Structural Biology Of Macromolecular Complexes

Arthritis, Musculoskeletal, Skin Dis

Investigators

Linked publications, trials & patents

Abstract

Many important cellular functions are performed by large complexes which operate like macromolecular machines. Complexes also play primarily structural roles as biomaterials in many tissues, including skin and muscle. The goals of this project are to elucidate the structures, assembly properties, and interactions of complexes of both kinds, with close attention to the functional connotations. We pursue three subprojects. (i) Energy-dependent Proteases. Background: Protein quality control is essential for eliminating and recycling aberrant proteins that would otherwise pollute the cell, for example by amyloid formation. This activity is largely carried out by energy-dependent proteases, which consist of two subcomplexes - a peptidase and a chaperone-like ATPase. Our studies focus on the Clp proteases of E. coli, which offer a relatively simple model system. Earlier, we showed that peptidase ClpP consists of two apposed heptameric rings and the cognate ATPase - either ClpA or ClpX - is a single hexameric ring. ClpA/X stack axially on one or both faces of ClpP to form active complexes. We went on to study the interaction of ClpAP and ClpXP with model substrates. In both cases, these proteins initially bind to distal sites on the ATPase and are then translocated along an axial pathway into the digestion chamber inside ClpP. Current Objectives. AAA ATPases form a family of proteins whose AAA modules drive diverse functions. For ClpA with two AAA modules and ClpX with one, that function is unfolding and translocation of substrates. (1) We are extending the resolution of our cryo-EM density maps of ClpA and ClpAP and using them to incorporate crystal data on AAA domains and ClpP to create quasi-atomic models of functional complexes. (2) ClpX and ClpA both recognize several different signals on target proteins. In the ClpXP system, we have compared the binding modes of differently signalled substrates and investigated cooperativity in its translocation of substrates. Results. (1) We reconstructed the ClpA hexamer to 1.2 nm resolution. The two ATPase domains (D1, D2) form two tiers with a cavity between. In the D2 ring, which contacts ClpP, there is tenuous contact between adjacent domains. In contrast, the D1 domains are closely knit together, consistent with the prior observation that D1 controls hexamer formation. At most, vestigial density may be assigned to the 17 kDa N-terminal domains, indicating that they are connected to D1 by a highly flexible linker. (2) We found that substrates with N-terminal or C-terminal recognition signals bound to the same site or closely adjacent sites on ClpX, and both substrates are internalized when ATP is added. Some substrates are internalized, albeit ~ 100-fold more slowly, with ATPgS. Time-course experiments and experiments with compound substrates indicate that in doubly loaded ClpXP complexes (i.e. with substrate bound at both ends), translocation takes place almost exclusively from one end at a time. Conclusions. (1) Our analysis reveals that there is considerable flexibility in certain strategic parts of the ClpAP complex. Loose coupling between adjacent D2 domains should allow them to move them over the symmtry-mismatched surface of ClpP whereby the respective rings rotate relative to each other without losing contact. Such rotation has been proposed to occur during processive processing of substrates. (2) There are two explanations for the predeliction of ClpXP to translocate from one end at a time. First, initiation of translocation may be the rate-limiting step, making the probability of simultaneous translocation from both ends low. Second, there may be negative cooperativity, the presence of translocating substrate by the cis end generating an inhibitory signal that blocks translocation from the trans end. (ii) Structure and Assembly of Cornified Cell Envelopes (CEs) in Terminally Differentiated Keratinocytes. Background. The CE is a covalently cross-linked layer of protein that lines the cytoplasmic surface of terminally differentiated keratinocytes. CEs are thought to contribute physical resilience and impenetrability to these tissues. We study their biogenesis, and have applied a variety of EM approaches, both to isolated CEs and in situ. Including compositional inferences based on mathematical modeling of their amino acid compositions, we developed a model of CEs as monolayers of molecules of the protein, loricrin, cross-linked both directly and via minor CE proteins. We envisage the CE as a ?composite? biomaterial with a matrix substance (loricrin) and cross-linkers (the minor proteins). Current Objectives. As noted above, loricrin is the main constituent of epidermal CEs but there is also a growing number of other components, including proteins that substitute for loricrin in knockout mice to assemble normal-looking and normally functioning CEs. The most recently discovered components are the family called Late Expression Proteins (LEPs). We are investigating their status as bona fide CE components and characterizing their distribution in the epidermis by EM. Results and Conclusions. By immunogold-EM of cryosections, we found that the cornified envelopes in newborn mouse skin labeled positive for LEPs, as did granules in the stratum granulosum. In their biosynthetic pathway and ultimate destination, LEPs resemble loricrin. (iii) Amyloid formation by the yeast prion protein, Ure2p. Amyloid is fibrous aggregates of protein(s) in protease-resistant, beta-sheet-rich, and usually non-native conformations. Amyloid accumulates in a number of disease situations including rheumatoid arthritis. Prions (infectious proteins) are transmissible amyloids that have been implicated in certain neuropathies, including the spongiform encephalopathies. To investigate the structure of amyloids and the mechanisms that underlie their formation, we study yeast prions. Unlike mammalian prions, their phenotypes are expressed as lack of metabolic functions rather than cytopathic effects. Our studies have focused on Ure2p, a protein normally involved in nitrogen metabolism. Its prion phenotype presents as the ability to grow on poor nitrogen sources. In earlier work, we demonstrated the filamentous nature of Ure2p amyloid polymers formed in vitro and the presence of filamentous aggregates in prion-infected cells. Current Objectives. Ure2p has an N-terminal ?prion domain? that is necessary for filament formation and a C-terminal domain that performs its nitrogen regulation function. We have formulated a ?backbone? model of Ure2p filaments whereby the prion domains form an amyloid backbone which ar surrounded by the C-terminal domains. In FY02, we performed experiments to test this model and to determine whether prion-form Ure2p is inactivated by unfolding or simply by steric blocking of its binding to the transcription factor Gln3p. Results. We fused the Ure2p [rion domain with several enzymes (barnase, carbonic anhydrase, glutathione S-transferase, green fluorescent protein) that have small substrates. In each case the constructs were found to form filaments. We showed that the appended enzymes were still active in filaments. Two of these proteins are diffusion-limited and their activities were reduced in the filaments by a moderate factor consistent with diffusion effects. Conclusions. The ability of each fusion protein to form filaments fulfils the prediction of the backbone model, providing strong support for this model. The functional measurements imply that the enzyme moieties are affected at most very slightly by entering the filamentous state, suggesting that Ure2p also retains its fold upon entering the filamentous state (prion conversion) and is inactivated by being sterically restrained from binding Gln3p.