BNL-108388-2015-JA

The structure of the Caenorhabditis elegans

MnSOD-3-Azide complex

Gary J. Hunter1, Chi H. Trinh2, Rosalin Bonetta1, Emma E. Stewart2, Diane E.

Cabelli3, Therese Hunter1*

1 Department of Physiology and Biochemistry, Faculty of Medicine and Surgery,

University of Malta, Malta

2 Astbury Centre for Structural Molecular Biology, Institute of Molecular and Cellular

Biology, University of Leeds, Leeds, United Kingdom

3 Chemistry Department, Brookhaven National Laboratory, Upton, New York, United

States

*Correspondence to : Therese Hunter, Department of Physiology and Biochemistry,

Faculty of Medicine and Surgery, University of Malta, Msida, MSD 2080, Malta.

Tel: +356 2340 2916. Email: [email protected]

Running title: C.elegans MnSOD azide complex Abstract: C.elegans MnSOD-3 has been implicated in the longevity pathway and its mechanism of catalysis is relevant to the aging process and carcinogenesis. The structure of MnSOD-3 determined at room temperature provides crystallographic

evidence for the first time of a dynamic region of helix (residues 41 to 54) that plays

roles in structural integrity through the tetrameric interface and in substrate steering.

We have also determined the structure of the MnSOD-3-azide complex to 1.77 Å

resolution. Analysis of this complex shows that the substrate analog, azide, binds end-

on to the manganese center as a sixth ligand and that it ligates directly to a third and

new solvent molecule also positioned within interacting distance to the His30 and

Tyr34 residues of the substrate access funnel. This is the first structure of a eukaryotic

MnSOD-azide complex that demonstrates the extended, uninterrupted hydrogen-

bonded network that forms a proton relay incorporating three outer sphere solvent

molecules, the substrate analog, the gateway residues, Gln142, and the solvent ligand.

This configuration supports the formation and release of the hydrogen

product in agreement with the 5-6-5 catalytic mechanism for MnSOD. The high

product dissociation constant k4 of MnSOD-3 reflects low product inhibition making this efficient even at high levels of superoxide.

Keywords: Superoxide dismutase; conformational variation; MnSOD-3-azide complex; catalytic mechanism; product inhibition

Introduction

During , the molecular in the cell can readily acquire an

1 electron to form superoxide radicals O2- . These are rapidly converted to other

(ROS) such as the hydroxyl . Unless removed, these

ROS will damage cellular macromolecules2. Both prokaryotic and eukaryotic cells

produce superoxide dismutases (SODs, EC 1.15.1.1), a family of metalloenzymes that

eliminate superoxide radicals by dismutation to oxygen and hydrogen peroxide3,4.

This reaction involves the cycling of the active site metal cofactor that may be manganese, iron, nickel or copper5-7 (Scheme 1).

Scheme 1. Dismutation by SOD

3+ - 2+ M + O2 ⟶ M + O2

2+ - + 3+ M + O2 + 2H ⟶ M + H2O2

The free-living nematode, Caenorhabditis elegans, has five distinct sod genes, which

express two cytosolic Cu/ZnSODs (SOD-1 and SOD-5), an alternatively spliced extracellular Cu/ZnSOD (SOD-4) and two mitochondrial MnSODs (SOD-2 and SOD-

3)8-12. The MnSODs are both synthesized as precursor molecules with a N-terminal mitochondrial-targeting sequence and are assembled into homotetramers containing one manganese per subunit, a process that requires interaction with hsp60/1013-15.

Sod-2 is constitutively expressed while the sod-3 gene is inducible. The latter is

expressed only in the dauer diapause stage and is a direct target of the DAF-16/FOXO

transcription factor16,17. Interest in sod-3 grew when its over-expression was observed in the long-lived daf-2 and age-1 mutants. Both these mutants exhibit the Age

phenotype16,18-23 and have altered expression of certain components of the

insulin/IGF-1 signaling pathway24-27. The precise role of MnSOD-3 in life extension

has been a topic of debate. As superoxide and both act as signaling

molecules28,29, it is likely that SOD-3, possibly together with SOD-2, serves as a modulator of longevity and other cellular pathways30-36, by fine-tuning the cellular

levels of superoxide and hydrogen peroxide.

Catalysis is a cyclic two-stage process, with the manganese ion switching from the

oxidised MnIII to the reduced MnII state in a ‘ping-pong’ mechanism37,38. At high superoxide concentrations, the human MnSOD undergoes an initial burst of catalysis that lasts only a few milliseconds followed by a region of zero order decay of superoxide indicating product inhibition39. This inhibition is caused by the formation of the peroxy complex of the manganese ion40,41. The McAdam scheme37 describes

the MnSOD catalytic cycle in four reactions that show the dismutation of superoxide

occurring through two simultaneous pathways (Scheme 2). Equation 1 and 2 represent

the outer-sphere pathway whereby the reduction of superoxide to hydrogen peroxide

is instantaneous. The second or inner-sphere pathway shown in equations 3 and 4

involves the formation and dissociation of the product-inhibited complex7. The degree

of product inhibition is determined by the gating ratio (k2/k3) which gives a measure

of the amount of superoxide that is reduced by the outer-sphere pathway relative to the inner-sphere pathway and k4, the measure of dissociation of the product inhibited

complex39,42.

Scheme 2. Mechanism of catalysis of MnSOD + k1 H

3+ - - 2+ Mn SOD(OH ) + O2 Mn SOD(H2O) + O2 (Eq.1)

+ k2 H

2+ - 3+ - Mn SOD(H2O)+ O2 Mn SOD(OH ) + H2O2 (Eq.2)

k 3

2+ - 3+ 2- Mn SOD(H2O)+ O2 Mn SOD(H2O)(OO ) (Eq.3)

+ k4 H

3+ 2- 3+ - Mn SOD(H2O)(OO ) Mn SOD(OH ) + H2O2 (Eq 4)

To date there are ten known structures of native homotetrameric MnSOD13,43-47. In the active sites of each, the manganese cofactor is coordinated to three histidinyl, one aspartyl and either a hydroxyl or water molecule, forming a trigonal bipyramidal geometry that is highly conserved. A hydrogen bond network, involving the second sphere residues of the active site, extends to include residues from a neighboring subunit to form a dimer interface13,43,47,48. This network permits the proton transfer needed to reduce superoxide to hydrogen peroxide. In order to have a clearer understanding of how this enzyme functions, we have determined the first molecular structure of a eukaryotic MnSOD in complex with its substrate analog. The structure of the MnSOD-3-azide complex suggests how the superoxide is positioned in the active site and helps elucidate the catalytic mechanism. This structure is a good model for MnSOD structure/function studies as the only other structural data of the MnSOD- azide complex is that of the T.thermophilus which has a different tetrameric assembly to the known eukaryotic MnSODs49.

Results

Structural data of 293 K MnSOD-3 and MnSOD-3-azide complex

Structural data of the 293 K MnSOD-3 was determined at a resolution of 1.77 as was the 100 K MnSOD-3-azide complex. Both formed pink crystals belonging to the space group P41212, with unit cell parameters of a=b=81.5,c=138.0 Å (293 K

MnSOD-3)13 and a=b=82.1,c=138.0 (MnSOD-3-azide complex). The asymmetric unit-cell consists of two subunits designated A and C. Overall the tetrameric structure is very similar to that of the human MnSOD forming a dimer of dimers and shares the same monomeric fold of mononuclear SODs, having a predominantly α-helical hairpin N-domain and an α/β C-domain. Superimposition of the CA atoms of

MnSOD-3 with human MnSOD gives a rmsd value of 0.8 Å50. The data collection

and refinement statistics for the 293 K MnSOD-3 and 100 K MnSOD-3-azide

complex are presented in Table 1.

Table 1. Data collection, processing and refinement statistics for MnSOD-3 293 K

and MnSOD-3-azide complex.

MnSOD-3 293 K MnSOD-3-azide 100 K

Source Rigaku RUH3R Daresbury SRS 10.1 rotating anode Wavelength (Å) 1.541 0.97 Resolution range (Å) * 70.19–1.77 (1.87– 32.4–1.77 (1.86– 1.77) 1.77)

Space group P41212 P41212 Unit-cell parameters (Å) a=b=81.5,c=138.0 a=b=82.1,c=138.0 No. of observed reflections 190572 654522 No. of unique reflections 45989 46922 Redundancy 4.1 (4.0) 13.9 (4.0) Completeness (%) * 99.9 (99.9) 99.9 (99.9) < I/σ(I) >* 15.6 (2.3) 16.7 (13.3)

Rmerge (%)§* 6.2 (33.8) 9.0 (51.7)

Rpim (%)¥* 3.4 (19.4) 2.5 (14.5) Resolution range for refinement (Å) 70.19–1.77 30.6-1.77 R factor (%) 18.1 18.7

Rfree (%)† 21.4 21.6 No. of protein non-H atoms 3427 3178 No. of water molecules 296 384 No. of manganese 2 2 No. of sulfate ions 1 1 No. of malonate-ion 1 1 No. of acetate ion - 1 No. of azide - 2 R.m.s.d bond lengths (Å) ξ 0.013 0.011 R.m.s.d bond angles (˚) ξ 1.5 1.4 Average overall B factor (Å2) Protein 26 31 Water 30 36 Manganese ions 16 19 Sulfate ions 74 48 Malonate-ion 32 48 Acetate ion - 49 Azide - 16 Ramachandran analysis, the percentage of residues in the regions of plot (%) ‡

Favored region 97.4 96.9 Outliers 0 0 PDB code 4X9Q 5AG2

*Values given in parentheses correspond to those in the outermost shell of the resolution range.

§ Rmerge = ∑hkl ∑i|Ii(hkl)-[I(hkl)]|/ ∑hkl ∑ Ii(hkl)

421/2 ¥ Rpim = ∑hkl {1/[N(hkl)-1]} ∑i|Ii(hkl) – [I(hkl)]|/ ∑hkl ∑i Ii(hkl)

† Rfree was calculated with 5% of the reflections set aside randomly. ξ Based on the ideal geometry values of Engh and Huber51. ‡ Ramachandran analysis using the program MolProbity52.

Structural mobility of the N-domain α-hairpins

The interaction between the four N-domain α-hairpins forms the tetrameric quaternary structure consisting of two four-helix bundles as seen in other eukaryotic

MnSODs (Fig. 1A). Interestingly, structural mobility is visible in the electron densities of both the 100 K and 293 K MnSOD-3 structures as well as the MnSOD-3- azide complex. There are two conformations (designated A and B) of the N-domain

α - helical region composed of the 14 residues from Ile41 to Leu54, in a region that forms symmetrically apposed helices of the tetrameric α-hairpin interface (Fig.1B).

For clarity however, we have modeled only one conformer for each structure submitted to PDB. The 100 K MnSOD-3 and the MnSOD-3-azide structures were modeled with conformer A and the 293 K structure with conformer B.

Figure. 1. The structure of MnSOD-3 from C. elegans. A. The tetrameric assembly of

MnSOD-3 with subunits labeled A (blue), B (yellow), C (green) and D (orange).

Manganese atoms are shown as purple spheres and indicate the positions of the active sites. Subunits A and B form a biologically active dimer as do C and D. The tetrameric interface forms mostly between the hairpin helical segments seen in the center of the figure. Subunits A and C illustrate one of the two four-helix bundles in this interface (blue and green subunits). Red side chains shown on subunit A highlight

the position of the alternate conformer B. B, the tetrameric interface between subunits

A (blue) and C (green) illustrating the position of conformer B (red). Side chains

indicative of those amino acids where interactions may be enhanced due to the

position of conformer B are labeled. C. The salt bridges unique to MnSOD-3 formed

between subunits A and D (illustrated) and B and C. All figures were drawn using

PyMOL53. Protein illustrations were prepared from PDB entry 3DC5 with conformer

B from PDB entry 4X9Q. Of significance is the increase in the number of possible hydrophobic interactions

between subunits A and C (which form the tetramer) of conformer B (Fig. 1B). 19F-

tyrosine NMR and amide H/D exchange studies of the human MnSOD have been

reported showing a high degree of structural dynamic mobility in the same region;

between residues Val40 and Ile5854. This suggests that the conformational variability

observed in all MnSOD-3 structures is not an artifact of crystal packing or the

temperature at the time of data collection. The two 14-residue conformers (A and B)

were resolved with a rmsd value of 2.0 Å when superimposed, diverging by more

than 3.5 Å at Lys51CA (Fig. 1B). Superimposition with human MnSOD indicates that

the position this region adopts in the crystal structure of the human MnSOD

approaches that of conformer B. In , the tetrameric interface has been shown to

have a direct influence on the dimer interface that forms the catalytically active site55.

Unique to MnSOD-3 are salt bridges that form at the cross chain interface A/D (and

B/C) between Arg109 in the first β-sheet of one subunit and Asp131 in the sixth α- helix of the opposite subunit of the tetramer (2.9 to 3.1 Å) (Fig. 2C). Human MnSOD has a unique A/D and B/C bond between Arg132NH1 and Leu135O where this arginine residue is in the equivalent position to Asp131 of MnSOD-3.

Active site geometry of MnSOD-3

The inner sphere arrangement of MnSOD-3 is, as expected, the characteristic five coordinate trigonal bipyramidal configuration. In the active site, the manganese ion is coordinated to five ligands that include residues from both N and C-domains of the subunit: His26 and His74 from the N-terminal hairpin helices, Asp155 and His159

from the C-domain, and a solvent molecule (designated WAT1). The latter may be

either a water or hydroxyl molecule.

Figure 2. Stereo representation of the structural alignment between MnSOD-3 (green) and human MnSOD (pink). Metal ligands are His26, His74, His159 and Asp155

(MnSOD-3 numbering). Major components of the hydrogen-bonded network are shown, although the structure of Tyr162 has been omitted for brevity. WAT2 and

WAT3 (red) in MnSOD-3 replace a single molecule (pink) in human MnSOD and bridge the gap between the gateway residues His30 and Tyr34. The position of WAT1

(the solvent ligand to the metal) is slightly closer to the gateway residues in MnSOD-

3 (red sphere). The metal ion is illustrated as a sphere (purple, MnSOD-3, red, human

MnSOD). Bonding is illustrated by black dashes (MnSOD-3, 3DC5; human MnSOD,

1LUV).

Overall the active site residues including the gateway residues Tyr34 and His30 positioned at the top of the solvent access funnel of MnSOD-3 and human MnSOD superimpose well with only negligible deviation, with a Tyr34OH to His30ND1 distance of 5.4 Å (Fig. 2). Despite this similarity however, a noteworthy observation is the presence of two water molecules in both 100 K and 293 K MnSOD-3

(designated WAT2 and WAT3) that participate in the hydrogen bonding between

Tyr34 and His30 (Fig. 2). The bond lengths of Tyr34OH-WAT2-WAT3-His30NE2 are 2.7 Å, 2.9 Å and 2.8 Å respectively. No hydrogen-bonded network between Tyr34

and His30 involving only one of the water molecules is feasible. This two-water

arrangement is present in both the A and C subunits of the asymmetric unit of the 100

K and the 293 K MnSOD-3 structures (3DC5 and 4X9Q) but has not been observed in any other native MnSOD structures.

Visible absorption spectra and azide inhibition

The as-isolated MnSOD-3 was predominantly in the Mn3+ oxidized state giving the

characteristic 480 nm peak in the visible spectrum (Fig. 3). Incubation with potassium

permanganate showed a small increase in absorbance at this wavelength with an

accompanying increase at lower wavelengths (400 nm), while the presence of DTT

abolished the 480 nm peak as the manganese was reduced to the Mn2+ state. The

formation of the MnSOD-3-azide complex caused a shift in absorption from 480 to

410 nm that is indicative of a six-coordinate complex56. The occurrence of a six-

coordinate configuration was verified by the structural data of the MnSOD-3-azide

complex. The effectiveness of azide as a competitive inhibitor was demonstrated by

the loss of 45% and 61% SOD activity when the enzyme was assayed in the presence

of 4 mM and 10 mM sodium azide respectively. Without azide the enzyme activity

was 3261 ± 26 U/mg/Mn.

Figure. 3. Visible absorbance spectra of MnSOD-3 in potassium phosphate buffer (pH

7.8) in oxidized form (top, dashed line), as isolated (dotted line) or reduced (bottom, solid black line). Also shown is the blue shift observed on binding the inhibitory molecule, azide (grey).

Structure of MnSOD-3-azide complex

The crystal structure of the MnSOD-3-azide complex (PDB entry 5AG2) shows that in both chain A and chain C the azide binds directly to the manganese to form a six- coordinate distorted octahedral active site. This is supported by previous spectroscopic work on other SODs56-59. In this protein, the azide binds to the

manganese ion in an end-on manner (Fig. 4B), in an arrangement that is similar to that

observed in the E.coli and P. haloplanktis FeSOD azide complexes (PDB entries

1ISC and 3LJ9 respectively)49,60 and supported by theoretical studies61,62. In MnSOD-

3, azide is accommodated into the active site with relatively small perturbations of

surrounding residues and has an average B factor value of 16 Å2 in both chain A and

chain B active sites. The azide lies directly opposite the Asp155OD1 ligand and in

line with the phenolic ring of Tyr34 in the cavity formed by the metal ion, the

imidazole ring of His74, the backbone of His30 and the NE2 of Gln142. In subunit A,

N1 of the azide is positioned end-on 2.8 Å from the manganese ion at an angle of 94˚

(Mn-N1-N3) in comparison to the 117˚ and 123˚ found in the dimeric E.coli and P.

haloplanktis FeSODs respectively49,60.

In the MnSOD-3-azide complex the manganese ion and solvent ligand WAT1 are

displaced slightly towards azide N1. This places the solvent ligand 3.5 Å away from

the azide N1. The distance however, between the solvent ligand and the manganese

ion remains unchanged at 2.3 Å. The His74-Mn-His159 angle increases by 21˚ to

145˚ and both His74NE2 and His159NE2 are within bonding distances to the azide N3 (3.0 Å) and azide N1 (2.8 Å) respectively. The most perturbed residues are the

outer sphere Tyr34 and His30. The phenolic group of Tyr34 that lies adjacent to N1

of the azide moves away by 0.5 Å and is 5.7 Å from the manganese ion. Positioned at

3.7 Å away from N1, the Tyr34OH is too far from the azide to interact effectively.

Twisting of the imidazole ring of His30 creates a difference in the torsion angle of 5˚.

Consequently the Tyr34OH-Mn-His30ND1 angle widens from 56˚ to 61˚. The

hydrogen bonds between the two water molecules and Tyr34 and His30 at the top of the solvent access funnel change only slightly, with the Tyr34-WAT2 and WAT2-

WAT3 remaining unchanged and the bond between WAT3 and His30 stretching slightly to 3.0 Å (from 2.8 Å). The major difference in the arrangement of the water molecules at the solvent funnel access is the appearance of a third and new water molecule (designated WAT4) identified in a position where it can hydrogen bond directly with Tyr34 (2.9 Å), His30 (2.9 Å), WAT3 (2.3 Å) and azide N1 (2.5 Å) (Fig.

4). Tyr34 and His30 are therefore connected to each other by two water molecules

WAT2 and WAT3 as seen in the MnSOD-3 structure (PDB entry 3DC5 and 4X9Q) and to the azide (and therefore the peroxy adduct) via WAT4. An identical arrangement is also observed in chain C. This configuration has not been previously

reported in the literature and lends support to a 5-6-5 catalytic mechanism47,49,63. The

30% occupancy of the azide shown by the electron density map compares well with

the occupancy achieved by Lah et al for T. thermophilus MnSOD, the only other

MnSOD-azide structure known49. Azide soaking gave better diffraction data

compared to co-crystallization of MnSOD-3 with azide and was therefore the preferred method.

Figure. 4. Stereo figure of the active site of native MnSOD-3 (A) and MnSOD-3- azide complex (B). The four amino acid ligands to the manganese atom are shown with thin bonds while those involved in the hydrogen-bonded network are shown with thicker bonds. The Mn metal (purple) and water molecules (red) are represented by spheres and the azide in (B) is blue; Glu158 and Tyr162 (grey) are contributed by the dimeric partner, subunit B. Bonding is shown as black.

Kinetic studies by pulse radiolysis

A solution of 32.6 μM MnSOD-3 (100% metallated) pH 7.5 was pulse irradiated such that 2 or 4 μM of superoxide was formed and the change in absorbance with time was followed at 350-650 nm. The formation of the initial spectrum formed within 100 microseconds (Fig. 5A closed circles). This was followed by the first order conversion of the initial spectrum into the final species (Fig. 5A open circles) at a rate

of 300 s-1. This final spectrum is that of the oxidized Mn3+SOD. Additionally, the

- 3+ decay of 50 μM superoxide using 6:1 and 13:1 ratios of O2 :Mn SOD was followed

at 260 nm to determine how the enzyme behaves at relatively low and high levels of

superoxide (Figure 5B). The disappearance of superoxide by MnSOD-3 as a function

9 -1 -1 of time (1.2 x 10 M s ) is comparable to that observed for human MnSOD (human

9 -1 -1 8 -1 - MnSOD, 1.5 x 10 M s ). The gating ratio k2/k3 of 1 (both k2 and k3 are 5 x 10 M s

1) for MnSOD-3 is again similar to that of human MnSOD, meaning that the amount of superoxide reduced by the outer-sphere pathway is equivalent to that reduced by the inner-sphere pathway. The two however, differ in the rate of

dissociation of the Mn-peroxy complex. In MnSOD-3 this protonation off occurs

-1 more rapidly (k4 of 300 s ) compared to that observed for the product inhibited

- 42 human MnSOD (k4 of 120 s ) . As the catalytic rates of superoxide dismutation are

close to the limit of the rate of diffusion, a 2.5 fold increase in a kinetic constant has a

considerable effect on the catalysis of the enzyme.

Figure. 5. A. The Mn3+ SOD-3 spectrum obtained from pulse radiolysis. The sample

contained 32.6 µM manganese (protein concentration 0.75 mg/mL) in potassium phosphate buffer, pH 7.5. The closed circles represent the spectrum formed at the rate

constant of 1.2 x 109 M-1s-1 when 32.6 µM Mn2+SOD reacts with either 2 or 4 µM

superoxide. The open circles show the spectrum formed at a rate of 300 s-1 and is the

spectrum of Mn3+SOD. B. Observed disappearance of 50 µM superoxide in the presence of 8.15 µM or 3.26 µM MnSOD (black and blue traces respectively). Both

traces were fitted to the McAdam scheme to derive the rate constants k1 to k4.

Discussion

Evidence is emerging that mitochondrial MnSOD is not just an but also a

tumor suppressor protein as it affects the levels of superoxide and hydrogen peroxide

in the cell64. We show three structural features that have not been previously observed

in native MnSODs. The first is evidence in the crystal structure of conformational

variation (designated as conformer A and conformer B, Fig. 1) of the α-helical segment of the N-terminal hairpin that forms the major component of the tetrameric interface and leads to the substrate access funnel. This was observed in the 100 K

MnSOD-3, the 293 K MnSOD-3 and in the MnSOD-3-azide complex showing that this protein is a dynamic, breathing molecule and may serve as a model for the study of molecular dynamics of MnSODs in general. The flexibility of this segment of the

α-hairpin may therefore support both stability and catalysis. Firstly, hydrophobic interactions between chain A and chain C are increased in conformer B (Fig. 1B).

This is a direct enhancement of the stability of the tetrameric interface. Secondly, as superoxide anions are steered towards the solvent access funnel, in part, along the length of the α-hairpin (Fig. 1A), this effect is likely to be more pronounced in conformer A. Movement away from the B conformation would serve to increase this.

Therefore the deduced flexibility of the 14-residue segment near the turn of the α- hairpins may contribute to enhance catalysis through stabilization of the quaternary structure and increased superoxide flux towards the active site.

The second feature is the presence of two and not one, second sphere water molecules involved in the vital network between Tyr34 and His30, designated

Tyr34OH-WAT2-WAT3-His30NE2. This arrangement occurs in the two subunits of the asymmetrical unit in both 100 K and 293 K MnSOD-3 structures. Furthermore

hydrogen bonding between Tyr34 and His30 is not feasible using a single water

molecule. As the catalytic reaction relies on protonation, the number and arrangement

of the water molecules in the proton relay pathway leading to the manganese center is

of significance. All other MnSODs described to date have only one water molecule

linking these gateway residues.

And finally we present the structure of the MnSOD-3-azide complex, where the azide

is end-on to the manganese center. Despite a low occupancy of 0.3, the azide is

observed in the same arrangement in both subunits of the asymmetrical unit and is not

bound non-specifically to other parts of the structure. In addition an average azide B

factor of 16 indicates that the azide molecules present are firmly bound in the active

site. In T. thermophilus, azide was shown to be bound side-on to the manganese and

within bonding distance to Tyr34 (PDB entry 1MNG). Given the similarity of human

MnSOD to C.elegans MnSOD-3, the MnSOD-3-azide structure is probably a better

model to use for the study of substrate binding to human MnSOD than the

T.thermophilus-azide complex. Due to the absence of structural data of eukaryotic

MnSOD-azide complexes, previous studies have relied on superimposition of eukaryotic MnSODs with the T.thermophilus MnSOD-azide complex and postulated hypotheses regarding the mechanism of catalysis derived from the position of the substrate complex and the arrangement of the active site solvent molecules62,65. In an arrangement observed for the first time in MnSOD, the azide in MnSOD-3 is

linked directly to the Tyr34 hydrogen-bond network through an additional water molecule (WAT4) introduced at this strategic point in the proton relay system (Fig.

3B). This creates an original bonding network directly between the gateway residues, the substrate analog and the metal center of the enzyme Although it is not certain whether all three solvent funnel access water molecules (WAT2, WAT3, WAT4) occur in the structure simultaneously, it is likely that the network involving the gateway residues and the three waters does in fact form during catalysis. We propose that this provides a constant and rapid source of protons directly to the peroxy-adduct intermediate thereby facilitating the release of hydrogen peroxide.

Mutations of Tyr34 have been the subject of many studies to investigate whether the phenolic OH is the donor of the hydrogen needed for protonation off to release the peroxy adduct from the active site during product turnover62,63,66-68. In human

MnSOD mutations Y34A, Y34N and Y34V, the rate of hydrogen product dissociation

was faster than in the wild type human MnSOD42. In these cases a water molecule in a

similar position to that observed for WAT4 can be seen in one subunit of each

asymmetrical dimer (PDB entry 1ZSP, 2P4K and 1ZUQ respectively). As few

structures of SOD-azide complexes are currently available, the coordination and

position of the substrate has been a point of contention for many years whereby even

the crystal structure itself has been questioned. Consequently MnSOD-3 is an elegant

model to study the mechanism of catalysis and product inhibition particularly in view

of the tumor suppressor property of MnSOD .

Materials and Methods

Protein expression and purification

The E. coli strain OX326A (∆sodA, ∆sodB)69 was used throughout for protein

expression. The coding region of the mature MnSOD-3 was expressed using the

pTrc99A expression system and MnSOD-3 was purified as previously described13,70.

Protein concentrations were determined by optical absorption at 280 nm using an extinction coefficient of 43340 M-1cm-1 and the subunit molecular weight of MnSOD-

3. Measurement of metal content by inductively-coupled plasma-mass spectrometry

(ICP-MS) was performed by ALS Analytica AB, Sweden.

Protein characterization

Optical absorption spectra of MnSOD-3 (3 mg/mL) were collected using a Beckman

DU7500 diode array spectrophotometer equipped with 50 µL micro cells. The as-

isolated MnSOD-3 was oxidized by incubation with potassium permanganate (0.1

mM, 0.66:1 ratio with MnSOD-3) and reduced by 50 mM DTT. The MnSOD-3-azide

complex was formed using 100 mM azide.

Superoxide dismutase assays were performed essentially as described by McCord and

Fridovich3 and Ysebaert-Vanneste and Vanneste71 using as detector and

as superoxide generator. The inhibitory effect of azide on SOD

activity was determined by adding 4 mM or 10 mM sodium azide in 10 mM potassium phosphate buffer pH 7.8 to the SOD assay mixtures before the addition of xanthine oxidase66.

Crystallization, data collection, structure solution and refinements

Crystallization, data collection, structure solution and refinements for MnSOD-3 at

100 K (PDB accession code 3DC5) have been described previously13. Although some

preliminary data for the 293 K MnSOD-3 was previously published13, the structure

was not deposited in the Protein Data Bank (PDB). Further rounds of refinement were

performed to 293 K MnSOD-3 prior to depositing in the PDB. Crystals for the

MnSOD-3-azide complex were grown using the hanging-drop vapor diffusion method

with drops consisting of 2 µl of the MnSOD-3 (100% metallated, confirmed by ICP-

MS) and 2 µl of the reservoir liquor. Optimal protein crystallization was achieved

using 0.1 M bicine pH 9.2 and 2.7 M ammonium sulfate for MnSOD-3 (8 mg/ml).

The MnSOD-3-azide complex was obtained by soaking a crystal in 100 mM sodium

azide overnight. The crystal was flash-frozen in 50% (v/v) 3.4 M sodium

malonate/100 mM azide solution and a data set was recorded at 100 K at Daresbury

SRS (Synchroton Radiation Source) station 10.1 on an ADSC Quantum charge-

coupled device (CCD) detector)72. Data reduction and subsequent calculations were

carried out using the CCP4 program suite. The diffraction data was integrated using

MOSFLM73 and scaled with SCALA74. Five percent of the data (the same set used for

both MnSOD-3at 100 K and 293 K) was excluded from the refinement to constitute

75 the Rfree set using FREERFLAG . The MnSOD-3-azide complex was determined using the MnSOD-3 structure as a starting model for the refinement in REFMAC576.

Ten cycles of rigid body refinement (resolution range 12.0-3.0 Å) were followed by ten cycles of restrained refinement (whole resolution range). Reiterated model building was performed using the program Coot77 and water molecules were added to

the Fo-Fc map peaks greater than 3.0 r.m.s.d. The azide ions for chains A and C were added after placement of all the water molecules of the active site using 2Fo-Fc and Fo-Fc maps. TLS and restrained refinement along with occupancy refinement of the azide ions were carried out using REFMAC576. The geometry of the final model was

checked using MOLPROBITY52. The final structure of the MnSOD-3-azide complex

was refined to R=18.7% and Rfree=21.6%.

Pulse radiolysis

The pulse radiolysis experiments were performed using a 2 MeV Van de Graaff

accelerator at Brookhaven National Laboratory as described78. In brief, superoxide

anions were produced in situ in an air-saturated aqueous solution of 30 mM sodium

formate (as a scavenger). All UV/visible spectra were recorded on a

Cary 210 spectrophotometer with a path length of 2.0 or 6.1 cm, at a constant

temperature of 25˚C. The catalytic rates were measured by two methods4. In the first

44 method hydrogen peroxide (2:1 [H2O2]:[MnSOD-3]) was used to reduce the enzyme

2+ to the Mn state. This reduced form of the enzyme was then oxidized (32.6 µM in

potassium phosphate pH 7.5, 30 mM sodium formate and 10 µL EDTA) with substoichiometric amounts of superoxide. The respective appearance/disappearance of Mn3+SOD was followed by recording the change in absorbance in the visible range

(350-650 nm) of the spectrum. Mn3+SOD exhibits an absorbance band at 480 nm. In

- the second method the decay of 50 mM superoxide using 6:1 and a 13:1 ratio of O2

:Mn3+SOD was followed at 260 nm. The rate constants were calculated by fitting the

data obtained to the mechanism given in Scheme 2 using the Chemical Kinetics

program in PRWIN79 as described previously78,80, taking into consideration the

manganese concentration measured by ICP-MS (rather than the protein concentration)

assuming that all the manganese is specifically bound at the active site.

Atomic coordinates The atomic coordinates and structural factors for 100 K MnSOD-3, 293 K MnSOD-3

and MnSOD-3-azide complex have been deposited in the Protein Data Bank under the

accession code 4X9Q and 5AG2 respectively.

Acknowledgments

This study was supported by a grant (PHBRP02) awarded to TH by the University of

Malta and funding by the Dean’s Research Award to TH and RB (MDSIN08-19 and

MEDIN08-01 respectively). The pulse radiolysis studies are carried out at the

Accelerator Centre for Energy Research, supported by the U.S. Department of

Energy, Office of Science, Office of Basic Energy Sciences, Division of Chemical

Sciences, Geosciences, and Biosciences, under Contract DE-AC02-98CH10886.

References

1. Harman D (1956) Aging: a theory based on free radical and radiation chemistry. J Gerontol 11:298-300. 2. Halliwell B, Gutteridge JMC (1985) The importance of free radicals and catalytic metal ions in human diseases. Mol. Aspects Med. 8:89-124. 3. McCord JM, Fridovich I (1969) Superoxide dismutase. An enzymic function for erythrocuprein (hemocuprein). J. Biol. Chem. 244:6049-55. 4. Sheng Y, Abreu IA, Cabelli DE, Maroney MJ, Miller AF, Teixeira M, Valentine JS (2014) Superoxide dismutases and superoxide reductases. Chem Rev 114:3854-918. 5. Fridovich I (1995) Superoxide radical and superoxide dismutases. Annu Rev Biochem 64:97-112. 6. Youn HD, Kim EJ, Roe JH, Hah YC, Kang SO (1996) A novel nickel- containing superoxide dismutase from Streptomyces spp. Biochem J 318 889-96. 7. Abreu IA, Cabelli DE (2010) Superoxide dismutases-a review of the metal- associated mechanistic variations. Biochim Biophys Acta 1804:263-74. 8. Giglio AM, Hunter T, Bannister JV, Bannister WH, Hunter GJ (1994) The /zinc superoxide dismutase gene of Caenorhabditis elegans. Biochem Mol Biol Int 33:41-4. 9. Larsen P (1993) Aging and Resistence to oxidative damage in Caenorhabditis elegans. PNAS 90:8905-8909. 10. Giglio MP, Hunter T, Bannister JV, Bannister WH, Hunter GJ (1994) The manganese superoxide dismutase gene of Caenorhabditis elegans. Biochem Mol Biol Int 33:37-40. 11. Suzuki N, Inokuma K, Yasuda K, Ishii N (1996) Cloning, Sequencing and Mapping of a Manganese Superoxide Dismutase Gene of the Nematode Caenorhabditis elegans. DNA Research 3:171-174. 12. Fujii M, Ishii N, Joguchi A, Yasuda K, Ayusawa D (1998) A novel superoxide dismutase gene encoding membrane-bound and extracellular isoforms by alternative splicing in Caenorhabditis elegans. DNA Research 5:25-30. 13. Trinh CH, Hunter T, Stewart EE, Phillips SEV, Hunter GJ (2008) Purification, crystallization and X-ray structures of the two manganese superoxide dismutases from Caenorhabditis elegans. Acta Crystallographica Section F 64:1110-1114. 14. Magnoni R, Palmfeldt J, Hansen J, Christensen JH, Corydon TJ, Bross P (2014) The Hsp60 folding machinery is crucial for manganese superoxide dismutase folding and function. Free Radic Res 48:168-79. 15. Hunter GJ, Hunter T (2013) GroESL protects superoxide dismutase (SOD)- Deficient cells against and is a chaperone for SOD. Health 5:1719-1729. 16. Honda Y, Honda S (1999) The daf-2 gene network for longevity regulates oxidative stress resistance and Mn-superoxide dismutase gene expression in Caenorhabditis elegans. FASEB J. 13:1385-93. 17. Jones SJ, Riddle DL, Pouzyrev AT, Velculescu VE, Hillier L, Eddy SR, Stricklin SL, Baillie DL, Waterston R, Marra MA (2001) Changes in gene expression associated with developmental arrest and longevity in Caenorhabditis elegans. Genome Res. 11:1346-52. 18. Honda Y, Honda S (2001) Life span extensions associated with upregulation of gene expression of antioxidant enzymes in Caenorhabditis elegans, studies of mutation in AGE-1, PI3 kinase homologue and short- term exposure to hyperoxia. J. Amer. Aging Assoc. 24:7. 19. Murphy CT, McCarroll SA, Bargmann CI, Fraser A, Kamath RS, Ahringer J, Li H, Kenyon C (2003) Genes that act downstream of DAF-16 to influence the lifespan of Caenorhabditis elegans. Nature 424:277-83. 20. Lee SS, Kennedy S, Tolonen AC, Ruvkun G (2003) DAF-16 target genes that control C. elegans life-span and metabolism. Science 300:644-7. 21. McElwee J, Bubb K, Thomas JH (2003) Transcriptional outputs of the Caenorhabditis elegans forkhead protein DAF-16. Aging Cell 2:111-21. 22. McElwee JJ, Schuster E, Blanc E, Piper MD, Thomas JH, Patel DS, Selman C, Withers DJ, Thornton JM, Partridge L and others (2007) Evolutionary conservation of regulated longevity assurance mechanisms. Genome Biol 8:R132.1-R132.16. 23. Yanase S, Yasuda K, Ishii N (2002) Adaptive responses to oxidative damage in three mutants of Caenorhabditis elegans (age-1, mev-1 and daf- 16) that affect life span. Mech Dev 123:1579-87. 24. Kenyon C, Chang J, Gensch E, Rudner A, Tabtiang R (1993) A C. elegans mutant that lives twice as long as wild type. Nature 366:461-4. 25. Morris JZ, Tissenbaum HA, Ruvkun G (1996) A phosphatidylinositol-3-OH kinase family member regulating longevity and diapause in Caenorhabditis elegans. Nature 382:536-539. 26. Kimura KD, Tissenbaum HA, Liu Y, Ruvkun G (1997) daf-2, an insulin receptor-like gene that regulates longevity and diapause in Caenorhabditis elegans. Science 277:942-6. 27. Guarente L, Kenyon C (2000) Genetic pathways that regulate ageing in model organisms. Nature 408:255-62. 28. Yang W, Hekimi S (2010) A mitochondrial superoxide signal triggers increased longevity in Caenorhabditis elegans. PLoS Biol 8:e1000556. 29. Sies H (2014) Role of metabolic H2O2 generation: redox signaling and oxidative stress. J Biol Chem 289:8735-41. 30. Honda Y, Tanaka M, Honda S (2010) Redox regulation, gene expression and longevity. Geriatr Gerontol Int 10 S59-69. 31. Hekimi S, Lapointe J, Wen Y (2011) Taking a "good" look at free radicals in the aging process. Trends Cell Biol 21:569-76. 32. Johnston AD, Ebert PR (2012) The Redox System in C. elegans, a Phylogenetic Approach. J Toxicol 2012:546915-546935. 33. Back P, Braeckman BP, Matthijssens F (2012) ROS in aging Caenorhabditis elegans: damage or signaling? Oxid Med Cell Longev 2012:608478. 34. Hughes BG, Hekimi S (2011) A mild impairment of mitochondrial electron transport has sex-specific effects on lifespan and aging in mice. PLoS One 6:e26116. 35. Karkas MD, Akermark T, Johnston EV, Karim SR, Laine TM, Lee BL, Akermark T, Privalov T, Akermark B (2012) Water oxidation by single- site ruthenium complexes: using ligands as redox and proton transfer mediators. Angew Chem Int Ed Engl 51:11589-93. 36. Sohal RS, Orr WC (2012) The redox stress hypothesis of aging. Free Radic Biol Med 52:539-55. 37. McAdam ME, Fox RA, Lavelle F, Fielden EM (1977) A pulse-radiolysis study of the manganese-containing superoxide dismutase from Bacillus stearothermophilus. A kinetic model for the enzyme action. Biochem J 165:71-9. 38. Bull C, Fee JA (1991) Kinetic studies of superoxide dismutases: Properties of the manganese containing protein from Thermus thermophilus. J.Am.Chem.Soc. 113:4069-4076. 39. Hsu JL, Hsieh Y, Tu C, O'Connor D, Nick HS, Silverman DN (1996) Catalytic properties of human manganese superoxide dismutase. J Biol Chem 271:17687-91. 40. Hearn AS, Tu C, Nick HS, Silverman DN (1999) Characterization of the product-inhibited complex in catalysis by human manganese superoxide dismutase. J Biol Chem 274:24457-60. 41. Porta J, Vahedi-Faridi A, Borgstahl GE (2010) Structural analysis of peroxide-soaked MnSOD crystals reveals side-on binding of peroxide to active-site manganese. J Mol Biol 399:377-84. 42. Perry JJ, Hearn AS, Cabelli DE, Nick HS, Tainer JA, Silverman DN (2009) Contribution of human manganese superoxide dismutase tyrosine 34 to structure and catalysis. Biochemistry 48:3417-24. 43. Borgstahl GE, Parge HE, Hickey MJ, Beyer WF, Jr., Hallewell RA, Tainer JA (1992) The structure of human mitochondrial manganese superoxide dismutase reveals a novel tetrameric interface of two 4-helix bundles. Cell 71:107-18. 44. Sheng Y, Stich TA, Barnese K, Gralla EB, Cascio D, Britt RD, Cabelli DE, Valentine JS (2011) Comparison of two yeast MnSODs: mitochondrial Saccharomyces cerevisiae versus cytosolic Candida albicans. J Am Chem Soc 133:20878-89. 45. Fluckiger S, Mittl PR, Scapozza L, Fijten H, Folkers G, Grutter MG, Blaser K, Crameri R (2002) Comparison of the crystal structures of the human manganese superoxide dismutase and the homologous Aspergillus fumigatus allergen at 2-A resolution. J Immunol 168:1267-72. 46. Nakamura T, Torikai K, Uegaki K, Morita J, Machida K, Suzuki A, Kawata Y (2011) Crystal structure of the cambialistic superoxide dismutase from Aeropyrum pernix K1--insights into the enzyme mechanism and stability. FEBS J 278:598-609. 47. Ludwig ML, Metzger AL, Pattridge KA, Stallings WC (1991) Manganese superoxide dismutase from Thermus thermophilus. A structural model refined at 1.8 Angstrom resolution. J. Mol. Biol 219:335-358. 48. Edwards RA, Baker HM, Whittaker MM, Whittaker JW, Jameson GB, Baker EN (1998) Crystal structure of Escherichia coli manganese superoxide dismutase at 2.1-Å resolution. J. Biol. Inorg. Chem. 3:161-171. 49. Lah MS, Dixon MM, Pattridge KA, Stallings WC, Fee JA, Ludwig ML (1995) Structure-function in Escherichia coli iron superoxide dismutase: Comparisons with the manganese enzyme from Thermus thermophilus. Biochemistry 34:1646-1660. 50. Shindyalov IN, Bourne PE (1998) Protein structure alignment by incremental combinatorial extension (CE) of the optimal path. Protein Eng 11:739-47. 51. Engh RA, Huber R (1991) Accurate bond and angle parameters for X-ray protein structure refinement. Acta Crystallographica Section A 47:392- 400. 52. Chen VB, Arendall WB, 3rd, Headd JJ, Keedy DA, Immormino RM, Kapral GJ, Murray LW, Richardson JS, Richardson DC (2010) MolProbity: all-atom structure validation for macromolecular crystallography. Acta Crystallogr D Biol Crystallogr 66:12-21. 53. Schrodinger, LLC. The PyMOL Molecular Graphics System, Version 1.3r1. 2010. 54. Quint P, Ayala I, Busby SA, Chalmers MJ, Griffin PR, Rocca J, Nick HS, Silverman DN (2006) Structural mobility in human manganese superoxide dismutase. Biochemistry 45:8209-15. 55. Sheng Y, Durazo A, Schumacher M, Gralla EB, Cascio D, Cabelli DE, Valentine JS (2013) Tetramerization reinforces the dimer interface of MnSOD. PLoS One 8:e62446. 56. Tabares LC, Cortez N, Hiraoka BY, Yamakura F, Un S (2006) Effects of substrate analogues and pH on manganese superoxide dismutases. Biochemistry 45:1919-29. 57. Whittaker JWMM (1991) Active site spectral studies on manganese superoxide dismutase. Journal of the American Chemical Society 113:5528-5540. 58. Jackson TA, Karapetian A, Miller AF, Brunold TC (2005) Probing the geometric and electronic structures of the low-temperature azide adduct and the product-inhibited form of oxidized manganese superoxide dismutase. Biochemistry 44:1504-20. 59. Miller AF, Sorkin DL, Padmakumar K (2005) Anion binding properties of reduced and oxidized iron-containing superoxide dismutase reveal no requirement for tyrosine 34. Biochemistry 44:5969-81. 60. Merlino A, Russo Krauss I, Castellano I, De Vendittis E, Rossi B, Conte M, Vergara A, Sica F (2010) Structure and flexibility in cold-adapted iron superoxide dismutases: the case of the enzyme isolated from Pseudoalteromonas haloplanktis. J Struct Biol 172:343-52. 61. Abreu IA, Rodriguez JA, Cabelli DE (2005) Theoretical studies of manganese and iron superoxide dismutases: superoxide binding and superoxide oxidation. J Phys Chem B Condens Matter Mater Surf Interfaces Biophys 109:24502-9. 62. Sheng Y, Butler Gralla E, Schumacher M, Cascio D, Cabelli DE, Valentine JS (2012) Six-coordinate manganese(3+) in catalysis by yeast manganese superoxide dismutase. Proc Natl Acad Sci U S A 109:14314-9. 63. Edwards RA, Whittaker MM, Whittaker JW, Baker EN, Jameson GB (2001) Outer sphere mutations perturb metal reactivity in manganese superoxide dismutase. Biochemistry 40:15-27. 64. Holley AK, Dhar SK, St Clair DK (2013) Curbing cancer's sweet tooth: is there a role for MnSOD in regulation of the Warburg effect? 13:170-88. 65. Emmler T, Ayala I, Silverman D, Hafner S, Galstyan AS, Knapp EW, Buntkowsky G (2008) Combined NMR and computational study for azide binding to human manganese superoxide dismutase. Solid State Nucl Magn Reson 34:6-13. 66. Hunter T, Ikebukuro K, Bannister WH, Bannister JV, Hunter GJ (1997) The conserved residue tyrosine 34 is essential for maximal activity of iron- superoxide dismutase from Escherichia coli. Biochemistry 36:4925-33. 67. Guan Y, Hickey MJ, Borgstahl GE, Hallewell RA, Lepock JR, O'Connor D, Hsieh Y, Nick HS, Silverman DN, Tainer JA (1998) Crystal structure of Y34F mutant human mitochondrial manganese superoxide dismutase and the functional role of tyrosine 34. Biochemistry 37:4722-30. 68. Tabares LC, Cortez N, Un S (2007) Role of tyrosine-34 in the anion binding equilibria in manganese(II) superoxide dismutases. Biochemistry 46:9320-7. 69. Steinman HM (1992) Construction of Escherichia coli K-12 strain deleted for manganese and iron superoxide dismutase genes and its use in cloning the iron superoxide dismutase gene of Legionella pneumophila. Mol. Gen. Genet. 232:427-430. 70. Hunter T, Bannister WH, Hunter GJ (1997) Cloning, expression, and characterization of two manganese superoxide dismutases from Caenorhabditis elegans. J Biol Chem 272:28652-9. 71. Ysebaert-Vanneste M, Vanneste WH (1980) Quantitative resolution of Cu,Zn- and Mn- superoxide dismutase activities. Anal. Biochem. 107:86- 95. 72. Winn MD, Ballard CC, Cowtan KD, Dodson EJ, Emsley P, Evans PR, Keegan RM, Krissinel EB, Leslie AG, McCoy A and others (2011) Overview of the CCP4 suite and current developments. Acta Crystallogr D Biol Crystallogr 67:235-42. 73. Leslie AG (2006) The integration of macromolecular diffraction data. Acta Crystallogr D Biol Crystallogr 62:48-57. 74. Evans P (2006) Scaling and assessment of data quality. Acta Crystallogr D Biol Crystallogr 62:72-82. 75. Brunger AT (1997) Free R value: cross-validation in crystallography. Methods Enzymol 277:366-96. 76. Vagin AA, Steiner RA, Lebedev AA, Potterton L, McNicholas S, Long F, Murshudov GN (2004) REFMAC5 dictionary: organization of prior chemical knowledge and guidelines for its use. Acta Crystallogr D Biol Crystallogr 60:2184-95. 77. Emsley P, Lohkamp B, Scott WG, Cowtan K (2010) Features and development of Coot. Acta Crystallogr D Biol Crystallogr 66:486-501. 78. Hearn AS, Stroupe ME, Cabelli DE, Ramilo CA, Luba JP, Tainer JA, Nick HS, Silverman DN (2003) Catalytic and structural effects of amino acid substitution at histidine 30 in human manganese superoxide dismutase: insertion of valine C gamma into the substrate access channel. Biochemistry 42:2781-9. 79. Schwatz H Pulse Radioloysis Program Brookhaven National Laboratory. 80. Cabelli DE, Guan Y, Leveque V, Hearn AS, Tainer JA, Nick HS, Silverman DN (1999) Role of tryptophan 161 in catalysis by human manganese superoxide dismutase. Biochemistry 38:11686-92.

This study was supported by a grant (PHBRP02) awarded to TH by the University of Malta and funding by the Dean’s Research Award to TH and RB (MDSIN08-19 and MEDIN08-01 respectively). The pulse radiolysis studies are carried out at the

Accelerator Centre for Energy Research, supported by the U.S. Department of

Energy, Office of Science, Office of Basic Energy Sciences, Division of Chemical

Sciences, Geosciences, and Biosciences, under Contract DE-AC02-98CH10886.