compared to those treated with control siRNA, and the increase was not observed in cells that were co‐transfected with both Syvn1 and Ppargc1b siRNAs (Fig 5D). Furthermore, transcript levels of medium chain acyl‐coenzyme A dehydrogenase and mitochondrial ATP synthase β subunit, two known PGC‐1β target genes (Rodriguez‐Calvo et al, 2006; Shao et al, 2010), were upregulated in Adipoq‐Cre;Syvn1flox/flox (Fig 5E and F) and CAG‐Cre‐ER;Syvn1flox/flox compared to control mice (Supplementary Fig S5A). An electron microscopic analysis revealed increases in both the number and size of mitochondria in CAG‐Cre‐ER;Syvn1flox/flox compared to controls (Fig 5G). To investigate whether the mitochondria in CAG‐Cre‐ER;Syvn1flox/flox mice were functional, O2 consumption in a single‐cell suspension of primary mouse adipocytes was measured and was found to be higher in cells from CAG‐Cre‐ER;Syvn1flox/flox than from control mice (Supplementary Fig S5B). Moreover, mitochondrial respiration and activity were higher in CAG‐Cre‐ER;Syvn1flox/flox than in control mice (Fig 5H). The high respiration rate was also observed in Syvn1 knockout pre‐adipocyte cells compared to Syvn1 WT cells, and the high respiration was not observed by treatment of Syvn1 knockout cells with Ppargc1b siRNAs (Supplementary Fig S5C). These results indicated that the respiratory phenotype of Syvn1 knockout is PGC1b dependent. Finally, the basal energy expenditure was increased in CAG‐Cre‐ER;Syvn1flox/flox relative to control mice (Fig 5I). These results indicate that Syvn1 deletion enhances mitochondrial activity in vivo and strongly suggests that PGC‐1β is a functional target of SYVN1.
Selective SYVN1 inhibition attenuates weight gain
We previously demonstrated that LS‐102 selectively inhibits the E3 Ub ligase activity of SYVN1 (Yagishita et al, 2012) and suppresses rheumatoid arthritis (Yagishita et al, 2012), liver cirrhosis (Wu et al, 2014), and sarcoglycanopathy (Bianchini et al, 2014) in a mouse model. Here, it was used to examine the effects of SYVN1 inhibition on obesity. LS‐102 suppressed the regulation of PGC‐1β function by SYVN1, as evidenced by reduced ubiquitination of PGC‐1β (Fig 6A) and activation of PPARα‐mediated transcription in HEK 293 cells (Fig 6B). In addition, LS‐102 induced the expression of PGC‐1β in WT MEFs, but not SYVN1 KO MEFs (Supplementary Fig S6). Next, C57BL/6J mice were treated with the vehicle control dimethylsulfoxide (DMSO) or LS‐102 and their body weight was monitored over a 2‐month treatment period (Fig 6C). Mice treated with LS‐102 showed no weight gain as a result of normal food intake, while food intake itself was unaffected (Fig 6D); analysis by dissection revealed a reduction in WAT fat mass of the epididymis and fewer lipid droplets in these mice compared to controls (Fig 6E and F). Similar to the WAT of CAG‐Cre‐ER;Syvn1flox/flox mice, C57BL/6J mice treated with LS‐102 had greater numbers of mitochondria (Fig 6G). The effect of LS‐102 on obesity was examined in db/db mice. The body weight of LS‐102‐treated animals gradually decreased to approximately 85% of that of DMSO‐treated mice (Fig 6H and I), but there was no effect on food intake (Fig 6J). LS102 also significantly decreased blood glucose in db/db mice compared to controls (Fig 6K). These findings indicate that the inhibition of SYVN1 can suppress weight gain in a mouse model of obesity.
Gene knockout technology is a useful method for assessing gene function and often yields unexpected results, such as the reductions in body weight and WAT accumulation observed in post‐neonatal Syvn1 knockout mice in the present study (Fig 1). This weight loss was also observed in crosses with two genetically obese mouse lines, ob/ob and db/db, as well as in adipose tissue‐specific Syvn1 knockout mice (Figs 2 and 3). Taken together, these results indicate that the loss of Syvn1 causes peripheral activation of energy expenditure in WAT, eventually leading to weight loss.
Obesity is a risk factor for other chronic diseases such as cardiovascular disorders and diabetes (Wickelgren, 1998). Recently, the relationship between obesity and chronic inflammation—especially the contributions of adipokines such as TNFα and IL‐1 that link obesity to rheumatic diseases (Abella et al, 2014)—has been closely examined (Johnson et al, 2012). For instance, recent studies have shown that obesity impairs the efficacy of anti‐TNFα therapy in rheumatoid arthritis (RA) patients (Gonzalez‐Gay & Gonzalez‐Juanatey, 2012). We have identified SYVN1 as a causative factor for RA, and many studies have confirmed that SYVN1 is an important target of cytokines (Yamasaki et al, 2005; Gao et al, 2006; Toh et al, 2010). Our study is the first demonstration that SYVN1 is a key for understanding the common feature among obesity, chronic inflammatory, and RA.
The thermogenic transcriptional coactivator PGC‐1β was identified as a SYVN1‐interacting molecule (Fig 4). PGC‐1β, PGC‐1α, and PRC regulate mitochondrial function in the control of energy expenditure (Lin, 2001; Kamei et al, 2003; Liu & Lin, 2011). Although these coactivators share functional and structural identities the primary structure of PGC‐1β has several unique features—including an middle LXXLL motif (Lin, 2001)—, and knockout studies have suggested functional differences among these proteins (Scarpulla, 2008). In the present study, the aa 195–367 region of PGC‐1β encompassing the middle LXXLL motif was found to interact with and serve as a target of negative regulation by SYVN1 (Fig 4). This novel regulatory mechanism was also confirmed in vivo in global and tissue‐specific Syvn1 knockout mice (Figs 4 and 5). The selective regulation of PGC‐1β by SYVN1 could be evidenced for the unique role of PGC‐1β compared to PGC‐1α. Thus, a novel function for SYVN1 in energy metabolism could be exerted via negative regulation of PGC‐1β. Recently, knockout mice of Sel‐1 suppressor of lin‐12‐like protein (Sel1L), which is an adaptor protein for SYVN1 and functions in the ERAD pathway, were shown to be resistant to diet‐induced obesity (Sha et al, 2014). Therefore, these studies point to a new direction for research on obesity that includes a role for the ERAD pathway in energy control.
Nuclear receptors are transcription factors involved in various metabolic processes and are therefore attractive targets for therapeutic interventions. Several studies on PPARs—the main regulators of lipid, glucose, and energy metabolism—have been published in the past two decades and have led to the development of PPAR‐related drugs. Although many PPAR agonists are clinically efficacious, their use is associated with many side effects (Swanson et al, 2013). The PPAR coactivators PGC‐1α and β are also possible drug targets, and adequate modulation or activation of PGC‐1 expression has great potential in the treatment of diseases associated with mitochondrial dysfunction and dysregulation of oxidative metabolism (Liu & Lin, 2011). Although targeting a coactivator in new drug development is attractive, the pharmacological upregulation of nuclear coactivators is challenging. In the present study, LS‐102—a selective inhibitor of the E3 ligase activity of SYVN1—induced activation of PGC‐1β and prevented weight gain and WAT accumulation in mice (Fig 6). In addition, SYVN1 was more highly expressed in obese (ob/ob) mice than in their non‐obese (ob/+) counterparts (Fig 2). These results provide novel insight into the mechanistic basis of obesity, since the pharmacological inhibition of SYVN1 enhances energy expenditure by preventing PGC‐1β degradation. In addition, LS‐102 treatment improved blood glucose before stimulating weight loss (Fig 6). Although further studies are needed to clarify the role of SYVN1 in glucose metabolism, SYVN1 is an important drug candidate for obesity and associated metabolic diseases. In conclusion, this study demonstrated a novel role for SYVN1 in the control of energy metabolism via negative regulation of PGC‐1β.
Materials and Methods
All procedures involving animals were performed in accordance with institutional and national guidelines for animal experimentation and were approved by the Institutional Animal Care and Use Committee of Tokyo Medical University (#S‐24021). Mice were kept in SPF under conditions (20–26°C temperature; 40–65% humidity) on a 12‐h light/12‐h dark cycle. F‐1 Foods (5.1% fat, 21.3% protein) were purchased from Funabashi farm (Chiba, Japan). All mice used in the study were of the C57BL/6J background. Conditional Syvn1loxP/loxP mice were generated through homologous recombination in embryonic stem (ES) cells (Fig 1). The mouse Syvn1 locus was cloned from a BAC clone, and the targeting construct was linearized and transfected into ES cells by electroporation. Recombinant ES cell clones expressing the neomycin gene were selected in medium supplemented with G418 and injected into C57BL/6 mouse‐derived blastocysts using standard procedures. The neomycin selection cassette was flanked by Frt recombination sites and excised in vivo by crossing with the general FLP deleter strain (Jackson Laboratories, West Grove, PA, USA). Floxed heterozygous Syvn1flox/+ and heterozygous CAG‐Cre‐ER mice (Jackson Laboratories) were crossed to generate double heterozygous CAG‐Cre‐ER;Syvn1flox/+ mice, which were bred with homozygous Syvn1flox/flox mice to produce conditional Syvn1 homozygous (CAG‐Cre‐ER;Syvn1flox/flox) and Syvn1 heterozygous (CAG‐Cre‐ER;Syvn1flox/+), homozygous Syvn1flox/flox, and heterozygous Syvn1flox/+ mice. Control mice lacking the Cre transgene (Syvn1flox/flox or Syvn1flox/+) received Tam injections to eliminate any potential effects of Tam. To generate the CAG‐Cre‐ER;Syvn1flox/flox;ob/ob and CAG‐Cre‐ER;Syvn1flox/flox;db/db genotypes, CAG‐Cre‐ER;Syvn1flox/flox mice were bred with ob/+ or db/+ mice (Jackson Laboratories), yielding the compound heterozygotes CAG‐Cre‐ER;Syvn1flox/+;ob/+ and CAG‐Cre‐ER;Syvn1flox/+;db/+. These were interbred to obtain CAG‐Cre‐ER;Syvn1flox/flox;ob/ob, Syvn1flox/flox;ob/ob, and Syvn1flox/+;ob/ob mice or CAG‐Cre‐ER;Syvn1flox/flox;db/db, Syvn1flox/flox;db/db, and Syvn1flox/+;db/db mice. The genotypes Syvn1flox/flox;ob/ob, Syvn1flox/+;ob/ob, Syvn1+/+;ob/ob, Syvn1flox/flox;db/db, Syvn1flox/+;db/db, and Syvn1+/+;db/db were used as control mice. To generate adipose‐specific Syvn1 knockout mice, Syvn1flox/flox and Adipoq‐Cre mice (Jackson Laboratories) were crossed to generate Adipoq‐Cre;Syvn1flox/+ compound heterozygotes, which were mated with Syvn1flox/flox mice. For blinding and randomization, genotyping and measurement of body weight were separately performed by different person, respectively. In addition, treatment with LS‐102 and measurement of body weight were also performed by different person, respectively. We randomly allocated mice and estimated sample size as the number was more than 3 at least. We excluded mice with sickness or injury from experiment before starting the experiment.
Tamoxifen, LS‐102 administration
Tamoxifen stock solution was prepared using 100 mg tamoxifen suspended in corn oil (both from Sigma, St. Louis, MO, USA). Starting from 7 to 8 weeks after birth, mice were injected intraperitoneally with 125 mg/kg tamoxifen per day for five consecutive days. LS‐102 solution (50 mg/kg) or DMSO was injected intraperitoneally once a day.
Plasmids and antibodies
Coding sequences of full‐length Ppara, Pparg, Ppargc1a, and Ppargc1b were PCR‐amplified from mouse 3T3‐L1 cDNA. Fragments of Ppargc1b deletion mutants were obtained by PCR amplification, and these along with full‐length Ppargc1b were inserted into the pcDNA3 HA plasmid (Invitrogen, Carlsbad, CA, USA) for GST pull‐down and transient transfection assays. Plasmid sequences were confirmed by sequencing. PPRE X3‐TK‐luc was purchased from Addgene (Cambridge, MA, USA). SYVN1 plasmids were previously described (Amano et al, 2003; Yamasaki et al, 2007), and Syvn1 point mutants were obtained by PCR amplification. The following antibodies were used: anti‐FLAG (M2), anti‐tubulin (both from Sigma), and anti‐HA (12CA5 and 3F10; Roche, Indianapolis, IN, USA). The anti‐SYVN1 rabbit polyclonal antibody that was used was previously reported (Yamasaki et al, 2007). Polyclonal antiserum against PGC‐1β was generated by immunizing rabbits with purified GST‐PGC‐1β (1–367 aa).
GST pull‐down assay
GST fusion proteins were expressed and purified using glutathione sepharose beads (GE Healthcare, Little Chalfont, UK) (Aratani et al, 2001; Fujita et al, 2003). Cell extracts were incubated with each GST fusion protein bound to resin in 1 ml buffer A [20 mM Tris–HCl, pH 8.0; 100 mM NaCl; 1 mM ethylenediaminetetraacetic acid (EDTA); 1 mM dithiothreitol (DTT); 0.1% Nonidet P‐40 (NP‐40); 5% glycerol; 1 mM Na3VO4; 5 mM NaF; 1 μg/ml aprotinin; and 1 μg/ml leupeptin] for 4 h at 4°C. After washing with buffer A, bound proteins were fractionated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) followed by Western blotting.
HEK 293T cells were transfected with HA‐PGC‐1β and SYVN1/FLAG expression vectors; 24 h later, cells were lysed in 200 μl lysis buffer (20 mM Tris–HCl, pH 8.0; 100 mM NaCl; 1 mM EDTA; 1 mM DTT; 1% NP‐40; 5% glycerol; and protease inhibitors). Lysates were diluted in binding buffer (of the same composition as lysis buffer, except that it contained 0.1% NP‐40) and mixed with 2 μg anti‐FLAG (M2) antibody conjugated to protein G‐sepharose beads. After a 4‐h incubation at 4°C, beads were washed three times with binding buffer. Bound proteins were fractionated by SDS–PAGE and analyzed by Western blotting. To detect the interaction between endogenous SYVN1 and PGC‐1β, HEK 293 cells were lysed in a solution composed of 100 mM Tris–HCl, 80 mM NaCl, 1 mM EDTA, 5 mM ethylene glycol tetraacetic acid, 5% glycerol, 2% (w/v) digitonin, 0.1% Brij 35, protease inhibitor cocktail, and 20 mM MG132. Immunoprecipitation was carried out with anti‐SYVN1 antibody or control IgG, followed by Western blot using anti‐PGC‐1β antibody.
In vitro ubiquitination assays
The in vitro ubiquitination assay was performed as previously described (Yamasaki et al, 2007). Briefly, GST‐PGC‐1β (aa 1–367) was incubated with 0.75 μg HA‐Ub, 125 ng E1 (Biomol International, Plymouth Meeting, PA, USA), 150 ng UBcH5C, and 150 ng maltose‐binding protein‐tagged SYVN1ΔTM‐His in reaction buffer (50 mM Tris–HCl, pH 7.5; 5 mM MgCl2; 0.6 mM DTT; and 2 mM ATP) at 37°C for 2 h. Glutathione sepharose was then added, and the mixture was washed with GST wash buffer (50 mM Tris–HCl, pH 7.5; 0.5 M NaCl; 1% Triton X‐100; 1 mM EDTA; 1 mM DTT; and protease inhibitors). The ubiquitinated PGC‐1β was analyzed by Western blotting with anti‐HA antibody. For inhibition of ubiquitination by LS102, FLAG‐PGC‐1β was purified from HEK 293T cells transfected with FLAG‐PGC‐1β with an anti‐FLAG antibody and incubated with 1.5 μg HA‐Ub, 250 ng E1, 300 ng UBE2G2, and 150 ng GST‐SYVN1 (aa 233–338) in reaction buffer (50 mM Tris–HCl, pH 7.5; 5 mM MgCl2; 0.6 mM DTT; and 2 mM ATP) at 37°C for 2 h. The ubiquitinated PGC‐1β was re‐immunoprecipitated with anti‐FLAG antibody and analyzed by Western blotting with anti‐HA antibody.
In vivo ubiquitination assay
HEK 293T cells were transfected with HA‐PGC‐1β, Ub/FLAG, SYVN1, or SYVN1 3S expression plasmids; 24 h later, 20 μM MG‐132 was added, and cells were lysed in lysis buffer (50 mM HEPES, pH 7.9; 150 mM KCl; 1 mM phenylmethanesulfonyl fluoride; 1% Triton X‐100; 10% glycerol; and protease inhibitors). Lysates were mixed with 1 μg anti‐HA antibody conjugated to protein G‐sepharose beads. After a 4‐h incubation at 4°C, beads were washed three times with lysis buffer. Bound proteins were fractionated by SDS–PAGE and analyzed by immunoblotting.
Cells cultures and transient transfection assay
HEK 293, HEK 293T, and 3T3‐L1 cell lines; MEF; and skin fibroblasts were cultured in Dulbecco's modified Eagle's medium as previously described (Yamasaki et al, 2007). Transient transfection was performed with Lipofectamine 2000 or LTX according to the manufacturer's protocol (Invitrogen). Cells were lysed with cell lysis buffer (Promega, Madison, WI, USA) 24 h after transfection, and luciferase activity was measured. To ensure equal amounts of DNA, empty plasmids were added to each transfection.
The detailed procedure has been described in detail elsewhere (Fujita et al, 2003, 2005). Briefly, cells were transfected with plasmids and fixed 24 h later with 3.7% formaldehyde, permeabilized with 0.2% Triton X‐100, and blocked with 2% bovine serum albumin. Cells were incubated with rat anti‐HA (1:1,000) or rabbit anti‐FLAG (1:1,000) primary antibody, followed by staining with Alexa Fluor 594 anti‐mouse or Alexa Fluor 498 anti‐rat secondary antibody (1:1,000; Molecular Probes, Eugene, OR, USA).
Measurement of protein half‐life
The assay was performed using a previously described method with some modifications (Yamasaki et al, 2007; Bernasconi et al, 2010). MEFs derived from Syvn knockout mice (Yamasaki et al, 2007) were transfected with 1 μg pcDNA3 Syvn/FLAG, empty vector, 1.5 μg pcDNA3 SyvnΔSyU/FLAG, and 0.75 μg pcDNA3 HA‐PGC‐1β; 48 h later, cells were treated with 40 μM cycloheximide for various times, then lysed with buffer (10 mM Tris–HCl, pH 8.0; 150 mM NaCl; 1 mM EDTA; 1% NP‐40; 1 mM DTT; and protease inhibitors), and analyzed by immunoblotting with antibodies against PGC‐1β, SYVN1, or α‐tubulin (loading control). Each experiment was performed at least three times.
The assay was performed as previously described (Chakravarti et al, 1996; Nakajima et al, 1996). HEK 293 cells were transiently transfected with 12.5 ng PPRE X3‐TK‐luc reporter plasmid, 25 ng pcDNA3 HA‐PPARα, 25 ng pcDNA3 HA‐PGC‐1β, 0.1 ng pRL‐CMV, and 15 nM Syvn1 siRNA. For SYVN1 overexpression, 12.5 ng PPRE X3‐TK‐luc reporter plasmid, 25 ng pcDNA3 HA‐PPARα, 75 ng pcDNA3 HA‐PGC‐1β, 0.1 ng pRL‐CMV, and 50 or 100 ng SYVN1 expression vector were transfected. After 12 h, cells were treated with 10 μM WY‐14643 for 6 h and lysed with cell lysis buffer, followed by measurement of luciferase activity. Each experiment was performed at least three times.
RNA interference and real‐time PCR
SiRNAs for human SYVN1 have been previously described (Yamasaki et al, 2007). SiRNAs for mouse Syvn1 and Ppargc1b were purchased from Ambion Inc. (Austin, TX, USA). SiRNA transfection was performed with Lipofectamine 2000. Total RNA from adipose tissue was purified 7 days after Tam injection using ISOGEN (Nippon Gene, Tokyo, Japan) according to the manufacturer's instructions and reverse‐transcribed using ReverTra Ace with random primers (Toyobo, Osaka, Japan). Real‐time PCR was performed using LightCycler 480 Probes Master (Roche Diagnostics, Mannheim, Germany). Expression levels were determined relative to that of 18s rRNA. Primers and probes used in this study are shown in Supplementary Table S2.
WAT was collected from control and CAG‐Cre‐ER;Syvn1flox/flox or Adipoq‐Cre;Syvn1flox/flox mice. The detailed procedure for measurement of mitochondrial respiration is described elsewhere (Sjovall et al, 2013). Pre‐adipocyte cells were obtained from CAG‐Cre‐ER;Syvn1flox/flox mice. White adipose tissues were minced and incubated with collagenase solution for 45 min at 37°C. Then, cell suspension was filtered twice. Cells were cultured in DMEM/F12. Cells were treated with DMSO or tamoxifen, and transient transfection of siRNA was performed 5 days before experiment.
Measurement of basal metabolism
Oxygen consumption and carbon dioxide production were measured using an Oxymax Equal Flow System (Columbus Instruments, Columbus, OH, USA) during the rest period after a 4‐h fast 7 days after Tam administration. In addition, motor activity (number of movements) was measured using the DAS system (Neuroscience, Inc, Tokyo, Japan).
One‐way analysis of variance was used to determine correlations among individual genotypes of mice and body weight at each time point. The non‐paired Student's t‐test was used to analyze mean differences between control and LS‐102‐treated mice in luciferase activity, tissue weight, basal metabolism, and body weight change at each time point. The Mann–Whitney U‐test was used in the mitochondrial respiration assay. A P value < 0.05 was considered statistically significant.
HF, NY, SA, and TN conceived the project and designed the experiments. HF, NY, SA, KS, HK, TS‐F, IH, SM, NH, and TN performed the experiments and analyzed the data. HF and TN wrote the manuscript. All authors discussed the results and commented on the manuscript.
Conflict of interest
The authors declare that they have no conflict of interest.
Supplementary Figure S1
Supplementary Figure S2
Supplementary Figure S3
Supplementary Figure S4
Supplementary Figure S5
Supplementary Figure S6
Supplementary Table S1
Supplementary Table S2
Supplementary Figure Legends
We thank Dr Allan M Weissman for scientific advice. We also thank S. Shibata, H. Nagai, K. Takahashi, Y. Suzuki, T. Kasai, K. Nakamura, E. Hiratsu and M. Chijiiwa for the technical assistance. We also thank all members of Dr Nakajima's laboratory, and OYC and Charles River Japan for their assistance in animal production, and also all members of the Bozo Research Center for pathological analysis. In addition, we thank E. Morita and S. Yoshihara of the Mitsubishi Chemical Medicine Corporation for basal metabolism analysis of mice. We thank Dr Bruce Spiegelman for providing PPRE X3‐TK‐luc plasmid. This work was funded in part by grants from the Naito Foundation, Natural Science Scholarship Daiichi‐Sankyo Foundation of Life Science, Mitsubishi Tanabe Pharma Corporation, Bureau of Social Welfare and Public Health, Takeda Science Foundation, AstraZeneca R&D Grant 2013. This work was also supported in part by Health Labour Sciences Research Grant (201229044A) and Japan Society for the Promotion of Science KAKENHI Grant Numbers 20249052, 23659176, 23659502, 22790947, 24791006, 26670479 and 26461478.
- © 2015 The Authors