The COP9 signalosome coerces autophagy and the ubiquitin-proteasome system to police the heart

ABSTRACT

We demonstrated for the first time that the COP9 signalosome (COPS) controls the degradation of a surrogate and a bona fide misfolded protein in the cytosol of cardiomyocytes likely via supporting ubiquitination by CUL/cullin-RING ligases, and that Cops8 hypomorphism exacerbates cardiac proteinopathy in mice, in which autophagic impairment appears to be in play. It will be extremely imprtant to investigate cardiac ablation of another Cops gene to decipher whether COPS8 deficiency phenotypes are attributable to the COPS or unique to COPS8.

By minimizing the level and toxicity of malfolded proteins in the cell, protein quality control (PQC) is vital to proper function and the survival, especially in postmitotic cells (e.g., cardiomyocytes and neurons). PQC inadequacy is implicated in the pathogenesis of a large subset of heart diseases; thus it is imperative to improve our understanding of cardiac PQC. The autophagic-lysosomal pathway (ALP) helps PQC mainly by removing aberrant protein aggregates, while individual terminally misfolded proteins are degraded primarily by the proteasome. For both mechanisms, ubiquitination of misfolded proteins is often required. However, the ubiquitin (Ub) ligases responsible for the degradation of cytosolic misfolded proteins in mammals remain obscure.

The COP9 signalosome (COPS), an evolutionarily conserved protein complex consisting of 8 unique protein subunits (GPS1/COPS1/CSN1, and COPS2 through COPS8), is well known for its CUL/cullin deneddylation activity, which regulates the catalytic dynamics of CUL-RING ligases (CRLs). CRLs are the largest family of Ub ligases, responsible for ˜20% of Ub-dependent degradation of cellular proteins. The activation of CRLs requires covalent conjugation of a Ub-like protein NEDD8 (neural precursor cell expressed, developmentally downregulated 8) to CUL via a ubiquitination-like process known as neddylation. The opposite process, deneddylation, is performed by the COPS. By simple logical deduction, deneddylating a CUL inactivates the CRL; hence, removing the COPS should enhance CRL activity, which was indeed demonstrated in earlier in vitro biochemical studies, but was opposed by subsequent in vivo studies using genetic ablation of COPS subunits. Ablation of COPS genes does not enhance but rather impairs the degradation of CRL substrates, indicating that the COPS is required for the proper functioning of CRLs. The underlying mechanism remains elusive. Prior to our recent report, neither the COPS nor CRLs had been demonstrated to degrade misfolded proteins.

We had shown in earlier reports that cardiomyocyte-restricted knockout of Cops8 in mice (COPS8^CKO) accumulates neddylated CULs and exhibits a rather delayed impairment of myocardial Ub-proteasome system (UPS) performance. The most surprising and striking findings were that before the UPS malfunction became discernible, severe ALP impairment occurred in CSN8^CKO hearts, as reflected by decreased autophagic flux, massive accumulation of autophagosomes and Ub-positive aggregates, and increased abundance of the autophagic substrate protein SQSTM1/p62. It turns out that the fusion of autophagosomes with lysosomes (also referred to as autophagosome maturation) is compromised by COPS8 deficiency. As a result, COPS8^CKO mice display massive cardiomyocyte necrosis, develop rapidly deteriorated dilated cardiomyopathy, and uniformly die prematurely. These studies not only for the first time demonstrate that COPS8/COPS is essential to postnatal cardiomyocyte survival and cardiac functioning but also reveal that COPS8/COPS regulates both the UPS and autophagy.

Premature death of COPS8^CKO mice makes them unsuitable for long-term studies. To investigate whether COPS8/COPS regulates the ubiquitination and degradation of a bona fide misfolded protein in vivo, we used the COPS8 hypomorphic mouse (COPS8^hypo) developed elegantly by our collaborator Dr. Ning Wei of Yale University. The COPS8^hypo mice show ˜80% reduction of myocardial COPS8 proteins and moderate increases in neddylated forms of CULs, without apparent ALP defects at the baseline. However, the myocardial capacity of degrading a surrogate misfolded protein is discernibly impaired. The transgenic mice with cardiac overexpression of a bona fide misfolded protein, the human disease-linked missense (R120G) mutation of CRYAB (CRYAB^R120G), are a widely used animal model to study cardiac proteinopathy. When introduced into the proteinopathic mice, COPS8^hypo aggravates CRYAB^R120G-induced restrictive cardiomyopathy and shortens the life span of CRYAB^R120G mice, which is associated with increased neddylated proteins, reduced levels of total ubiquitinated proteins and LC3-II, and augmented levels of protein aggregates in the heart. In cultured cardiomyocytes, ubiquitination and degradation of CRYAB^R120G but not of native CRYAB is suppressed by either COPS8 knockdown or CRL inactivation using a NEDD8 activating enzyme inhibitor, resulting in accumulation of protein aggregates and exacerbation of CRYAB^R120G cytotoxicity. These findings demonstrate that COPS8/COPS and CRLs contribute to the degradation of a bona fide cytosolic misfolded protein by the UPS and perhaps autophagy.

Regarding the regulation of autophagy by COPS8/COPS, this recent study showed that the increases of LC3-II but not of SQSTM1 in CRYAB^R120G-overexpressing hearts were significantly attenuated by COPS8^hypo and that in cultured cardiomyocytes overexpressing CRYAB^R120G, which was previously shown to activate autophagy, siRNA-mediated COPS8 knockdown increases SQSTM1 but not LC3-II, while decreasing both LC3-II flux and SQSTM1 flux, suggesting that COPS8/COPS can regulate autophagosome formation in addition to autophagosome maturation. Presently, it is unclear how COPS8 accomplishes these regulations: through CUL deneddylation and through other unknown function(s) unique to COPS8 are both possible. To differentiate among them, it will be extremely informative to knock out the gene for at least one other COPS subunit in the heart and compare the effect with COPS8^CKO. To this end, COPS5 would be the best candidate because COPS deneddylase resides in COPS5 although all the 8 COPS subunits are required for COPS deneddylation activity. A similar strategy should also be very helpful in sorting out whether massive cardiomyocyte necrosis is unique to COPS8 deficiency. At least one other mouse model (Mcl1 knockout) with impaired autophagy also displays cardiomyocyte necrosis, implying that cardiomyocyte necrosis in COPS8^CKO mice might result from impaired autophagy but it is formally possible that loss of an undocumented COPS8/COPS function, such as transcriptional repression of a gene critical to a regulated necrosis pathway, is responsible. After all, COPS8 was originally identified in Arabidopsis as a transcription repressor and some subsequent evidence also supports a role of the COPS in regulating transcription via a protein degradation-independent mechanism.

Despite moderate impairment of CUL deneddylation, COPS8^hypo mice seem to live a normal life for at least their first 6 mo, indicating that the remaining 20% of COPS8 proteins are sufficient to meet the essential basal need of cardiac tissues and that normal hearts have a large COPS8 reserve that is important under stress conditions for handling increased proteotoxic stress as evidenced by exacerbation of CRYAB^R120G-based cardiomyopathy by COPS8^hypo. The CRYAB^R120G mice are a model of a rare form of human disease; hence, it will be very important to test the impact of COPS8^hypo on a common cardiac disorder with augmented proteotoxic stress, e.g., myocardial ischemia/reperfusion injury.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Funding

This work is in part supported by NIH grants R01HL085629, R01HL072166 (to X.W.) and R01HL124248 (to H.S.) and by the National High Technology Research and Development Program of China 2006AA02Z4B5 (to J.L).