We have continued our studies of the mechanism and control of protein degradation in bacterial and human cells. Intracellular protein degradation is highly specific and serves both regulatory and protein quality control purposes. Most intracellular protein degradation is carried out by multi-component, self-compartmentalized ATP-dependent proteases, which selectively screen potential targets and control access to the proteolytic sites. Our work has focused on the structure and biochemical properties of the ATP-dependent Clp and Lon proteases. Progress in the study of Clp proteases was made in two areas. In collaboration with Alasdair Steven (NIAMS, NIH), we have analyzed by electron microscopy the changes in enzyme structure and the location of bound substrates during the protein unfolding and degradation cycle. We have shown that ternary complexes of ClpA and ClpX bound to the protease, ClpP, are readily formed in vitro and are functional, suggesting that both chaperones may influence the degradation of proteins and the peptide products that are produced. We have analyzed ClpAP complexes by three-dimensional reconstruction and provided evidence for the mobility of the N-domains. Biochemical studies show that the N-domains, while not required for protein degradation, nonetheless influence the activity against different classes of substrates. ClpAP with or without its N-domain degrades substrates bearing specific degradation motifs with virtually the same Km, whereas non-specific substrates have much higher Kms when the N-domains of ClpA are missing. ClpA-deltaN has a higher Vmax than full-length ClpA, and activity is proportional to the number of N-domains in mixed hexamers of ClpA and ClpA-deltaN. ClpS inhibits activity of ClpA by binding to the N-domains, and inhibition is also proportional to the number of N-domains bound, suggesting that the N-domains act independently to influence activity. Our model assigns a role to the N-domains in binding weakly and reversibly to exposed hydrophobic regions of proteins, increasing the local concentration of substrates but also preventing non-productive interactions with the specific binding sites on the apical ring surface of ClpA, leaving them accessible to specific protein targets. In collaboration with Dr. Di Xia (LCB, NCI), detailed analysis of the crystal structure of ClpA N-domain complex with the adaptor protein, ClpS, showed that the hydrophobic N-terminal peptide of ClpS binds to a large hydrophobic pocket in the N-domain. Binding occurred in two different but reproducible configurations, suggesting that the non-specific binding sites in the N-domain are malleable and adjust to the bound substrate. In studies on human ClpXP, we have completed the crystal structure of the human ClpP protein at 2.1 angstroms and identified a number of unique features. A combination of the crystal structure and mutagenesis studies have led to a new model in which we propose that the N-terminal peptide plays a heretofore unreported role in substrate translocation. We also obtained evidence for regarding ClpX binding site on the surface of hClpP. Chemical modification of a specific histidine residue affects ClpX activated protein degradation, and a mutation in this site in eClpP leads to loss of activity and possible loss of interaction with ClpX. Our studies of Lon protease have led to the first crystal structure of a functional domain of Lon protease. In collaboration with Alex Wlodawer (LMC, NCI-Frederick) the structure of the protease domain of E. coli Lon has been solved. The structure confirms the unusual serine-lysine dyad at the active site of Lon. The protease domain has a unique fold and assembles into a hexamer in the crystal, suggesting that the oligomeric form of Lon, which has been variously reported to have four to eight subunits, is indeed a hexamer. We also obtained a crystal structure of the small helical domain within the Lon ATPase domain, sometimes referred to as the sensor and substrate discriminating domain, confirming that it has the highly conserved fold expected of proteins in the AAA superfamily. Biochemical studies with fusion proteins suggest that Lon can specifically recognize substrates bearing sequence tags that serve as degradation motifs, but that the ability to unfold and completely degrade the fusion protein is variable and may depend on the thermodynamic stability of the fusion protein or local structural stability of the protein surrounding the region where the degradation tag is attached.
