INFORMATION TO USERS

This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.

The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.

In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book.

Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.

A Bell & Howell Information Company 300 North Zeeb Road. Ann Arbor. Ml 48106-1346 USA 313/761-4700 800/521-0600 Kinetic and Spectroscopic Studies on Dissimilatory and Assimilatory-Type Sulfite

Reductases from Desulfovibrio vulgaris (Hildenborough)

Dissertation

Presented in Partial Fulfillment of the Requirements

for the Degree of Doctor of Philosophy in the Graduate School

of The Ohio State University

by

Aileen G. Soriano

I********

The Ohio State University

1995

Dissertation Committee: Approved by:

Dr. Bruce E. Bursten

Dr. Sheldon Shore Adviser

Dr. James A. Cowan Department of Chemistry UMI Number: 9544694

UMI Microform 9544694 Copyright 1995r by UMI Company. All rights reserved.

This microform edition is protected against unauthorized copying under Title 17# United States Code.

UMI 300 North Zeeb Road Ann Arbor# MI 48103 To My Family Acknowledgements

I wish to express my sincere appreciation to Dr. James A. Cowan for his guidance and encouragement throughout my research and for giving me the opportunity to learn

NMR spectroscopy as well as techniques in the area of molecular biology. This dissertation owes a lot to the valuable work of Siu Man Lui, Wen Liang, Jian Tan and

Bonnie Wolfe. I would like to thank Dr. Marco Sola for introducing me to NMR spectroscopy; Dr. Chuck Cottrel for his help in using the NMR spectrometers; Don

Ordaz, at the O.S.U. Fermentation lab, for his valuable help in growing D. vulgaris cells;

Brian Becknel, Chris Black, Francis Ford, Tsuyoshi Ohyama, and Ronnie Tungol for their valuable help while I was growing D. vulgaris cells in the lab; Dawei Li for helping me understand 2D NMR pulse programs and for his initial guidance in setting-up 2D NMR experiments; and to Chris Black for his help with the use of the computers and softwares in the lab and with the stopped-flow instrument. My sincere appreciation also goes to

Anshu Agarwal for making my molecular biology work easier by sharing her expertise

(and materials) in preparing HIPIP mutant cells; Ruby Casareno, Mesut Eren and

Jianchao Zhang for showing me techniques in molecular biology and for sharing their reagents and equipment; Kshama Natarajan and Shiumin Bian for always being ready to help in the lab whenever they can. I would like to thank Francis Ford for his help in my attemps at organic synthesis and for taking over my teaching class during emergencies. I would like to thank Peggy May Hwang for helping to proof-read the text and together with Lucy Lapar, Elsa Olaer, and Tina Bautista, I thank them for their support and encouragement. To my parents, and to my sisters Adel, Nennette, and Susan, I thank them for their unconditional support throughout my “student days” from grade school to graduate school. And to the rest of my family, I thank them for their love and faith in me. Vita

May 27, 1966 ...... Bom, Cabanatuan City, Philippines

1987 ...... B. S., University of the Philippines, Diliman, Quezon City, Philippines

1989-1991, 1994-1995 ...... Graduate Teaching Associate Department of Chemistry The Ohio State University

1992-1994 ...... Graduate Research Associate The Department of Chemistry The Ohio State University

List of Publications

“Sulfite Reductase: Active Site Residues are “Non-catalytic”. Comparison of Reaction Energetics for Enzyme- and Siroheme-Catalyzed Reduction of Inorganic Substrates” Soriano, A.; Cowan J. A. J. Am. Chem. Soc.,1995, 117,4724-4725.

“Functional Expression and Characterization of the Assimilatory-Type Sulfite Reductase from Desulfovibrio vulgaris (Hildenborough)”. Tan, J.; Soriano, A.; Lui, S. M.; Cowan, J.A. Arch. Biochem. Biophys., 1994, 312, 516-523.

“Electronic properties of dissimilatory sulfite reductase from Desulfovibrio vulgaris (Hildenborough): comparative studies of optical spectra and relative reduction potentials for [Fe 4S4]-siroheme prosthetic centers”. Lui, S. M.; Soriano, A.; Cowan, J. A. Biochem. J.i, 1994, 304, 441-447.

“Enzymatic Reduction of Inorganic Anions. Variable-Temperature Steady-State Kinetics Experiments to Map the Energy Profile of an Enzymatic Multielectron Redox Reaction. v Application to the Dissimilatory Sulfite Reductase from Desulfovibrio vulgaris (Hildenborough)”. Lui, S. M.; Liang, W.; Soriano, A.; Cowan, J. A. J. Am. Chem. Soc. 1994, 116, 4531-4536.

“Enzymatic Reduction of Inorganic Anions. Pre-Steady-State Kinetic Analysis of the Dissimilatory Sulfite Reductase (Desulfoviridin) from Desulfovibrio vulgaris (Hildenborough). Mechanistic Implications”. Lui, S.M.; Soriano, A.; Cowan, J. A. J. Am. Chem. Soc., 1993, 115, 10483-10486.

Fields of Study

Major Field: Chemistry

Studies in Inorganic/Bioinorganic Chemistry

Prof. S. Shore, Prof. A. Wojcicki, Prof. B. E. Bursten, Prof. J. A. Cowan, Dr. R. Ziebart List of Tables

Table Page

1 Composition of Baars’ Medium for Growing D. vulgaris Cells (ATCC medium no. 1249) ...... 16

2 Composition of LS Medium for Growing D. vulgaris Cells ...... 17

3 Comparison of Optical Characteristics for Oxidized and Reduced Sulfite Reducing Enzymes from Prokaryotic and Eukaryotic Sources ..... 39

4 Pre-Steady-State Kinetic Parameters for Desulfoviridin ...... 53

5 Summary of Reaction Rate Constants at Room Temperature ...... 61

6 Steady-State Activation Parameters ...... 67

7 Breakdown of Steady-State Parameters ...... 6 8

8 Summary of Factors Contributing to the Activation Free Energy ...... 69

9 Pre-Steady-State Bond Cleavage Energy ...... 70

10 Steady-State Activation Parameters for Siroheme-Catalyzed Substrate Reduction ...... 8 6

11 NMR Parameters for the Hyperfine-Shifted Signals in Oxidized Assimilatory- type Sulfite Reductase from D. vulgaris ...... 1 0 1

12 Hyperfine-Shifted Resonances of Cysteine P-CH2’s in Iron-Sulfur Containing Proteins at 298 K ...... 109

vii 13 NMR Parameters for the Fully Reduced Assimilatory-Type Sulfite Reductase (SiR) from D. vulgaris (Hildenborough) at 298 K

viii List of Figures

Figure Page

1 The Biological Sulfur Cycle ...... 2

2 Adenosine-5’-phosphate (APS) ...... 3

3 The dissimilatory-sulfate reduction pathway ...... 5

4 A proposed scheme for the bioenergetics of growth of Desulfovibrio in lactate and sulfate ...... 6

5 Structure of siroheme, an isobacteriochlorin ...... 9

6 The magnetically coupled [Fe 4S4]-siroheme prosthetic centers in the active site of E. coli sulfite reductase ...... 12

7 The proposed model for the prosthetic centers in the active site of desulfoviridin (DV) and assimimilatory-type sulfite reductase (SiR) from D. vulgaris ...... 14

8 Spectra of deazaflavin (oxidized and reduced) and DV (oxidized and reduced) in 50 mM potassium phosphate buffer, pH 7.6 ...... 24

9 A diagram of the stopped-flow apparatus used for the pre-steady-state kinetic measurements ...... 26

10 Absorption spectra of (a) oxidized and fully reduced DV; and (b) reduced DV with 10 mM As02’ ...... 38

11 Electronic absorption spectra of (a) nitrite and (b) sulfite reductases from spinach; sulfite reductase from E. coli (c); and SiR from D. vulgaris (d) ...... 44 ix 12 Typical fit to one-exponential rate profile for non-substrate ligand binding ...... 49

13 Typical fit to rise-fall rate profile for binding and subsequent reduction of substrate anion ...... 50

14 Plot of kob» versus [L] for L = As0 2 (a) and L = HS' (b) ...... 51

15 Proposed mechanism for sulfite reduction catalyzed by sulfite reductase ...... 55

16 Proposed mechanism for reduction of inorganic substrates by sulfite reductase ...... 56

17 General free energy profiles for substrate turnover showing the components of AGr\ AG0t, AGd, and AGt* ...... 64

18 Steady-state data. A plot of Rln[(kanh)/KJcT)] versus \/T for NH2OH ...... 6 6

19 Pre-steady-state data. Plot of R\n{kJilkT) versus MT for NH2OH ...... 72

2 0 Isokinetic plot of AHr* versus ASr* ...... 75

21 Initial veleocity plot for turnover of sulfite by free siroheme at pH 6 ...... 82

22 Initial velocity plot for turnover of (a) N 02' and (b) NH2OH by free siroheme at pH 7 ...... 84

23 Circular dichroism and electronic absorption spectra of assimilatoty-type sulfite reductase (SiR) ...... 89

24 A 300 MHz *H NMR spectrum of recombinant SiR in 50 mM potassium phosphate buffer in D 20, pH 7 at 296 K ...... 90

25 Diagram of the siroheme-[Fe 4S4] unit in SiR showing the peripheral acetate and propionate side chains in siroheme as well as the a and 0 protons of the cluster cysteines. The proposed ligand in the sixth axial position, histidine, is also shown ...... 92

26 *H NMR spectrum of oxidized SiR showing labelled hyperfine-shifted resonances ...... 99

x 27 (Top) A 300 MHz *H NMR reference spectrum of oxidized SiR. (Bottom) NOE difference spectrum obtained by saturating signal a ...... 100

28 A 500 MHz *H NMR reference spectrum (top) and NOES Y map of oxidized SiR emphasizing the crosspeak between signals./and / (bottom) at 296 K ...... 102

29 Magnitude COSY maps (500 MHz) of oxidized SiR. The maps emphasize the through-bond correlation between signals j and / (a); signals f and g; and signals g and h (b) ...... 103

30 A 500 MHz WEFT-NOESY map of oxidized SiR. The map shows the correlation of signals j and / with resonances within the diamagnetic region ...... 104

31 A 500 MHz ‘H NMR reference spectrum and NOES Y map of oxidized SiR at 298 K. The NOESY map shows the correlation among the hyperfine-shifted signals in the downfield region of the spectrum ...... 105

32 500 MHz reference spectrum and magnitude COSY map of oxidized SiR. Crosspeaks are shown between signals c and d and signals g and i ...... 106

33 A 600 MHz TOCSY map of oxidized SiR showing correlations between signals belonging to the same spin system in the downfield region of the spectrum ...... 107

34 300 MHz 'H NMR reference spectrum of SiR (top) and NOE difference spectrum obtained by saturating signal b (bottom) ...... I ll

35 A portion of a 500 MHz WEFT-NOESY map of oxidized SiR showing crosspeaks between fast-relaxing resonances inside the diamagnetic region ...... 113

36 500 MHz paramagnetic *H NMR spectrum of (a) fully oxidized; (b) partially oxidized; (c) fully reduced SiR ...... 114

37 *H NMR spectrum of the fully reduced pentacoordinate high-spin sulfite reductase from E. coli ...... 116

3 8 Temperature dependence of the hyperfine-shifted resonances in fully reduced SiR, pH 7.5 ...... 117

xi NOE difference spectrum obtained by saturating the resonance at 113 ppm in fully reduced SiR ...... List of Abbreviations

ID One-dimensional

2D Two-dimensional

AQ Acquisition time

ATP Adenosine 5-triphosphate BRL Bethesda Research Laboratory

CD Circular dichroism

COSY 2D correlated spectroscopy

Cys Cysteine

DEAE Diethylaminoethyl

DFL Deazaflavin

DMF Dimethylformamide

DV Desulfoviridin

DW Dwell time

EDTA Ethylenediaminetetraacetate

ENDOR Electron nuclear double resonance

EPR Electron paramagnetic resonance

FPLC Fast protein liquid chromatography

KP Potassium phosphate buffer

MeV+- Methyl viologen radical Mr Molecular weight

xiii NH^OMe methoxylamine

NMR Nuclear magnetic resonance

NOE Nuclear Overhauser effect or nuclear Overhauser enhancement

NOESY 2D Nuclear Overhauser and exchange spectroscopy

PAGE Polyacrylamide gel electrophoresis

ROESY Rotating frame nuclear Overhauser enhancement spectroscopy

SDS Sodium dodecyl sulfate

SiR Assimilatory-type sulfite reductase

SiR-HP Sulfite reductase heme protein, P subunit (E. coli)

SW Sweep width or spectral width

TOCSY 2D total correlated spectroscopy

WEFT Water eliminated Fourier transform

xiv Table of Contents

Dedication ...... ii

Acknowledgements ...... iii

Vita ...... v

List of Tables ...... vii

List of Figures ...... ix

List of Abbreviations ...... xiii

Chapter Page

I. Overview ...... 1

II. Materials and Methods ...... 15

2.1 General Materials ...... 15

2.2 Bulk Growth of Desulfovibrio vulgaris (Hildenborough) ...... 15

2.3 Enzyme Purification ...... 18

2.3.1 Isolation of crude extracts from D. vulgaris ...... 18 2.3.2 Purification of assimilatory-type sulfite reductase (SiR) from crude cell extracts ...... 19

xv 2.3.3 Purification ofDesulfoviridin (DV) ...... 21 2.3.4 Determination of Enzyme purity ...... 21

2.4 Measurement of Optical Spectra of Reduced and Oxidized Forms of DV and Ligand Adducts ...... 22

2.4.1 Photoreduction of samples ...... 22 2.4.2 Instrumentation and Methods ...... 23

2.5 Pre-Steady-State Kinetic Studies on Desulfoviridin(DV) ...... 23 2.5.1 Measurement of binding rate constants (k2) for non-substrate ligands ...... 23 Preparation of samples ...... 23 Instrumentation and Methods ...... 25

2.5.2 Enzymatic reduction of inorganic substrates: Measurement of binding (k2) and reaction (fa) rate constants ...... 27 Preparation of samples ...... 27 Instrumentation and Methods ...... 27

2.5.3 Enzymatic reduction of inorganic substrates: Variable temperature measurement of reaction rate constants (kr) ...... 28 Sample preparation ...... 28 Instrumentation and Methods ...... 28

2.6 Steady-State Kinetic Studies on Siroheme ...... 28

2.6.1 Sample preparation ...... 28 2.6.2 Instrumentation and Methods ...... 29

2.7 Studies on the Assimilatory-type Sulfite Reductase (SiR) ...... 30

2.7.1 Kinetic and spectral characterization of recombinant SiR ...... 30 Instrumentation and Methods ...... 30

2.7.2 Paramagnetic ‘H NMR Studies on OxidizedSiR ...... 31 Sample Preparation ...... 31 Instrumentation ...... 31 NMR Methodology ...... 32

2.7.3 Paramagnetic NMR Studies on Fully Reduced SiR ...... 34 Sample Preparation ...... 34 Variable Temperature Experiment ...... 35 NOE Experiment ...... 35 xvi III. Electronic Properties of the Dissimilatory Sulfite Reductase (DV) from Desulfovibrio vulgaris (Hildenborough): Electronic spectral measurements of DV and ligand adducts. Comparative studies with related enzymes ...... 36

3.1 Introduction ...... 36

3.2 Results ...... 37

3.2 Discussion ...... 42

3.3 Summary...... 45

IV. Enzymatic Reduction of Inorganic Anions. Pre-Steady-State Kinetic Analysis of the Dissimilatory Sulfite Reductase (DV) from Desulfovibrio vulgaris (Hildenborough). Mechanistic Implications ...... 46

4.1 Introduction ...... 46

4.2 Results and Discussion ...... 48

Stopped-flow experiment ...... 48 Binding and reaction rates ...... 52 Implications for enzyme mechanism and a molecular understanding of the reaction pathway ...... 54 Role of proton transfer ...... 58 Summary ...... 59

V. Enzymatic Reduction of Inorganic Ions. Variable Temperature Steady-State and Pre-Steady-State Kinetic Experiments to Map the Energy Profile of an Enzymatic Multielectron redox Reaction. Application to the Dissimilatory Sulfite Reductase from Desulfovibrio vulgaris (Hildenborough) ...... 60

5.1 Introduction ...... 60

5.2 Results ...... 62

Steady-State Variable-Temperature Data ...... 62 Pre-Steady-State Variable-Temperature Data ...... 71

5.3 Discussion ...... 71

Activation Parameters ...... 73 Implications for Understanding the Reaction Mechanism ...... 76 Role of Siroheme Cofactor ...... 77 xvii Summary ...... 78

VI. Sulfite Reductase: Active site residues are “noncatalytic”. Comparison of reaction energetics for enzyme- and siroheme-catalyzed reduction of inorganic substrates...... 79

6.1 Introduction ...... 79

6.2 Results and Discussion ...... 80

VII. Studies on the Assimilatory-type Sulfite Reductase (SiR) from Desulfovibrio vulgaris (Hildenborough) ...... 87

7.1 Introduction ...... 87

7.2 Characterization of Recombinant SiR from D. vulgaris ...... 87

Enzymatic Activity ...... 8 8 Spectral Characterization of Recombinant Holoenzyme ...... 8 8

7.3 Paramagnetic JH NMR Studies on SiR ...... 91

7.3.1 General Principles ...... 93

7.3.2 Establishing Connectivities Between the Hyperfine-shifted Resonances in Oxidized SiR ...... 96

Results and Discussion ...... 97

7.3.3 Paramagnetic *H NMR Studies on the Fully Reduced SiR ...... 112

Results and Discussion ...... 112

References ...... 120

Appendix ...... 128

SUWEFT.AU ...... 128 Pulse/Processing Programs Used for NOE Experiments ...... 129

xviii Chapter I

Overview

The biological reduction of sulfate is an important pathway in the biological sulfur cycle, the balance of which is critical to life on earth [Figure 1]. In the biosphere, sulfate is the most abundant and energetically stable source of sulfur for organisms. All green plants, fungi and most bacteria reduce sulfate to the level of sulfide. The reduction of sulfate plays an important role in two equally important processes: respiration and production of sulfur containing nutrients. The respiratory pathway utilized by some types of anaerobic bacteria uses sulfate as a terminal electron acceptor in a manner analogous to the use of oxygen by aerobes. The end product, sulfide, is excreted in the environment.

In this process, the reduction of sulfate is called "dissimilatory sulfate reduction". In the second process, sulfide is used to make amino acids and their further metabolites essential to living organisms. Here, the reduction of sulfate is called "assimilatory sulfate reduction". Humans, animals and those bacteria that cannot reduce sulfate obtain essential sulfur containing nutrients through the food chain. On the other hand, the metabolic degradation processes in higher organisms or decomposing bacteria, help refurbish the sulfate supply in the biosphere by reoxidizing reduced sulfur from organic compounds.

1 2

Coloured sulphur bacteria, Thiobacillus. etc.

Sulphur oxidation

S°

Sulphur reduction: Desulfuro mom. s sulphate Campylobacter, reduction Desulfovibrio ) Desulfotomaculum / Dissimilatory Sulphide Desulfurococcus etc. oxidation Desulfobacter etc. ) Bcggiatoa, Yeast, bacteria,) . . > Assimilatory some thiobacillC plants coloured sulphur bacteria, etc. Sulphur Putrefaction assimilation

R S II

Proteolysis Protein synthesis

Protein

Figure 1. The biological sulfur cycle [Postgate, 1984], The dissimilatory sulfate reducing bacteria provide the substrate, S2‘, needed by sulfide-oxidizing bacteria which in turn converts S2* back to sulfate. In assimilatory sulfate reduction sulfide becomes incorporated into an amino acid (RSH) used to make plant or microbial protein which is then consumed via the food chain. Mammals take part in the oxidative half of the cycle.

Putrefaction is performed by bacteria and fungi [Huxtable, 1986; Postgate, 1984], 3 NH 2

O O CKp 1 O—1=0

a

OH OH

Figure 2. Adenosine-5'- phosphosulfate (APS), the activated form of sulfate in the sulfate reduction pathway.

Sulfate Reduction: Enzymatic pathways and bioenergetics. Assimilatory

and dissimilatory sulfate reduction use similar enzymes, co-factors, and prosthetic groups in each step of the reduction pathway. Because of the inertness of sulfate, the reduction process starts with the activation of sulfate to adenosine-5'-phosphosulfate, APS [Figure

2], The reaction is catalyzed by APS-sulfurylase. In equation (1) the formation

ATP + SO4 2- AlPtSujfiirylase.^ ysjjg + PPi ( 1) of pyrophosphate (PPi) is thermodynamically unfavorable. Thus, the reaction needs to be pulled to the right by another reaction that will consume PPi. This reaction is catalyzed by pyrophosphatase.

H2 O + PPi pyrophosphatase ) 2Pi (2) 4 APS is then reduced to sulfite as catalyzed by APS-reductase. Finally, sulfite is reduced to sulfide as catalyzed by sulfite reductase. Figure 3 shows a diagram of the dissimilatory reduction pathway in Desulfovibrio.

Schemes for the bioenergetics of sulfate reduction, although tentative and open for future research, have been proposed for certain groups of organisms based on current data available. Figure 4 shows the scheme proposed for the dissimilatory sulfate reduction of a well studied group of anaerobic bacteria belonging to the genus Desulfovibrio. The scheme involves a proton gradient that provides the driving force for the production of

ATP needed for the respiration process to proceed. The proton gradient is generated as follows. Hydrogen is first formed in the cytoplasm from lactate and pyruvate. The hydrogen then diffuses across the cytoplasmic membrane. Outside the cytoplasmic membrane it then gets oxidized by hydrogenase requiring cytocrome C 3 . The electrons produced by the oxidation process are then shuttled back inside the cytoplasm leaving the protons behind. In the cytoplasm, the electrons are used in the reduction of sulfate to sulfide. The proton gradient is thus established without necessarily translocating protons across the membrane . This scheme is supported by observations of H 2 production during growth in Desulfovibrio [Traore et al., 1981; Tsuji and Yagi, 1980], enzyme locations in the cell [Odom and Peck, 1981; Badziong and Thauer, 1980], and observation of vectorial electron transfer [Odom and Peck, 1981; Kroger, 1978; Krouse and

McCready, 1979], 5

H- 2 H

Oxidized ferredoxin Reduced ferredoxin or flavodoxin or flavodoxin

Cytochrome-c3 (Fe ) Cytochrome-c3 (Fe )

ATP +2e +6e SO APS SO \ AMP 2Pi

Figure 3. The dissimilatory sulfate reduction pathway [Togo and Oae ,1992] 6

MCM6RANC CYTOPLASM PERIPLASM

flH *+ S 0 (,=

8 e“

2 LACTATE

EC P 2 PYRUVATE

hyorogenase HYOROGENASE

AH

2 ACETATE AOP + Pi

2 CO ATP

Figure 4. A proposed scheme for the bioenergetics of growth of Desulfovibrio in lactate and sulfate [LeGall and Fauque, 1988]: hydrogen, produced by an internal hydrogenase, diffuses across the membrane and is recycled by a periplasmic hydrogenase; electrons are the only ones being allowed to cross the membrane. The proton gradient drives the synthesis of ATP. C3, tetraheme cytochrome C3 ; ECP, electron carrier protein. 7 Oxidoreductases. Oxidoreductases are metalloenzymes involved in redox reaction catalysis and are usually complex bioinorganic systems. They usually have more

than one prosthetic group in order to fulfill two distinct operations: a suitable sink for

reducing electrons and a site to bind and activate substrate(s) for the subsequent

combination with stored electrons. The strategic grouping of metal ions to form multi­

center systems increase the versatility of these metalloenzymes as redox compounds.

Oxidoreductases are the principal components of respiratory pathways and play

key roles in anabolic and catabolic reactions. Important examples of reactions catalysed

by this class of metalloenzymes include the following conversions: N 2 —> NH3

(nitrogenase); tT -> H 2 (hydrogenase); C0 2 -» CH4; S042' -» SO32' -»S2' (sulfate or

sulfite reductases); N 03' -> N 0 2‘ ->NH 3 or N 2 (nitrate or nitrite reductases); and

0 2 -» H20 (cytochrome c oxydase) [Bastian et al., 1988; Godden et al., 1991; Shapleigh

et al., 1992; Lui et al., 1993, 1994].

Sulfite reductase. Sulfite reductase as indicated above is an oxidoreductase and is

involved in the reduction of SO32' to S2' in both assimilatory and dissimilatory sulfate

reduction pathways. A variety of sulfite reductases have already been isolated. They

have been classified to three groups according to the sulfate reduction pathway they are

involved in [Moura et. al.,1988]. The "dissimilatory" sulfite reductases include

desulfoviridin from Desulfovibrio (D.) vulgaris, D. gigas, and D. salexigens [Lee and

Peck, 1971; Lee et al. ,1973; Czechowski et al., 1986]; desulforubidins from D.

baculatus strains Norway and DSM 1743 [Lee et al., 1973]; P-582 from

Desulfotomaculum (Dt.) ruminis and Dt. nigrificans [Trudinger, 1970; Akagi and Adams,

1973]; and desulfofiiscidins from D. thermophilus and Thermodesulfobacterium

commune [LeGall and Fauque,1988; Hatchikian and Zeikus, 1983]. "Assimilatory"

sulfite reductases include the sulfite reductases from Escherichia coli [Siegel et al.,1973] 8 and spinach [Krueger and Siegel, 1982]. The third group of sulfite reductases is called

"assimilatory-type" sulfite reductase. They are found in several strictly anaerobic bacteria

and are distinguished from the other two types of sulfite reductases because of their

monomeric, low molecular weight (~25 kDa) polypeptide chain and possesion of a low-

spin siroheme, a prosthetic group common to sulfite reductases. Examples are the

"assimilatory-type" sulfite reductases from/), vulgaris (Hildenborough) [Lee et al, 1973],

Methanosarcina barkeri [Moura et al., 1982], and Desulfuromonas acetoxidans [Lino et

al., 1985]. In vitro, in the presence of reduced methyl viologen as electron donor, both

the "assimilatory" and "assimilatory-type" sulfite reductases can catalyze the six electron

reduction of SO32' to S2' without formation of free intermediates [Siegel and Kamin,1968;

Prabhakararo and Nicholas, 1969]. The "dissimilatory" sulfite reductase enzymes, on the

other hand, reduce SO32' to S2‘ with accumulation of trithionite and thiosulfate [Kobayashi

et al., 1972; Lee and Peck, 1971], However, as cited by Huynh et al., there is evidence

which suggests that the production of these intermediates may not occur in intact cells

[Huynh et al., 1984].

Sulfite Reductase Prosthetic Center. All sulfite reductases so far isolated posses a

common prosthetic center: a siroheme (the site of substrate binding and activation)

coordinatively linked to a [Fe 4 S4] cluster (postulated to act as an electron sink) [Siegel et

al., 1974]. Siroheme is an iron-isobacteriochlorin [Figure 5 ]. The isobacteriochlorin

ring is doubly reduced and has eight carboxylic acid side chains [Scott et al., 1978], A

striking feature of isobacteriochlorin is its ease of oxidation and difficulty of reduction

compared to porphyrins or chlorins [Richardson et al. ,1979]. This raises the possibility that isobacteriochlorin in siroheme may participate as an electron source in the

multielectron reduction of SO32' to S2' [Richarson et al., 1979], The coupled siroheme-

[Fe4 S4] prosthetic site is also found in some nitrite reducing enzymes [Lancaster et

al.,1979; McRee et al., 1986; Janick et al., 1983]. COOH

CH, COOH

CH, CH5

CH, I CH, HOOC CH,

COOH HC CH

H CH;

HgC CH, I h 2c COOH I CH. HOOC HOOC

COOH

Figure 5. Structure of siroheme, an iron isobacteriochlorin (adjacent pyrrole rings reduced) [Scott et al., 1978], 10 Extensive studies on E. coli sulfite reductase provided a clear picture of the

coupled siroheme-[Fe 4S4] prosthetic centers. The enzyme is a complex hemoflavoprotein

(Mr~685,000) with subunit structure a gP4 [Siegel and Davis, 1974]. It is capable of catalyzing the six-electron reduction of SO32* to S2' and NOa'to NH3 [Siegel et al.,1974].

The p subunit, SiR-HP (M f- 64000), contains the siroheme-[Fe4S4] prosthetic group;

SiR-HP can be isolated from the rest of the hemoflavoprotein [Siegel and Davis, 1974],

It can catalyze the six-electron reduction of substrates in the presence of electron donors like reduced methyl viologen. The coordinative linkage between siroheme and [Fe 4 S4] cluster was demonstrated by Mossbauer [Christner et al.,1981; 1983], ENDOR (Cline et al. 1985; 1986] and EPR spectroscopy [Janick and Siegel,1982; 1983]. The data showed evidence of magnetic exchange coupling between the siroheme and the [Fe4S4] cluster in frozen samples of SiR-HP. The coupling persisted in various redox states of the enzyme in the presence or absence of exogeneous ligands. Evidence of coupling was also shown in liquid solutions via *H NMR spectroscopy [Kaufman et al., 1993 a,b]. EPR data is also consistent with a high-spin pentacoordinate siroheme. A model of the active site was first proposed by Janick and Siegel wherein a cluster Fe is coordinatively linked to the siroheme via the S of a cysteinyl cluster ligand [Janick and Siegel, 1982], X-ray crystallographic data at 2.8 A resolution are consistent with this model [McRee et. al.

1986]. The sixth coordination position appears unoccupied which makes it the most likely binding site for the substrate. The crystallographic data also indicates that the edge of the siroheme macrocycle is in Van der Waals contact with a cubane S atom of the cluster. This may provide another means by which the71 orbitals of the siroheme can interact with the cluster [Kaufman et al., 1993]. The data though lacks sufficient resolution to identify the bridging ligand; the electron density at or near the bridge atom site was consistent with a sulfur or oxygen but is not compatible with a large side chain from residues such as tyrosine or glutamic acid. Nitrogen bridging has been eliminated by

ENDOR results [Cline et al., 1985, 1986]. However, resonance Raman data suggests 11 that a sulfur atom maybe bridging the heme Fe to the cluster [Madden et al.,1989]. Recently, 'H NMR studies on unligated SiR-HP showed support for the sulfur of a cystein

acting as the bridging ligand [Kaufman et. al., 1993b]. Ostrowski et al. provided a more

explicit model of the active site consistent with x-ray and all spectroscopic data available

(Figure 6 ) [Ostrowski et al.,1989].

Studies in our laboratory and those of other workers (as discussed below) on

sulfite reductases from D. vulgaris (Hildenborough) fully support the generality of the coupled siroheme-[Fe 4S4] unit. Two types of sulfite reductases can be isolated from D. vulgaris: the assimilatory-type sulfite reductase (SiR) and the dissimilatory sulfite reductase, disulfoviridin (DV). Both can catalyze the six electron reduction of S 0 3 ^_ to

S^- and NO 2 " to NH 3 . SiR (M,-~ 23,500) is monomeric and contains one siroheme-

[Fe4 S4] center [Huynh et al, 1984], DV (Mr~ 224000) is hexameric (a2 p2y2) and contains two siroheme-[Fe 4S4] centers; the distribution of these centers over the a, P and y subunits are uncertain [Wolfe et al., 1994]. Evidence has been put forward for magnetic exchange coupling between siroheme and the [Fe 4S4] cluster on the active sites of both DV (EPR power saturation experiments) and SiR (EPR and Mossbauer data) [Huynh et al., 1984;

Sola and Cowan, 1990; Lui, 1994], EPR results are consistent with a high-spin pentacoordinate siroheme in DV [Wolfe et. al.,1994]. On the other hand, *H NMR, EPR and Mossbauer data for SiR are consistent with a low-spin hexacoordinate siroheme

[Cowan and Sola, 1990; Huynh et al., 1984]. The type of coordination as indicated by the spin state of siroheme is supported by model studies on synthetic complexes demonstrating that pentacoordinate ferric isobacteriochlorins are high spin and hexacoordinate ferric isobacteriochlorins are low spin [Stolzenbergh et al., 1981]. 'H

NMR studies on SiR suggest histidine as the sixth ligand on siroheme [Cowan and Sola,

1990], On the basis of the analytically determined Fe:S ratio « 1, it was suggested that inorganic sulfide may be the ligand bridging the siroheme and the [Fe 4 S4 ] cluster in SiR 12

Cys HN UN

Fe /I ■Fe

Fe

Fe

Fe

Cys

Figure 6 . The magnetically coupled [Fe 4 S4 j-siroheme prosthetic centers in the active site of E. coli sulfite reductase [Ostrowski et al., 1989]. [Huynh et al., 1984]. This is in contrast to the bridging cysteine thiolate proposed for the£.coli sulfite reductase [Madden et al., 1989; McRee et al., 1986; Ostrowski et al.,

1989], Results from 35S2' exchange experiments in our laboratory supports the idea of a sulfide bridge in SiR [Tan and Cowan, 1991]; it was found that one sulfur atom underwent exchange more rapidly than others, an observation supporting the presence of a unique sulfide functioning as a bridge. However, Tan and Cowan pointed out that the identity of the bridge unit is unlikely to be important with regard to the details of enzymatic reaction chemistry in the active site of SiR.

On the basis of available data presented above, our laboratory proposed a general structure for the prosthestic unit in the assimilatory-type sulfite reductase (SiR) and the dissimilatory sulfite reductase (DV) from D. vulgaris as illustrated in Figure 7. Figure 7. The proposed model for the prosthetic centers in the active site of desulfoviridin (DV) and assimilatory-type sulfite reductase (SiR) from D. vulgaris.

Desulfoviridin contains a pentacoordinate high-spin siroheme with no protein derived ligand, although the possibility of a coordinated water is not excluded. The axial ligand

(L) in the case of SiR is likely to be His [Cowan and Sola, 1990], In both cases the bridging ligand X is most likely S2' or HS' [Tan and Cowan, 1991; Wolfe et al., 1994], Chapter n

Materials and Methods

2.1 General Materials

Unless otherwise indicated, all chemicals were purchased fron Sigma or Aldrich

(molecular-biology grade). DNase I and streptomycin sulfate were purchased from

Boeringer Mannheim. Measurements of solution pH were made with an Accumet model

910 pH-meter equipped with a Coming semi-micro combination pH-electrode. Sephadex

G-200 gel filtration material was obtained from Sigma, and DE-52 ion-exchange resin was from Whatman. Hydroxyappatite was synthesized by following published methods

[Methods of Enzymology, vol. 21]. Deazaflavin was synthesized as described by Janda and Hemmerich (1976).

2.2 Bulk Growth of Desulfovibrio vulgaris (Hildenborough) Two types of media preparation were used to grow D. vulgaris cells: enriched Baar's

Medium [American Type Culture Collection medium no. 1249] (Table 1) and LS Medium

[Rapp and Wall, 1987] (Table 2). The first was used in our earlier preparations and the latter is now regularly used in our laboratoiy. Growth of cells was started by innocuiating stock cell sample in 1 0 -ml media contained in test tubes tightly fitted with rubber stoppers to prevent diffusion of O 2 (the tubes were degassed by purging for 15 minutes with argon or nitrogen prior to innoculation). The cells were incubated at 37 °C and allowed to grow for 24 hours (or more depending on how well the medium was degassed). The 10-ml cultures were then transferred to 2-liter screw-capped 16

Table 1. Composition of Baars' Medium for growing D. vulgaris cells [ATCC medium no. 1249]

Reagent Amount in grams per 1 L of solution

MgS04 1.0 Sodium Citrate 5.0 CaS04 1.0 NH4CI 1.0 Sodium Lactate 5.8 (60 % solution) Yeast extract 1.0 Fe(NH4)2(S0 4). 6H20 0.15 17

Table 2. Composition of LS medium for growing D. vulgaris cells [Rapp and W all, 1987]

Reagent Amount in grams per 1 L of solution

MgSC>4 1.0 NH4CI 2.0 K2HP04 0.50 Na2S04 4.0 CaCl2.2H20 0.04 EDTA 0.003 Yeast extract 0.125 Sodium Lactate (60% solution) 12.5 Na2S.9H20 0.25 18 Erlenmeyer flasks containing media degassed for at least 30 minutes. The cells were incubated until a healthy growth was seen [usually 24 hours if the medium was well degassed]. The 2-liter cell growth was then used to innoculate 50-liter carboys. Usually one liter cell growth was enough for one carboy. Each carboy was degassed for one hour prior to innoculation. It was possible to bypass degassing if the medium was immediately innoculated when it cooled down to ~37 °C after sterilization [at this point O 2 has not yet diffused into the medium; temperatures higher than 37 °C may kill the cells]. The carboys were capped tightly and incubated at 30-32 °C for 50 hours [the carboys were shaken several times during this period]. All media preparations contain either kanamycin (200 mg/ml) or neomycin (lOOmg/ml) to select growth of D. vulgaris cells. The cells were harvested using an Amicon DC20L concentrator. Typically, 280 to 320 g wet cells were isolated from 4 carboys. Wet cell pastes were stored in -80 °C refrigerator.

2.3 Enzyme Purification

2.3.1 Isolation of crude extracts from D. vulgaris cells

The initial workup of recombinant D. vulgaris cells followed literature procedures

[Huynh et al., 1984] with some modifications. Typically, 300 g of cell paste was processed as follows: For ease of handling, the cell paste was split into 150 g portions and placed in a 250-ml plastic beaker. A minimum amount of 50 mM Tris-HCl buffer

(pH = 7.6) was added (approximately up to the 150 ml mark) and the beaker was kept in ice at all times. The mixture was sonicated using a Fisher sonic dismembrator Model 300 with power set at 90%. Sonication was kept within one minute followed by cooling of the sonicator probe and the cell mixture for a few minutes. The procedure was repeated several times until the mixture becomes very viscous. DNase I was then added (5 mg/300 g cell paste). Sonication was continued until the mixture became veiy smooth and homogeneous (presence of clumps of cells was checked using a spatula). Streptomycin 19 sulfate (5 g/300g cell paste) was dissolved in a minimum amount of 50 mM Tris-HCl

buffer. The solution was neutralized to pH 7 and then added to the sonicated cell

mixture. The mixture was stirred for 15 minutes at 4 °C. The mixture was then

centrifuged at 13,200 x g in a Sorvall RC-5B refrigerated centrifuge for one to two hours.

The supernatant was removed and cetrifuged in a Beckman L5-75b ultracentrifuge at 144,000 x g for 1.5 hours at 4 °C. Meanwhile the cell pellet was resuspended in a

minimum amount of 50 mM buffer and sonicated to make sure all cells were broken. The

cell mixture was again centrifuged. When the supernatant was seen to be still highly

colored, the cell pellet was washed with equal volume of buffer and centrifuged a third

time. All supernatant were pooled together and ultracentrifuged.

2.3.2 Purification of assimilatory-type sulfite reductase (SiR) from crude cell extracts

The cell extracts, after utracentrifugation, were loaded in a DE-52 column

( 6 x 30 cm) equilibrated with 10 mM Tris-HCl buffer (pH = 7.6 at 4 °C). The column

was then washed with1 liter of 10 mM Tris-HCl to remove unbound proteins ( a grayish- green band and a red-orange band [cytocrome]). The buffer concetration was then

stepped up to 50 mM to remove a second cytocrome; this usually takes 1.5 to 2 liters of buffer. The second cytochrome must be removed completely because it binds tightly to the second column, hydroxyappatite (HA), and will be eluted together with SiR. A gradient of 100 to 150 mM Tris buffer [4 liters total volume] was then applied to the column to elute proteins that included SiR. The gradient was stopped after the third cytochrome came out. This ensured that all SiR had come out of the column. SiR may

spread out in all the colored fractions that came out of the column, thus, all colored fractions were pooled together and concentrated in an Amicon with a 10 K membrane.

The buffer concentration was lowered to at least 50 mM by exchanging with 10 mM Tris-

HCl and the volume of solution lowered to less than 50 ml. Meanwhile, a small HA 20 column was prepared. The column had to be prepared carefully to prevent tight packing

of HA that would slow down the running of the column (which usually starts in the middle

of loading the sample). First, HA was resuspended in 10 mM potassium phosphate (KP)

buffer. Only the particles that settled down very quickly at the bottom of the container

were used to make the column. The HA slurry was then carefully loaded in the column

and allowed to settle by gravity. Disturbance of the column was kept at a minimum. When cracks started to develop during settling of the H A , the top of the column was

carefully stirred and resuspended; it was then allowed to settle by gravity. Prior to

loading the sample, buffer was allowed to flow at a slow rate to check if HA had settled

completely. The protein sample was transfered to the column using a glass pipette. The

sample was allowed to flow at a rate a bit slower than 1 drop per second. One way that

slow down of flow rate was prevented, specially if the quality of the HA was not good,

was to stir the top of the column carefully and mix some of the HA with sample to allow

proteins to immediately bind to HA and then allow settling by gravity. The HA with bound protein prevented tight packing on top of the column; also the solution on top of the column became less viscous and flow became easier. This technique was good if the

sample volume was small. Another way of preventing tight packing of HA was to immediately mix a small amount of HA to the sample before transferring to the column . Again this was good if the sample volume was small (20 - 30 ml). A yellow colored protein come out during loading of sample. After loading 50 mM KP was applied to the column to wash down the remaining unbound protein and the third cytochrome that came out from the DE column. The buffer concentration was stepped up to 100 mM KP to remove other contaminants. A 300 mM KP buffer was then used to elute SiR. Most of the time, a greyish-green band was eluted with SiR. This made the color of SiR fraction greenish-brown. This protein also absorbs strongly at 590 nm and quantitation of SiR after HA can sometimes be misleading. The grayish-green band can be removed after 21 FPLC. The FPLC step used a gradient of 0-1 M NaCl in 10 mM Tris-HCl (pH = 7) on a Mono-Q column (1 x 10 cm) with a total running time of 40 min.

2.3.3 Purification of Desulfoviridin (DV)

The purification of DV was a continuation of the procedure above. Application of

150 mM Tris-HCl buffer was continued to remove brown colored fractions. The concentation was then stepped up to 250 mM to elute most of the DV (green fractions).

Some DV may still remain in the column but increasing the concentration of buffer to try to recover them will only result in elution of more contaminants. The DV fractions were concentrated by use of an Amicon with a 100 K membrane. The salt concentration was lowered and DV was again loaded onto a DE-52 column equilibrated in 50 mM KP. The column was washed with 100 mM KP to further remove the brown colored contaminants.

The DV was then eluted using 300 mM KP. The green fractions were pooled and DV was precipitated at 50% saturation of (NH^SO.*. The green precipitate was isolated by centrifugation. It was then redissolved in a minimum amount of 50 mM KP and then applied to a G-200 column equilibrated with 50 mM KP. DV was further purified by

FPLC on a Mono Q column (1 x 10 cm). This resulted in the resolution of 3 bands. A gradient method was used using two stocks of degassed KP buffer (A = 50 mM, B = 500 m M ), pH 7.6. The total running time was 40 minutes (5 min 11 % B, 17 min 11-100 % B, and 10 min 100% B).

2.3.4 Determination of enzyme purity

After FPLC purification, the purity of SiR or DV was determined by running SDS-

PAGE on a Phast electrophoresis system (Pharmacia/LKB). An aliquot of concentrated protein (2-3 ml) was treated with an SDS gel loading buffer solution (2 ml, prepared according to manufacturer's protocols) and boiled at 1 0 0 °C for 5 minutes. A 1 ml sample was applied to each lane of the gel using an automated gel loading applicator 22 (Pharmacia/LKB). Running and staining (Coomasie) conditions followed recommended preprogramed operating conditions. Molecular weight determinations were referenced to

commercially available standards (BRL, high molecular weight range).

2.4 Measurement of Optical Spectra of Reduced and Oxidized Forms of

Desulfoviridin (DV) and Ligand Adducts

2.4.1 Photoreduction of samples All samples were reduced using the deazaflavin /EDTA method of Massey and

Hemmerich (1978). Solutions containing 20 mM EDTA, 10 mM ligand , 6-17 mM DV, and 24-68 mM deazaflavin were used. The ratio of deazaflavin to DV was kept at 4:1 in all the samples. The samples were degassed with 0 2 -free argon in a 5-ml round bottom flask with a stirrer and fitted with a rubber septum. The solutions were degassed for 10 minutes and then transferred to an EPR tube or a glass cuvette using a gas tight syringe

(Hamilton). The syringe was made anaerobic by flushing with argon several times prior to withdrawing samples under a positive pressure of argon. Samples were then injected into an argon-purged EPR tube or cuvette capped with a rubber septum . The samples were further degassed for about 5 minutes after which the septum was sealed with grease and secured with parafilm to prevent O 2 diffusion. The EPR tube or cuvette was then transferred to a glassbeaker filled with ice/water and irradiated (1000 W lamp; 90% power) for 5-30 minutes. Reduction was monitored by electronic absorption, and irradiation was continued until no further change was observed in the optical specrum.

Control solutions of deazaflavin/EDTA and desulfoviridin (without ligand) were also prepared and reduced in the same way. Spectra of the oxidized solutions were taken before photoreduction. 2.4.2 Instrumentation and Method Optical spectra were measured on a Hewlett-Packard 8452A spectrophotometer

(run by software from On-Line Instrument Systems) using either a 1-cm-path-length glass

cuvette or EPR tube [average effective pathlength is 0.30 cm, determined using oxidized

desulfoviridin ( 8 5 3 2 5.6 x 10* M em-1) as a control]. For optical measurements the

concentration of desulfoviridin was - 6 mM when using cuvettes, or 17 mM when using

EPR tubes. In each case the ligand concentration was 10 mM. The EPR tube was

positioned in the light path using a holder fabricated from an aluminum block with a slit

width of 3 mm. Optical data obtained from either cuvette or EPR tube were found to be

consistent. The relative positions of the inserted EPR tube and the holder were marked

to achieve uniformity in the baseline position for each spectrum (a slight change in the

position of the EPR tube can lead to a change in the position of the entire baseline). The

spectrum of oxidized and reduced deazaflavin control solutions were subtracted from the

spectrum of oxidized and reduced desulfoviridin solutions respectively. Figure 8 shows

the spectra of deazaflavin (oxidized and reduced form) and desulfoviridin (oxidized and

reduced form without ligand). It is assumed that the absorbance of the reduced

deazaflavin in the enzyme sample will not differ significantly from the control sample. In

each case the concentration of EDTA was in great excess [EDTA/desulfoviridin ratio -1000-3000],

2.5 Pre-Steady-State Kinetic Studies on Desulfoviridin

2.5.1 Measurement of binding rate constants (k2) for non-substrate ligands.

Preparation o f samples. Titanium(III)citrate was used to reduce desulfoviridin.

A 6 mM stock solution of titanum(III)citrate was prepared anaerobically following the

method ofZehnder and Wuhrmann [1976], The stock solution was kept under positive pressure of 0 2-free argon. A 6 ml solution containing 12 pM desulfoviridin in 50 mM Figure Figure n5 Mptsimpopaebfe,p 7.6. buffer,pH phosphate potassium mM 50 in

ABSORBANCE 0 • 8 45 75 IS . Spectra of deazaflavin (oxidized and reduced) and DV (oxidized and reduced) reduced) and (oxidized DV and deazaflavinreduced) of and (oxidized Spectra . 5 30 5 60 750 600 450 300 150 H T G N E L E U A U 24 potassium phosphate buffer was degassed in a 25-ml-pear-shaped flask by purging with argon for 15 minutes with constant stirring. A gas-tight Hamilton syringe was flushed several times with argon and then used to transfer 4 ml of titanium(UI)citrate stock solution to the enzyme sample which is kept under positive argon pressure. Ligand solutions were prepared the same way as the enzyme. Buffer solutions used to prepare the samples were degassed prior to preparation of samples to minimize degassing time and prevent a significant change in concentration due to evaporation.

Instrumentation and Method. Data were obtained with an OLIS (On-Line Systems, Inc.) stopped-flow apparatus. A broadband 75-W xenon arc lamp source

(Ischio) powered by an OLIS-XL150 power supply was filtered through a monochromator (Model H10 by Instruments Sa.) with resolving power of 8 nm/mm. A photomultiplier tube (Homatsu) usable between 185 and 900 nm was mounted linearly from the source to detect at 438 or 554 nm. The piston gas (nitrogen) was delivered at a rate between 9 and 14 ml/s.

Two gas-tight syringes (Hamilton) were loaded with 10 ml each of argon-purged reactants as described earlier. Titanium(III)citrate was present in excess in both ligand and enzyme solutions to maintain the two-electron reduced enzyme and to keep both solutions 0 2 -free. Absorption changes were monitored at 498 and/or 554 nm. Rate constants were determined by use of the OLIS Operating System software (version 12.05) by fitting to proprietary software for one-exponential kinetics. Figure 9 shows the setup for the stopped-flow experiment. 26

Trigger S t o p -d >

Reservoir Oscilloscope

Syringes Photomultipier

3 -W ay t a p

✓

abs

t i m e

Figure 9. A diagram of the stopped-flow apparatus used for the pre-steady-state kinetic measurements [ Burgess, 1978]. 27 2.5.2 Enzymatic reduction of inorganic substrate anions: Measurement of binding

(k2) and reaction rate constants (kr).

Preparation o f Samples. A 5-ml volume of a solution containing 60 fiM

enzyme, 120 |iM deazaflavin and 15 mM EDTA in 50 mM potassium phosphate buffer

(pH 7.6) was degassed in a 10-ml pear-shaped flask by purging the surface of the stirred

solution for 30 minutes with 0 2-free argon. A gas-tight Hamilton syringe was pre­ flushed with argon and loaded with the degassed solution under a positive pressure of argon. The syringe mouth was fitted with a small serum stopper to prevent Oj-diffusion and subsequently immersed in ice-water and irradiated (1000-W lamp, 90% power) for 20 minutes. During preliminary experiments, reduction was monitored by electronic absorption spectroscopy and irradiation was continued until no further change was observed in the optical spectrum. A second gas-tight syringe was loaded in the same manner with the appropriate substrate concentration, without subsequent irradiation.

The stopped-flow apparatus was flushed with argon-purged buffer prior to each experiment. During the course of the experiment, the gas-tight Hamilton syringe containing the enzyme was irradiated for 2 minutes to ensure retention of a high fraction of the two electron reduced enzyme during the prolonged experiments. A water filter was placed between the syringe and the lamp.

Instrumentation and Method. Data was obtained with an OLIS stopped-flow apparatus as described above. Two gas-tight syringes (Hamilton) were loaded with 5 ml each of argon-purged reactants as described earlier. During the course of the experiment the syringes were occasionally irradiated in situ. Deazaflavin photoreduction of DV maintained the two-electron-reduced enzyme. Reactants were pre-equilibrated at 20 °C prior to mixing. Absorption changes were monitored at appropriate wavelengths. Rate 28 constants were determined by use of the OLIS Operating System software (version

12.05) by fitting to proprietary software for rise-fall kinetics.

2.5.3 Enzymatic reduction of inorganic substrates: Variable temperature

measurement of reaction rate constant (kr).

Sample Preparation. A 10-ml solution of deazaflavin/EDT A reduced

desulfoviridin was prepared exactly as described previously. The same earlier method

was also used to prepare substrate solutions.

Instrumentation and Method. The stopped-flow apparatus used was as

described earlier. The apparatus was flushed with argon-purged buffer prior to each

experiment. Reactants in the drive syringes were pre-equilibrated at the appropriate

temperature prior to mixing. Absorption changes were typically observed at 438 nm.

During the course of the experiment, the gas-tight syringe holding the enzyme was

irradiated for 2 min to ensure retention of a high fraction of the two-electron reduced

enzyme during the prolonged experiments. A water filter was placed between the syringe

and the lamp.

Rate constants were determined by use of the OLIS Operating System software

(version 12.050) by fitting to proprietary software for rise-fall kinetics.

2.6 Steady-State Kinetic Studies on Siroheme

2.6.1 Sample Preparation

Siroheme was extracted following standard methods [Siegel et al., 1978; Kang et al., 1987] with some modifications. A 1 ml sample of concentrated desulfoviridin in 50 mM KP (pH 7.6) was degassed by purging with 0 2-free argon. Nine volumes of ice- cold acetone/HCl (0.016 N HC1) were added while mixing vigorously. The mixture was kept in the dark for 5 minutes (inside the freezer —4 °C) and then quickly centrifuged at high speed to remove precipitated proteins. To keep contact with air to a minimum, the above preparation was done in a nalgene centrifuge bottle with a stirring bar and fitted with a rubber septum; the setup was kept at all times under argon. The rubber septum was replaced by a fitted cap prior to centrifugation. Acetone was removed from the siroheme solution under a stream of argon taking care not to dry the siroheme. A minimum amount of degassed 50 mM potassium phosphate buffer was then added and the final solution kept under argon throughout the course of the experiment. The concentration of siroheme in the stock solution was determined using the extinction coefficient (S 5 9 5 nm = 1-5 x 10 4 M ’ 1 cm-1) for the cyanide adduct [Wolfe et al., 1994],

2.6.2 Instrumentation and Method

Electronic absortion measurement and kinetic assay were done on a Hewlett-

Packard 8452A spectrophotometer run by software from On-Line Instrument Systems.

Kinetic data was obtained at 298 K by monitoring the decrease in absorbance at 600 nm from the radical cation methyl viologen (MeV+ ) used as an electron source for siroheme during turnover. The required amounts of stock solutions of siroheme and substrate were injected in a septum-stoppered glass cell under a stream of 0 2 -free argon; degassed buffer was added to the cell to give a working volume of 1.5 ml after the addition of MV+-. The solution was purged with argon for 20 minutes prior to addition of MV* . The volume of

MV+' was sufficient to give an initial absorbance at 600 nm o f-2.5 (e _ = 1.3 x 104 ° ' 600nm

M' 1 cm-1 [Thomeley, 1974]). The MV+> was prepared by zinc reduction of MV2+ : 35 mg of methy viologen and small pieces of zinc were placed on a 1 0 -ml round bottom flask;

8 ml of degassed 50 mM potassium phosphate buffer (at the desired pH) was added under a stream of argon. Kinetic measurement was made by following the rate of change of the 30 absorbance monitored at 600 nm. The resulting decay profile was analyzed using On-

Line Instrument Systems. Michaelis-Menten parameters (k ^ , were determined using commercially available software (Origin) to fit initial velocity plots versus substrate concentration. Two sets of control measurements were also made to make sure that the kinetic profile obtained is due only to the reduction of substrates by siroheme (control 1 : siroheme/buffer plus MV+-, control 2: substrate/buffer plus MV+ ).

2.7 Studies on the Assimilatory-type Sulfite Reductase

2.7.1 Kinetic and Spectral Characterization of Recombinant SiR

Instrumentation and Method

Circular Dichroism, Electronic Absorption and NMR spectroscopy. Circular dichroism spectra were obtained on a Jasco J-500C spectrometer, interfaced to a personal computer loaded with standard Jasco operational software. Sample concentrations were

9.6 mM enzyme in 20 mM Tris-HCl buffer (pH 7.6). To improve signal to noise ratio, each spectrum is the average of eight scans. An equivalent number of baseline scans were taken against Tris buffer. Data was obtained at 298 K. Electronic absorption spectra were measured on a Hewlett-Packard 8452A spectrophotometer run by software from

On-Line Instrument Systems. NMR spectra were measured on a 500 MHz Brucker AM

500 spectrometer using 16 K data points over a 125-KHz bandwidth. The super-WEFT sequence (180-t-90-AQ, see appendix for pulse program) [Inobushi and Becker, 1983] was used to suppress the water signal with recycle time of 70 ms and x of 60 to 65 ms.

A line broadening of 25 Hz was applied prior to Fourier transformation. Peaks are referenced to the residual water peak at 4.8 ppm. 31 Kinetic Data. Kinetic data was obtained at 298 K by monitoring the decrease in absorbance at 600 nm from the MeV+- radical anion as an electron source for the enzyme

turnover. The method used is exactly the same as discussed earlier for siroheme turnover.

Saturating concentrations of substrate were chosen to yield turnover rate directly. The

resulting decay profile was analyzed using software from On-Line Instrument Systems.

2.7.2 Paramagnetic *H NMR Studies on Oxidized SiR

Sample Preparation

SiR was isolated and purified following procedures discussed earlier. To keep the residual water at a minimum, the enzyme sample (in 50 mM potassium phosphate buffer, pH 7.6) was lyophilized without application of heat. The lyophilized sample was then re­ dissolved in 0.4 ml D20 (99.9%). The sample was then carefully transferred to a NMR tube with a screw-cap. To minimize enzyme degradation, the sample was degassed by connecting the NMR tube to an 0 2 -free argon line and the house vacuum line. The tube was alternately filled and evacuated (about 5 min each time) with argon while carefully agitating the enzyme sample during the whole process. This was repeated 5 to 10 times.

An additional 5 to 10 minutes was spent flushing the solution continuously with argon.

The cap was sealed with grease and wrapped with parafilm. The sample concentration for each experiment were between 0.7 to 1 mM.

Instrumentation

Spectral measurements were obtained on a 300 MHz Bruker AC-300 [NOE and 2D

COSY/NOESY], a 500 MHz Bruker AM-500 [NOE, 2D NOESY/COSY] and a 600

MHz Bruker Avance DMX-600 [TOCSY] NMR spectrometers. The instruments were carefully tuned before starting an experiment. The parameteres for data acquisition were included in figure legends (Chap. VII). 32

NMR Methodology ]H NOE Experiment. !H NOE measurements were obtained using a program utilizing the super-WEFT sequence (180-X-90-AQ + delay) (see appendix for pulse programs) and cycling the frequency of the decoupler. The decoupler was kept on during the delay time t and the frequency of the decoupler cycled according to the scheme a - (co

+ 5i) - a - (a - 6 2) where a is the frequency of the irradiated signal and 5 the off- resonance offset (5i and § 2 may or may not be equal). S 3 (power used to saturate a particular resonance) used was between 12L to 17L for resonances with linewidths between 150 to 200 Hz. For linewidths wider than 300 Hz, S 3 values between 1L to 3L were used. The NOE response was determined by taking the difference between the on- and off-resonance spectra. The reference spectrum was obtained by summing the off- resonance spectra. The programs used for the experiment are included in the appendix.

The data was processed by applying a 25 Hz line broadening prior to Fourier transformation.

COSY Experiment. Magnitude COSY experiments have been shown to be one of the best sequences for detecting connectivities between paramagnetically shifted resonances

[La Mar and de Ropp, 1993], The COSY crosspeak in a paramagnetic system is weak because fast T 2 relaxation short-circuits the full buildup of coherence. Partial cancellation of the broad components of the antiphase crosspeaks adds to the weakening of the crosspeak intensity. To detect the weak crosspeaks, emphasis was set on increasing sensitivity as follows : ( 1 ) maximum possible number of scans was taken per fj block which depends on the NMR time available; (2) fastest repetition rate was used that will allow the relaxation of the slowest relaxing hyperfine resonance; usually relaxation times

(AQ + delay during presaturation) equal to 1.5 Ti to 2Ti were used where Tj is the logitudinal relaxation time of the slowest relaxing line (the narrowest hyperfine-shifted peak); and (3) the fastest aquisition time was employed; usually acquisition time was set

equal to 2 T2 as the fastest limit; T2 is the transvere relaxation time of the slowest relaxing peak (the narrowest linewidth).

Data processing. Coherence develops during the detection period and maximum

coherence is optimally detected during an acquisition time equal to 2 T2 for both fi and f 2

dimensions where T 2 is the transverse relaxation of the faster relaxing proton of a pair [La

Mar and de Ropp, 1993], Optimum processing was obtained by determining the number

of ti blocks (Nb) and t2 points (Np) corresponding to tAQ = 2T2 for the faster relaxing

proton of a pair and applying a pseudo-echo filter over this fraction of the collected data.

Apodization over this fraction improved crosspeak intensity for proton pairs with similar

T2 ’s. Since T 2 varies over a wide range among the hyperfine-shifted signals, different sets

of ti blocks and t 2 points were processed in order to detect all crosspeaks. A 0°-sine-belI- squared apodization was used over all the fractions processed. Np and Nb were determined using equations 3 and 4

2T2 = AQ (t2) = (DW2) Np = 1/2SW2 (Np) (3)

2T2 = AQ (tO = (DW,)Nb = 1/2SW! (Nb) (4)

where SWi and SW2 were the sweep widths in the fi and f2 dimensions, respectively.

To determine the aquisition time, T 2 was determined according to the relationship

A = l/7tT2 (5)

where A is the linewidth at half height of the broader line of a proton pair. 34 NOESY experiment. A phase sensitive NOESY experiment was used with

presaturation to eliminate residual water signal. The dipolar correlation develops during

the mixing time, t™, and optimal intensity is observed if t„, is of the order of Tj, the

relaxation time of the faster relaxing proton of a pair. As in the case of T2, T! values vary over a wide range in paramagnetic systems. In SiR it varies from 2 ms to 20 ms. Thus,

several mixing times had to be used in order to detect all of the crosspeaks. In the case of

SiR, all crosspeaks were detected at tm equal to 8 ms. At tm = 15 ms, crosspeaks between

fast-relaxing lines (4 to 6 ms) were weakened. The data were processed using 30°-, 45°-

or 60° -sine-bell-squared window function prior to Fourier transformation.

WEFT-NOESY. The pulse sequence for the NOESY experiment above was

slightly modified by replacing the presaturation sequence with the WEFT sequence (180 -

x - 90) at the start of the pulse program. The value for x was determined by initially

acquiring a ID spectrum using the super-WEFT sequence. The mixing time used was 10

ms. The data was processed as described earlier for the NOESY experiment.

2.7.3 Paramagnetic NMR Studies on Fully Reduced SiR

Sample Preparation.

The SiR sample was prepared as discussed earlier the only difference was the addition of deazaflavin (DFL ~ 3mM) and 10 mM EDTA to reduce the protein. To increase the solubility of DFL, the stock solution was prepared using deuterated DMF as solvent.

After thorough degassing, the sample was placed in a 4-liter beaker filled with ice/water and irradiated using a 1000 W lamp with the power set at 90% . The irradiation was continued until there was no change in the intensity of the absorption at -700 nm. The whole spectrum cannot be monitored because the concentration of the sample was too high. Typically it takes 2 hours to fully reduce the sample. 35 Variable Temperature Experiment !H NMR spectra were measured on a 500 MHz Bruker AM 500 spectrometer. The data was acquired using the super-WEFT sequence over a 125 KHz banwidth using 16K data points. The x used was between 60 to 65 ms and the relaxation sequence (AQ + delay) was equal to 60 to 70 ms. The recycle time was around 140 ms. The behavior of the paramagnetic resonaces (shifts) was monitored with varying temperature

(from 5 to 25 °C). The sample was allowed to equilibrate to the desired temperature before data acquisition. Typically, 80,000 scans were required to get a good signal to noise ratio. Data was processed by applying a 25 Hz line broadening prior to Fourier transformation.

NOE experiment. The method used was as described earlier for oxidized SiR. Chapter m

Electronic Properties of the Dissimilatory Sulfite Reductase (DV) from Desufovibrio vulgaris (Hildenborough): Electronic spectral measurements of DV and ligand adducts. Comparative studies with related enzymes.

Introduction The electronic spectral characteristics of the assimilatory sulfite reductase from E. coli have been extensively characterized, and similar optical spectra have been reported for the sulfite- and nitrite-reducing enzymes from Salmonella typhimurium and spinach, respectively [Siegel et al., 1973; Murphy et al., 1974; Janick and Siegel, 1983]. There is substantial amino-acid-sequence similarity among the enzymes from E. coli, S. typhimurium and spinach [Ostrowski et al., 1989a,b]. Recently, our laboratory reported the primaiy protein sequence for the assimilatory-type sulfite reductase from D. vulgaris

[Tan et al., 1991], This smaller enzyme (Mr ~ 23500) displayed no strong sequence similarity with that of theE. coli class of enzymes and also possessed distinct optical characteristics (Huynh et al., 1984). Also, the optical spectrum could be distinguished from the larger dissimilatory enzyme (DV) also isolated from D. vulgaris [ Lee et al.,

1973a, b]. In each of these enzymes there is strong evidence for one or more of an indentical set of coupled ^ 4 8 4 ]- siroheme units [Siegel et al., 1982; Cowan and Sola,

1990; Tan and Cowan, 1991; Wolfe et al., 1994]; however, there are substantial variations in electronic properties over the series. 36 In this chapter, we compare and contrast the electronic spectral properties of the

dissimilatory sulfite reductase, desulfoviridin (DV), from the sulfate reducing bacterium

D. vulgaris with published results for related enzymes, and the E. coli sulfite reductase in particular. The coupled redox center in such enzymes displays a remarkable diversity of optical features and redox chemistiy and so no single enzyme can be used as a satisfactory paradigm for understanding the electronic character of the rather unique redox center in this class. The influence of protein side chains on heme spectra is well documented in the literature of heme proteins [Smith and Williams, 1970; Cowan and Gray, 1989; Difeo and

Addison, 1991], and it is likely that the observed differences in spectral behavior and redox potentials are induced by variations in the protein environment surrounding each coupled center. The implications of this observation for the comparison of functional activity have yet to be resolved.

Results

In addition to the primary absorbance bands around 380-420 nm and 580-630 nm,

Figure 10a shows that desulfoviridin exhibits a weak absorbance ~ 702 nm

(e = 11 mM ! cm '). By analogy with E. coli sulfite reductase we ascribe this to a ligand- metal charge transfer [Janick and Siegel, 1982], This shoulder disappears after two- electron reduction, but is still prominent (e = 7.8 mM 'em') in the one-electron-reduced enzyme [reference spectra obtained from EPR samples judged to be one electron reduced

( work in our laboratory by S.M. Lui)]. A low-intensity band is seen around 498 nm as a shoulder on the absorbance tail from the Soret region. As described in more detail below, this band became more prominent after reduction (Table 3). A very weak absorption at

540 nm also appears after reduction. The 498 and 540 nm bands most likely correspond to weaker components from the Q-band absorption envelope (Figure 10a) [Adar, 1978;

Gouterman, 1978]. 1.05 Oxidized 0.90

0.75

0.60 x 3

Reduced v 0.45

0.30

0.15

300 450 600 750 1.4

1.2 Reduced + AsOo 1.0

0.8 x 3

0.6

Reduced V 0.4

0.2

300 450 600 750 W avelength

Figure 1 0 , Absorption spectra of oxidized and fully reduced DV (a), and reduced DV with 10 mM ASO2 " (b). Spectra were taken in a 1-cm pathlength optical cuvette with 6 mM enzyme in 50 mM potassium phosphate buffer, pH 7.6, at 298 K. Table 3. Comparison of Optical Characteristics for Oxidized and Reduced Sulfite Reducing Enzymes from Prokaryotic

and Eukaryotic Sources

ox or

a Enzyme (source) red X (nm) [e (mM^cm*1)] Reference

DV (D. vulgaris) ox 392 [130], 410 [140], 584 [33], 632 [56], 702 [11] This work

red 382 (sh)b [120], 390 [120], 410 [140], 498 [31], 540 [25], 584 [29], 632 [56] This work

DV-A s0 2- red 382 [120], 392 [120], 410 [150], 498 [31], 540 [28], 584 [22], 632 [53] This work DV-CN- red 392 [115], 510 [128], 498 [33], 588 [30], 632 [55] This work

C DV-C1- red 382 [120], 392 [130], 410 [140], 498 [31], 588 [29], 632 [57] This work

DV-I- red 382 [120], 392 [120], 410 [120], 498 [34], 588 [32], 632 [55] This work

DV (A gigas) ox 380(sh)b [172], 390 [182], 408 [200], 583 [42], 628 [84] d

red Results not available

SiR (E. coli) ox 278 [107], 386 [ 6 6 ], 547 [12], 591 [18], 714 [ 6 ] e

red 397 [49], 608 [14] e

SiR-CN- (E. coli) ox 386 [ 6 6 ], 549 [12], 596 [18], 714 [ 6 ] f

red 401 [48], 416 [49], 544 [39], 578 (sh) [38] f Table 3 continued

SiR (spinach) ox 278 [117], 384 [59], 542 [12], 587 [18], 712 [ 6 ] g

red 396 [42], 612 [13] g

NiR (spinach) ox 276 [72], 386 [40], 573 [10], 640 (sh) [5], 690 [4] h

red 394 [ 8 ], 557 [2], 588 [2] h

* b Note: DV, desulfoviridin (dissimilatory sulfite reductase); SiR, assimilatory sulfite reductase; NiR, nitrite reductase; sh, shoulder; C with the exception of I" the other halides (Br‘, F") gave similar absorbance results; these ligands are unlikely to bind to siroheme; d , * . ■ r g Lee and Peck (1971); Janick and Siegel (1982); Janick and Siegel, 1983; data interpolated from reported absorbance ratios in h Krueger and Siegel (1982 a, b); data interpolated from spectra in Wilkerson et al. (1982) and Vega and Kamen (1977).

Abbreviations: ox, oxidized; red, reduced. 41 There is minimal change in the absorbance of the oxidized chromophores after the

addition of the exogenous ligands (NO 2 ', S032', Se0 3 2', N3-, Cl-, CN-, As02-, B r, I-,

NCS-). Relatively small reductions (< 2%) in the absorbance of all bands (including the

weaker charge transfer and Q-bands) were noted. EPR measurements demonstrate that

these perturbations in optical spectra do not arise from axial ligation to the free

coordination site (Figure 7, Chap. I ), since the siroheme remains high spin [Lui et al.,

1994], Accordingly, we attribute these facts to random weak binding to the surface that

result in minor conformational perturbations in the vicinity of the siroheme.

Desulfoviridin was fully reduced within 1 0 minutes irradiation ( 1 0 0 0 W lamp, 90% power)

in the presence of deazaflavin/EDTA. Although the optical and EPR spectra of oxidized

enzyme were essentially unchanged over a period of several hours in the presence of

ligands noted above, rapid ligand binding has been observed in the reduced form (as will

be shown in Chapter IV). Reduction results in several notable changes in the optical

spectra that include the disappearance of the shoulder at 702 nm (charge-transfer band;

Figure 10 a, b). The absorbance bands in the 380-420 nm region also decrease in

intensity relative to the reduced native enzyme. In particular, relative to the oxidized

enzyme, the reduced form shows a very pronounced decrease in the absorbance tail from

the Soret region that extends to the Q-bands (either with or without exogenous ligand).

This is a particularly useful region to probe the kinetics of oxidation/reduction and ligand

binding by optical methods as demonstrated in the following chapters. The absorbance

of the 580 nm Q-band is reduced by 19% and is broadened in the reduced form. The Q-

bands noted previously as shoulders on the Soret absorbance also decrease slightly in

prominence. However, binding of As02‘ resulted in several deviations that are

characteristic for 7 t-acceptor ligands bound to reduced hemes (Table 3; Figure 10b). In

Figure 10b, binding to the reduced form is accompanied by a 16% increase in the

A 4iq/A 3 92 ratio relative to reduced native enzyme. Also, a 14% increase in intensity is

observed for the 540 nm band. Inasmuch as As02‘ is a reasonable ligand analogue of 42 SO3 2* and NC^-, it is likely that a similar spectral change accompanies binding of the

substrate anions. Overall the changes reflect increased 7 C-7 C character in the transitions for

the reduced chromophore and are commonly observed in heme spectra; for example, in

contrasting the optical spectra from ferric myoglobin with the ferrous derivative and

ferrous-CO adduct, there is a prominent loss of charge-transfer character with n -n

transitions from the heme ligand dominating [Adar, 1978; Gouterman, 1978],

Discussion

Comparison of the data in Table 3 illustrate the sharp contrasts in optical

characteristics for a variety of enzymes that carry the coupled -siroheme prosthetic

center. The family of enzymes represented by the E.coli enzyme show significant spectral

changes when comparing oxidized versus reduced, and ligated versus unligated enzyme

[Siegel et al., 1973; Janick and Siegel, 1982, 1983). The broad lines and uncharacteristic

absorption bands (relative to the metal-free isobacteriochlorin) reflect substantial n-d

(ligand-metal) orbital interactions and charge transfer character. There have been some

reports that certain enzymes in this class may contain demetallated siroheme

(sirohydrochlorin) [Moura et al., 1988; Pierik and Hagen, 1991]; however, work from our laboratory would suggest that this conclusion is flawed [Wolfe et al., 1994]. We restrict our discussion to those cases where it is known through the work of others and ourselves that the enzyme contains fully metallated siroheme. This includes desulfoviridin and sulfite reductase isolated fromZ). vulgaris, D. gigas, E. coli and S. typhimurium, and spinach sulfite and nitrite reductases. The spectrum of enzyme-free siroheme displays prominent changes with pH and ligands [Kang et al., 1987]. These variations are not obvious for the enzyme-bound complex, since the bridging axial ligand on one face is fixed

(sulfide or cysteinate), while the peripheral carboxylate groups are undoubtedly involved in the formation of salt bridges (Figure 7, Chap. I). In contrast with E. coli sulfite reductase, where reduction results in significant changes in the absorption bands [Janick 43 and Siegel, 1982], although without added ligand there is no sharp distinction between the one- and two-electron reduced samples, desulfoviridin is relatively insensitive to both oxidation and coordination states.

The optical spectrum of the assimilatory-type sulfite reductase (SiR) from D. vulgaris shows the greatest similarity with the corresponding assimilatory enzymes isolated from E. coli, S. typhimurium and spinach (Figure 11), despite the fact that SiR posses a low-spin six-coordinate siroheme and the others are high-spin five coordinate

[Janick and Siegel, 1982; Huynh et al., 1984]. All show broadened optical transitions with ill-resolved shoulders. The spectra of these assimilatory enzymes stand in contrast with those observed for dissimilatory enzymes, of which desulfoviridin (DV) is a typical example. Again, these variations do not result from distinct coordination states inasmuch as almost all assimilatory and dissimilatory enzymes possess a five-coordinate siroheme.

There is no evidence for ligand binding to the oxidized siroheme in DV, even over a period of one week. The optical spectrum of the reduced enzyme (in any coordination state) is relatively insensitive to ligand binding, and only slight changes are observed in the absorption characteristics of the enzyme. The appearance of the optical spectra suggest minimal "7 t-d" orbital overlap for the siroheme. The ligand-metal charge-transfer band

[-702 (shoulder) nm], characteristic of high-spin heme, disappears after two-electron reduction. EPR spectra show no evidence of features at unusual g-values for either oxidized or reduced enzymes [Lui et al.,1994]. o.e

oflj S 0.6 § .ata e Xiw < 0.4 -<

0.2

0.0 300 400 500 600 700 Wavelength, nm Wavelength, nm

0.8

3 .5 0.6 o u aO ccc c «c .au £ 0.4 ocn < X!

0.2

380

Wavelength, nm

Figure 11. Electronic absorption spectra of nitrite (a) and sulfite (b) reductases from spinach; sulfire reductase from K coli (c); and assimilatory-type sulfite reductase (SiR) from D. vulgaris [Siegel et al., 1982; Krueger and Siegel, 1982; Vega and Kamin, 1977; this work]. 45 Summary

There are obvious differences in both the electronic and coordination states of the

siroheme-[Fe4 S4 ] prosthetic site in this class of enzyme, and so no one enzyme can be used as a satisfactory paradigm for understanding the electronic character of the rather unique redox center in this class.

DV gives rise to optical characteristics that are relatively insensitive to both the oxidation and coordination state of the siroheme in contrast to E. coli sulfite reductase. Its oxidized spectrum differs from E. coli sulfite reductase although both have the same spin and coordination states. On the other hand, SiR, although different in spin and coordination state, shows a greater similarity with the optical spectrum of E.coli sulfite reductase and related enzymes. Clearly the physicochemical properties of the coupled siroheme-[Fe4 S4 ] redox unit common to this enzymes are varied by the local protein environment in a manner which has yet to be established. Chapter IV

Enzymatic Reduction of Inorganic Anions. Pre-Steady-State Kinetic Analysis of

the Dissimilatory Sulfite reductase (Desulfoviridin) from Desulfovibrio vulgaris

(Hildenborough). Mechanistic Implications.

Introduction

The field of mechanistic inorganic biochemistry remains at an early stage of

development. Mechanisms for selective binding and activation of inorganic anions or

gaseous molecules by metalloenzymes may differ in several important respects relative to

common examples of nonredox enzymatic catalysis on organic substrates. The

metalloredox prosthetic center both binds and catalytically activates the substrate, while

neighboring protein side chains may regulate and optimize the electronic and coordination

properties of the prosthetic metal sites toward these functions. Ionizable residues may further contribute to substrate recognition, binding, and activation either by serving as proton donors or providing electrostatic stabilization of charged substrates or intermediates.

With this issues in mind, our laboratory has targeted the p^S^-siroheme-containing sulfite-reducing enzymes (SO 32* —> HS") from Desulfovibrio vulgaris (Hildenborough) for detailed study. This enzyme class also catalyzes the six-electron reduction of 46 47 N 0 2‘ -» NH3, although a larger range of prosthetic redox centers and pathways are available for the reduction of nitrite: [I^S^-siroheme (N 02* -» NH3), hexaheme

(N 02" NH3), copper ion (N 02‘ —> N2, N 2 0 ) [Camack et al., 1978; Moura et al., 1986;

Averill and Tiejde, 1982; Godden et al., 1991; Vega et al., 1977; Henry and Bessieres,

1984; Petratos et al, 1986]. The chemistry of the electron-rich iron isobacteriochiorin

(siroheme) has been addressed in model studies [Richardson et al., 1979; Chang and Fajer,

1980; Barkigia et al., 1982; Chang et al.,1981; Procyk and Bocian, 1991; Stolzenberg et al., 1980, 1981; Strauss and Holm, 1982; Sullivan et al., 1991; Melamed et al., 1991; Kang et al., 1987], Inasmuch as the N 02' reduction pathway proceeds via intermediates that can be studied as discrete substrates, this provides a useful probe of the general mechanistic features associated with this class of enzyme. Previous studies have focused extensively on structural issues; especially elucidation of the coordination details of the prosthetic redox centers [Averill and Tiejde, 1982; Godden et al., 1991; Vega and Kamen,

1977; Henry and Bessieres, 1984; Petratos et al., 1986; McRee et al., 1986; Christner et al., 1983; Janick et al., 1983], Mechanistic schemes have been proposed only for the copper-containing dissimilatory nitrite-reducing enzymes from denitrifying bacteria, following use of isotopically labelled substrates to elucidate reaction pathways [Averill and

Tiejde, 1982; Godden et el., 1991; Vega and Kamen, 1977; Henry and Bessieres, 1984;

Petratos et al.,1986; Weeg-Aerssens et al., 1988; Scott et al.,1989]. To our knowledge such schemes have not been tested by kinetic studies. In this work, we measured microscopic rate constants for ligand-binding and bond-cleavage steps obtained from a pre-steady-state kinetic analysis of the reductive pathway for the desulfoviridin catalyzed reaction of both S032- and N 02‘ and putative reaction intermediates. 48 Results and Discussion

The following examines the reaction pathway for reduction of substrate molecules possessing N -0 and S-0 bonds under pre-steady-state conditions and evaluate substrate- binding and bond-cleavage chemistry. Work involving non-substrate ligands were done by myself while work involving substrate ligands were shared with S. M. Lui.

Stopped-flow experiment. Absorbance wavelengths used in stopped-flow experiments give rise to prominent change following ligand binding for non-substrate ligands (Figure 12 , Table 3 [Chap. Ill]) and ligand binding plus subsequent redox reaction for substrate ligands (Figure 13).

Addition of excess titanium(III)citrate to 12 pM DV yielded the 2 electron reduced enzyme that readily binds exogenous non-subtrate ligands like As02" and HS". Ligand binding resulted in absorbance changes that were in turn used to monitor the binding event. Excess ligand was reacted with reduced enzyme in the mixing chamber of a stopped-flow instrument under pseudo-first-order conditions. Anaerobic conditions were maintained by the presence of excess titanium(UI)citrate in all solutions before and after the binding event. Second-order ligand binding rate constants (k2) were evaluated by variation of the observed rate constant, k ^ , with ligand (L) concentration (equation 3).

Figure 14(a,b) shows examples of k ^ vs. [L] plots.

kobs = k2[L] (6)

Deazaflavin photoreduction of 60 pM desulfoviridin also yielded the two-electron- reduced enzyme [Massey and Hemmerich, 1978; Janda and Hemmerich, 1976; Yoneda et al., 1976], Unlike the reduction method described above, the present method allowed 49

1 ------— ------,—

h J j 0 W nil 11 L l \ h AA A ^ i f v v w y - V V V i / v V 10,~3

-4 • I 1 4

2 ou z

X - 3 10 J -4 0 2 4 6 8

SECONDS

Figure 12. Typical fit to a one-exponential rate profile for non-substrate ligand binding.

The data shown was taken for a final [DV] = 6 pM and [AsQ 2 -] = 10 mM. N0-= lOOmM 02-]= [N substrate anion. The data shown were taken for a final [DV] = 30 pM and pM 30 = final a [DV] for taken shown were data The anion. substrate Figure 13. Typical fit to a rise-fall rate profile for binding and subsequent reduction of of reduction rise-fallsubsequent binding profile a for and fitrate Typical to 13. Figure Absorbance Residuals (x103) . 6 3 -.8 .S 7 - .s r • 07S -IS T Time (sec) Time . . Z2S T T —«'YVA'< • 37S 50 40 51

30

w 20 H £

0.000 0.002 0.004 0.006 0.008 O.OtO 0 .0 1 2

( A s 0 2 ], M

2.0 H

1.5 H

I 09 1.0 H

0.5 -I

o.o H

0.0 0.2 0,4 0.6 0.8 1.0 1.2 (HS ] , M

Figure 14. Plot of versus [L] for L = As02" (a) and L = HS" (b). The pre-steady- state kinetic measurements were taken with final (DV] = 6 pM. a limited supply of reducing electrons which made it possible to monitor elementary

steps of the reduction pathway. Substrate binding and subsequent reaction to regenerate

oxidized siroheme result in absorbance changes that were used to monitor the two-

electron reduction of SC> 3 2' and N 0 2 - and intermediates. Excess subtrate was reacted

with reduced enzyme in the mixing chamber of a stopped -flow instrument under pseudo-

first-order conditions. Both binding (k 2 ) and reductive (kr) steps were directly monitored

and the resulting optical traces fit by rise-fall kinetic profile (Figure 13). Second order

rate constants (k2) were evaluated as described earlier.

Binding and reaction rates. The kinetic parameters detailed in Table 4 suggest that substrate binding to reduced enzyme is generally rapid and not rate limiting. An

exception is found with lower concentrations of NH 2 OH. The dependence of on-rates

(k2) on ligand nucleophilicity suggests that binding to the siroheme is associative in character: k2 (S032) ~ 4.3 x 103 M ' 1 s'1, k2 (N02-) ~ 3.6 x 103 M' 1 s'1, k2 (NO) ~ 7 x 10 5

M’ 1 s_I, k2 (NH2 0 H) - 24 M' 1 s '1. Moreover, if H 2 O is bound to the axial site of the

"pentacoordinate" high-spin siroheme (Figure 7, Chap. I), its release is not rate limiting.

On-rates for ASO 2 " and HS’ [k 2 (As0 2 ’) ~ 3 x 1 0 3 M' 1 s'1, k2 (HS") ~ 1.8 M ' 1 s'1] are consistent with the dominance of 7t-acceptor or o-donor properties, respectively, of substrate molecules. Although ASO 2 ' has the ability to both a-donate and 7 t-accept, the latter appears to dominate . 1 The electron-rich siroheme ring promotes x-backbonding through the d-orbitals of the ferrous ion. It is likely, in fact, that 7 t-backbonding of nitrite and sulfite plays a crucial role in the molecular mechanism of catalysis: promoting tight binding of substrate relative to product, while population of antibonding N -0 or S-0 orbitals results in a weakening of the bonds that are to be reductively cleaved. The large

1 This is consistent with the coordination behavior of other potential o-donor/ 7t-acceptor ligands such as CN~ and CO [unpublished results from our laboratory]. 53

Table 4. Pre-Steady-State Kinetic Parameters for Desulfoviridin3

Substrate / Ligand k2 (M_1 s -l)b M s -1)

S032- 4.3 x 103 (1) 12 (2’) n o 2- 3.6 x 103 (1) 14 (2) NO 7x105 (3) 6.5 (4) n h 2o h 24 (5) 9 (6) A s0 2- 3x103 HS" 1.8

a Stopped-flow kinetic data for substrates and ligands were obtained as described in the legends of Figures 12 and 13 with the following solution conditions: N 0 2", 438 nm,

1-200 mM; NO, 438 nm, 0.1-1 mM; NH 2 OH, 438 nm, 1-250 mm; SO 3 2", 438 nm, 10- 250 mM; As02", 554 and 438 nm, 0.3-10mM; HS", 438 nm, 10-1000 mM. Errors in each measurement are estimated to be 50 ± %. b Numbers in parenthesis after each rate constant correspond to the reaction step indicated in Figure 16. value of k 2 (NO) ~ 105 M' 1 s' 1 for NO binding is consistent with on rates determined for

other heme proteins [hemoglobin(a/p-subunits) ~3 x 10 7 M_1s*1] [Olson, 1981],

Steady-state kinetic parameters for N 02‘ and NH2OH have been determined

previously from our laboratory [Wolfe et al., 1994] and here we extend this to other

substrate molecules. With the exception of the case of NH 2 OH, reaction rates (kr) show

minimal correlation with steady- state turnover [kr(S 0 3 2') ~ 12 s '1, kcat(S032') ~ 0.3 s'1;

kr(N 02') ~ 14 s'1, kcatCNO^) - 0.04 s'1; kr(NH2 OH) ~ 9 s' 1 ^ ( N H ^ H ) ~ 30 s'1]

[Wolfe et al., 1994] suggesting that the bond-cleavage steps are not rate limiting in the 2 early stages of the reaction. For hydroxylamine, kr ~ k ^ and the enzyme saturates when

[NH2 OH] > 250 mM (Km ~ 46 mM) [Wolfe et al., 1994]. Both observations are

consistent with the second-order binding constant, k 2 ~ 24 M' 1 s'1.

Implications for enzyme mechanism and a molecular understanding o f the reaction pathway. In the context of the reaction previously outlined for S032' reduction to HS'

(Figure 15) [Tan and Cowan, 1991], the corresponding model for reduction of N0 2‘ is illustrated in Figure 16. That each of the substrates employed (N 02', NO, NH 2 OH) reacts readily lends credence to this mechanistic interpretation of the results. We view the reaction as arising through three two-electron reductive cleavages of N-0 (or S-O) bonds

[Tan and Cowan, 1991], As detailed later, there is a requirement for a number of distinct proton-transfer steps. The reaction of the nitrosyl radical most likely reflects the availability of an additional reducing equivalent in the isobacteriochlorin ring ( Figure 16c) rather than turnover of a bona fide intermediate during nitrite reduction

2 Rapid-quench EPR experiments yielded kr(S 0 3 2") ~ 95 s '1, and kr(NC> 2 ') ~ 115 s' 1 for E. coli sulfite reductase [Janick etal., 1983] 55

Fe2+ Fe2+ Fe3+ Fe2+ \ \ \ S S \ 2- S S / S0L / / / — Fe2< —Fe2‘- - —Fe3* - —Fe2*- I FT HO-S-O- OH* o=s—o L II 0 = S —OH O

»2- -OH*

Fe Fe3+ Fe2+ Fe3+ \ \ \ \ S S S S 2e' / -OH* / 2e* l •—Fe2* _ • • -F e 3* —Fe2±- —Fe3t- • • • « FT • • :s: :s: :s: :s: i ii OH o

Figure 15. Proposed mechanism for sulfite reduction catalyzed by sulfite reductase [Tan and Cowan, 1991], 56

(A) Q f;o(li) Refill) ' Fe(ll) Q X X X -HO* _2e* 4 -Fe(IIH— -Fe(lll)- -F e (ll)- h+,sq 3 0*?0H o*”~o- 0*§-o- Fe(ll) X —Fe(ll)-

(B) Fe(ll) (C) 0 , -An- 0 *no Fe(ll) Fe(lll) Fe(ll) Fe(lll) 0 X -HQ- X H \ 2e* x X —Fe(ll)- —Fe(lll)^ H+, N0 ?> ^ — * - —Fe(ll) -Fe(ll>- ^NH 3 0'-N*0H 0 '* H Fe(ll) o'-N-h X -Fe(ll) H+ H* 0 h-n -h H \H * .2 e - Q Fe(lll) Fe(ll) - r Fe(HI) -I Fe(lll) X .HO' *X H+, 2e" X —Fe(lll)--*— —Fe(ll)- -< — —Fe(lll)- —Fe(lfl)— ;N-H HO'N'H H H HO'N‘H HO'N*H 0 + nh 2oh Fe(ll) X —Fe(ll)-

Figure 16. Proposed mechanism for reduction of inorganic substrates by sulfite reductase. [Richardson et al., 1979; Chang and Fajer, 1980; Barkigia et al., 1982; Chang et al.,

1981], Nevertheless, a transient siroheme (Fe2 +-NO) adduct can be detected by EPR after addition of N02" to reduced DV (charateristic coupling patterns are observed with

14N 02‘ and 15N 02‘) [Janick et al., 1983; Lui and Cowan, unpublished results]. In related enzymes this evidence has been used to argue for nitric oxide as an obligatory intermediate [Janick et al., 1983]; however, this would require only a one-electron reduction. Subsequent reductive steps would require either a further three-electron reduction to produce NH2OH or a series of one electron and/or two electron transfers.

Although this cannot be discounted a priori, we find this less appealing inasmuch as it is difficult to rationalize the need for one-electron reductive steps.

Scheme I illustrates an alternative explanation for formation of this nitrosyl complex by invoking back-electron-transfer from "[NO*]" after the initial reductive addition. No additional electrons are added to push the reaction forward, and so the thermodynamically most stable form of the intermediate is adopted.

Scheme I

Fve(ll) Fxe(lll)

OH* — Fe(ll)- Fe(lll)- • • * * — Fe(ll)- • • N: N* Q' v OH O 58 Role o f proton tranfer. We have formulated a mechanism that implicitly demands

participation by ionizable residues in substrate binding and proton delivery. Clearly,

reduction of either SC> 3 2' or N 02’ (summarized below, assuming a reaction pH of 7.6)

requires an efficient pathway for transfer of proton equivalents to the active site.

S032- + 6 e- + 7H+ -> HS- + 3H20 (7)

N 02- + 6 e- + 8 H+ -> NH4+ + 2H20 (8)

In this regard resonance Raman studies of E. coli sulfite reductase have suggested

hydrogen bond formation to enzyme-bound CN' [Han et al., 1989]. There are however

two significant differences in the reactions represented by the above equations. First, the

formal reaction products are formally regarded as an acid (H 2 S) or as a base (NH 3 ),

respectively. Second, the number of protons that must be delivered to the Fe-bound N or

S atom (at pH 7.6) is distinct [one for SO 3 2' and three (or four) for N 02*]. This is a

simple consequence of the inability of nitrogen, unlike sulfur, to accomodate these

additional electron lone pairs. We have found the turnover number (kcat) for N 0 2' and

NH2OH to be essentially independent of pH over the range 5-10. This contrasts with

steady-state kinetic data for E. coli sulfite reductase that demonstrated pH optima of 7.9,

8 .6 , and 9.5 for SO 3 2*, N 02", and NH 2 OH, respectively [ Siegel et al., 1974], The

dependence of the pH optimum on substrates suggests that for that enzyme the data is not tracking ionization of an amino acid side chain, while the results show no obvious correlation with the pKa values listed below for these molecules. For desulfoviridin, proton transfers are apparently non-rate-limiting. Regular solution pKa's for substrate and product molecules are as follows: H 2 SC>3 (pKj = 1 .8 , pK2 = 6.9); HN 02 (pK = 3.4);

NH3 OH+ (pK = 6.0); H2S (pKj = 12.0, pK2 = 7.0); NH4+ (pK = 9.3) [CRC Handbook of

Chemistry and Physics, 1990]. Accordingly, the protonation states of substrate anions and intermediates shown in Figure 16 are appropriate for the pH of 7.6 employed in

experiments and simply represent bookeeping. The precise timing of the proton- transfer

steps indicated in Figure 16, relative to substrate binding and electron transfer, remains unclear at this time. For example, does SC> 3 2‘ or HSC> 3 " bind initially, and is two-electron reductive bond cleavage facilitated by further protonation of the oxygen atom to produce

H20 or is HO* released? Clearly these points require further evaluation.

Summary. We have described a pre-steady-state analysis of an enzymatic multielectron redox reaction of an inorganic anion and outlined a preliminary mechanistic model that rationalizes these and other published data. The structural and stereoelectronic features of the catalytic apparatus that promote this reaction and the rate- limiting factors in the early stages of sulfite and nitrite reduction are unclear at this time but form the basis for future investigation. Chapter V

Enzymatic Reduction of Inorganic Ions. Variable Temperature Steady-state and Pre-steady-state Kinetics Experiments to Map the Energy Profile of an Enzymatic

Multielectron Redox Reaction. Application to the Dissimilatory Sulfite Reductase

from Desulfovibrio vulgaris (Hildenborough).

Introduction

In Chapter IV we discussed the first part of a detailed study of the structural and electronic features of the substrate and enzyme that control ground- and transition-state

energies for enzyme-catalyzed reduction of inorganic anions. The present chapter

describes the second part of this study.

The model of the reaction pathway discussed in Chapter IV (Figure 16) shows

enzymatic reduction arising through a sequence of three two-electron reductive cleavages

of X-0 bonds (X = S or N). By comparing the steady-state values with pre-steady-

state rate constants for bond cleavage reactions ( kx), it is clear that only in the case of the

intermediate species NH2OH does ~ kr (Table 5), suggesting rate-limiting bond

cleavage. The structural and stereoelectronic factors that control this reaction, differentiate substrates and intermediates, and promote substrate binding and catalytic activation are unclear. This lack o f detailed mechanistic insight is generally valid for the broad spectrum of oxido-reductase enzymes required for electron transfer to substrate anions and gaseous molecules. 60 61 Table 5. Summary of Reaction Rate Constants at Room Temperature*

Substrate

s o 32- n o 2- NO n h 2o h

k r (S-1) 12 14 6.5 9

kcat (S '1) 0.3 0.04 30

a Results discussed in Chapter IV

In addition to the bond cleavage step, the activation barriers for each step of the

reaction reflect contributions from binding, steric barriers from protein side chains, and

conformational motion of the protein backbone. Evaluation of these discrete

contributions to the activation energy and understanding their relative magnitudes for a

variety of substrate molecules represent a significant challenge. The availability of

kinetics methods to monitor both steady- and pre-steady-state rate profiles for a number of

substrates and reaction intermediates afforded us the opportunity to address this problem

in quantitative detail.

In this chapter we will discuss results from variable temperature experiments

conducted under steady-state and pre-steady-state conditions that account for the

activation energies at each stage of the multistep reduction reaction. The free energy

profiles resulting from this analysis offer insight on the catalytic mechanism and the importance of the special siroheme cofactor for mediating this reaction pathway. 62 The discussion incorporates data gathered by three workers in our laboratory: steady-

state kinetics experiments were done by Wen Liang and pre-steady-state kinetic

experiments were done by Siu Man Lui (SO 32', NO, N 02') and myself (NH 2OH,

NH2OMe) [Lui et al, 1994].

Results

To delineate the energy profile for the multistep redox reaction catalyzed by the

dissimilatory sulfite reductase, we have examined the temperature dependence of both

steady-state and pre-steady-state kinetic rate constants. The latter principally reflect bond

cleavage chemistry, while the former also reflect additional activation barriers on the

reaction pathway. Moreover, only for the pre-steady-state study (and also the steady-

state kinetics of hydroxylamine reduction) can we investigate individual two-electron reductive bond cleavage steps along the reaction pathway. That is, steady-state turnover data obtained for S 0 3 2" or NOj" reduction include the complete sequence of three two- electron reductive steps.

Steady-State Variable-Temperature Data. Initial studies in our laboratory have shown that the steady-state kinetics of the reaction can be reasonably considered in terms of a Michaelis-Menten model. In this model £cat reflects transition-state energies and Km reflects ground-state binding energies [Fersht, 1985]. The assumption that Km ~ Kd (to a reasonable approximation) appears justified inasmuch as the ATm's determined for substrates are very similar to KJs evaluated for inhibitor analogues (CN* and HS*), while the likely error range would influence niether the discussion nor conclusions presented 63 below .3 For steady-state turnover, several simple relationships between standard

jji parameters and and activation energies are readily derived, where AGj and AGt are the ground- and transition-state contributions to the activation free energy (AG )

[Figure 17]. These profiles can be developed more thoroughly by including the analysis of the temperature dependence of pre-steady-state rates (discussed below) since each discrete bond breaking step can be independently analyzed. The magnitude of AG can be directly determined at any temperature from equation 9, which can be rewritten in the form of equation 10, where k, R, and h are the Boltzmann, gas, and Planck constants, respectively.

*cat = (kT/h)exp-(AG*/RT) (9)

R H k^hlkT ) = AS* - AH*/7* (10)

ifc sjc The latter equation was used to evaluate the values of AH and AS (Table 6 ) from the temperature dependence of kcat and plots directly analogous to that shown in Figure 18.

3 Substrate or ligand anions of similar bonding capabilities appear to have similar binding affinities. For example, compare tf^SC^2') ~ 59 mM with A'd(CN') - 200 mM, and /fm(NH2 0 H) ~ 59 mM with k^(HS") ~ 21 mM [Liang and Cowan, 1994]. Minor systematic errors in will have no effect on the conclusions reached in our discussion of the data, while major discrepancies are unlikely. A 10-fold error in introduces an error of 1.4 kcal mol ' 1 in AG,j. Since < Km, only if the magnitude of Ktj for NH 2 OH is underestimated by ~ 4 kcal mol ' 1 (that is, by 103 -fold) are the general conclusions compromised. The on-rate constant for NH2 OH has been determined (~ 24 M ' 1 s'1) [Chap. IV] and so on-rate of ~ 1.2 x 10 ' 3 s' 1 would be required for binding affinity in the 50 mM range. This is inconsistent with estimates of off- rates for related ligands [k 0 g(NH 3 ) > 30 s' 1 and k ^ H S '1) > 0.3 s ' 1 [Liang and Cowan, 1994; Lui et al., 1993] and suggests that for NH 2 OH will not be significantly smaller than Km(NH2 OH). 64

ES"

E + S AG AG, free energy (G) AG

ES

reaction pathway

Figure 17. General free energy profiles for substrate turnover showing the components of AGr , AG0f , AG

be estimated in a similar fashion from l n ^ ^ / ^ ) - \n(kT!h) (equation 1 la), and the

enthalpic and entropic components, from the temperature dependence defined by equation

1 lb. For each substrate, the temperature dependence o f Km was evaluated and the

appropriate value was used. Estimates of AGd are made from relationship 13, assuming

R m i k ^ / K J = RTln(kT/h) - AGt* (11a)

i?ln [(k^h)/(KmkT)] = AS* - AHt*/T (1 lb)

AG* = AGt* + AGd ( 1 2 )

AGd = -/?71nATm (13)

Figure 18 illustrates a typical plot obtained from variable-temperature steady-state experiments by use of equations 10 and lib, respectively. The activation parameters thereby determined are listed in Tables 7-9. Overall, there is a general decrease in the magnitude of AG* from S 0 3 2' and N 02‘ to NH 2 OH. This arises not through a decrease in the magnitude of AGt*, which in fact increases over the series, but rather through a decrease in the magnitude of AGd, reflecting stronger binding by the 7 t-acceptor ligands

N 0 2" and S032'. The lower AGt values for these 7 t-acceptor ligands presumably arise through population of the antibonding N-0 (or S-O) orbitals, which weakens the bond toward reductive cleavage [Jolly, 1984]. necp codn oeuto b[.Lagi u t a. 1994],1 al.,[W.Liang in et. Lui equation lb to and according intercept slope the from (AHe* obtained ASt*) were components and entropic and Enthalpic Figure 18. Steady-state data. Plot of of Plot data. Steady-state 18. Figure • ln(kcat/Km)-ln(kT/h) 5 I “I -50 48- 8 -4 -42 .00,..01 .02 .03 .04 .05 0.0036 0.0035 0.0034 0.0033 0.0032 ,..0.0031 0.0030 ----- 1 ------1 ----- 1 ( -i) (K r / RlnKk^h^KJcT)] 1 -- « ----- 1 ------ess 1 versus 1 ----- IT o NH for 1 2 OH. 66 67

a Table 6. Steady-State Activation Parameters

AH* AS* AG* substrate (kcal mol'1) (K^mol*1) (kcal mol'1)

2- so3 2.6 -51.7 18.0 n o 2' -6.8 -79.8 17.0 n h 2o h 3.1 -41.3 15.4

a + Kinetics data were obtained using saturating concentrations of substrate and MeV - [Data by W. Liang in Lui et al., 1994]. The method for kinetic assay is similar to that described in section 2.7, Chap. II. Samples were incubated for 15 min at the appropriate temperature, over the range 10-45 °C, before the addition of M eV \ Reaction conditions were as follows: [DV] =140 nM, [SO3'2] = 1 mM, [NO2'] = 200 |iM, [NH2 OH] = 150 mM. The activation free energy at 298 K was determined from equation 10 and the enthalpic and entropic components were determined from equation 11. Errors in each experiment are estimated to be on the order of + 0.5 kcal mol ' 1 for AH*, ± 2 K *1 cal mol ' 1 for AS , and ± 0.2 kcal mol for AG . Any temperature dependence of the constant Km was accounted for over the temperature range employed. Nitric oxide proved to be unsuitable as a substrate for steady-state kinetics experiments. 68

Table 7. Breakdown of Steady-State Parametersa

substrate AHt* ASt* AGt*

(kcal mol'1) (cal K' 1 mol'1) (kcal mol'1)

S032' -5.8 -60.5 1 2 . 2 N 02' -5.6 -53.8 10.4

n h 2o h 1 2 . 6 -3.9 13.8

a Kinetics data were obtained as described in the legend to Table 6 . The activation free energy at 298 K was determined directly from equation 11a, and the enthalpic and entropic components were determined from equatio^ 1 lb. Errors in each measurement are estimated to be on the order of ± 0.5 kcal mol for AHt*, ± 2 cal K ' 1 mol ' 1 for ASt\ and ± 0 . 2 kcal mol ' 1 for AGt*. 69

a Table 8 . Summary of Factors Contributing to the Activation Free Energy

Substrate AGr* a g „; AG* AGt* AGd(kcal mol*1)

S0 3 2- 15.6 2.4 18.0 1 2 . 2 5.8 n o 2- 16.2 0 . 8 17.0 10.4 6 . 6 NO 16.0 ^17.0 -1.9 (est) -14.1 (est.) n h 2o h 14.9 0.5 15.4 13.8 1 . 6

a i All data are in units of kcal mol . Kinetics parameters were determined from equations 6-13 and using the data from Tables 6 , 7, 9. Errors in each measurement are estimated be on the order of ± 0.2 kcal mol . Other than AGf*, data for NO have been estimated owing to the problems associated with steady-state turnover of this reactive substrate. 70

Table 9. Pre-Steady-State Bond Cleavage Dataa

AHr* ASr* AGf*

Substrate (kcal mol'1) (cal K^mol'1) (kcal mol'1)

S032' 12.2 -11.4 15.6

N 0 2' 17.4 4.0 16.2

NO 12.5 -11.9 16.0

NH2OH 5.0 -33.1 14.9

Stopped-flow kinetics data for substrates were obtained as previously described by monitoring the change in absorbance at 438 nm [section 2.5, Chap. II]. The solution conditions are detailed as follows: [DV] = 30 pM, [S032'] = 50 mM, [N02‘] = 50 mM, [NO] = 1 mM, [NH?OH] =100 mM. Parameters were determined from equations 11 and 12. Samples were incubated for 15 min at the appropriate temperature over the range 8 - 45 °C prior to mixing. Errors in epch npeasurement are estimated to be in the order of ± 0.3 kcal mol ’1 for AHr\ ±1.5 cal K mol for ASr*, and ± 0.4 kcal mol ' 1 for AGr* [Data for SO32', NO2', and NO by S.M. Lui; Data for NH2OH by A. Soriano (data for NH2OMe was not included because of difficulty in fitting the kinetic profiles)]. Pre-Steady-State Variable-Temperature Data. The temperature dependence of the Arrhenius rate constant kt is given by equation 14, which can be rewritten in the form of (15) that is of greater utility for variable-temperature

kT = (kT/h)exp-(AG*/RT) (14)

Rln^kjh/kT) = A S* - AH */T (15)

jji experiments since it accounts for the inherent temperature dependence of AGr . This form of the rate equation has been used to analyze pre-steady-state rate data. Figure 19 illustrates a typical plot obtained from variable-temperature stopped-flow data using equation 15. The activation free energy (AGr*) is calculated directly from equation 14, and the enthalpic (AHr ) and entropic (ASr ) components, from the temperature dependence of kr defined by equation 15. The results from these measurements are given in Table 9. We will consider first the data for the family of nitrogenous subtrates before commenting later on the result for sulfite. The free activation barrier (AGr*) for two- electron reduction of nitrite (16.2 kcal mol'1) is similar to that for nitrous oxide (16.0 kcal mol'1) but differs from the value for hydroxylamine (14.9 kcal mol*1). There is, however, a more systematic change in the enthalpic and entropic components, which is discussed in a later section. In particular we observe a decrease in the magnitude of the enthalpic barrier for the more reduced substrates, which may correlate with substrate bond energy, and also an increase in the entropic barrier.

Discussion

Previous work in our laboratory have demonstrated a small apparent kcat for S032' and N 0 2' relative to that for NH2OH reduction (Table 5), suggesting that bond cleavage steps in the early stages of the reaction might be rate limiting. This would follow Enthalpic and entropic components (AH/ and ASr*) were obtained from the slope and ASr*)andand slope the from (AH/ obtained components were entropic and Enthalpic necp codn oeuto 15. equation to according intercept of Plot data. Pre-steady-state 19. Figure R ln(kr.h/k.T) -216 - 0 1 2 0.0031 “

0.0032

0.0033

1/T (K-1) 1/T R ln(kjh/kT) R 0.0034

vru 1/7' NH for versus 0.0035

0.0036 2 OH. 72 73 expectations based on relative S-0 and N-0 bond energies for each substrate. That is, the resonance-stabilized sulfite and nitrite anions react more slowly than hydroxylamine.

However, this hypothesis is not supported by the similarity in the rate constants for bond

cleavage determined from pre-steady-state kinetic measurements for two-electron

reductions of SO 3 2-, N 02*, NO, and NH2OH [*r(S032') ~ 12 s"1, *r(N 02-) ~ 14 s'1,

£r(NO) ~ 6.5 s_1 and £r(NH2 OH) ~ 9 s-1]. This dichotomy can be explained if rc-acid

ligands S032’ and N 02* should exhibit weaker S-0 and N -0 bonds, respectively, after binding to the electron-rich siroheme as a result of 7C-backbonding and population of the

antibonding S-0 and N -0 orbitals. In this work, we have described a series of experiments that lend quantitative support to this hypothesis and construct a free energy profile for the catalytic reduction of N02" to NH3. This provides considerable insight on the factors underlying the enzymatic reduction of these inorganic species.

Activation Parameters. For N 02‘, S032*, and NO reduction, we have demonstrated through pre-steady-state kinetic measurements that £cat does not reflect reductive bond cleavage, and so the activation parameters noted in Figure 17 for N 02‘ reduction include other factors (AG0f ) that so far remain ill-defined. That is, the

AG*(N02') indicated in Figure 17 contains contibutions from both AGr* and AG 0 f*

(equation 16). The contribution to AG from bond cleavage

AG*(N02-) = AGr*(N02‘) + AG 0f* (N02‘) (16)

(AGr*) can be independently determined from the pre-steady-state variable-temperature data for the two-electron reductions of N02‘, S032‘, NO and NH 2OH. Using the AG* values determined from steady-state rate measurements, the additional contributions

(AG0f ) from enzyme conformational changes, etc., can be determined (Table 8 ). The results demonstrate that bond cleavage gives rise to the dominant barrier for enzymatic substrate reduction (that is, AGr » AGof ). In the specific case of hydroxylamine we

might expect AG0f (NH2 OH) ~ 0 since ~ kr. It is therefore reassuring to note that

this is indeed experimentally verified [AG*(NH 2 0 H) ~ 15.4 kcal mol-1; AGr*(NH 2 0 H) ~

14.9 kcal mol-1; AG 0 f*(NH 2 OH) ~ 0.5 kcal mol -1 confirming the conclusions reached earlier that bond cleavage is the dominant rate-limiting step for this substrate. For earlier

reaction intermediates there is apparently a significant contribution from AG 0 f* (but much

smaller than AGr ). The origin of this barrier remains unclear at this time but may involve

conformational changes of the protein or siroheme ring. Figure 17 and Tables 6-9

summarizes our current evaluation of the free energy profiles thst can be constructed from

this data.

For a series of similar reactions there can exist a linear relationship between the activation enthalpies and entropies [Isaacs, 1987]. By writing the free energy change

(AG*) in a revised form (AH* = TAS* + AG*), it is seen that a linear plot of AH* versus

AS (obtained from variable-temperature studies on a family of related reactions) yields a temperature T (isokinetic temperature) at which the Arrhenius plots for each data set would intersect. In theory such a linear relationship suggests the dominance of one stereoelectronic parameter in controlling the relative rates of a number of substrates

[Isaaks, 1987]. The isokinetic relationship also reflects a balancing of the change in one activation parameter by another. That is, a more negative enthalpic component is offset by a more negative entropic component, and vice versa. Figure 20 shows the isokinetic l|t )|| plot obtained from the variation of AHr and ASr , which yielded an isokinetic temperature o f334 K (correlation coefficient of 0.998). As noted earlier, there is a general decrease in the magnitude of AHr*, which most likely reflects the difference in bond energies. Since related plots for AH* versus AS* and AHt* versus ASt* are obtained over a narrower range of values, the deviations from experimental error are more pronounced but the general trends are similar. NO, and NH and NO, temperature of 334 K (correlation coefficient of 0.998). Data for SO for Data 0.998). coefficient of (correlation K 334 of temperature Figure 20. Isokinetic plot of AHr* versus ASr*. The gradient yields an isokinetic isokinetic an yields gradient The ASr*. versus AHr* of plot Isokinetic 20. Figure

AH r* (kcal mole-1) 20 i 20 2 OH were taken from Table 9.fromTable taken were OH 40 -4 30 -3 *(a K1 mole-1) K-1 r* (cal S A -20 -10 0 3 2-,NC> 10 2 “, 75 Implications for Understanding the Reaction Mechanism. In the discussion that

follows, emphasis will be placed on the enzymatic reduction of N02* to NH 3 rather than

the reaction of S032*. This simply reflects the availability of reaction intermediates and

substrate analogues for the former with which to carry out detailed kinetics studies. It is

reasonable to assume that the arguments and conclusions presented below are equally

valid for S032" reduction. We also note that there is an implicit assumption that the magnitude of AG for any given substrate mainly reflects the activation barrier for the first

two-electron reductive cleavage. That is, AG for N 02" reflects the first two-electron

step for N 02“ -» "NO"", etc., rather than reduction of a later intermediate. We feel that this is justified inasmuch as there is a decrease in the magnitude of AG* moving from

N 0 2" to NH2OH as the substrate.4

For nitric oxide, only the magnitude of AGr* could be determined experimentally.

However, estimates of AG*, AGj*, and AGj could be readily evaluated 4 The following trends in activation free energies may be noted from the data in Tables 2-4. First, the magnitude of AG* decreases as S032- > N 02‘ ~ NO > NH 2 OH. Second, AGt* increases, with S032' and N 0 2" < NH20H. Third, the binding affinity AG

N 02‘ > NH2 OH. Note that NO yields anomalous AGt* and AGj values as a result of the unusually strong binding to siroheme .4 In the discussion that follows we will show how the observed trends in free energy components lend considerable insight on the design of the catalytic site and, in particular, the choice of siroheme as the catalytic cofactor.

4 The binding energy Kj may be estimated from kofl/kon- We previously determined kon(NO) ~ 7 x 105 M -1 s’ 1 [Chap. V]. We have no estimate of for the NO complex of desulfoviridin; however, Olson has reported k 0 g-~ 3 x 10-^ s’* for binding in the a- and p-subunits of hemoglobin [Olson, 1981], Using this as an estimate yields K

and N 02‘ to NH2OH reflects the change in bonding as we progress from the strong 71-

acceptors SO 3 2* and N 02‘ to the purely a-donor NH 2 OH. We have noted in Chapter IV

how the electron-rich siroheme ring has a particularly high affinity for71 -acceptor ligands.

We shall now argue that the bonding mode that results in tight binding by N 0 2~ is also

responsible for the reversal in the relative magnitudes of AGt along the same series.

Population of the antibonding S-0 and N -0 orbitals as a result of 7 t-back-bonding

weakens the chemical bonds that are to be reductively cleaved [Jolly, 1984]. This is

reflected in the smaller values of AGt* for SO 3 2' and N 02‘ relative to the corresponding

parameter for NH2OH (Table 8 ). We see then that the relative magnitudes of the AGj

and AGt components are in opposition with regard to defining the overall activation

barrier AG . Although the absolute magnitude of the AGt component is substantially

greater than that of AGj, the relative change in the magnitude of AG^, comparing

substrates and reaction intermediates, is larger than the corresponding change in the AGt

term. The observed trend in the magnitude of the AG* values therefore reflects the

dominance of the binding term (AG^). The design of the prosthetic center therefore

accomodates two important requirements: (1) Strong binding of the substrate and

weaker binding of the product is promoted by the dominance of 7 r-back-bonding. These

factors are also manifest by the kinetic rate constants for release of the substrate and

product [Lui et al.,1993; Christner et al., 1983], (2) The aforementioned scheme serves

to weaken the chemical bond that is to be reductively cleaved by populating an

antibonding orbital in 7t-acceptor substrates, thereby lowering the transition-state

contribution to the activation free energy [JollY, 1984], This is carried to an extreme in

the case of nitric oxide , 4 although the intermediate (HNO) to be expected during nitrite turnover would most likely fit in the overall trends noted above for AG*, AGj, and AGt*

in the series N 02‘ -> HNO -» NH2OH -> NH3. We note that we have not discussed the contributions of proton-transfer steps, which appear to be rapid and do not contribute significantly to AG*.

Summary. In this chapter we have evaluated the component activation energies for each step of the catalytic reduction of an inorganic anion by variable- temperature experiments conducted under steady-state and pre-steady-state conditions. The free energy profiles resulting from the analysis offer insight on the catalytic mechanism and the importance of the special siroheme cofactor for mediating this reaction pathway. This strategy should be of general value for the analysis of multistep enzymatic reductions of other inorganic substrates. Chapter VI

Sulfite Reductase: Active site residues are "noncatalytic". Comparison of reaction

energetics for enzyme- and siroheme-catalyzed reduction of inorganic substrates

Introduction

Enzymatic catalysis of biological transformations is a well-studied phenomenon. A

lowering of the activation barrier for key reaction steps can be achieved either by ground-

state destabilization of the enzyme- substrate complex or by lowering of the transition-

state energy [Fersht, 1985]. In co-factor mediated reactions, the relative contributions of the cofactor and the active site protein residues to enzyme catalysis may be evaluated by comparison of kinetic parameters for the cofactor- mediated reaction relative to the enzymatic reaction. Current wisdom dictates that metal-cofactor-mediated chemistry

(both redox and nonredox) commands significant involvement from the protein environment. For example, metal ion dependent nuclease enzymes exhibit substrate hydrolysis rates that are considerably greater than those of metal ion activation alone

[Burstyn and Deal, 1993]. Also, functional models for redox enzymes typically show turnover rates that are several orders of magnitude lower than those of the enzyme- catalyzed reaction and often require extremes of temperature and pH or addition of other active species to effect turnover at acceptable rates. Several examples of competent functional models and turnover rates have been reported including the following: molybdo-enzyme mimic [Berg and Holm, 1985]; methane monooxygenase mimic [Leising et al., 1993]; nitrogenase mimic [Coucouvanis et al., 1993]; and manganese catalase mimic 79 [Gelasco and Pecoraro, 1993]. As mentioned earlier, siroheme serves as the principal catalytic center in the active site of sulfite reductases, while the cluster contributes as an electron trapping and storage unit. Isolated siroheme can itself catalyze the reduction of inorganic substrates [Seki et al., 1981; Kang et al., 1987; Soriano and Cowan, 1995].

Thus, comparison of catalytic parameters of free and enzyme-bound siroheme allows a direct evaluation of the role of active site residues in modulating both ground- and transition-state energies. In this chapter we present a detailed evaluation of the kinetic parameters for catalytic reduction of inorganic substrate molecules by isolated siroheme.

The results suggest an unusual noncatalytic role for the active site protein residues in this enzyme class. This conclusion may well be generally valid for a wider range of metal- cofactor-mediated enzyme reactions.

Results and discussion

Free siroheme was isolated from the dissimilatory sulfite reductase (disulfoviridin) from the sulfate reducing bacterium Desulfovibrio vulgaris (Hildenborough) by standard procedures [Kang et al., 1987; Siegel et al.,1978] with some modifications (as discussed in

Chap. II) It was quantitated using the published extinction coefficient

(S595nm = * ^ x M 1 cm *) for the cyanide adduct [Wolfe et al., 1994]. Following isolation, the siroheme was demonstrated to be fully metallated by standard analytical procedures [Wolfe et al., 1994], Initial velocity calculated from changes in absorbance measurements in the MeV+- assay (i.e. vQ = A[AjJ/At) were corrected to concentration values by the absorption coefficient of MeV+ ; for the cell length of 1 cm v0 = (A[ApJ/At) x 1 lsy Also, we have taken account of the difference in reducing equivalents for S 0 3 2', N 02‘ and NH2OH ( 6 , 6 , and 2 electrons respectively). 81 The following conversions were used [Wolfe et al., 1994]:

v0 (S032' or NOa ) = A[S032' or NC^/At

= (1/6) x (A[MeV+ ]/At

= (l/6 )x(A[A 600]/A t)x(l/e600) (17)

v 0(NH2OH) = A[NH2OH]/At

= (1/2) x (A[MeV+ ]/At)

=(l/2)x(A[A 6J /A tx ( l/e 600) (18)

At pH 6 , a plot of initial velocity versus sulfite concentration showed an inverted shape

(Figure 21) that had been observed previously but not explained [Kang et al. 1987], We rationalize this plot as an example of uncompetitive substrate inhibition, which is manifest in terms of a reaction mechanism (Scheme II) that has been previously proposed to explain the high levels of trithionate product obtained from the siroheme-catalyzed reduction of sulfite [Tan and Cowan, 1991], Figure 21 also shows a fit to the data using a standard

Scheme II.

H* 2e-.2H* 2e* 2tr,2SO s -Fe(ll)- — Fe(ll)- • • -HzO -HjO « • O* L"OH o- 82

0.5

0.4

o fH 0,3 VI g O > 0.2

0.1

0.0 0.0 0.05 0.10 0.15 0.20 0.25 0.30 [S032 ] (mM)

Figure 21. Initial velocity plot for turnover of sulfite by free siroheme at pH 6 . At pH 7, substrate inhibition by SO 3 2- is significantly reduced (see text). The data were fitted to equation 19, yielding the following fitted parameters: V , ^ = 1.4 x 10 -6 M (mole of siroheme) -1 s-1, Km = 0.1 mM, and Kj = 0.02 pM with [siroheme] = 0.6 pM (note that the low K; value indicates a higher affinity of sulfite to the siroheme-intermediate complex than to siroheme alone). 83 equation (19) for substrate inhibition [Palmer, 1985], which has been modified to account for the involvement of two SC> 3 2‘ species in formation of a trapped trithionate intermediate.

v0 = {*Jsiroheme][S 0 3 2-]}/{*m + [S032](1 + [SOs2]2^ ' 1)} (19)

At pH ~ 7, substrate inhibition is less significant, although k ^ t and km parameters are similar to those obtained at lower pH (Figure 21). This observation is consistent with a previous report that desulfoviridin-catalyzed sulfite reduction leads to significantly less trithionate by-product formation at pH 7 than at pH 6 [Jones and Skyring, 1975] and is fully consistent with an increase in reactivity of S032- toward the bound intermediate as a result of protonation of either sulfite or bound thiosulfite at pH levels below 7.5 Free siroheme was also found to catalyze the reduction of NO 2 ' and NH 2 OH substrates. In contrast to the inverted plot observed for sulfite reduction, the reaction of nitrite and hydroxylamine exhibits standard Michaelis-Menten kinetics (Figure 22), since neither is sufficiently nucleophilic to react with siroheme-bound intermediates.

Steady-state kinetic parameters can be reasonably considered in terms of a

Michaelis-Menten model, where k ^ reflects activation energies and Km reflects ground state binding energies [Fersht, 1985; Tan and Cowan, 1991; Lui et al., 1994], The magnitude of AG can be directly determined at any temperature from equation 2 0 , where k, R, and h are the Boltzmann, gas, and Planck constants, respectively.

*cat = (kT/h) exp(-AG*//?7) ( 2 0 )

pKa's for HjSO^: pKj 1.8 and pK 2 6.9 . The protonation state of HS03'most likely influences its reactivity towards siroheme-bound intermediates. 84 0 2

©

OJ JS

QD

0.00 0.01 0.02 0.03 0.04 0.05

[no 2j ,m 6

5

o K 4

S 3 >C

2

0 0.00 0.02 0.04 0.06 0.08 0.10 [NHjOH], M

Figure 22. Initial velocity plot for turnover of NC> 2 “ (a) and NH 2 OH (b) by free siroheme at pH 7. The data were fitted to equation 19 yielding the following fitted parameters: For N 02' reduction, V , ^ = 1.9 x 10 *7 M(mole siroheme ) ' 1 s '1, Km = 0.4 mM with [siroheme] = 0.76 nM. For NH2OH reduction, Vmax = 5.8 x 10 "6 M(mole siroheme) ' 1 s"1, K« = 3 mM with [siroheme] =0 . 8 nM. 85 Evaluation of the energy barriers for the turnover of each substrate listed in Table 10 reveals that for free siroheme the and values are generally smaller and larger, respectively, than for the enzyme bound cofactor .6 That is, free siroheme is a more efficient catalyst than the enzyme. The only exceptions are catalytic reduction of NO 2 ' and NH2OH by the low-spin assimilatory-type sulfite reductase (SiR) [Moura et al.,

1986], these species being 7-fold and 70-fold more active than free siroheme, respectively.

However, this is a significantly lower activity enhancement than is observed for normal enzyme catalysis. Significantly in all cases, SC ^" turnover is mediated by free siroheme at least as well as by the enzymes studied. The unexpected conclusion from these experiments is that the protein matrix apparently does not contribute significantly to catalysis, relative to the siroheme center, but serves simply to afford protection for reduction intermediates from the physiological substrate (SOj2*) and prevent side reaction of the sort illustrated in Scheme II. Finally, these results suggest that, for other enzymes carrying inorganic redox cofactors, the catalytic role of the protein may be overestimated and formation of a protective reaction pocket may in fact be the dominant function.

6 The relationship ATm ~ has been previously justified [Tan and Cowan, 1991; Lui et al., 1994; Liang and Cowan, 1994]). 8 6 Table 10. Steady-State Activation Parameters for Siroheme- Catalyzed Substrate Reduction a

fccat(substrate Km ^ cat^m x lO"'* AG* substrate heme'V1) (mM) (s-’M*1) (kcal mol’1) ref.

Siroheme S 032- 2.36 0.1 23.6 16.9 this work

n o 2- 0.25 0.4 0.63 18.2 this work n h 2o h 7.31 3.0 2.43 16.3 this work DV

S 032- 0.31 0.06 5.16 18.1 b

n o 2_ 0.038 0.028 1.35 19.4 b n h 2o h 29 48 0.60 15.4 b SiR S 032- 0.21 0.05 4.2 18.4 c n o 2- 20 4.7 4.2 15.7 c n h 2o h 2300 14 164 12.9 c

a DV (desulfoviridin), dissimilatory sulfite reductase and SiR, assimilatory-type sulfite reductase are from D. vulgaris (Hildenborough). Kinetic data were taken at pH 7 at 298 K following procedures described in Chap. II, section 2.7. Reaction conditions were as follows: (S032- turnover) [siroheme] = 0.6 pM, [SO 3 2”] = 2-300 pM; (N02“ turnover) [siroheme] = 0.76 pM, [N02‘] = 0.1-49 mM; (NH2OH turnover) [siroheme] = 0.8 pM, [NH 2 OH] = 0. l-100mM. Errors in each measurement are estimated to be on the order of ± 0.2 kcal mol ' 1 for AG*. & Wolfe et al., 1994. c Tan et al., 1994. Chapter VII

Studies on the Assimilatory-Type Sulfite Reductase from Desulfovibrio vulgaris

(Hildenborough)

7.1 Introduction

The assimilatory-type sulfite reductase (SiR) from the sulfate-reducing catalyzes the six-electron reduction of SC> 3 2" to S2' and NO 2 " to NH3. It contains the coupled siroheme-[Fe4 S4 ] prosthetic center common to this class of enzymes. By virtue of its size

[Mr ~ 23500], solubility, and accessibility to direct physicochemical measurements [ ID and 2D NMR, electrochemistry, among others], SiR represents an outstanding model system for understanding the structural basis of recognition, bond activation, and catalytic redox chemistry of inorganic anions and gaseous substrates. In this chapter we will present and discuss the characterization and paramagnetic *H NMR studies on recombinant SiR from D. vulgaris. Our laboratory has recently reported the primary protein sequence and over-expression of SiR [Tan et al., 1991, 1994]. The expression system now provides the opportunity for mutagenesis studies and detailed mechanistic and spectroscopic studies on SiR.

7.2 Characterization of Recombinant SiR from D. vulgaris

The SiR gene has been overexpressed by utilizing a triparental conjugation method that allowed transformation of a high copy number broad-host-range plasmid pDSK519 into the native Desulfovibrio source to give an expression vehicle that provided 87 a greater than 5 O-fold increase in enzyme production relative to the native strain [Tan et al., 1994], The expressed protein was characterized by N-terminal sequencing and amino acid analysis [Tan et al., 1994]. The data were in satisfactory agreement with results previously published for native enzyme [Tan et al., 1991], Iron and sulfide analyses using standard chemical methods yielded 4.4 + 0.3 iron centers and 5.1 + 0.5 sulfides per mole of enzyme [Tan et al., 1994], Again this is in good agreement with the previous data on the native sulfite reductase [Huynh et al., 1984], The recombinant enzyme was also characterized by its activity and spectroscopic characteristics. Here we present the results of these experiments.

Enzyme Activity. Reaction with NH2OH is rapid and provides a convinient assay of enzymatic activity. The turnover rate observed under saturating substrate conditions

([NH2 OH] = 150 mM) was *cat(NH 2 OH) = 2.4 ± 0.2 x 103 NH2 OH/sec/heme. This value is similar to that observed for the native enzyme: £cat(NH 2 OH) = 2.3 + 0.2 x 103

NH2 OH/sec/heme [work in our laboratory by B. M. Wolfe].

Spectral characterization o f Recombinant Holoenzyme. The electronic, CD and high field *H NMR spectra of the recombinant enzyme [Figure 23 and 24] matched those obtained from the native enzyme [Huynh et al., 1984; Sola and Cowan, 1990], The EPR spectrum of oxidized recombinant enzyme (work in our laboratory by S.M. Lui [Tan et al.,1994]) also shows good agreement with published data on the native enzyme [Huynh et al., 1984]. These spectroscopic results indicate the incorporation of all prosthetic centers in the recombinant enzyme. 89

200 300 400 500 600 700 50.00

CD Mdeg

Abs 0. 8-

0 .4 -

200 300 400 500 600 700 X (nm)

Figure 23. Circular dichroism spectrum of assimilatory-type sulfite reductase (SiR) taken with 9.6 mM enzyme in 50 mM Tris buffer, pH 7.6. An electronic absorption spectrum is shown below. 90

/ - r I- I I I / / ■ - y / / 95 40 20 15 •15 PPM

Figure 24. A 300 MHz *H NMR spectrum of recombinant SiR obtained in 50 mM potassium phosphate buffer in D 2 0, pH 7.6 at 296 K. The SiR concentration was ~ 1 mM. The spectrum was measured over a 100 KHz bandwidth with 4K data points using the super-WEFT sequence (180-X-90-AQ + delay). The relaxation time (AQ + delay) used was 67 m s, t was 60 ms. The spectrum required 7,000 scans. 7.3 Paramagnetic NMR Studies on the Assimilatory-Type Sulfite Reductase 9 1

(SiR)

*H NMR spectra of paramagnetic proteins are rich in information content. In the presence of a paramagnetic prosthetic center, hyperfine-shifted resonances can be observed outside the poorly resolved diamagnetic region and can subsequently be used to probe the active site of the protein. The hyperfine-shifted resonances can provide unique information on the magnetic and electronic properties of the paramagnetic metal center that is inaccessible in diamagnetic systems [La Mar, 1979]. They can also be used to identify ionizable residues neighboring the paramagnetic prosthetic center by pH titration.

Metalloproteins that have been successfully studied by paramagnetic NMR include both heme and iron-sulfur containing enzymes.

Understanding the interaction of the coupled P^S^-sirohem e prosthetic center may provide insight to the magnetoelectronic phenomena and mechanistic chemistry exhibited by a variety of complex redox enzymes. To this end one of the goals in our laboratory is to characterize the electronic and magnetic properties of the individual cluster and siroheme in different redox states of SiR using the hyperfine-shifted resonances as probes.

The assignment of the paramagnetically-shifted signals to specific cluster-bound cysteines and the peripheral acetate and propionate side-chains of the siroheme ring is crucial to this goal (Figure 25). In this section, the results of paramagnetic ]H NMR studies on oxidized and fully reduced SiR will be presented. Results from ID NOE and 2D

NOESY, COSY, and TOCSY experiments will be shown; the results are aimed to assist future work to unambiguously assign the hyperfine resonances originating from the active site of oxidized SiR. Results of paramagnetic NMR studies on fully reduced SiR will also be presented and their machanistic implications discussed. 92

"CH CH S — CH2 Cl CH2^ S\ :Fe I P ^Pe-L

-c h : -S Fe 'CH,

CO,H

HO,C

CH 'CH

Figure 25. Diagram of the siroheme-[Fe 4 S4 ] unit in SiR showing the peripheral acetate and propionate side chains in siroheme as well as the a and P protons of the cluster cysteines. The proposed ligand in the sixth axial position, histidine, is also shown [Cowan and Sola, 1990J. 93 7.3.1 General principles

The chemical shift, longitudinal relaxation time (Tj), and transverse relaxation time

(T2 ) are important parameters in NMR experiments; the design of paramagnetic NMR

experiments is guided by these parameters. In a paramagnetic system, the presence of

unpaired electrons induce large effects on NMR parameters and in turn, these parameters

cany with them a wealth of information about the paramagnetic prosthetic center and its

immediate surroundings. The paramagnetically influenced resonances are sensitive to the

spin/oxidation state of the metal center(s) and the contact between the prosthetic unit and the apoprotein [La Mar ,1979]. The effect of unpaired electrons on NMR parameters will be described briefly and qualitatively in the following discussions.

The hyperfme shift. In diamagnetic compounds, two factors contribute to the chemical shift: the diamagnetic shift, which arise from the shielding of the nucleus from the external magnetic field as a result of the circulating electron density around the nucleus; and the paramagnetic shift due to excited electronic states [Bertini and Luchinat, 1986].

In paramagnetic systems, a third factor, the presence of unpaired electrons, contribute further to the chemical shift. The resonance position of a nucleus is shifted from its diamagnetic position by an amount defined as the hyperfine shift (Au/u0)hf in parts per million (ppm). This hyperfine shift may arise from two distinct physical interactions: the

Fermi contact shift and the dipolar shift. The Fermi contact shift, (Au/u)con, is due to delocalization of unpaired metal spin(s) to the ligand nucleus . In the ideal case of a single populated level for a paramagnetic metal center, the Fermi shift is expressed as [La

Mar, 1979]

-A gBs(s+1) (Au/u)con = ------(21)

2(y/2iz)kT whereA is the hyperfine coupling constant, s the spin quantum number, g the spectroscopic splitting factor, B the Bohr magneton, k the Boltzmann constant, and T the absolute temperature. On the other hand, the dipolar shift (A u /o )^ results in cases where the metal has an anisotropic magnetic susceptibility [Jesson, 1973].

(Au/u)dip = 1/3N {xz - 1/2 (xx + Xy) [(3cos 2 0 -1) r 3] + 3/2 (xx - Xy) sin 2 0 cos 2Q r 3} (22)

where xz> Xx> an<^ Xy are the principal components of the magnetic susceptibility tensor, 0 the angle between the metal nuclear vector and the z axis, r the length of this vector, and

Q the angle between the projection of r on the xy plane and the x axis. The first term is due to axial anisotropy, while the second is due to the rhombic or in-plane anisotropy .

Spin transfer can occur only to ligands directly coordinated to the metal [La Mar,

1973], thus, in the coupled p^S^-sirohem e unit, contact shifts should arise only for the ligands to the heme iron (the isobacteriochlorin ring and the axial ligands) and the cysteines coordinated to the p^S,*] cluster [Figure 25], The dipolar shifts varies as r -3 and decrease rapidly with distance from the metal. Thus, both contact and dipolar interactions predict that hyperfine shifts are observed primarily for nuclei close to the metal center, i. e., the prosthetic unit. However the dipolar term is mediated through space, and if large enough can affect the amino acid residues in the active site pocket such as those above and below the siroheme.

Ti and T2. The longitudinal relaxation time, T i, is related to the time necessary for an ensemble of nuclear spins, which are equally distributed among the degenerate spin states in the absence of a magnetic field, to reach Boltzmann equilibrium after a magnetic field, B 0 has been applied. Since longitudinal relaxation involves a change in the 95 population of the nuclear spin levels, and therefore a change in the total energy of the

system, only those mechanisms that allow a nucleus to switch spin state through energy

exchange with the environment are operative in such a process [Bertini and Luchinat,

1986], For this reason, T i is also called the spin-lattice relaxation time. On the other hand, the transverse relaxation time, T2, is related to the lifetime of the magnetization

component in the xy plane perpendicular to B0. Such a component is zero at equilibrium

since the component of the magnetic moment of the individual spins are randomly oriented

in the xy plane. But as a result of a perturbation of the spin system, a finite

magnetization can be created in a general direction prependicular to B0. T 2 is thus the

time constant for this component to decay to zero.

T2 and Tj are determined by the same mechanisms, i.e., coupling of the nucleus

with atoms, molecules, electrons, etc. in the immediate environment. After perturbation

by an external magnetic field, B0, a return to equilibrium of a nuclear spin system is

obtained through spontaneous transitions between the energy levels split by the external

magnetic field. These transitions are essentially induced by coupling of the spin system

within itself and with the environment. The random motions of the particles in the

environment can generate random fluctuating magnetic fields, which can induce the spin transitions responsible for the relaxation process [Band et al., 1991 and references therein].

The presence of unpaired electrons enhance nuclear relaxation, T f 1 and T 2 '1

(i.e., very short Tj and T 2 values), which also operates via dipolar and contact terms

[Swift, 1973], Electrons have magnetic moments three to five orders of magnitude larger than those of nuclei [Bertini and Luchinat, 1986; Band, 1993]. Therefore the local fluctuating magnetic fields produced by them are orders of magnitude larger than those produced by other magnetic nuclei and thus a nuclear magnetic moment coupled to 96 fluctuating electron magnetic field is provided with an efficient and dominant relaxation

pathway for the nucleus.

In the simplest case, the nuclear relaxation rates are related to the electron spin-lattice relaxation time T ie, as shown below [Swift, 1973].

T i-1 = T2 -1 = 4 y^g^B2 s(s+ l)/3 r* T ie + 2 s (s+l)A2/3 h2Tie (23)

In general, the dipolar interaction, which is the first term, dominates so that T j " 1 a r 6,

and the relative relaxation rates therefore yields relative values for r " 6 for nonequivalent

nuclei in the same complex [La Mar, 1979].

7.3.2 Establishing Connectivities Between the Hyperfine-shifted Resonances in

Oxidized Assimilatory-Type Sulfte Reductase (SiR)

The biggest obstacle to a full utilization of the paramagnetic NMR spectral

information is the unambiguous assignment of the hyperfine-shifted resonances since the

hyperfine shifts erase the functional group/chemical shift correlation established in

diamagnetic systems. As a consequence, assignment of the hyperfine shifted resonances

have relied on comparisons with model compounds [La Mar, 1979; Saterlee, 1985; Bertini

and Luchinat, 1986], analysis of differential paramagnetic relaxation [Bertini and

Luchinat, 1986; Emerson et al., 1988], and isotopic labelling [Saterlee, 1985; de Ropp et

al., 1984; Mayer et al., 1974]. Recently, however, successful applications of 2D NMR

experiments in combination with ID NOE experiments in paramagnetic metalloproteins have been reported in the literature [La Mar, 1993 and references therein; La Mar et a l., 1994]. 97 The general NMR methods of structural determination of diamagnetic proteins are well established [Wuthrich, 1986]. They involve, first, the identification of coupled spin systems using COSY/TOCSY-like methods and, second, the identification of sequential assignments by NOES Y/ROES Y-like methods. The extension of these methods to paramagnetic metalloproteins have been hindered by the belief that the rapid paramagnetic-induced relaxation would render cross peaks undetectable [de Ropp and La

Mar, 1991]. For bond correlation or COSY data, the broad lines (short T 2 ’s) result in both rapid decay of coherence and extensive cancellation of antiphase cross peaks [Bax,

1982; Ernst, 1987], while the short Ti’s severely short-circuit the buildup of nuclear

Overhauser effect or NOESY crosspeaks [Ernst, 1987; Neuhaus and Williamson, 1989],

However, with modifications (discussed in Materials and Methods, Chap, II), these pulse methods have been applied to paramagnetic systems with considerable success [Sadek et al., 1993 and references therein]. The results shown below demonstrate that SiR is amenable to these methods.

Results and Discussion

Figure (26) gives the spectrum of oxidized SiR showing the hyperfine-shifted resonances from the active site of the protein. The labelled resonances will be the focus of the following discussions.

Signals a, j, and I. Figure (27) shows the ID NOE difference spectrum of oxidized

SiR. The trace shows that saturating signal a produces NOE responses from signals j and /. Signal a was postulated to be a p proton on a cluster-bound cysteine because of its unusual chemical shift, which was attributed to the magnetic coupling of the cluster and siroheme [Cowan and Sola, 1990]. The proximity of signal a to signals j and I as demonstrated by NOE data and the relative hyperfine shifts of signals j and I (Table 11) suggest signal j to be the a-cysteine proton, and signal / the geminal partner of signal a. 98 NOESY experiment [Fig 28] shows j and I to be near each other and COSY data also suggests a through-bond coupling between the two (Figure 29a, the crosspeak is not well resolved because it is too close to the diagonal). Results of a WEFT-NOES Y experiment

(details discussed in Chap. II) shows that j and / are also near a signal located 6 . 8 ppm in the diamagnetic region (Figure 30). This signal may be coming from a residue near the cluster. Another possibility is that signals j, /, and the one located at 6 . 8 ppm are coming from residue(s) near the [Fe 4S4] cluster.

Signals c, d, and h. NOESY data (Figure 31) shows that signal c is in close proximity with signals d and h. This data is supported by ID NOE results. The relative position of the hyperfine-shifted resonances, T! values (Table 1 1 ) and intensity of the

NOESY crosspeak between c and d suggest that they are geminal partners. This is confirmed by COSY experiment [Fig 32]. The possibility that signals c, d and h belong to the same spin system is confirmed by TOCSY experiment [Fig. 33]. The connectivity pattern and chemical shifts of signals c, d, and h as well as the chemical shifts of signalsc and d (Table 12) suggests that they may be coming from one of the cysteines binding to the cluster; signals c and d being the geminal (3-cys protons and h the a-cys proton.

However the possibility that they may be coming from propionate functions around the siroheme cannot be excluded.

Signals/ g, and i. NOESY experiment (Figure 31) shows that signals/, g, and i are in close proximity with each other. The intensity of the NOESY crosspeak between/ and g , their hyperfine shifts relative to each other as well as their relative Ti values (Table

1 1 ) suggest that they are geminal partners. This is confirmed by a COSY experiment

[Fig 29b], although the crosspeak between them is not well resolved because of the width of the diagonal. Figure 29b also shows the through-bond connectivity between g and /'.

The possibility that signalsf g, and / belong to the same spin system is confirmed iwtiiwi

Figure 26. NMR spectrum of oxidized SiR showing labelled hyperfine-shifted resonances. The solution conditions and acquisition parameters are the same as in Figure 24. 1

yTV'

I i 1 i i'"t | "i ■■ r" i 1 (“ j- i i i ■"i" | ■ i ~ ~ ~ i i “ " r " p / / ”i—f i — i—i—r—]—r-1— i—i—f—i—r—t/ f-\—i^-i—i—i—p i—r 95 90 85 80 75 25 20 15 0 -5

Figure 27. A 300 MHz *H NMR reference spectrum of SiR (top) in 50 mM potassium phosphate buffer in D 2 0, pH 7.6, 296 K.

The trace below shows the NOE difference spectrum obtained by saturating signal a. © o 101

Table 11. NMR Parameters for the Hyperfine-Shifted Signals in Oxidized Assimilatory-

type Sulfite Reductae from D. vulgaris (Hildenborough).

signal S, ppm Ti, ms8 A,(linewidth in Hz)b T2, m sc

a 93 3 -614 -0.52 b 35 1 .2 d c 23.7 6 - 2 0 1 - 1 .6 d 21.5 4 -271 - 1 .2 e 18 6 -248 -1.3 f 15.5 9 -229 -1.4 g 14.1 6 -2 3 9 -1.3 h 1 1 . 2 i 1 0 . 8 j -2 . 8 15 k -3.2 19 - 187 -1.7

1 -4.2 14 m -13.3 4 -9 0 0 -0.35

8 Cowan and Sola, 1991; b this work, estimated by a Lorentzian fit;c calculated using the reationship A = 1 /7 cT2; d this work, estimated using the super-WEFT sequence and plotting initial points as In (l-Io) versus 1/x. 102

-4.5

—I

-2.5

PPM -2.5 -3.0-3.5 -4.0 -4.5 PPM

Figure 28. A portion of a 500 MHz NMR reference spectrum (top) and NOESY map of oxidized SiR at 296 K emphasizing the crosspeak between signals j and / in the upfield region of the spectrum. The NOESY map was collected with a mixing time of 8 ms (the same crosspeak was seen at tm = 15 ms) using 256 tj blocks (8,000 scans/block) over 25-KHz bandwidth using 2048 tj points at a repetition rate of ~12 s '1. The data was processed by a 45°-sine-beIl-squared window over 256 tj and 2048 12 points zero filled to 2048 x 2048 prior to Fourier transformation. 1 0 .0

- 6.0

-4 .0

15.0 L.

- 2.0

PPM TTTTT ▼ TTT 20.0 , 0 -4 .0 - 6.0 r rmf T 20.0 15.0 10 PPM

Figure 29. 500 MHz magnitude COSY maps of oxidized SiR in 50 mM potassium phosphate buffer in D 2 0 , pH 7.6, 296 K. The maps emphasize the through bond correlation between signals j and / (a); signals / and g ; and signals g and h (b). The maps were processed from different fractions of the same data set collected over 100 KHz bandwidth with 512 tj blocks (10400 scans each block) and 2048t2 points with repitition rate o f-14 s'1. Map (a) was processed with a 0°-sine-bell-squared apodization over 512 tj points and 1048t2 points zero filled to 1048 x 2048 prior to Fourier transformation. Map (b) was processed with a 0°-sine-bell-squared apodization over 400 tj points and 80012 points zero filled to 1048 x 2048 prior to Fourier transformation. 8 104

. -4.0

. - 2.0

. 0.0

. 2.0

. 4.0

. 6.0

PPM i ' — i • ■" ...... i , j , — i • • • • ...... 7.0 ’6.0 5.0 4.0 3.0 2.0 1.0 -.0 -1.0 -2.0 -3.0 -4.0 PPM

Figure 30. A portion of the 500 MHz WEFT-NOES Y map of oxidized SiR in 50 mM potassium phosphate buffer in D 2 0 , pH 7.5, 298 K. The map shows correlation of signals j and / with resonance within the diamagnetic region. The data was collected over a 25-KHz bandwidth with 240 tj blocks (8,000 scans /block) and 2048 t 2 points with a mixing time of 10 ms. The data was processed using 22.5°-sine-bell-squared apodization over 2401 \ points and 2048 t 2 points zero filled to 2048 x 2048 prior to

Fourier transformation. 105

qs

O

10.0

15.0

20.0 ------T« & ! &-V.X.

25.0

PPM T T I... T .... I..... T 24.0 22.0 20.0 1B.0 16.0 14.0 12.0 10.0 PPM

Figure 31. A 500 MHz ‘H NMR reference spectrum and NOESY map of oxidized SiR at 298 K. Sample conditions and acquisition parameters are the same as in Figure 28.

The map shows the correlations among hyperfine -shifted signals in the downfield region of the spectrum. The data was processed using a 60°-sine-bell-squared apodization over

256 1 } points and 2048 tj points zero filled to 2048 x 2048 prior to Fourier transformation. i 106

10.0

15.0

20.0

25.0

PPM 24.0 22.0 20.0 18.0 16.0 14.0 12.0 10.0 PPM

Figure 32. 500 MHz reference spectrum and magnitude COSY map of oxidized SiR in

50 mM potassium phosphate buffer in D 2 0, pH 7.5, 296 K. Crosspeaks are shown between signals c and d and signals g and The crosspeak between f and g cannot be resolved because of the broad diagonal. The 2D data was collected over 100 KHz bandwidth with 256 1 1 blocks (16,000 scans/block) and 20481 2 points with repetition rate o f-14 s"1. The data was processed with 0°-sine-bell-square apodization over 256 tj points and 60012 points zero filled to 1048 x 2048 prior to Fourier transformation. 107

T

| " i— I" i v—i—i— r—v—i— |— i— i—i— i—i—i— i—i— i— p — ?— i i i — i— i— i—i — i— |— 25 20 15 10 ' 108

Figure 33. A 600 MHz TOCSY map of oxidized SiR in SO mM potassium phosphate buffer in D 2 0, pH 7.5, 298 K showing correlations between signals in the downfield region of the spectrum belonging to the same spin system. The data was collected over

48 KHz bandwidth with 256 tj blocks (6,144 scans/block) and 20481 2 points. MLEV-

17 pulse train was used. A mixing (spin lock) time of 1.5 ms and a trim pulse length of

500 ps were used. The data was processed using 0°-sine-bell-square apodization over

256 tj and t 2 points. A severe window function had to be used in order to see the crosspeaks. This in turn made the phase of the crosspeaks hard to correct (the crosspeaks should all be in the absorptive mode). The acquisition parameter values were not optimal for this experiment. 109 Table 12. Hyperfine-Shifted Resonances of Cystein P-CH 2 S in Iron-Sulfur Containing

Proteins at 298 K.

source 8 , ppm reference

[Fe4 S4 ]2+ coupled to Low-spin siroheme

E c o li SiR-HP-CN" 16.21, 15.96, 11.88, 11.66, 11.51,-13.28 a

E. coli SiR-HP-S032' 25.2, 25.2, 17.0,14.2, -13.45 a

E. coli SiR-HP-N02' 17.32, 15.921, 15.921,-14.411 a

Oxidized Ferredoxins [Fe 4 S4 ]2+

C. pasteurianum 17.3, 16.2, 15.9, 14.9, 13.6, 12.3, 12.0, 11.2, b

9.3, 9.4, 9.0, 8.0, 6 .8 , 6.2, 5.0,4.9,

C. acid urici 16.2, 15.3, 15.1, 13.3, 12.1, 11.1, 10.8 c

Reduced HIPIPs [Fe 4 S4]2+

C. vinosum 16.8,15.9, 12.7, 11.6, 10.7, 7.7, 7.2, 5.3 d

C. gracile 16.2, 16.2, 1 2 .6 , 18.8, 1 0 . 0 e

a SiR-HP-X, sulfite reductase heme protein subunit with ligand (X= CN", SO 3 2", NO2 ") [Kaufman et al, 1993];b Sadek et al., 1993; c Packer et al., 1977; d Bertini et al., 1991; e Sola et al.,1989 (P-CH 2 and a-CH not resolved). 110 by TOCSY experiment (Figure 33). The pattern of connectivities between the three signals [NOESY and TOCSY results] and the position of the geminal partners/and g

(Table 12) strongly suggest that they are coming from a cystein cluster ligand.

Signals b and k. Figure 34 shows the NOE difference spectrum on saturating signal b which gave a response from proton(s) under signal k. The result needs to be confirmed by improving signal to noise ratio. It has been postulated that there is at least one methyl group under signal k which may be coming from a residue near the siroheme [Cowan and

Sola, 1990].

Signal e. ID NOE and 2D NOESY and COSY experiments failed to show connectivities between signal e and other hyperfine-shified resonances.

Signal m. COSY and ID NOE experiments failed to reveal the correlation between signal m and other resonances in the SiR spectrum. On the other hand, parameters used in the TOCSY and NOESY experiments were not tailored to detect correlations involving signal m. Initial XH NMR studies on SiR suggest that signal m maybe coming from a histidine residue acting as the sixth axial ligand in siroheme (Figure

25) [Cowan and Sola, 1990].

The WEFT-NOESY Method. The utility of this method to monitor and assign amino acid residues near the paramagnetic prosthetic center has been demonstrated in the literature [La Mar et al., 1994], The method effectively suppresses the slow relaxing signals coming from residues not experiencing the paramagnetic effect of the prosthetic group via the WEFT sequence at the beginning of the pulse program. This enables the relatively fast relaxing residues in the active site of the protein to be seen in the otherwise crowded diamagnetic region. The application of this method to SiR shows promise “i—r -i—i—i—r " I_l "» 1 I 1 ~ l 1----1----1----1----1----1----1----1---- — |— i— i— i— i— j— i— i— r i i i i i ■ • r 40 35 30 25 20 15 10 0 -5 PPM

Figure 34. A 300 MHz NMR reference spectrum of SiR (top) in 50 mM potassium phosphate buffer in DjO, pH 7.6,296 K.

The trace below shows the NOE difference spectrum obtained by saturating signal b. 112 as shown in Figures 30 and 35. In contrast to the conventional NOESY program that uses presaturation to suppress the solvent signal, WEFT -NOESY can show

connectivities between signals in the diamagnetic region coming from residues near the

paramagnetic center experiencing enhanced relaxation. It also enables the observation of connectivities between hyperfine shifted signals and signals within the diamagnetic region

(Figure 30).

7.3.3 Paramagnetic NMR Studies on the Fully Reduced SiR

The spectrum of the fully reduced SiR is shown in trace c .Figure 36. The hyperfine-

shifted resonances range from 113 ppm to - 60 ppm (Table 13). The wide spread and

significant broadening of the hyperfine-shifted resonances indicate that siroheme is high-

spin in the fully reduced form of SiR. The spectrum is very similar to that of the fully

reduced high-spin pentacoordinate sulfite reductase from E.coli (Figure 37) [Kaufman et

al., 1993], These observations supports the idea that the sixth ligand to the siroheme

(proposed to be a histidine residue) dissociates from the siroheme upon reduction clearing

the way for binding of substrate and its subsequent reduction. Figure 38 shows the temperature dependence of the hyperfine-shifted resonances. All signals follow a Curie temperature dependence except for two signals at -75 ppm and -16 ppm which shows

slight anti- Curie temperature dependence. NOE data also shows that saturating the

resonance at 113 ppm gives an NOE response from the resonance at 36.5 ppm (Figure

39). A peak at almost the same position is present in the E. coli spectrum (Figure 37); this peak is unambiguously assigned to a P-proton on a cluster-bound cysteine. In comparison to this observation, the similar signal in SiR may also be a P-cysteine proton with the resonance at 36.5 ppm corresponding to its geminal partner. Partial oxidation of the fully reduced SiR showed a mixture of resonances that correspond to both fully oxidized and fully reduced SiR. In addition, peaks that do not correspond to either fully oxidized or fully reduced SiR were also seen (Figure 36, trace b). These peaks PPM

•■■I"" "T • " I ' 12.0 10.0 B.O 6.0 4.0 2.0 PPM

Figure 35. A portion of the 500 MHz WEFT-NOESY map of oxidized SiR showing crosspeaks between fast-relaxing resonances inside the diamagnetic region. Experimental conditions were the same as in Figure 30. t

its V A- T ¥ IS •AO MM* 1 -

Figure 36. 500 MHz paramagnetic ^H NMR spectrum of (a) fully oxidized; (b) partially reduced; and (c) fully reduced SiR in 50 mM potassium phosphate buffer in D 2 O, 298 K. The data was collected using the super-WEFT sequence over a 125 KHz bandwidth using 16 K points. The relaxation (AQ + delay) time was 79 ms and x was between 60 to 65 ms. A line broadening of 25 Hz was applied prior to Fourier transformation. 115

Table 13. NMR Parameters for the Fully Reduced Assimilatory-Type Sulfite Reductase (SiR) from D. vulgaris (Hildenborough) at 298 K

signal 3 §, ppm signal 5, ppm

1 113.5 8 1 2 . 2

2 75.4 9 -3.2

3 36.6 1 0 -5.0

4 35.2 11 -13.5

5 26.5 1 2 -54.9

6 15.9 13 -60.9

7 13.0

3 numbering starts with the most hyperfine-shifted signal in the downfield region 116

f I i y -] — i— i— r— j— i— i— i— |— i— i— i— |— i— i— i— j— i— i i | 1 i < " i 1 "" I 1 r i < « | i 11 j i 120 100 80 60 40 20 -2 0 -to -60

Figure 37. *11 NMR spectrum of the fully reduced pentacoordinate high-spin sulfite reductase from E. coli [Kaufman et al., 1993], Figure 38. Temperature dependence of the hyperfine-shifted resonances in folly reduced follyin reduced resonances hyperfine-shifted the of dependence Temperature 38. Figure SiR, pH 7.5. pH SiR,

Chemical Shift, ppm 100 120 -60 - 20 .0 .5 .0 .5 .0 .5 .0 3.65 3.60 3.55 3.50 3.45 3.40 3.35 3.30 ; / 103 x 1/T —■ ▼— 117 Figure 39. NOE difference spectrum obtained by saturating the resonance at 113 ppm in fully reduced SiR. The trace below is the *H NMR reference spectrum of fully reduced

SiR. Experimental conditions are the same as in Figure 36. may be attributed to the partially reduced (one- electron reduced) SiR. A similar observation was reported for sulfite reductase fron K co li (Kaufman et al., 1993). References

Adar, F. (1978) in The Porphyrins, vol. 3, pp. 176-209, Academic Press, New York.

Akagi, J. M., Adams, V. (1973)7. Bacteriol. 116, 392-396.

Averill, B. A., Tiejde, J. M. (1982) FEBSLett. 138, 8-12.

Badziong, W , Thauer, R. K , (1980) Arch. Microbiol. 125, 167-169.

Banci, L., Bertini, I., Luchinat, C. (1991), Nuclear and Electron Relaxation: The Electron Nucleus Hyperfine Coupling in Solution, VCH, Weinhein, Germany.

Barkigia, K. M., Fajer, J., Chang, C. K., Hanson, G. J.B. (1982) J. Am. Chem. Soc. 104, 315-317.

Bastian, N. R., Wink, D. A., Wacket, L. P., Livingston, D. J., Jordan, L. M. Fox, J. Orme-Johnson, W. H., Walsh, C. T. (1988) in The Bioinorganic Chemistry o f Nickel, pp. 227-247 VCH, New York.

Bax, A., (1982), Two Dimentional Nuclear Magnetic Resonance in Liquids, D. Reidel Publishing Co., Dordrecht, Holland.

Berg, J. M., Holm, R. H. (1985) J. Am. Chem. Soc. 107, 925.

Bemardi, G. “Chromatography of Nucleic Acids on Hydroxyapatite”, in M ethods o f Enzymology, vol. 21, 95-102.

Bertini, I., Briganti, F., Luchinat, C., Scozzafava, A., Sola, M.(1991)7 Am Chem. Soc. 113, 1237-1245.

Bertini, I., Luchinat, C. (1986,1 NMR o f Paramagnetic Molecules in Biological Systems, The Benjamin/Cummings Publishing Co., Inc., Menlo Park, CA.

Burgess, J. (1978) Ions in Solution: Basic Principles o f Chemical Interactions, Ellis Horwood Limited, West Sussex, England

Burstyn, J. N., Deal, K. A. (1993) Inorg. Chem 32, 3585-86.

Camack, R., Hucklesby, D., Hewitt, E. J. (1978) Biochem. J. 171, 519-526. 120 121 Chang, C. K., Fajer, J. (1980) 7. Am. Chem. Soc. 102, 848-851.

Chang, C. K., Hanson, L. K., Richardson, P. F., Young, R., Fajer, J. (1981) Proc. Natl. Acad. Sci. U. S. A. 78, 2652-2656.

Chen, Z., deRopp, J. S., Hernandez, G., LaMar, G. N., (1994) J. Am. Chem. Soc. 116, 8772-8783.

Christner, J. A., Munck, E., Janick, P. A., Siegel, L. M. (1981)./. Biol. Chem. 256, 2098- 2101.

Christner, J. A., Munck, E., Janick, P. A., Siegel, L. M. (1983)7. Biol. Chem. 258, 11147-11156.

Cline, J. F., Janick, P. A., Siegel, L. M., Hoffman, B. M. (1985) Biochemistry 24, 7942- 7947.

Cline, J. F., Janick, P. A., Siegel, L. M., Hoffman, B. M. (1986) Biochemistry 25, 4647- 4654.

Coucouvanis, D., Mosier, P. E., Demadis, K. D., Patton, S., Malinak, S. M., Kim, C. G.,

Tyson, M. A., (1993)7. Am. Chem. 115, 12193-12194.

Cowan, J. A., Gray, H. B. (1989 )Inorg. Chem. 28, 4554-5556.

Cowan, J. A., Sola, M. (1990), Inorganic Chemictry 29, 2176-2179.

CRC, Handbook of Chemistry and Physics (1990) 71st ed., pp. 8-37, CRC Press, Inc., Boca Raton, FI.

Czechowski, M., Fauque, G., Galliano, N., Dimon, B., Moura, I., Moura, J. J. G., Xavier, A. V., Barata, B. A. S., Lino, A. R., LeGall, J. (1986)7. Ind. Microbiol 1, 139- 147. de Ropp, J. S., La Mar, G. N., (1991) 7. Am. Chem. Soc. 113, 4348-4350. de Ropp, J. S., LaMar, G. N., Smith, K. M., Langry, K. C. (1984) 7 Am. Chem. Soc. 106, 4438-4444.

DiFeo, T. J., Addison, A. W. (1991) Inorg. Chem. 30, 1151-1153.

Emerson, S. D., Lecomte, J. T. J, LaMar, G. N., (1988)7. Am. Chem. Soc. 1 1 0 , 4176- 4182. 122 Ernst, R. R., Bodenhausen, G., Wokaun, A., (1987), Principles o fNuclear Magnetic Resonance in One and Two Dimensions, Clarendon Press, Oxford, England.

Fersht, A. (1985,) Enzyme Structure and Mechanism, Freeman, New York

Gelasco, A., Pecoraro, V. L., (1993)7. Am. Chem. Soc. 115, 7928-29.

Godden, J. W., Turley, S., Teller, D. C., Adman, E. J., Lui, M. Y„ Payne, W. S., LeGall, J. (1991) Science 253, 438-442.

Gouterman, M. (1978) in The Porphyrins, vol. 3, pp. 1-165, Academic Press, New York.

Han, S., Madden, J. F., Siegel, L. M., Spiro, T. D. (1989 ) Biochemistry 28, 5477-5485.

Hatchikian, E. C., Zeikus, J. G. (1983) J. Bacteriol 153, 1211-1220.

Henry, Y., Bessieres, P. (1984) Biochimie 266, 259-289.

Huxtable, R. J. (1986) Biochemistry o f Sulfur, vol. 6 , Biochemistry of the Elements Series, Friedman, E., Ed., Plenum Press, New York.

Huynh, B. H., Kang, L., DerVartanian, D. V., Peck, H. D., Jr., LeGall, J. (1984) J. Biol. Chem. 259, 15373-15376.

Inobushi, T., Becker, E. D. (1983)7. Mag. Reson. 51, 128-133.

Isaacs, N. S., (198 1) Physical Organic Chemistry, p. 102, Longman, New York,.

Janda, M., Hemmerich, P. (\916) Angew, Chem. Int. Ed. Engl. 15, 443-444.

Janick, P. A., Rueger, D. C., Krueger, R. J., Barber, M. J., Siegel, L. M. (1983) Biochemistry 22, 396-408.

Janick, P. A., Siegel, L. M. (1982) Biochemistry 2 1 , 3538-3547.

Janick, P. A., Siegel, L. M. (1983) Biochemistry 22, 504-515.

Jesson, J. P. (1973), "The Paramagnetic Shift " in NMR o f Paramagnetic Molecules, pp. 2-45, La Mar, G. N., Horrocks, W. DeW., Jr., Holm, R. H., Eds., Academic Press, Inc., New York.

Jolly, W. L. (1984) Modem Inorganic Chemistry, pp. 365-370, 438-444, McGraw-Hill, New York. Jones, H. E., Skyring, G. W. (1975) Biochem. Biophys. Acta 377, 52-60.

Kang, L., LeGall, J.m Kowal, A.T., Johnson, M. K. (1987) J. Inorg. Biochem. 30, 273- 290.

Kaufman, J., Siegel, L. M., Spicer, L. D. (1993a) Biochemistry 32, 8782-8791.

Kaufman, J., Spicer, L. D., Siegel, L. M. (1993b ) Biochemistry 32, 2853-2867.

Kobayashi, K., Takahasi, E., Ishimoto, M. (1972) J Biochem. (Tokyo) 72, 879-887.

Kroger, A. (1978) Biochim. Biophys. Acta 505, 129-145.

Krouse, H. R., McReady, R. G. L. (1979) “Biogeochemical Cycling of Sulfur” in Biogeochemical Cycling o f Mineral-Forming Elements, Trudinger, P. A., Swaine, D. J., Eds., Elsevier, Amsterdam.

Krueger, R., Siegel, L. M. (1982) Biochemistry 21, 2892-2904.

Krueger, R., Siegel, L. M. (1982) Biochemistry 21, 2905-2909.

La Mar, G. N., de Ropp, J. S., (1993) "NMR Methodology for Paramagnetic Proteins," in NMR o f Paramagnetic Molecules, Biological Magnetic Resonance Series, vol. 12, Berliner, L. J., Reuben, J., Eds., Plenum Press, New York.

LaMar, G. N. (1973) "Spin Delocalization and Electronic Structure" in NMR o f Paramagnetic Molecules, LaMar, G. N., Horrocks, W. DeW., Jr., Holm, R. H., Eds., Academic Press, Inc., New York.

LaMar, G. N. (1979) "Model Compounds as Aids in Interpreting NMR Spectra of Hemoproteins" in Biological Applications o f Magnetic Resonance, Shullman, R. G., Ed., Academic Press, New York.

Lancaster, J. R., Vega, J. M., Kamin, H., Orme-Johnson, W. H., Krueger, R. J., Siegel, L. M. (1979) J. Biol. Chem. 254, 1268-1272.

Lee, J. P., LeGall, J., Peck, H. D., Jr. (1973a) J. Bacteriol. 115, 529-542.

Lee, J. P., LeGall, J., Peck, H. D., Jr. (1973b) J. Bacteriol. 115, 453-455.

Lee, J. P., Peck, H. D. (1971) Biochem. Biophys. Res. Commun. 45, 583-589.

Lee, J. P., Yi, C. S., LeGall J., Peck H. D., Jr. (1973) J. Bacteriol. 115, 529-542. LeGall, J., Fauque, G. (1988) “Dissimilatory Reduction of Sulfur Compounds” in 124 ' Biology o f Anaerobic Microorganisms, Zehder, A. J. B., Ed., John-Wiley and Sons, Inc., New York.

Leising, R. A., Kim, J., Perez, M. A,, Que, L. J., (1993)7. Am. Chem. Soc, 115, 9524- 9530.

Liang, W., Cowan, J. A. (1994) Inorg. Chem. 33, 4604-4606.

Lino, A. R., Moura, J. J. G., Xavier A. V., Fauque, G., LeGall, J. (1985) Rev. Port. Quim. 27, 215.

Lui, S.M. (1994) Ph. D. Dissertation.

Lui, S. M., Liang, W., Soriano, A., Cowan, J.A. (1994) 7. Am. Chem. Soc. 116, 4531 - 4536.

Lui, S. M., Soriano, A., Cowan, J. A. (1993a) 7. Am. Chem. Soc. 115, 10483-10486.

Lui, S. M., Soriano, A. Cowan, J. A., (1994b ) Biochem. 7. 304, 441-447.

Madden, J. F., Han, S., Siegel, L. M., Spiro, T. G. (1989) Biochemistry 28, 5471-5477.

Massey, V., Hemmerich, P. (1978), Biochemistry 17, 9-17.

McRee, D. E., Richardson, D. C., Richardson, J. S., Siegel, L. M. (1986)7. Biol. Chem. 261, 10277-10281.

Melamed, D., Sullivan, E.P., Prendergast, K., Strauss, S. H., Spiro, T. G. (1991) Inorg. Chem. 30, 1308-1319.

Meyer, A., Ogawa, S., Shullman, R. G., Yamane, T., Cavaleiro, J. A. S., Rocha Gonsalves, A. M. d'A., Kenner, G. W., Smith, K. M., (1974)7. Mol. Bio. 8 6 , 749-756.

Moura, I., LeGall, J., Lino, A. R., Peck, H. D., Fauque, G., Xavier, A. V., DerVartanian, D. V., Moura, J. J. G., Huynh, B. H. (1988)7. Am. Chem. Soc. 110, 1075-1082,

Moura, I., Lina, A. R., Moura, J. J. G., Xavier, A.V., Fauque, G., Peck, H. D., LeGall, J. (1986)7. Biochem. Biophys. Res. Commun. 141, 1032-1041.

Moura, J. J. G., Moura, I., Santos, H., Xavier, A. V., Scandellari, M., LeGall, J. (1982) Biochim. Biophys. Res. Commun. 108, 1002-1009.

Murphy, M. J., Siegel, L. M., Kamin, H. (1974)7. Biol. Chem. 249, 1610-1614. 125 Neuhaus, D., Williamson, M., (1989) The Nuclear Overhauser Effect in Structural and Conformational Analysis, VCH Publishers, New York.

Odom, J. M., Peck, H. D., Jr., (1981) J. Bacteriol. 147, 161-169.

Olson, J. S. (1981 ), Methods in Enzymol. 76, 631-651.

Ostrowski, J., Barber, M. J., Ruegers, D. C., Miller, B. E., Siegel, L. M., Kredich, N. M. (1989a) J. Biol. Chem. 264, 15726-15737.

Ostrowski, J., Wu, J.Y., Ruegers, D. C., Miller, B. E., Siegel, L. M., Kredich, N. M. (1989b) J. Biol. Chem. 264, 15796-15808.

Packer, E. L., Sweeney, W. V., Rabinowitz, J. C., Stemlicht, H., Shaw, E. N. (1977) J. Biol. Chem. 252, 2245-2253.

Palmer, T. (1985), Understanding Enzymes, pp. 161-163, Ellis Horwood Publishers, New York.

Petratos, K., Beppu, T., Banner, D. W., Tsemoglou, D. (1986)7. Mol. Biol. 190, 135.

Pierik, A. J., Hagen, W. R. (1991) Eur. J. Biochem. 195, 505-516.

Postgate, J. R. (1984) The Sulphate-Reducing Bacteria, Cambridge University Press, New York.

Prabhakararo, K., Nicholas, D. J. D. (1969)Biochim. Biophys. Acta 180, 253-263.

Procyk, A. D., Bocian, D. (1991) J.Am. Chem. Soc. 113, 3765-3773.

Rapp, B. J., Wall, J. D. (1987) Proc. Natl. Acad. Sci., U.S.A. 84, 9128-9130.

Richardson, P. F., Chang C. K., Spaulding, L. D., Fajer, J. (1979)7 Am. Chem. Soc. Commun. 1 0 1 , 7736-7738.

Sadek, M., Brownlee, T.T.C., Scrofani, S.D.B., Wedd, A. G. (1993)7 Mag. Reson.llll

Satterle, J. D. (1985 )Annu. Rep. NMR Spectrosc. 17, 79-178.

Scott, A. I., Irwin, A. J., Siegel, L. M., Shoolery, J. N. (1978,) 7 Am. Chem. Soc. 100, 316-318.

Scott, R. A., Wumft, W. G., Coyle, C. L., Doolie, D. M. (1989) Proc. Natl. Acad. Sci. U. S. A. 8 6 , 4082-4086. 126 Seki, Y., Ishimoto, M. (1979) J. Biochem. (Tokyo) 8 6 , 273-276.

Seki, Y., Sogawa, N., Ishimoto, M., (1981)/. Biochem. (Tokyo) 90, 1487-1492.

Shapleigh, J. P., Hosier, J. P., Tecklenburgh, M. M. J., Kin, Y., Babcock, G. T., Gennis, R. B., Fergusson-Miller, S. (1992) Proc. Natl. Acad Sci. U.S.A. 89, 4786-4790.

Siegel, L. M., Kamin, H. (1968) in Flavins and Flavoproteins, pp. 15-40, Yagi, K., Ed., University Park Press, Baltimore.

Siegel, L. M., Davis, P. S. (1974)/. Biol. Chem. 249,1587-1598.

Siegel, L. M., Davis, P. S., Kamin, H. (1974)/. Biol. Chem. 249, 1572-1586.

Siegel, L. M., Murphy, M. J., Kamin, H. (1973), /. Biol. Chem. 248, 251-264.

Siegel, L. M., Murphy, M. J., Kamin, H., (1978) Methods in Enzymol. ,vol. 52, 436-447.

Siegel, L. M., Rueger, D.C., Barber, M. J., Krueger, R. J., Orme-Johnson, N. R., Orme- Johnson, W. H. (1982)/. Biol. Chem. 257, 6343-6350.

Smith, D. W., Williams, R. J. P. (1970) Struct. Bond. 7, 1-45.

Sola, M., Cowan, J. A., Gray, H. (1989) J Biochemistry 28, 5261-5268.

Soriano, A., Cowan J. A. (1995) /. Am. Chem. Soc. 117,4724-4725.

Stolzenberg, A.M., Spreer, L. O., Holm, R. H. (1980)/. Am. Chem. Soc. 102, 364-370.

Stolzenberg, A. M., Strauss, S. H., Holm, R. H. (1981)/. Am. Chem. Soc. 103, 4763- 4778.

Strauss, S. H., Holm, R. H. (1982) Inorg. Chem. 21, 863.

Sullivan, E. P., Grantham, J. D., Thomas, C.S., Strauss, S. H. (1991)/. Am. Chem. Soc. 113, 5264-5270.

Swift, T. J., (1973), "The Paramagnetic Linewidth" in NMR o f Paramagnetic M olecules, LaMar, G. N., Horrocks, W. DeW., Jr., Holm, R. H., Eds., Academic Press, Inc., New York.

Tan, J., Cowan, J. A. (1991) Biochemistry 30, 8910-8917.

Tan, J., Helms, L., Swenson, R. P., Cowan, J. A.(1991) Biochemistry 30, 9900-9907. 127 Tan, J., Soriano, A., Lui, S. M., Cowan, J. A., (1994) Arch. Biochem. Biophys. 312, 516- 523

Togo, H., Oae, S. (1992) “Reduction of Sulfur Compounds” in Organic Sulfur Chemistry: Biochemical Aspects, Oae, S., Okuyama, T., Eds., CRC Press, Boca Raton, FI.

Traore, A. S., Hachikian, J. P., Belaich, J. P., LeGall, J. (1981) J. Bacteriol. 145, 191- 199.

Trudinger, P. A. (1970) J. Bacteriol. 104, 158-170.

Tsuji, K., Yagi, T. (1980) Arch. Microbiol. 125, 35-42.

Vega, J. M., Kamin, H. (1977) J. Biol. Chem. 252, 896-909.

Weeg-Aerssens, E., Tiedje, J.M., Averill, B. A. (1988) J. Am. Chem. Soc. 1 1 0 , 6851- 6856.

Wilkerson, J. O., Janick, P. A. and Siegel, L. M. (1983) Biochemistry 22, 5048-5054.

Wolfe, B. M., Lui, S.M., Cowan, J. A. (1994) Eur. J. Biochem. 223, 79-89.

Wuthrich, K. (1986)NMR o f Proteins and Nucleic Acids, Wiley, New York.

Yoneda, F., Sakuma, Y., Ichiba, M., Shinomura, K. (1976) J. Am. Chem. Soc. 98, 830-

835

Zehnder, A. J. B., Wuhrmann, K. (1976) Science, 1165-1116. Appendix

; SUWEFT.AU ( in Broker system)

;(Delay - 180 - x - 90 - AQ)

1 ZE

2 D1

3 PI * 2 PHI

4 D 6

5 P1PH2

6 GO = 2 PH3

EXIT

PHI = 0 2 2 0 1 3 3 1

PH2= 00 221133

PH3 = R0 R0R2R2R1 R1 R3 R3

; D1 = delay

; D6 = x

; D1 + AQ = total relaxation time for hyperfine-shifted signals; usually 3 T r STi where

Ti corresponds to the slowest relaxing peak.

128 129 Pulse/Processing Programs Used for NOE Experiments (in Bruker system)

;NOE BOTH.AU ;(NOE with super-WEFT to suppress strong solvent signals; writes ;both “ON”resonance and “OFF” resonance spectra)

1 ST AO ZE STA ZE 2 D1 DO S3 3 0 2 4 PI *2 5 D3 6 HG 7 D3 * 39 8 DO 9 PI 1 0 GO = 2 ST AO 1 1 WR # 1 1 2 IF # 1 STA WR#2 IF # 2 13 IN = 1 EXIT

;set NS = 2 * N ;set NE = number of pairs of blocks to be acquired ;set ASTI = 1 ;set NBL = 2 ;set CP ;set first filename as “OFF” ;set second filename as “ON” ;FL = frequency list /02 list [ fi, f 2, ft(= fi), ft] ;use NOESUM. AU to get reference spectrum ;use NOEDIF. AU to get difference spectrum ;determine D1 and D 6 by acquiring ID spectrum using super-WEFT program ;D3 = D6/40 ;NOESUM.AU

ZE

WR # 2 1 R E # 1 2 IF # 1 3 AT # 2 WR # 2 4 IN = 1 EXIT

;set NE as in NOEBOTH. AU ;set DC = 1/NE ;set ASTI = 0 ;set filename # 1 = filename of “OFF” spectra to be summed. ;set filename # 2 = filename of reference spectrum (result of summation)

; NOEDIF.AU

1 ZE 2 W R# 3 3 R E # 1 4 IF# 1 5 NM 6 AT #3 WR # 3 7 RE # 2 8 IF # 2 9 AT #3 1 0 IN = 2 1 1 W R# 3 EXIT

;set NE as in NOEBOTH. AU ;set DC = 1 ;set ASTI = 0 ;filename # 1 = filename of “OFF” spectra ;filename # 2 = filename of “ON” spectra ;filename # 3 = filename of difference spectrum