Biochemical Society Transactions

BioScience2004

Structural and functional studies of mutations affecting the UBA domain of SQSTM1 (p62) which cause Paget's disease of bone

R. Layfield, B. Ciani, S.H. Ralston, L.J. Hocking, P.W. Sheppard, M.S. Searle, J.R. Cavey

Abstract

Mutations affecting the UBA (ubiquitin-associated) domain of SQSTM1 (Sequestosome 1) (p62) are a common cause of Paget's disease of bone. The missense mutations resolve into those which retain [P392L (Pro392→Leu), G411S] or abolish (M404V, G425R) the ability of the isolated UBA domain to bind Lys-48-linked polyubiquitin. These effects can be rationalized with reference to the solution structure of the UBA domain, which we have determined by NMR spectroscopy. The UBA domain forms a characteristic compact three-helix bundle, with a hydrophobic patch equivalent to that previously implicated in ubiquitin binding by other UBA domains. None of the mutations affect overall folding of the UBA domain, but both M404V and G425R involve residues in the hydrophobic patch, whereas Pro-392 and Gly-411 are more remote. A simple model assuming the isolated UBA domain is functioning as a compact monomer can explain the effects of the mutations on polyubiquitin binding. The P392L and G411S mutations do however have subtle local effects on secondary structure, which may become more relevant in full-length SQSTM1. Identification of the in vivo ubiquitylated substrates of SQSTM1 will be most informative in determining the functional significance of the SQSTM1–ubiquitin interaction, and consequences of the disease-associated mutations.

  • osteoporosis
  • p62
  • Paget's disease of bone
  • SQSTM1
  • ubiquitin
  • ubiquitin-associated (UBA) domain

Introduction

PDB (Paget's disease of bone) affects between 1 and 3% of individuals over 55 years of age in Caucasian populations [1,2]. The disorder is characterized by focal increases in bone turnover, reflecting aberrant osteoclastic activity, which leads to debilitating symptoms including bone pain, deformity and susceptibility to pathological fractures in up to a third of sufferers. Although the etiology of PDB is unknown, there is a genetic predisposition for the disease, and recent positional cloning studies have shown that chromosome 5-linked forms of PDB are caused by mutations in the SQSTM1 (Sequestosome 1; also known as p62) gene [3,4]. Over the past 2 years, 11 separate mutations have been identified, all of which affect the UBA (ubiquitin-associated) domain of SQSTM1 [58] (Figure 1), a region of the protein which interacts with ubiquitin. SQSTM1 is a scaffold protein in a number of signalling pathways, which ultimately lead to activation of NF-κB (nuclear factor κB), including the RANK–TRAF6–NF-κB pathway (where RANK stands for receptor activator of NF-κB and TRAF6 for tumour-necrosis-factor-receptor-associated factor 6), which is an important mediator of induced osteoclastogenesis [9]. Whereas the precise functional significance of the SQSTM1 UBA domain is unknown, the domain-specific clustering of clinically similar mutations indicates that the disease may be caused by a common mechanism of action involving the ubiquitin-binding properties of SQSTM1.

Figure 1 Surface representation of the UBA domain (residues 387–436) of SQSTM1 [10] and position of the amino acids affected by PDB-associated missense mutations

The central white area indicates the hydrophobic patch which is implicated in ubiquitin binding. Met-404, Ser-399 and Gly-425 are involved in this surface patch, whereas Pro-387, Pro-392 and Gly-411 are located outside. The blue line denotes the position of the Cα backbone of the UBA domain, with residue 3 corresponding to Pro-387. The tabulated data summarizes the functional and structural consequences of the UBA domain mutations. + indicates Lys-48-linked polyubiquitin-binding, − indicates loss of Lys-48-linked polyubiquitin-binding and n/a indicates mutation not assessed in the present study.

Functional consequences of PDB-associated SQSTM1 mutations

UBA domains are short (∼50 residue) sequences, which bind non-covalently to polyubiquitin chains and in some cases to monomeric ubiquitin. Towards an understanding of how the SQSTM1 UBA domain mutations predispose to PDB, we have determined the effects of the mutations on ubiquitin binding. Four of the mutations generate proteins with truncated UBA domains [4,5,8] and are predicted to ablate ubiquitin binding. The remaining seven missense mutations result in amino acid substitutions [P387L (Pro387→Leu), P392L, S399P, M404V, M404T, G411S and G425R] in the SQSTM1 UBA domain [38]. Using in vitro assays, we first showed that the isolated UBA domain (residues 387–436) is capable of precipitating Lys-48-linked polyubiquitin chains (oligomers which when attached to target proteins signal proteasomal degradation) containing two or more ubiquitins, and that the most common PDB-causing mutation (P392L) does not affect the ability of the isolated UBA domain to bind Lys-48-linked polyubiquitin [10]. Recently, we have shown that other missense mutations in the isolated UBA domain resolve into those which retain (G411S) or abolish (M404V and G425R) the ability of the UBA domain to bind Lys-48-linked polyubiquitin ([8]; Figure 1).

Structural rationalization

Analysis of one-dimensional proton NMR spectra showed that all of the mutant UBA-domain polypeptides form native-like folded structures and that the mutations are not significantly affecting the overall tertiary structure of the UBA domain [8]. The loss of ubiquitin-binding associated with the M404V and G425R mutations can, however, be rationalized with reference to the solution structure of the SQSTM1 UBA domain, which we have determined by NMR spectroscopy [10]. The UBA domain forms a compact three-helix bundle characteristic of UBA domains from other proteins, with Met-404 being located in the extended loop between helices 1 and 2, and Gly-425 within helix 3. Both Met-404 and Gly-425 are involved in the hydrophobic patch on the UBA domain (Figure 1), equivalent to the surface previously implicated in ubiquitin binding by other UBA domains [11]. Whereas the M404V mutation maintains surface hydrophobicity, substitution of the methionine side chain for that of valine modifies the putative ubiquitin-binding van der Waals surface by introducing a small surface cavity, which presumably perturbs the surface complementarity necessary for the UBA domain–ubiquitin interaction [8]. The G425R mutation reduces ubiquitin affinity by substituting a highly polar arginine side chain in the same hydrophobic patch [8]. Consistent with a model in which the isolated UBA domain is functioning as a compact monomer, Pro-392 and Gly-411 are located outside of the hydrophobic patch (Figure 1), which would account for the binding of Lys-48-linked polyubiquitin comparable with wild-type seen in the P392L and G411S mutant UBA domains. The P392L and G411S mutations do however have subtle local effects on UBA-domain structure, with the former extending helix 1 by four residues [10] and the latter extending helix 2 by a single residue. Although we have not yet assessed the effects of the P387L, S399P and M404T mutations, we predict that S399P, which is located within helix 1, will adversely affect protein folding, and that similar to the M404V mutation, M404T will disrupt the hydrophobic patch (in this case a polar side chain is also introduced); furthermore, we predict both mutations probably affect Lys-48-linked polyubiquitin binding. It is noteworthy that the P387L mutation falls outside the structured region of the UBA domain (helix 1 begins at residue 392) and is unlikely to affect the structure of the minimal UBA domain.

Future studies

In summary, we have found that clinically similar PDB-associated SQSTM1 mutations can be resolved into those which retain (P392L and G411S) or abolish (M404V and G425R) the ability of the isolated UBA domain to bind Lys-48-linked polyubiquitin, and have provided a structural rationalization for these effects [8,10]. Genotype–phenotype analysis showed that there is no correlation between the Lys-48-linked polyubiquitin-binding properties of the different mutant UBA domains and disease occurrence or extent [8], indicating that the mechanism of action is unlikely to involve a simple common loss of binding to such ubiquitin chains. Our functional studies have however focused solely on unanchored Lys-48-linked polyubiquitin chains with UBA domains in isolation, and it will be of particular interest to extend these to ubiquitin chains linked through other lysine residues (e.g. Lys-29 and Lys-63), where the mechanism of ubiquitin-recognition may differ to that for Lys-48-linked polyubiquitin, and also to substrate-conjugated ubiquitins, as well as investigations of full-length proteins, with the ultimate goal being the identification of in vivo ubiquitylated substrates of SQSTM1.

Acknowledgments

This work was supported by the Wellcome Trust, BBSRC and the Arthritis Research Campaign.

Footnotes

  • Structure Related to Function: Molecules and Cells: A Focus Topic at BioScience2004, held at SECC Glasgow, U.K., 18–22 July 2004. Edited by D. Alessi (Dundee, U.K.), T. Cass (Imperial College London, U.K.), T. Corfield (Bristol, U.K.), M. Cousin (Edinburgh, U.K.), A. Entwistle (Ludwig Institute for Cancer Research, London, U.K.), I. Fearnley (Cambridge, U.K.), P. Haris (De Montfort, Leicester, U.K.), J. Mayer (Nottingham, U.K.) and M. Tuite (Canterbury, U.K.).

Abbreviations: NF-κB, nuclear factor κB; PDB, Paget's disease of bone; SQSTM1, Sequestosome 1; TRAF6, tumour-necrosis-factor-receptor-associated factor 6; UBA, ubiquitin-associated

References

View Abstract