Dimer destabilization in superoxide dismutase may result in disease-causing properties: Structures of motor neuron disease mutants Michael A. Hough*, J. Gu¨nter Grossmann*, Svetlana V. Antonyuk*, Richard W. Strange*, Peter A. Doucette†, Jorge A. Rodriguez†, Lisa J. Whitson‡, P. John Hart‡, Lawrence J. Hayward§, Joan Selverstone Valentine†, and S. Samar Hasnain*¶ *Molecular Biophysics Group, Council for the Central Laboratory of the Research Councils, Daresbury Laboratory, Warrington, Cheshire WA4 4AD, United Kingdom; †Department of Chemistry and Biochemistry, University of California, Los Angeles, CA 90095; ‡Center for Biomolecular Structure Analysis, University of Texas, 7703 Floyd Curl Drive, San Antonio, TX 78229-3900; and §Department of Neurology, University of Massachusetts Medical School, 55 Lake Avenue North, Worcester, MA 01655 Edited by Gregory A. Petsko, Brandeis University, Waltham, MA, and approved January 26, 2004 (received for review August 11, 2003)
More than 90 point mutations in human CuZn superoxide dismutase lead to the development of familial amyotrophic lateral sclerosis, known also as motor neuron disease. A growing body of evidence suggests that a subset of mutations located close to the dimeric interface can lead to a major destabilization of the mutant enzymes. We have determined the crystal structures of the Ala4Val (A4V) and Ile113Thr (I113T) mutants to 1.9 and 1.6 Å, respectively. In the A4V structure, small changes at the dimer interface result in a substantial reorientation of the two monomers. This effect is also seen in the case of the I113T crystal structure, but to a smaller extent. X-ray solution scattering data show that in the solution state, both of the mutants undergo a more pronounced conformational change compared with wild-type superoxide dismutase (SOD) than that observed in the A4V crystal structure. Shape reconstructions from the x-ray scattering data illustrate the nature of this destabilization. Comparison of these scattering data with those for bovine CuZn SOD measured at different temperatures shows that an analogous change in the scattering profile occurs for the bovine enzyme in the temperature range of 70 – 80°C. These results demonstrate that the A4V and I113T mutants are substantially destabilized in comparison with wild-type SOD1, and it is possible that the pathogenic properties of this subset of familial amyotrophic lateral sclerosis mutants are at least in part due to this destabilization. human superoxide dismutase 兩 crystal structure 兩 x-ray solution scattering 兩 neurodegenerative disease
A
myotrophic lateral sclerosis (ALS; Lou Gehrig’s disease; motor neuron disease) is a fatal neurodegenerative condition characterized by the progressive, selective loss of motor neurons in the spinal cord, brainstem and motor cortex (1). Sporadic ALS (SALS), which makes up ⬇90% of all ALS cases, arises without family history and has no known cause (2). However, some 10% of cases are familial ALS (FALS), and of these cases, ⬇25% are linked to mutations in sod1, the gene encoding copper, zinc superoxide dismutase (SOD1) (3). SALS is clinically indistinguishable from FALS, but the average age of onset in FALS is somewhat earlier (4). Since the link between SOD1 and FALS was first established ⬇10 years ago, ⬎90 FALS-linked SOD1 mutations have been discovered (5) Whereas the underlying mechanism for motor neuron death in both SALS and FALS is unknown, it is expected that an understanding of SOD1-linked FALS may also have direct implication for our understanding of SALS. Several experimental approaches have demonstrated that FALS is not caused by a reduction in SOD activity. SOD1 knockout mice exhibit decreased fertility and recovery from axonal injury, yet they do not develop motor neuron disease (6). However, transgenic mice that overexpress FALS mutant human 5976 –5981 兩 PNAS 兩 April 20, 2004 兩 vol. 101 兩 no. 16
SOD1 develop a severe motor neuron disease (7), whereas those overexpressing the wild-type enzyme demonstrate only mild motor dysfunction at an advanced age (8, 9). In addition, the vast majority of FALS mutant SOD1s display identical, and in some cases, even elevated SOD activity, compared with that of the wild-type SOD1 (wtSOD1) (10). These results strongly suggest that the gain of a toxic property, rather than decreased SOD1 function, is responsible for ALS pathogenicity. SOD1 is a 32-kDa homodimeric enzyme that forms a critical component of the cellular defense against reactive oxygen species (11). This ubiquitous enzyme catalyses the dismutation reaction of superoxide to hydrogen peroxide and oxygen through the cyclic reduction and reoxidation of copper (12). The protein fold is described as an eight-stranded Greek-key -barrel with three extended loop regions (13, 14). FALS-linked mutations have been found in every major structural element throughout the dimeric enzyme, including the inter-monomer interface, -barrel, loop regions, disulfide bond, and both the copper and zinc sites. FALS-linked nonsense mutations that cause Cterminal truncations have also been identified (5). The Ala4Val mutant of SOD1 (A4V) is the most common SOD1 mutation discovered to date, accounting for ⬇50% of SOD1-linked FALS cases (15). A4V is also a particularly severe mutation in that it induces a rapid rate of disease progression, resulting in death within an average of 1.2 years after the onset of symptoms (16). In light of these clinical statistics, as well as the availability of a large body of existing research on this mutant, the high-resolution structural characterization of A4V is of considerable importance. The Ile113Thr mutant of SOD1 (I113T) mutation is less common than A4V and has been reported as causing FALS with a rapid onset, but highly variable duration of disease. This mutation is notable for the observation of massive neurofilament aggregation in affected motor neurons (17). Intriguingly, the I113T mutation has also been identified in some cases of apparent SALS (18). These cases may suggest that some SALS cases are in fact SOD1-related FALS with a very low penetrance. Recent biophysical studies (19, 20) have placed the A4V and I113T mutants in a category of ‘‘wild-type-like’’ mutants that possess similar metal-binding and catalytic activity This paper was submitted directly (Track II) to the PNAS office. Abbreviations: ALS, amyotrophic lateral sclerosis; FALS, familial ALS; SALS, sporadic ALS; SOD, superoxide dismutase; wtSOD, wild-type SOD; BSOD, bovine SOD; A4V, the Ala4Val mutant of SOD1; I113T, the Ile113Thr mutant of SOD1. Data deposition: The atomic coordinates and structure factors have been deposited in the Protein Data Bank, www.pdb.org [PDB ID codes 1UXM (A4V) and 1UXL (I113T)]. See Commentary on page 5701. ¶To whom correspondence should be addressed at: CCLRC Daresbury Laboratory, Keckwick
Lane, Warrington WA4 4AD, United Kingdom. E-mail:
[email protected]. © 2004 by The National Academy of Sciences of the USA
www.pnas.org兾cgi兾doi兾10.1073兾pnas.0305143101
I113T
Space group Resolution range, Å Unit cell dimensions, Å
P21 50.0–1.9 a ⫽ 112.4, b ⫽ 145.6, c ⫽ 112.5,  ⫽ 120.1
Unique reflections Completeness, % I兾(I) Rmerge*, % Rwork†, % Rfree‡, % No. of protein atoms No. of water molecules Average B factor: Protein, Å2 Water, Å2 rms deviations Bond distances, Å Bond angles, ° Estimated standard uncertainty, Å
245,475 96.7 (70.7) 8.3 (3.1) 6.3 (25.6) 22.7 24.9 13,368 1,230
C2221 50–1.6 a ⫽ 166.0, b ⫽ 203.5, c ⫽ 144.0 317,774 99.6 (99.7) 21.0 (2.4) 5.1 (42.0) 16.1 18.8 13,719 2,488
16.6 43.6
18.1 49.5
0.016 1.78 0.13
0.016 1.60 0.062
*Rsym ⫽ 兺h兺i兩I(h) ⫺ I(h)i兩兾兺h兺iI(h)i, where I(h) is the mean intensity after rejections. †R work ⫽ 兺h储Fobs(h)兩 ⫺ 兩Fcalc(h)储兾兺h兩Fobs(h)兩, where Fobs(h) and Fcalc(h) are the observed and calculated structure factors, respectively. No F兾(F) cutoff was applied to the data. ‡The 11954 (A4V) and 15958 (I113T) reflections were excluded from refinement to calculate Rfree. Nos. in parentheses refer to the outermost-resolution shell (1.95–1.90 Å; A4V) (1.66 –1.60; I113T).
own promoter. The proteins were expressed in Saccharomyces cerevisiae SOD1-EG118 strain lacking the endogenous yeast sod1 gene. Cells were grown in yeast extract兾peptone兾dextrose media for 36 h, were resuspended in 200 mM Tris䡠HCl buffer, pH 8.0, containing 0.1 mM EDTA, 50 mM NaCl, 0.045 M PMSF (Sigma), and were lysed with 0.5-mm glass beads by using a blender. The pH value of the lysate was monitored throughout the lysis procedure to ensure that it remained ⬎7.0. Lysate was centrifuged at 8000 ⫻ g for 1 h at 4°C and the supernatant was slowly brought to 60% ammonium sulfate saturation, and was incubated on ice for 30 min. It was subsequently centrifuged at 8000 ⫻ g for 1 h at 4°C and the supernatant was applied to a fast f low phenyl-Sepharose (Pharmacia) column (2.5 ⫻ 25 cm) preequilibrated with 2.0 M ammonium sulfate兾150 mM NaCl兾50 mM potassium phosphate, pH 7.0 (buffer A). The column was washed with one volume of buffer A and the protein was eluted by using a step gradient of decreasing concentrations of ammonium sulfate (0.1 M every 45 ml). Fractions were analyzed by SDS兾PAGE and those containing SOD were combined and concentrated. The protein was transferred into 2.25 mM potassium phosphate buffer, pH 7.0, containing 160 mM NaCl (high salt buffer) and was subjected to gel filtration chromatography on a Sephadex G-75 superfine (Pharmacia) column equilibrated with the same buffer. Some preparations were subsequently subjected to additional anion exchange chromatography on a DEAE Sephadex A-50 (Pharmacia) column. The purity of the protein was verified by SDS兾PAGE and electrospray ionization MS. Crystallization, Data Collection, Structure Solution, and Refinement.
to wtSOD1. However, differential scanning calorimetry studies have indicated significantly decreased thermal stability of the apoprotein in a number of FALS-related SOD1 mutants compared with wtSOD1, including A4V (21). In addition, it has been demonstrated that A4V exhibits an increased susceptibility to disulfide reduction (22) and peroxidase activity, whereas the I113T mutant has been shown to have an increased susceptibility to glycation (23). The half-life of several mutants in COS-1 cells was shown to be decreased compared with that in wild type (10). Whereas all mutants studied showed a decrease in half-life, there was wide variation between mutants. I113T showed a half-life of 20 h, while that of A4V was 7.5 h. We have determined the crystal structures of the A4V and I113T mutants by x-ray crystallography to 1.9- and 1.6-Å resolution, respectively, revealing a significant alteration in the relative orientation of the two monomers making up the enzyme dimer in the mutants. Moreover, we obtained low-resolution shapes of these mutants in solution from x-ray scattering data. These structures reveal a much greater destabilization, namely alteration of relative subunit orientation, of both mutants in solution to that observed in the crystal structures. The nature of this structural change is very similar to what we find when the bovine CuZnSOD protein in solution is heated to some 70–80°C, suggesting an overall destabilization of the mutant proteins. We note that recently a destabilization of the dimer interface has been suggested in the Parkinson’s disease-associated mutation L166P in DJ-1 in which case the loss of function or inactivation of DJ-1 has been assigned to monomerization after this destabilization (24). Methods Protein Expression and Purification. The DNA fragments encoding
mutant A4V and I113T human SOD1 were generated by PCR and were ligated into the YEP351-hSOD plasmid, which directs the expression of the protein under the control of its
Hough et al.
Crystals were grown by using the vapor diffusion hanging drop method with a protein concentration of 10 mg兾ml. The reservoir solution consisted of 15% polyethylene glycol 2000, 0.2 M calcium acetate, and 0.1 M Tris pH 8.0 (A4V) or 2.4 M ammonium sulfate, 100 mM NaCl, 100 mM Tris䡠HCl, pH 7.5 (I113T). Crystals grew within a week in space groups P21 (A4V) and C2221 (I113T). Crystals of A4V (I113T) were soaked in a cryoprotectant solution comprised of 25% ethylene glycol (glycerol) in the appropriate mother liquor before flash-cooling in the cryostream. X-ray data were collected at 100 K on an Area Detector Systems Corporation Quantum 4 charge-coupled device detector using 0.978-Å x-ray radiation on station 14.2 at the Synchrotron Radiation Source, Daresbury Laboratory. The data resolution was assessed as 1.9 (A4V) and 1.6 Å (I113T), respectively, based on an I兾(I) ⬎ 2.0 in the outermost-resolution shell (Table 1). The data were indexed, scaled, and merged by using HKL2000. The structures were solved by using MOLREP (25) using the human wtSOD dimer (PDB ID codes 1SOS and 1HL5) as the search models for A4V and I113T, respectively. A total of six (A4V) and five (I113T) SOD dimers were located in the crystallographic asymmetric unit. The initial model was refined by the maximum likelihood method implemented in REFMAC5 (26) as part of the CCP4 program suite and rebuilt interactively by using 2Fobs ⫺ Fcalc and Fobs ⫺ Fcalc electron density maps in the program O (27). Five percent of the data were set aside from refinement to calculate the Rfree. Each monomer was refined independently without the application of noncrystallographic symmetry restraints. At no point in the refinement were restraints applied to the copper ligand distances or bond angles. A bulk solvent correction was applied and solvent molecules added by using ARP/WARP (28). In the latter stages of refinement, translation-libration-screw parameters were determined (29). Refinement converged to final R factors of 22.7% (Rfree ⫽ 24.9%; A4V) and 16.1% (Rfree ⫽ 18.8%; I113T), respectively. PNAS 兩 April 20, 2004 兩 vol. 101 兩 no. 16 兩 5977
SEE COMMENTARY
A4V
BIOPHYSICS
Table 1. Crystallographic data
Solution X-Ray Scattering Data Collection and Analysis. Scattering data from the A4V and I113T mutants and human wtSOD, measured at 4°C, as well as native bovine SOD (BSOD), recorded between 4°C and 80°C, were collected at station 2.1 (30) of the Synchrotron Radiation Source, Daresbury Laboratory. Samples for wild-type and mutants were prepared at protein concentrations of 1 and 15 mg兾ml in an identical buffer to that used for crystallization. Lyophilized BSOD was purchased from Boehringer Mannheim (Mannheim Germany) and prepared at concentrations of 1 and 20 mg兾ml in 50 mM glyclglycine buffer, pH 6.5. All data were collected and analyzed according to established procedures (31). Briefly, a sample-todetector distance of 1.25 m was used covering the momentum transfer interval 0.025 Å⫺1 ⱕ兾q ⱕ 0.62 Å⫺1. The modulus of the momentum transfer is defined as q ⫽ 4 sin⌰兾, where 2⌰ is the scattering angle and is the wavelength (1.54 Å). The q range was calibrated by using silver behenate powder and wet rat tail collagen (based on diffraction spacings of 58.38 and 670 Å, respectively). The radius of gyration, the forward scattering intensity, and the intra-particle distance distribution function p(r) were evaluated with the indirect Fourier transform program GNOM (32). Particle shapes were restored ab initio from the experimental scattering profiles with the dummy atom procedure, as implemented in the program DAMMIN (33) imposing two-fold molecular symmetry.
Results and Discussion Description of the Structures. The overall fold observed for wtSOD
(12, 13, 34) is largely preserved in the A4V and I113T mutants. Each monomer of the dimeric enzyme forms an eight-stranded antiparallel -barrel. Two extended loop regions, the Zn loop and electrostatic loop, form the walls of a channel from the enzyme surface to the active site. Several charged residues within the electrostatic loop, as well as the catalytically important Arg-143, are involved in the electronic guidance of the substrate to the active site and contribute to the high specificity of CuZnSOD for the superoxide substrate. The individual monomers of the mutant structures and wtSOD1 superimpose well and changes to the structure are localized to loop regions, the mutation sites, and to the intermonomer interface. Comparison of the A4V and I113T metal centers with those in wtSOD (13) indicate that no significant change occurs to either the copper or zinc sites as a result of the mutations. This finding is consistent with the results of extended x-ray absorption fine structure analyses (R.W.S, unpublished results), which show no deviation from the wild-type spectra for copper or zinc, nor do we observe any significant changes in the conformation of the active-site cavity. Crystal Packing: The Trimer of Dimers. The molecules of A4V and I113T mutant SOD1 pack into the crystals in a so-called ‘‘trimer-of-dimers’’ motif (Fig. 5, which is published as supporting information on the PNAS web site), involving primarily interactions between residues 128–131 of the electrostatic loop in one molecule and Thr-88, Ser-96, and Ser-98 from the Zn loop of a second SOD1 molecule. This motif is also observed in crystal structures of human wtSOD1 in two different crystal forms (13) and for the L38V FALS-related mutant (S.A. and R.W.S., unpublished results). The observation of this packing motif in all of these structures of metallated human SOD1, in several space groups and significantly different crystallization conditions, suggests that it is a strongly preferred means of interaction between metal-loaded SOD1 molecules. Notably, the crystal structures of apo wild-type (14) and several metal-deficient mutants of SOD1 showed distinctly different modes of crystal packing in which so-called linear, zig-zag, and helical fibril-like arrangements were observed (35). This observation was proposed to be the result of a loss of order in the electrostatic and Zn loops in the metal5978 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0305143101
deficient SOD1 proteins, leading to the accessibility of a gain-of interaction interface, due to the deprotection of edge strands in the -barrel. Dimer Interface and Mutation Sites. In wtSOD, the monomers
forming the dimeric molecule are linked by means of an extensive interface consisting of four hydrogen bonds plus a large number of hydrophobic and water-mediated interactions. The main residues involved are residues 50–53, 114, 148, and 150–153 at the C terminus. In a SOD1 dimer with subunits designated as A兾B, hydrogen bonds are formed between AGly51N–BIle151O (⬇ 2.7 Å) and AGly114O–BIle151N (⬇ 2.8 Å), as well as the symmetrical A151N–Gly114O (⬇ 2.8 Å) and Ile151O–Gly51N (⬇ 2.8 Å). Residue 4 occupies a position close to the interface, but is not directly involved. However, the side-chain C forms a close hydrophobic contact to the C␦2 atom of Ile-113 (Fig. 1a), neighbor to Gly-114, which is a key hydrogen-bonding residue at the interface. The dimer–interface interactions in wtSOD are largely maintained in the A4V structure with similar bond lengths. The most prominent changes occur at the mutation site and around the N terminus. The mutation of Ala-4 to Val is well defined in the electron density (Fig. 1b). The mutation results in a shift in the position of Ile-113 due to the replacement of the single hydrophobic interaction present in wtSOD with two others, namely from Val-4 C␥1 to the C␥1 and C␦1 atoms of Ile-113. Also, a close contact is formed between Val-4 C␥2 and Ile-149 C␥2. Notably, the two cysteines mutated in the A4V– C6A–C111S protein (36) lie only two residues away from the mutated Ala-4 and Ile-113.储 In addition, an extensive hydrophobic contact region is formed between the side chains of residues Val-5 and Val-7, close to the N terminus and residues Gly-51 to Thr-54. In wtSOD, the N terminus of the protein chain is stabilized by an extensive series of hydrogen bonds linking N and O backbone atoms from residues 2–5 with those of residues 20–22 and 150–152. In A4V, this hydrogen bond network is maintained as far as the backbone oxygen atom of Thr-2, which makes an H-bond to the backbone N of Gln-22. At this point, the peptide adopts a different conformation such that the Thr-2N to Gln-22O hydrogen bond is no longer intact. It is not clear whether the new orientation of residue 1 in the mutant is significant or merely a result of crystal packing forces, although it may possibly increase the exposure of the N terminus to bulk solvent and hence to proteolytic digestion. The C terminus itself (residue Gln-153) exhibits no significant change in position or side chain orientation to that observed in wtSOD in either mutant structure. This finding is consistent with the heavy involvement of the C-terminal region in the monomer–monomer interface. In the I113T structure, the mutation of Ile-113 to Thr does not result in a significant structural change in surrounding residues. The major effect is the loss of the close hydrophobic contact between Ala-4C and Ile-113C␦2 (Fig. 1c). In the mutant structure, the separation of Ala-4C and the closest atom from residue Thr-113 is ⬇4.5 Å. The loss of this contact may well affect the dynamic behavior of the mutant. The major structural change in the I113T mutant structure occurs in the -turn region around residues 109–111 (Fig. 6, which is published as supporting information on the PNAS web site). However, these residues are not involved in the dimer interface and the effect of this structural change on the molecule is not clear. 储The
crystal structure of an A4V–C6A–C111S triple mutant of human CuZnSOD has been determined (36) and compared with the structure of its parent thermostable mutant (C6A–C111S). The interpretation of these results is complicated by the mutation of two cysteine residues at the dimer interface, in close proximity to the site of the A4V mutation, in both structures. The effect on the structure and stability of SOD1 as a result of the A4V mutation is unlikely to be the same when applied to the thermostable mutant, as in the case where this mutation is applied to wtSOD.
Hough et al.
SEE COMMENTARY
(rms deviation, 0.85Å) in the A4V dimer (Fig. 2, and Fig. 7, which is published as supporting information on the PNAS web site). The motion is complex, but may be approximated by a corkscrew-like twist around the normal to the twofold axis, and a small rotation of the plane formed by the interface. This result may indicate a more flexible structure in the mutant and the adoption of a different conformation between the monomers of the dimer as a result of the changes at the interface. In the I113T structure, the orientation of the two monomers is changed to a lesser extent (rms deviation, 0.57Å) compared with wtSOD1. Solution X-Ray Scattering and Shape Reconstruction. The x-ray
Fig. 1. Images of 2Fobs ⫺ Fcalc electron density maps, contoured at 1.5 are shown for wtSOD (a): the Ala-4 side-chain C forms a 3.6-Å hydrophobic contact to the C␦2 atom of Ile-113. (b) A4V: The mutation is clearly defined in the electron density. Hydrophobic close contacts are indicated by dashed lines. Steric constraints due to the bulkier Val side chain in the mutant structure cause a shift in the position of Ile-113 due to the replacement of the single 3.6-Å Ala-4C–Ile-113C␦2 hydrophobic interaction with two others, namely from Val-4C␥1 to Ile-113C␥1 (3.8 Å) and Ile-113C␦2 (3.9 Å). In addition, a contact is formed between Val-4C␥2 and Ile-149C␥2 (3.5 Å). (c) I113T: The close contact between Ala-4C and Ile-113C ␦2 is lost.
Shift in Monomer–Monomer Orientation. The relative orientation of
the two monomers in the dimer for the mutants was examined by superposition of the major interface region for all molecules in the asymmetric unit of each mutant and with those of the nine molecules comprising the 1.8-Å wtSOD structure. This comparison showed that the orientation of two monomers was changed
Hough et al.
solution-scattering profiles for wtSOD, A4V, and I113T are shown in Fig. 3. The most striking changes are associated with the characteristic scattering minimum around q ⫽ 0.25 Å⫺1 (see Inset), which can be associated with the compact ellipsoidal shape of the native dimer conformation. Furthermore, the structural parameters (radius of gyration, Rg, and maximum particle dimension, Dmax) are increased from 21.4, 68 (wildtype), to 22.0, 74 (A4V), and to 22.9, and 82 Å (I113T), respectively. These data indicate that there are significant changes in the overall molecular conformation, which are likely to result from different orientations of the monomers. The increase in Rg with a corresponding increase in Dmax clearly indicates an enlargement of the molecule, which would be consistent with significant reorientation of the subunits. We note that the introduction of a small population of monomers in the sample would result in a smaller Rg value. In fact, these data suggest that a more pronounced change takes place in solution between the native and mutant proteins than that represented in the crystal structure of A4V. Notably, the scattering profile of I113T shows larger differences to wtSOD than that of A4V. It seems likely that the nature of the crystalline lattice limits the dynamic behavior of the enzyme such that only a relatively small deviation, if any, from the wtSOD structure may be observed in the crystalline state. It is also possible that molecules with a particular conformation have been preferentially selected during the crystallization process. The scattering data also show that the mutant enzymes do not aggregate in solution under these experimental conditions, and in this respect behave in a similar manner to the native enzyme; in all three cases, the scattering data are consistent with a dimeric SOD1 molecule. PNAS 兩 April 20, 2004 兩 vol. 101 兩 no. 16 兩 5979
BIOPHYSICS
Fig. 2. Shift in the relative orientation of the two subunits of SOD on mutation of Ala-4 to Val. Dimers of A4V (red) and the wild-type structure (blue) were superimposed by least-squares superposition of the interface residues (114, 51–53, and 149 –153). The relative position of the two monomers is significantly altered in A4V. This shift takes the form of a small rotation and a larger corkscrew-like motion about the normal to the twofold pseudo symmetry axis defined between the two molecules. The six molecules of A4V express this shift to different levels. In general, molecules with higher thermal parameters (and hence higher mobility) show the larger shifts. The maximum shift is shown by dimer 5 in A4V (I and J) and this shift is shown in the figure.
mutant. The latter conformation seems to indicate a rather elongated dimeric structure that is likely to have a strikingly different monomer–monomer interface compared with the wildtype protein. Temperature-Dependence Study of Bovine CuZnSOD and Nature of Dimer Destabilization. To further explore the nature of the change
in molecular conformation of the two disease-causing human SOD mutants, x-ray scattering experiments were performed by using native BSOD at different temperatures between 4°C** and 80°C (Fig. 3b). BSOD represents a stable and well characterized dimeric structure, and is very similar in structure to wtSOD1. These data show that the changes observed for A4V and I113T in comparison with human wtSOD at 4°C appear to correlate with those seen for BSOD when it is heated to 80°C. It is of particular interest to focus again on the scattering minimum around q ⫽ 0.25 Å⫺1 (see Fig. 3 a and b Insets). By using the scattering profile of human wtSOD as a reference, the A4V scattering curve shows a very similar change when compared with the change in scattering of BSOD between low temperatures (4–50°C) and higher temperatures; i.e., between 60°C and 70°C. The minimum around q ⫽ 0.25 Å⫺1 becomes more shallow in the A4V mutant profile. In the case of the I113T mutant, this scattering minimum is reduced to a shoulder that compares well with the scattering data recorded for BSOD at nearly 80°C. This result suggests that the degree of destabilization of human SOD1 due to the A4V and I113T mutations can be mirrored in the bovine CuZnSOD when it is brought close to its melting temperature. We note that a significant reduction of protein half-life has been observed for FALS mutants in transfected cells, in comparison with wtSOD1 (10), although the extent of this effect was inconsistent across a range of mutants. Fig. 3. Solution x-ray scattering profiles of A4V and I113T mutants and human wtSOD, measured at 4°C (a), as well as native BSOD recorded at different temperatures (b). The profiles result from merging the low- and high-concentration data in the small- and wide-angle region.
To visualize the structural changes observed in the mutant proteins with respect to the wtSOD dimer, particle shapes were restored ab initio from the experimental scattering profiles. Several independent shape calculations have been performed for each of the three proteins. The models generated with the known twofold symmetry showed only minor variations illustrating the consistency of the ab initio reconstructions. Thus, the restored shape models displayed for wtSOD1 and the A4V and I113T mutants in Fig. 4 are reliable representatives. From these data, it is clear that the native conformation of the wild-type protein undergoes a small opening of the dimer interface in the A4V mutant and a much more pronounced change in the I113T
Fig. 4. Two perpendicular views of the shapes reconstructed from the wildtype, A4V, and I113T scattering profiles shown in Fig. 3a. Each shape reconstruction is performed with a dimer obeying twofold symmetry in which each subunit is represented by closely packed spheres. The wild-type crystal structure (PDB ID code 1HL5) has been superimposed in the scattering envelope of wtSOD1. 5980 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0305143101
Implications for FALS Pathogenicity of A4V and I113T. Several hypotheses have been developed in recent years to explain the mode of toxic action of SOD1 mutants that selectively targets motor neurons (20, 37, 38) and yet it remains largely unclear at this time how nearly 100 different mutations convert this otherwise protective, housekeeping enzyme into a toxic molecule. One hypothesis is that the copper ion at the active site of mutant SOD1 could catalyze harmful oxidation reactions by hydrogen peroxide (38) or peroxynitrite (39), as the result of greater access of abnormal substrates to the active site. However, our structure shows no significant changes in the active site itself or the cavity leading to it, and thus provides no evidence in support of this hypothesis. In addition, our extended x-ray absorption fine structure data show that the Cu and Zn sites in A4V and I113T are similar to those of wtSOD (metal-ligand bond distances are within 0.02 Å from extended x-ray absorption fine structure). Thus, our structural results described here provide no evidence for alterations at the metal-binding sites of these species that could result in decreased metal binding affinity. It has also been suggested that FALS mutant SOD1 may possess altered solubility and that this property may lead to the formation of aggregates containing SOD1 in affected tissues. Although the solution-scattering data do not provide evidence that the mutant enzymes oligomerize at the solution concentrations and buffer conditions used in our in vitro experiments, it is possible that an oxidatively damaged and兾or metal-deficient species is responsible for oligomerization or aggregation in vivo (40–42). In fact, recent structures of metal-deficient S134N and H46R FALS mutants (35) showed that these mutant SOD enzymes formed amyloid-like fibrils in the crystals. It remains to be determined whether the metallated A4V and I113T SOD1 mutants examined here share a common mechanism of cellular
**Scattering data for BSOD at 4 and 50° C are indistinguishable.
Hough et al.
Conclusions The crystal structures of the A4V and I113T mutants of SOD1 reveal a significant reorientation of the two subunits at the monomer–monomer interface, with the A4V structure depicting a more dramatic alteration. The structure of the metal sites is unaffected by the mutation, and this finding is confirmed by solution x-ray absorption fine structure data. X-ray scattering data confirm the presence of a large conformational change for these mutants compared with wtSOD in solution. In fact, the
We thank members of the International Consortium on Superoxide Dismutase and Amyotrophic Lateral Sclerosis for their interest and discussion. This work was supported by the Council for the Central Laboratory of the Research Councils and the Motor Neurone Disease Association (to S.S.H.); National Institutes of Health Grants NS39112 (to P.J.H.), GM28222 (to J.S.V.), and NS44170 (to L.J.H.); Robert A. Welch Foundation Grant AQ-1399 (to P.J.H.); grants from the ALS Association (to L.J.H., J.S.V., and P.J.H.); and a Ford Foundation Predoctoral Fellowship (to L.J.W.).
1. Brown, R. H., Jr. (1995) Cell 80, 687–692. 2. Maier, C. M. & Chan, P. K. (2002) Neuroscientist 8, 323–334. 3. Rosen, D. R., Siddique, T., Patterson, D., Figlewicz, D. A., Sapp, P., Hentati, A., Donaldson, D., Goto, J., O’Regan, J. P., Deng, H. X., et al. (1993) Nature 362, 59–56. 4. Gaudette, M., Hirano, M. & Siddique, T. (2000) Amyotroph. Lateral Scler. 1, 83–89. 5. Cleveland, D. W. & Rothstein, J. D. (2001) Nat. Rev. Neurosci. 11, 806–819. 6. Reaume, A. G., Elliot, J. L., Hoffman, E. K., Kowall, N. W., Ferrante, R. J., Siwek, D. F., Wilcox, H. M., Flood, D. G., Beal, M. F., Brown, R. H., Jr., et al. (1996) Nat. Genet. 13, 43–47. 7. Wong, P. C., Pardo, C. A., Borchelt, D. R., Lee, M. K., Copeland, N. G., Jenkins, N. A., Sisodia, S. S., Cleveland, D. W. & Price, D. L. (1995) Neuron 14, 1105–1116. 8. Jaarsma, D., Haasdijk, E. D., Grashorn, J. A. C., Hawkins, R., van Duijn, W., Verspaget, H. W., London, J. & Holstege, J. C. (2000) Neurobiol. Dis. 7, 623–643. 9. Gurney, M. E., Pu, H., Chiu, A. Y., Dal Canto, M. C., Polchow, C. Y., Alexander, D. D., Caliendo, J., Hentati, A., Kwon, Y. W., Deng, H. X., et al. (1994) Science 264, 1772–1775. 10. Borchelt, D. R., Lee, M. K., Slunt, H. S., Guarnieri, M., Xu, Z. S., Wong, P. C., Brown, R. H., Jr., Price, D. L., Sisodia, S. S. & Cleveland, D. W. (1994) Proc. Natl. Acad. Sci. USA 91, 8292–8296. 11. Fridovich, I. (1975) Annu. Rev. Biochem. 44, 147–159. 12. Tainer, J. A., Getzoff, E. D., Richardson, J. S. & Richardson, D. C. (1983) Nature 306, 284–287. 13. Parge, H. E., Hallewell, R. A. & Tainer, J. A. (1992) Proc. Natl. Acad. Sci. USA 89, 6109–6113. 14. Strange, R. W., Antonyuk, S., Hough, M. A., Doucette, P., Rodriguez, J., Hart, P. J., Hayward, L. J., Valentine, J. S. & Hasnain, S. S. (2003) J. Mol. Biol. 328, 877–891. 15. Cudkowicz, M. E., McKenna-Yasek, D., Chin, W., Geller, B., Hayden, D. L., Schoenfeld, D. A., Hosler, B. A., Horvitz, H. R. & Brown, R. H. (1997) Ann. Neurol. 41, 210–221. 16. Rosen, D. R., Bowling, A. C., Patterson, D., Usdin, T. B., Sapp, P., Mezey, E., McKenna-Yasek, D., O’Regan, J., Rahmani, Z., Ferrante, R. J., et al. (1994) Hum. Mol. Genet. 3, 981–987. 17. Kokubo, Y., Kuzuhara, S., Narita, Y., Kikugawa, K., Nakano, R., Inuzuka, T., Tsuji, S., Watanabe, M., Miyazaki, T., Murayama, S. & Ihara, Y. (1999) Arch. Neurol. (Chicago) 56, 1506–1508. 18. Andersen, P. M. (2001) Amyotroph. Lateral Scler. Other Motor Neuron Disord. 2, S37–S41. 19. Hayward, L. J., Rodriguez, J. A., Kim, J., M., Tiwari, A., Goto, J. J., Cabelli, D. D., Valentine, J, S. & Brown, J. H., Jr. (2002). J. Biol. Chem. 277, 15923–15931.
20. Valentine, J. S. & Hart, P. J. (2003) Proc. Natl. Acad. Sci. USA 100, 3617–3622. 21. Rodriguez, J. A., Valentine, O. S., Eggers, D. K., Roe, J. A., Tiwari, A., Brown, R. H. & Hayward, L. J. (2002). J. Biol. Chem. 277, 15932–15937. 22. Tiwari, A. & Hayward, L. J. (2003) J. Biol. Chem. 278, 5984–5992. 23. Takamiya, R., Takahashi, M., Myint, T., Park, Y. S., Miyazawa, N., Endo, T., Fujiwara, N., Sakiyama, H., Misonou, Y., Miyamoto, Y., et al. (2003) FASEB J. 17, U452–U469. 24. Wilson, M. A., Collins, J. L., Hod, Y., Ringe, D. & Petsko, G. A. (2003) Proc. Natl. Acad. Sci. USA 100, 9256–9261. 25. Vagin, A. A. & Teplyakov, A. J. (1997) J. Appl. Crystallogr. 30, 1022–1025. 26. Murshudov, G. N., Vagin, A. A. & Dodson, E. J. (1997) Acta Crystallogr. D 53, 240–255. 27. Jones, T. A., Zou, J. Y., Cowan, S. W. & Kjeldgaard, M. (1991) Acta Crystallogr. A 47, 110–119. 28. Lamzin, V. S. & Wilson, K. S. (1993) Acta Crystallogr. D 49, 129–149. 29. Winn, M. D., Isupov, M. N. & Murshudov, G. N. (2001) Acta Crystallogr. D 57, 122–133. 30. Towns-Andrews, E., Berry, A., Bordas, J., Mant, G. R., Murray, K., Roberts, K., Sumner, I., Worgan, J. S. & Lewis, R. (1989) Rev. Sci. Instrum. 60, 2346–2349. 31. Hasnain, S. S., Murphy, L. M., Strange, R. W., Grossmann, J. G., Clarke, A. R., Jackson, G. S. & Collinge, J. (2001) J. Mol. Biol. 311, 467–473. 32. Semenyuk, A. V. & Svergun, D. I. (1991) J. Appl. Crystallogr. 24, 537–540. 33. Svergun, D. I. (1999) Biophys. J. 76, 2879–2886. 34. Deng, H. X., Hentati, A., Iqbal, Z., Cayabyab, A., Hung, W. Y., Getzoff, E. D., Hu, P., Herzfeldt, B., Rood, R. P., Warner, C., et al. (1993) Science 261, 1047–1051. 35. Elam, J. S., Taylor, A. B., Strange, R., Antonyuk, S. V., Doucette, P. A., Rodriguez, J. A., Hasnain, S. S., Hayward, L. J., Valentine, J. S., Yeates, T. O. & Hart, P. J. (2003) Nat. Struct. Biol. 10, 461–467. 36. Cardoso, R. M. F., Thayer, M. M., DiDonato, M., Lo, T. P., Bruns, C. K., Getzoff, E. D. & Tainer, J. A. (2002) J. Mol. Biol. 324, 247–256. 37. Hand, C. K. & Rouleau, G. A. (2002) Muscle Nerve 25, 135–159. 38. Wiedau-Pazos, M., Goto, J. J., Rabizadeh, S., Gralla, E. B., Roe, J. A., Lee, M. K., Valentine, J. S. & Bredesen, D. E. (1996) Science 271, 515–518. 39. Yim, H. S., Kang, J. H., Chock, P. B., Stadtman, E. R. & Yim, M. B. (1997) J. Biol. Chem. 272, 8861–8863. 40. Valentine, J. S. (2002) Free Radical Biol. Med. 33, 1314–1320. 41. Goto, J. J., Zhu, H. J., Sanchez, R. J., Nersissian, A. M., Gralla, E. D. & Valentine, J. S. (2000) J. Biol. Chem. 275, 1007–1014. 42. Elam, J. S., Malek, K., Rodriguez, J. A., Doucette, P. A., Taylor, A. P., Hayward, L. J., Cabelli, D. E., Valentine, J. S. & Hart. P. J. (2003) J. Biol. Chem. 278, 21032–21039.
Hough et al.
PNAS 兩 April 20, 2004 兩 vol. 101 兩 no. 16 兩 5981
SEE COMMENTARY
scattering data indicate the presence of a more pronounced structural change in solution than in the crystals, suggesting that crystals have preferentially selected a population of molecules. These conformational changes are analogous to those observed in the solution scattering profile of bovine CuZnSOD on heating from 4°C to 80°C, advocating strongly that the mutants are significantly destabilized in comparison with human wtSOD. This destabilization of the dimeric interface may result in an increased tendency to unfold or lose metals in vivo. This observation is of significant importance in view of the recent proposal that a destabilization of the dimeric interface occurs in one of the Parkinson’s disease-causing mutants of DJ-1.
BIOPHYSICS
toxicity with that of the metal-deficient species. One possibility may be that the enhanced peroxidase activity and兾or monomermonomer reorientation observed in metallated A4V and I113T contributes to demetallation of the mutant protein in vivo. It is noteworthy that destabilization of a dimeric interface in the L166P-recessive mutation of the DJ-1 protein, associated with early-onset Parkinson’s disease, has also been postulated to be important to the mechanism of neurodegeneration in that disorder. In the case of DJ-1, the protein was rendered drastically less stable as a result of the mutation, leading to much lower protein levels and a subsequent loss-of-function disorder.
