Aggregation modulating elements in mutant human superoxide dismutase 1

https://doi.org/10.1016/j.abb.2010.07.027Get rights and content

Abstract

Mutations in superoxide dismutase 1 (SOD1) cause some forms of familial amyotrophic lateral sclerosis (fALS). Affected tissues of patients and transgenic mouse models of the disease accumulate misfolded and aggregated forms of the mutant protein. In the present study we have identified specific sequences in human SOD1 that modulate the aggregation of fALS mutant proteins. From our study of a panel of mutant proteins, we identify two sequence elements in human SOD1 (residues 42–50 and 109–123) that are critical in modulating the aggregation of the protein. These sequences are components of the 4th and 7th β-strands of the protein, and in the native structure are normally juxtaposed as elements of the core β-barrel. Our data suggest that some type of intermolecular interaction between these elements may occur in promoting mutant SOD1 aggregation.

Research highlights

► Mutations in human superoxide dismutase 1 induce aggregation of the protein. ► A cysteine residue at position 111 in concert with surrounding amino acids plays an important role in modulating aggregation. ► Additional amino acids between positions 42 and 50 interact with the region surrounding cysteine 111 to promote aggregation of mutant protein.

Introduction

Dominantly inherited mutations in superoxide dismutase 1 (SOD1)2 are linked to familial amyotrophic lateral sclerosis (fALS) [1]. To date, more than 140 mutations in SOD1 have been associated with fALS [2] (http://alsod.iop.kcl.ac.uk/). Because these mutations have varied effects on enzyme activity and stability, it is thought that the mutant enzymes acquire one or more toxic properties [3]. The majority of fALS mutations are point mutations that occur predominantly at highly conserved amino acids [2], [4]. A subset of fALS mutations produce shifts in the reading frame or early termination codons that produce truncated mutant protein [2]. The effects of fALS mutations on enzyme activity, turnover, and folding of the SOD1 protein vary considerably [3], [5], [6]. Enzyme activity ranges from undetectable to normal [5], [7], [8], [9], [10], and many mutants increase the susceptibility of SOD1 to disulfide reduction [11]. One property that may be shared by all mutants is a higher inherent propensity to form large sedimentable structures that are insoluble in non-ionic detergent [2], [12]. To date, the inherent aggregation propensity of more than 40 different fALS mutants of SOD1 has been examined in cell culture models and all have been found to generate aggregates [2].

The role of large aggregates of mutant protein in neurotoxicity is not well understood. Recent studies have revealed a relationship between the relative rate at which mutant SOD1 forms large aggregates and the rapidity with which the human disease progresses [2], [13]. For example, the A4V mutation is associated with rapidly progressing disease and a high inherent propensity to aggregate whereas the H46R mutation is associated with slowly progressing disease and a low propensity to aggregate [2]. In transgenic mouse models of ALS, the large sedimentable aggregates begin to accumulate to significant levels at the age at which symptoms are first noticeable and build in abundance as symptoms progress [14], [15]. However, in mice that express the G93A and G37R fALS mutants, it is possible to accelerate disease by increasing the levels of the copper chaperone for SOD1 (CCS) and in such cases the large sedimentable aggregates of mutant protein do not accumulate [16], [17]. Notably, increasing CCS levels has no effect on the course of disease in mice that express the G85R and L126Z fALS mutants [17]. Thus, although it is possible to induce ALS-like symptoms in mice expressing mutant SOD1 without generating aggregates, such aggregates have been described in multiple mouse models that express only mutant SOD1 [13], [18], [19], [20], [21], [22], [23].

The mechanisms involved in the aggregation of SOD1 are not completely understood. Considerable attention has been placed on the role of disulfide cross-linking in the formation of SOD1 aggregates [4], [22], [24], [25]. Human SOD1 encodes 4 cysteines at positions 6, 57, 111, and 146. Studies in vitro and in cell culture suggest that cysteine residues 6 and 111 participate in mutant SOD1 aggregation perhaps by mediating intermolecular disulfide bonds [22], [24] or by participating in other types of intermolecular interactions [25]. In symptomatic SOD1 transgenic mice, high-molecular-weight, disulfide cross-linked forms of human SOD1 are prominent in the detergent-insoluble protein fraction, which become more abundant as mice approach disease endstage [4], [14], [22]. However, we have demonstrated that SOD1 aggregates are not stabilized by disulfide cross-linking alone [14]. Moreover, missense mutations at cysteines 6, 111 and 146 cause fALS (http://alsod.iop.kcl.ac.uk/). In cell culture models, SOD1 variants with mutations at these cysteine residues aggregate robustly and when combined into one recombinant gene with an experimental mutation to eliminate cysteine 57, the resultant mutant SOD1 protein retains the ability to aggregate [25]. Lastly, fibrillar aggregates of human SOD1, formed in vitro, that resemble amyloid structures are not extensively cross-linked by disulfide bonding [26]. Overall, the weight of evidence indicates that disulfide cross-linking is secondary to other mechanisms of protein self-assembly in the formation of large aggregate structures.

In studies to examine the role of disulfide cross-linking in mutant SOD1 aggregation, described above, there has been much focus on the cysteine at position 111 as a possible mediator of cross-linking. In cell culture and in vitro models of mutant SOD1 aggregation, mutagenesis of this cysteine to serine has been shown to reduce the potential of human SOD1 harboring an fALS mutation to aggregate to a level similar to wild-type protein [22], [24], [25]. Notably, in NSC-34 cells that inducibly express mutant SOD1, the toxicity of the fALS SOD1 variants A4V, C6F, G93A, and C146R was diminished by mutation of cysteine 111 to serine, correlating with diminished generation of detergent-insoluble aggregates [24]. Although this finding is consistent with the idea that disulfide cross-linking, via cysteine 111, could be important in aggregation and toxicity, we have previously demonstrated that mouse SOD1 encoding the fALS mutation G85R shows a high potential to aggregate [25]; and transgenic mice that express mouse SOD1 with this mutation develop ALS [27]. Mouse SOD1 normally possesses only 3 cysteines (positions 6, 57, 146) and encodes serine at position 111. Mouse and human SOD1 differ at 25 amino acid positions, and thus it would seem that some elements in the mouse sequence obviate the need for cysteine at residue 111 in promoting aggregation.

In the present work we generated chimeric human/mouse SOD1 constructs to examine how sequences surrounding cysteine 111, as well as sequences located more distally to cysteine 111, may modulate aggregation in proteins derived from the two species. We demonstrate that chimeric SOD1 proteins in which amino acids 42–50 (4th β-strand) and 109–123 (7th β-strand) are of human origin show the highest potential to aggregate. In conjunction with studies by Seetharaman and colleagues (accompanying manuscript), sequence elements in SOD1 that may be important sites of non-native interactions in the formation of multimeric aggregate structures are identified.

Section snippets

Construction of SOD1 expression vectors

All SOD1 variants were expressed using the pEF-BOS vector [28]. The cDNA genes for chimeric mouse/human (N-M/C-H) and human/mouse (N-H/C-M) SOD1 of wild-type sequence (WT) were synthesized by Genescript (Piscataway, NJ, USA). Experimental and fALS associated mutations in human, mouse, and chimeric N-H/C-M SOD1 cDNAs were generated by standard PCR strategies with oligonucleotides that introduce the specific point mutations. All cDNA genes and pEF-BOS vectors encoding these cDNAs were verified by

Results

Human and mouse SOD1 mutated to encode the G85R fALS mutation show high propensities to aggregate in cell culture [25], and both mutants cause hindlimb paralysis when expressed in mice [27], [30]. Human and mouse SOD1 differ at 25 amino acids, but position 111, which encodes cysteine in human SOD1 and serine in mouse SOD1, has been of particular interest [25]. Note that a serine encoded at amino acid 111 is the more conserved residue at this position in mammals [4]. In human SOD1, when amino

Discussion

Our study was designed to identify regions in the human SOD1 protein that modulate aggregation of mutant SOD1 in an effort to improve our understanding of the role of cysteine 111 in modulating aggregation [24], [25]. Intriguingly, mouse SOD1 normally encodes a serine at position 111 and when mouse SOD1 is mutated to encode human fALS mutations, we observe robust aggregation [25]. To identify sequence elements in human and mouse SOD1 that account for the differing dependencies on a cysteine at

Acknowledgments

We thank Sai Seetharaman and P. John Hart for continued collaboration and invaluable insight into the structural aspects of SOD1 and the SOD1 chimeras described in this paper. We thank Hilda Brown for advice and assistance in generating the SOD1 expression constructs. We are grateful to Mercedes Prudencio and Guilian Xu for thoughtful discussion and advice throughout this study. This work was supported by National Institutes of Health Grant P01 NS049134 (Project 3 to DRB).

References (37)

  • L.J. Hayward et al.

    J. Biol. Chem.

    (2002)
  • A. Tiwari et al.

    J. Biol. Chem.

    (2003)
  • M. Son et al.

    Neurobiol. Dis.

    (2009)
  • M. Watanabe et al.

    Neurobiol. Dis.

    (2001)
  • J. Wang et al.

    Neurobiol. Dis.

    (2005)
  • M. Cozzolino et al.

    J. Biol. Chem.

    (2008)
  • C.M. Karch et al.

    J. Biol. Chem.

    (2008)
  • L.I. Bruijn et al.

    Neuron

    (1997)
  • B.F. Shaw et al.

    J. Biol. Chem.

    (2008)
  • D.R. Rosen et al.

    Nature

    (1993)
  • M. Prudencio et al.

    Hum. Mol. Genet.

    (2009)
  • J.S. Valentine et al.

    Proc. Natl. Acad. Sci. USA

    (2003)
  • J. Wang et al.

    J. Neurochem.

    (2006)
  • D.R. Borchelt et al.

    Proc. Natl. Acad. Sci. USA

    (1994)
  • S.Z. Potter et al.

    J. Biol. Inorg. Chem.

    (2003)
  • C.R. Nishida et al.

    Proc. Natl. Acad. Sci. USA

    (1994)
  • M. Wiedau-Pazos et al.

    Science

    (1996)
  • T. Ratovitski et al.

    Hum. Mol. Genet.

    (1999)
  • Cited by (0)

    1

    Present address: Department of Psychiatry, Washington University St. Louis, St. Louis, MO, USA.

    View full text