[5 Crystal Structures of Oxidized and Reduced Forms of NADH Peroxidase

By Joanne I. Yeh and Al Claiborne

Introduction

Modulations of protein function by oxidation-reduction reactions through co-factors such as FAD, iron-sulfur centers, heme prosthetic groups, and redox-active disulfides are well known. However, only more recently have cysteine-sulfenic acid (Cys-SOH) derivatives been recognized as novel moieties (cofactors) functioning in enzyme catalysis and redox regulation. A major reason for the delayed recognition of the functional importance of this moiety in proteins is that sulfenic acids are difficult to identify and have long been characterized primarily as transient intermediates because of their highly reactive and unstable chemical nature. There are three major simple organic oxyacids of sulfur: (1) sulfenic acids (RSOH), (2) sulfinic acids (RS02H), and (3) sulfonic acids (RSO3H).1 Of these, sulfonic acids are by far the most stable; sulfinic acids are somewhat unstable thermally,

and sulfenic acids are for the most part elusive reactive intermediates. In proteins, organic oxyacids are obtained by the oxidation of cysteine residues and, until more recently, the presence of these sulfur oxyacids, primarily as cysteine-sulfinic and cysteine-sulfonic acids, was thought to be nonnative, caused by the oxidation of surface cysteines by ambient oxygen or as experimental artifacts formed, for example, through oxidation by free radicals generated during X-ray data collection. In many enzymes that require sulfhydryls or sulfenic acids for catalysis, the oxidation to a sulfinic or sulfonic acid prohibits the formation of an "activated" form of the sulfur, such as a thiolate (S~) or sulfenate (SO-), and irreversibly inhibits the enzyme. However, the crystal structure of NADH peroxidase2 (Npx) led to an unambiguous identification of a reversibly reducible cysteine-sulfenic acid as a secondary redox center in this enzyme. This, along with substantial biochemical data, established this moiety as an intrinsic functional group in this protein. Npx utilizes an essential Cys-SOH and the reversibility of the Cys-SH -o- Cys-SOH redox cycle is inherent to its catalytic mechanism. In general, the facile redox potential of this reversible redox partner highlights its utilization in a variety of regulatory functions and other biological reactions. The noted reactivity of sulfenic acid, which until recently prevented its identification as important functional groups in proteins, is also a hallmark as Cys-SOH is considerably more reactive than the corresponding Cys-SS-Cys disulfide.3 This reactivity is advantageous when utilized in roles such as redox control; the identification of regulation via reversible SOH -o- SH oxidation substantiates this idea.4-7 Among the distinct protein systems that utilize cysteine-sulfenic acids for catalysis and/or regulation are FAD-dependent peroxide and disulfide reductases, peroxiredoxins, transcription factors, and transport proteins, among others.8 These diverse systems illustrate the functional significance of sulfenic acids in proteins and in proteinDNA complexes. Further descriptions of redox regulation can be found in other chapters in this volume and in reviews, which give additional consideration to the topic of redox regulation and catalysis via cysteine-sulfenic acids.3,8

Mechanistic Considerations

The defining property of Npx relative to other FAD-dependent disulfide reductases is the absence of a redox-active protein disulfide. Npx contains only one-half

2 J. I. Yeh, A. Claiborne, and W. G. J. Hoi, Biochemistry 35, 9951 (1996).

3 A. Claiborne, J. I. Yeh, C. Mallett, J. Luba, E. J. Crane III, V. Charrier, and D. Parsonage, Biochemistry 38, 15407(1999).

5 S.-R. Lee, K.-S. Kwon, S.-R. Kim, and S. G. Rhee,./. Biol. Chem. 273, 15366 (1998).

6 S. Veeraraghava, C. C. Mello, E. J. Androphy, and J. D. Baleja, Biochemistry 38,16115 (1999).

7 M. A. Gorman, S. Morera, D. G. Rothwell, E. de La Fortelle, C. D. Mol, J. A. Tainer, I. D. Hickson, and P. S. Freemont, EMBO J. 16, 6548 (1997).

8 A. Claiborne, T. C. Mallett, J. I. Yeh, J. Luba, and D. Parsonage, Adv. Protein Chem. 58,215 (2001).

-h20

+nadh

+nadh

Scheme 1

cystine per subunit and NADH reduction generates one Cys-SH.9 In Npx, the cycling between the reduced thiol or thiolate and the oxidized sulfenate or sulfenic acid of residue 42 is an intrinsic part of the catalytic cycle (species 1 and 3, Scheme 1) as the formation of the catalytically competent thiolate form (species 2, Scheme 1) precedes the binding of H2O2.

The structures of native, oxidized Npx to 2.1 A resolution and the NADH-bound reduced form at 2.85 A resolution have yielded insight into how elements of the protein help to stabilize and modulate the redox and catalytic properties of this enzyme. This chapter describes the experimental approaches used to capture the active, sulfenic acid form of Npx as well as the NADH-reduced form, to outline the steps that were taken to minimize oxidation to an inactive form during crystal growth and X-ray data collection.

When working with a protein that is suspected to contain a functional cysteine-sulfenic acid, it should be kept in mind that the presence of stable Cys-SOH derivatives in proteins implies that the respective Cys-SOH must exist in an unusual microenvironment, protected from solvent and preferentially associated with apolar regions of the protein structure (Fig. 1). All known structures of functional Cys-SOH-containing proteins have the Cys-SOH group effectively immersed within a cavity and isolated from other reactive groups.2'10'11 A point worth emphasizing is that the presence of Cys-SOH in proteins occurs when the surface topology of the protein prevents the formation of interchain disulfide bonds and there are no sulfhydryl groups in the vicinity to form intramolecular cystine bonds. This provides the environment for the formation of a sulfenic acid on reaction of a cysteinyl side chain with an oxidant.12

9 L. B. Poole and A. Claiborne, J. Biol. Chem. 261, 14525 (1986).

10 K. Becker, S. N. Savvides, M. Keese, R. H. Schirmer, and P. A. Karplus, Nat. Struct. Biol. 5, 267 (1998).

11 H.-J. Choi, S. W. Kang, C.-H. Yang, S. G. Rhee, and S.-E. Ryu, Nat. Struct. Biol. 5,400 (1998).

12 W. S. Allison, Accounts Chem. Res. 9, 293 (1976).

Fig. 1. Dimer interface of Npx. FAD and cysteine-sulfenic acid (Cys-SOH, denoted in figure as Cyo-42) residue indicate the active site of Npx, which is formed at the dimer interface. Residues from each monomer subunit at the interface contribute to the tight packing; numbered residues from one monomer subunit are differentiated from the second by primes. The tight packing shown limits solvent accessibility, stabilizing the labile cysteine-sulfenic acid (Cyo-42) by burying the group in an apolar environment of the protein. In addition, the hydrogen-bonding chain formed by Arg-307, Glu-14, Arg-303, His-10, and Cyo-42 stabilizes the sulfenate form (SO") of the cysteine-sulfenic acid, further enhancing the stability of this reactive group.

Fig. 1. Dimer interface of Npx. FAD and cysteine-sulfenic acid (Cys-SOH, denoted in figure as Cyo-42) residue indicate the active site of Npx, which is formed at the dimer interface. Residues from each monomer subunit at the interface contribute to the tight packing; numbered residues from one monomer subunit are differentiated from the second by primes. The tight packing shown limits solvent accessibility, stabilizing the labile cysteine-sulfenic acid (Cyo-42) by burying the group in an apolar environment of the protein. In addition, the hydrogen-bonding chain formed by Arg-307, Glu-14, Arg-303, His-10, and Cyo-42 stabilizes the sulfenate form (SO") of the cysteine-sulfenic acid, further enhancing the stability of this reactive group.

For a protein suspected of utilizing a cysteine-sulfenic acid, several approaches can be taken to determine the oxidation state of the suspected group. After examination of the sequence to determine the number of cysteines and whether cystine bonds are possible, based on a known structure or prediction of the fold, site-directed mutagenesis can be done to systematically map the function of each cysteine. If a cysteine is determined not to be in a disulfide bond, owing to the absence of a partner or because it is the lone cysteine in the protein, then modification of the cysteine can be done to determine whether the oxidation state of the residue affects protein function. In all the proteins that utilize Cys-SOH, the reversibility of Cys-SH -o- Cys-SOH is obligatory for activity. Consequently, the addition of an equivalent of an oxidizing reagent, such as H2O2, should have measurable effects on activity whereas addition of a reducing agent, such as dithiothreitol (DTT), should reverse the effects, as long as oxidation has not proceeded to the sulfinic or sulfonic acid state. One approach that exploits the intrinsic nucleophilicity of the Cys-SO" anion toward the amine and thiol reagent NBD-C1 (7-chloro-4-nitrobenzo-2-oxa-l,3-diazole) results in a protocol for Cys-SOH modification13 that can be analyzed in conjunction with UV-visible, fluorescence, and electrospray interface-mass spectrometric data. Even in the absence of chemical modification such as the NBD-C1 approach, spectroscopic analysis monitoring charge-transfer interactions between the Cys-SOH and a cofactor, such as FAD, can provide another means of determining the presence of such a catalytic pair if such interaction exists, as it does for Npx.14 Similarly, site-directed mutagenesis,15 l3C-based nuclear magnetic resonance (13C-NMR) analysis of labeled cysteines, and mass spectrometric analysis of reduced and oxidized forms of the protein under study can be done to ascertain whether oxidation occurs via 1,2-or 3-oxygen addition to a sulfhydryl group when an oxidizing agent is added, where the presence of these oxidized forms can be correlated to functional activity.914

An understanding of the redox mechanism can be helpful in monitoring the state of the enzyme or protein before structural analysis. For Npx, a simple spectro-photometric assay9 to determine the activity of the enzyme allowed us to assess the oxidation state of the crystals before data collection in most cases or to confirm the oxidation state of the enzyme in the crystals after structural analysis.

General Considerations

Although protein sulfenates are generally unstable, reactivities of sulfenic acids in proteins can span a wide range. For some proteins, severe limitation of exposure to oxygen and persistent presence of reducing agents are necessary to prevent significant oxidation whereas for others, oxidation/inactivation by ambient oxygen is a slow process and these proteins can be stored under ambient conditions for hours to days before a substantial population of the protein molecules is oxidized. Hence, it may be necessary to empirically determine the sensitivity of a protein through various means, as indicated in the preceding section. For Npx, we were able to use a spectrophotometric activity assay to ascertain whether freshly expressed and purified protein, stored protein, and protein obtained from dissolved crystals at various time points were active. These measurements gave us some general understanding of the time frame over which oxidation of the enzyme occurred and the expected oxidation state of the protein crystals before X-ray structure analysis.

13 H. R. Ellis and L. B. Poole, Biochemistry 36, 15013 (1997).

14 E. J. Crane III, D. Parsonage, and A. Claiborne, Biochemistry 35, 2380 (1996).

15 D. Parsonage and A. Claiborne, Biochemistry 34, 435 (1995).

A final unambiguous view of the oxidation state of the enzyme was obtained after our structural analysis.

To obtain a native, active, cysteine-sulfenic acid-containing crystal structure of Npx, several precautionary measures were taken. X-ray structure determination of a protein containing a cysteine-sulfenic acid may require additional precautions in order to prevent the formation of oxidation artifacts. With the advent of cryotech-niques during data collection, the production of free radicals that result in crystal decay, among other phenomena, has been minimized. These free radicals can also presumably be the cause of oxidation of sulfenic acids as the radicals can readily oxidize the acid to its more stable, higher oxidation states and, consequently, use of cryogenic temperatures during data collection can also minimize these oxidation events. However, in some systems, data collection at cryogenic temperatures alone may not be sufficient to prevent the oxidation of sulfenic acids as these groups are highly labile and could presumably react with any residual molecules, including ambient oxygen, before and during crystal growth. An earlier X-ray crystal structure of the Npx had shown oxidation of what is now known to be the sulfenic acid to a sulfonic acid.16 To maximize the possibility of crystallizing the active, sulfenic acid form of the enzyme, freshly prepared protein was used for crystallization setups and crystals were stored under argon, as described below.

It should be noted that the structures of other proteins/enzymes have indicated that the above-described precautionary measures were not necessary to limit oxidation of the cysteine-sulfenic acid. In human glutathione reductase, a sulfenate of residue 63 was found under fully aerobic conditions and at ambient temperatures.10 Here, the sulfenate was generated through oxidation of Cys-63 by .S-nitrosoglutathione, a physiological carrier of NO, and the formation of this reversible, inactive form of glutathione reductase may play a role in NO-based redox signaling, where interactions at the active site were speculated to be responsible for the stability of the sulfenic acid. In a human cysteine peroxidase enzyme, a cysteine-sulfenic acid was unexpectedly found at the active site although data were collected at room temperature.11 The authors propose that the unusual stability of the sulfenic acid was due to interaction with an Mg2+ ion, which stabilized the ionized sulfenate by charge interactions and blocked the narrow entrance of the active site pocket, inhibiting the access of additional O2 or H2O2. The stability of protein sulfenates can be attributed to specific interactions and factors based on the localized protein environment surrounding the sulfenic acid residue. Criteria that represent factors for stabilization include limited solvent accessibility, an apolar microenvironment, hydrogen bonding and/or ionization of -SOH to the sulfenate (-SO") form, and the absence of vicinal protein thiols.3 These criteria have been satisfied in all the proteins that have been structurally characterized, including Npx, glutathione reductase, and human cysteine peroxidase enzyme, in which a

16 T. Stehle, S. A. Ahmed, A. Claiborne, and G. E. Schulz, J. Mol. Biol. 221, 1325 (1991).

Cys-SOH is utilized in catalysis or regulation, highlighting the significance of the local protein environment in stabilization and function of this reactive moiety.

General Protocol

To minimize the possibility of oxidation under ambient conditions, crystals are grown from freshly prepared protein, thereby limiting aging time and exposure of the enzyme to atmospheric oxygen before crystallization. Typically the amount of time from crystallization to data collection is between 2 weeks and 2 months and crystallization trays are stored under inert argon, purged of oxygen. It is not known whether this measure is necessary; it is likely, for freshly grown crystals that are frozen, with data collection at 100 K soon after crystal formation, that this would be unnecessary. For crystals that are to be stored for more than a few weeks, the storing of these crystals under inert atmospheres prevents oxidation. It is known from cryogenic data collection on crystals grown and allowed to age for lengths of time spanning weeks to months under ambient oxygen conditions that this storage time contributes to oxidation. A sulfinic acid form of Npx has been obtained by cryogenic data collection methods,17 indicating that oxidation from the sulfenic acid occurred during crystal storage and that this is likely a stepwise event, depending on the amount of storage time. Storage of crystals for as long as 7 months before cryogenic data collection results in the sulfinic acid form of residue 42. The formation of a sulfonic acid may be due to long, ambient oxygen exposure of the protein and/or protein crystal, in conjunction with data collection at room temperature, which was the original data collection protocol.16

The setup for storage under inert atmospheres can be as sophisticated as a glovebox or as simple as oxygen-resistant bags constructed of polymers such as polyvinyl chloride, which with sufficient thickness can provide reasonable resistance to oxygen permeation. To purge the trays of oxygen, a gentle stream of argon is passed over the sitting drop tray before sealing it with tape. For hanging drop trays, each coverslip is lifted and argon is passed over the reservoir before reseating each drop. These crystallization trays are then placed into an oxygen-resistant bag and the bag is purged with argon. As argon is heavier than air, the argon displaces oxygen from the system. Care is taken to flush out the system on a routine basis to sustain an inert atmosphere and crystals of active Npx are maintained for more than 6 months by this method. A monitoring system is not used for Npx, but for more sensitive proteins oxygen sensors can be used to measure the amount of oxygen in the setup and care can be taken to maintain the oxygen levels at a low value.

Npx crystals can be readily dissolved and analyzed by a simple spectrophoto-metric method that monitors the decrease in 340-nm absorbance of NADH in the

17 J. I. Yeh et al., manuscript in preparation (2002).

presence of H202.9 This assay, which can be directly correlated to the cysteine-sulfenic acid oxidation state, is a useful and expedient means of monitoring oxygen-aging effects. A unit of activity is defined as the amount of enzyme that catalyzes the oxidation of 1 ¿xmol of NADH/min at 25°. This allows the monitoring of the redox state of Npx, because oxidation of the sulfenic acid at residue 42 to sulfinic or sulfonic acid forms can be distinguished in this assay as diminished or abolished activity of the enzyme. During the catalytic cycle, SOH must first be reduced to the thiolate by NADH; this precedes binding of H202 (see Scheme 1). When the sulfenic acid is oxidized to the sulfinic or sulfonic acid form, the initial "priming step" for the formation of the critical thiolate cannot occur and, hence, hydrogen peroxide cannot bind to the active site. Although NADH can still bind, this assay readily ascertains the oxidation state of the cysteine at position 42, as reductants such as dithionite or NADH can monitor the presence of redox intermediates through UV-visible and fluorescence properties.18 Reductive titrations of wild-type Npx generate a charge-transfer intermediate characteristic of the two-electron reduced (EH2) forms of most disulfide reductases.9 In Npx, this absorbance band is centered at 540 nm and is due to the electronic interaction between the nascent Cys-42-S~ and the oxidized FAD. When the sulfenic acid is oxidized to sulfinic or sulfonic acid forms, reduction to Cys-42-S~ is not possible and hence reduction or absence of the 540-nm absorbance band can be used to monitor the presence of these oxidized species. Similarly, NADH reduction experiments with intact crystals (described below) can be done, where color changes in the crystals can be simply visualized under a microscope to ascertain the oxidation state of the sulfenic acid at position 42.

Structural Analysis

The structural analysis of Npx utilizes conventional structure validation and assessment tools including SA-omit maps and iterative model building correlated to statistical factors and map correctness to confirm the final accuracy of the structure, especially in the assignment of a sulfenic acid at residue 42. For the 2.1 A native structure, confirmation that the Cys-42-SOH has not been oxidized to the sulfinic Cys-42-S02H or sulfonic Cys-42-S03H acid is achieved through refinements with several models that differ at residue 42. These models are generated with a Cys-42-SOH model previously refined to 2.8 A (Yeh et al.2) and the Cp substituent at position 42 is changed to include -SH, -S02H, and -S03H as well as a truncated form (-H; alanine model). Omit difference maps, calculated after simulated annealing to reduce model bias, are then used to ascertain the amount of electron density that is present after omitting a specific region from the model. Refined omit maps are calculated by omitting from the model a region of 8 A

18 C. H. Williams, Jr., L. D. Arscott, R. G. Matthews, C. Thorpe, and K. D. Wilkinson, Methods Enzymol. 62, 185 (1979).

around a residue and refining the remaining model. Atoms within a 3-A shell are restrained in order to avoid artificial movement into the omitted region. Each cycle includes manual adjustment and fitting into difference Fouriers [(Fsoak — Fnative) exp(!'acaic_native)] and additive Fouriers [(2Fsoak - Fmliw)e\p(iacaic_niitiye)] maps.19 These maps show unambiguously that the resulting difference electron density at position 42 best fits a Cys-42-SOH model (Fig. 2), whereas Cys-S03H results in two negative density peaks in the maps, indicating that the model contains too many atoms at that site.

To obtain the NADH-reduced structure, NADH is added to intact crystals; this is done with crystals of the cysteine-sulfenic acid and cysteine-sulfinic acid forms. In the case of the active, sulfenic acid form, addition of NADH immediately turns the yellow crystals red, signifying reduction of the sulfenic acid. To obtain this complex of Npx with NADH, a 100 mM solution of NADH is freshly prepared in Tris buffer, pH 6.8, and diluted 1:10 with the artificial mother liquor, for a final concentration of 10 mM. Crystals are soaked in this solution for several minutes, during which the crystals turn from yellow to red, signifying reduction of the sulfenic acid. Once the red color is observed, the crystal is flash-frozen to 100 K and data are collected to capture the reduced form. The frozen crystal remains red throughout data collection, indicating a charge-transfer complex between Cys-42-Sy, and the isoalloxazine of FAD is formed. When this same procedure is performed on crystals that have the inactive, sulfinic acid-oxidized form, the red shift is not observed, even with addition of a molar excess of NADH. The crystal structure of this sulfinic acid form shows that although NADH is bound, reduction of the sulfinic acid cannot occur17 and, hence, a charge-transfer complex does not form with the isoalloxazine ring of FAD.

Although soaking in NADH causes minor changes in cell parameters and overall structural conformation, considerable changes are found at the active site. The overall extent of structural changes that occurs on soaking is reflected in /?iso values, when compared with the apo form, of 19-25%. These values are appreciably higher than 7?merge values of 5-9% within apo forms or within NADH-bound forms. Structurally, these changes reflect reorientation of several side chain groups immediately at the active site of Npx, to accommodate bound NADH, as well as a shift in the isoalloxazine ring of FAD.

Structural analyses for the NADH-reduced form of Npx are done similarly, as described for the oxidized form of Npx. Models are generated as described above except that the Cp substituent at position 42 is changed to include -SH, -SOH, and a truncated form (-H; alanine) for the 2.85-A NADH-reduced native structure. Models are refined, and after every cycle of refinement maps are calculated to visually inspect the positive and negative residual densities in a erA-weighted

19 P. A. Karplus and G. E. Schulz, J. Mol. Biol. 210, 163 (1989).

Fig. 2. Residual electron density. A <TA-weighted F0 — Fc electron density map contoured at 3.5 a of the active site of Npx. The strong positive electron density identifies the position of the oxygen of the sulfenic acid residue (Cyo42). The map was calculated with a model containing a cysteine residue at position 42. The single, strong peak unambiguously identifies the residue at position 42 as a cysteine-sulfenic acid. Active site residues that are important for stabilizing the sulfenic acid include His-10 and Arg-303, whereas Glu-14 participates in forming the tight dimer interface that limits solvent accessibility, important for maintaining the oxidation state of the sulfenic acid.

Fig. 2. Residual electron density. A <TA-weighted F0 — Fc electron density map contoured at 3.5 a of the active site of Npx. The strong positive electron density identifies the position of the oxygen of the sulfenic acid residue (Cyo42). The map was calculated with a model containing a cysteine residue at position 42. The single, strong peak unambiguously identifies the residue at position 42 as a cysteine-sulfenic acid. Active site residues that are important for stabilizing the sulfenic acid include His-10 and Arg-303, whereas Glu-14 participates in forming the tight dimer interface that limits solvent accessibility, important for maintaining the oxidation state of the sulfenic acid.

difference (F0 — Fc) density map. After these first maps are examined, a cysteine is placed at residue 42 and another round of refinement is completed. The difference Fouriers clearly show densities corresponding to NADH and reveal areas of conformational change in the structures. Examination of the difference Fourier and additive Fourier maps allows assignment of the oxidation state of residue 42 to correspond to a sulfhydryl.

Summary

X-ray structural characterization of cysteine-sulfenic acid-containing proteins is one of the most defining approaches to characterizing this rapidly growing class of protein functional groups. Although outside the scope of this chapter, these structural analyses can lead to kinetic measurements in the crystal that allow intermediate states to be trapped, visualized, and studied. An understanding of the biochemistry of these reactive groups can be more fully gained by studying the localized protein environment in which these groups function. Increased perception of how elements of a protein can stabilize and contribute to modulation of function in these systems will allow novel means of enhancing or inhibiting function in important classes of protein molecules, including transcription factors and redox-regulated enzymes.

Acknowledgments

This work was supported by the National Institutes of Health NRSA Fellowship DK-09568 (J.I.Y.) and the National Institutes of Health Grant GM-35394 (A.c.).

0 0

Post a comment