Structural determination and functional annotation of ChuS and ChuX,

two members of the heme utilization operon in pathogenic

Escherichia coli O157:H7

By

MICHAEL DOUGLAS LEO SUITS

A thesis submitted to the Department of Biochemistry in conformity with the

requirements for the degree of Doctor of Philosophy

Queen's University

Kingston, Ontario, Canada

April, 2007

Copyright © Michael Douglas Leo Suits, 2007

ii

Abstract

For pathogenic microorganisms, heme uptake and degradation is a critical

mechanism for iron acquisition that enables multiplication and survival within hosts they

invade. While the bacterial proteins involved in heme transport had been identified at the

initiation of our investigation, the fate of heme once it reached the cytoplasm was largely

uncharacterized. Here we report the first crystal structures of two members of the heme

utilization operon from the human pathogen Escherichia coli O157:H7. These are the

heme oxygenase ChuS in its apo and heme-complexed forms, and the apo form of heme

binding protein ChuX. Surprisingly, despite minimal sequence similarity between the N-

and C-terminal halves, the structure of ChuS is a structural repeat. Furthermore, the ChuS

monomer forms a topology that is similar to the homodimeric structure of ChuX. Based

on spectral analysis and carbon monoxide measurement by gas chromatography, we

demonstrated that ChuS is a heme oxygenase, the first to be identified in any E. coli

strain. We also show that ChuS coordinates heme in a unique fashion relative to other

heme oxygenases, potentially contributing to its enhanced activity. As ChuS and ChuX

share structural homology, we extended the structural insight gained in our analysis of

ChuS to purport a hypothesis of heme binding for ChuX. Furthermore, we demonstrated

that ChuX may serve to modulate cytoplasmic stores of heme by binding heme and

transferring it to other hemoproteins such as ChuS. Based on sequence and structural

comparisons, we designed a number of site-directed mutations in ChuS and ChuX to

probe heme binding sites and mechanisms in each. ChuS and ChuX mutants were

analyzed through reconstitution experiments with heme and functional analyses,

including enzyme catalysis by ChuS and mutants, and in culture development during

heme challenge experiments by ChuX and mutants. Taken together, our results suggested

that ChuX acts upstream of ChuS, and regulates heme uptake through ChuX-mediated

heme binding and release. ChuS can degrade heme as a potential iron source or

antioxidant, thereby contributing directly to E. coli O157:H7 pathogenesis. Functional

implications that may be revealed from sequence and structure based information will be

addressed as they pertained to our evaluation of ChuS and ChuX.

iii

Acknowledgements

First and foremost, I would like to express my sincerest gratitude to my supervisor,

Dr. Zongchao Jia, for his guidance, support, and belief in my abilities. His clever insight

and unwavering encouragement have enabled me to attain numerous goals and have

instilled in me the drive to continue to seek greater scientific achievement. I would also

like to acknowledge the support of past and present members of the Jia research group,

including Dr. Melanie Adams for her crystallographic expertise and contagious drive, and

Drs. Gour Pal, Vinay Singh and Qilu Ye, who have allowed me to draw on their vast and

various scientific backgrounds. I am also grateful to Jarrett Adams, Daniel Lee, Peter

Iannuzzi, Dr. Allan Matte, Dr. Chris Udell, Andrew Wong, and Jimin Zheng for their

support, cooperation and technical expertise. Thank you to Drs. Kanji Nakatsu, Michael

Nesheim, Steve Smith, and Glenville Jones for professional support and mentoring. I

extend sincere appreciation to the Synchrotron facilities, notably the beamline scientists

at APS, NSLS, and CHESS for technical support and assistance in interpreting diffraction

data. I am especially grateful to Leonid Flaks and Howard Robinson, at NSLS beamlines

X8C and X29, respectively.

I would like to acknowledge my siblings, parents and grandparents, to whom I am

so appreciative for their unbounded love and encouragement. Their collective passion for

learning, quest for knowledge, and enduring support has inspired my own academic

career. Thank you, thank you, and thank you.

I would like to thank my wife Jocelyne for speaking the same language, and for

encouraging me through technical insights and exceptional writing abilities to better

convey ideas and concepts on paper. Thank you for sharing love and laughter in all

aspects of our life.

This research has been funded by the Canadian Institutes of Health Research and

various Queen’s University Fellowships including: Queen’s Graduate Award, R.S.

McLaughlin, E.G. Bauman, and Franklin Bracken fellowships and Graduate Dean’s

Travel Grant for Ph.D. Field Research. Further support was provided by an Ontario

Graduate Scholarship.

iv

TABLE OF CONTENTS

Abstract ii Acknowledgements iii List of Tables vii List of Figures viii List of Abbreviations x List of Protein Mutants xi Chapter 1: General Introduction 1

1.1. General introduction and relevance 2 1.2. Bacterial iron acquisition from heme sources 4 1.3. Role of heme and structural diversity of heme-proteins 12 1.4. Bacterial heme oxygenases 15 1.5. Heme utilization protein ChuS 17 1.6. ChuW, ChuX, ChuY and bacterial heme proteins 18

Chapter 2: Identification of an Escherichia coli O157:H7 heme oxygenase with

tandem functional repeats 22

2.1. Abstract 23 2.2. Introduction 24 2.3. Materials and methods 27

2.3.1 Sequence analysis, protein expression, and purification 27 2.3.2 Crystallization of full-length ChuS 28

2.3.3 Data collection and structure determination 28 2.3.4 Reconstitution with heme and absorbance measurements 29 2.3.5 CO activity assay and inhibition 30 2.3.6 Dynamic light scattering of C-ChuS and N-ChuS 30 2.3.7 EPR experiments 31

2.4. Results 31 2.4.1 Expression and purification 31 2.4.2 Structure determination and analysis 32

2.4.3 ChuS contains a structural duplication 36 2.4.4 Spectral properties of the heme-ChuS complex 38 2.4.5 Heme oxygenation by ChuS 43 2.4.6 Detection of CO release 45 2.4.7 Organization of the heme uptake operon 45 2.4.8 Dynamic light scattering 46 2.4.9 Inhibition by Sn(IV) mesoporphyrin 47

2.5. Discussion 51 2.6. Acknowledgements 54

v

Chapter 3: Structure of the Escherichia coli heme oxygenase ChuS in complex with heme and enzymatic inactivation by mutation of the heme coordinating residue His-193. 55

3.1. Abstract 56 3.2. Introduction 57 3.3. Experimental Procedures 59

3.3.1 Protein expression and purification, site-directed mutagenesis 59 3.3.2 ChuS-heme reconstitution 61 3.3.3 Crystallization of ChuS-heme 62 3.3.4 Data collection and structure determination 62 3.3.5 Spectral analysis and CO measurement 63

3.4. Results 64 3.4.1 Structure determination and analysis of ChuS-heme 64 3.4.2 Heme binding 68 3.4.3 Spectroscopic properties of ChuS, H73A, and H193N in complex with heme 71 3.4.4 Heme degradation by ChuS 71 3.5. Discussion 75 3.6. Footnotes 78

Chapter 4: Structure of the Escherichia coli O157:H7 heme binding protein ChuX

and its role in regulating heme uptake 80

4.1. Abstract 81

4.2. Introduction 82

4.3. Experimental procedures 85 4.3.1 Sequence analysis, protein expression and purification,

site directed mutagenesis Crystallization and Data Collection 85

4.3.2 Crystallization of ChuX 88

4.3.3 Data collection and structure determination 88

4.3.4 Heme reconstitution 89

4.3.5 Growth of ChuX and DM in heme 90

4.4. Results 90

4.4.1 Sequence analysis 90

4.4.2 ChuX protein fold and structural analysis 91

4.4.3 The ChuX structure may represent an open and closed

conformation 99

vi

4.4.4 ChuX-heme association 100

4.4.5 ChuX heme binding affinity 104

4.4.6 ChuX protection from heme toxicity 104

4.5. Discussion 107

4.6. Footnotes 111

Chapter 5: General discussion and summary 112

5.1. Gaining functional information from protein structures 113

5.2. ChuS and ChuX: structural and functional implications

for heme utilization 118

5.3. Inhibition of heme utilization as a potential mode for

treatment 123

5.4. Conclusion 127

Reference list 130 Appendix I: Crystal structure of the conserved Gram-negative

lipopolysaccharide transport protein LptA 144

A1.1. Abstract 145

A1.2. Introduction 146

A1.3. Materials and methods 149

A1.3.1 Sequence analysis, protein expression and purification 149

A1.3.2 Crystallization of LptA 150

A1.3.3 Data collection and structure determination 151

A1.4. Results and Discussion 153

A1.4.1 Sequence analysis 153

A1.4.2 Overcoming severe anisotropy in diffraction data 156

A1.4.3 LptA protein fold and structure analysis 159

A1.4.4 Conclusions 163

A1.4.5 Footnotes 163

A1.5. Appendix References 164

vii

LIST OF TABLES

Table 1.1 Summary of proteins encoded within the heme utilization

operon from pathogenic Escherichia coli O157:H7. 11

Table 2.1 Diffraction data and refinement statistics for ChuS. 34

Table 3.1 Diffraction data and refinement statistics for ChuS-Heme. 66

Table 4.1 Diffraction data and refinement statistics for ChuX. 94

Table A.1 Diffraction data and refinement statistics for LptA. 157

viii

LIST OF FIGURES Figure 1.1 Structure of Heme (protoporphyrin IX) and mechanism

of heme degradation catalyzed by heme oxygenases. 6 Figure 1.2 Schematic of the heme assimilation system from Gram-

negative bacteria at the initiation of our investigation and organization of the heme utilization operon from Escherichia coli O157:H7. 20

Figure 2.1 Ribbon diagram of ChuS. 35 Figure 2.2 The two halves of ChuS superimposed, with an overall rsmd 37

of 2.1 Å. Figure 2.3 Spectral analysis of ChuS reconstituted with heme. 40

Figure 2.4 EPR analysis of ChuS reconstituted with heme. 41

Figure 2.5 ChuS reconstituted with heme. 42 Figure 2.6 Spectral change of ChuS reconstituted with heme,

monitored over time following ascorbic acid addition. 44 Figure 2.7 Compiled average CO detected for ChuS, N-ChuS, and

C-ChuS. 46 Figure 2.8 Sequence alignment of ChuS with homologues from

various Gram-negative bacteria. 49 Figure 2.9 Inhibition of ChuS heme degradation by Sn(IV)

mesoporphyrin IX. 50 Figure 3.1 Ribbon diagram of ChuS in complex with heme. 67 Figure 3.2 Difference omit map for the heme moiety and two water molecules (blue) within the active site pocket of ChuS contoured to 2σ σ σ σ in the R3 space group. 70 Figure 3.3 Spectral change of ChuS reconstituted with heme,

monitored over time following ascorbic acid addition. 73 Figure 3.4 CO measurement via gas chromatography of ChuS, H73A,

H193 and heme following ascorbic acid addition and 25 min incubation (n=3). 74

ix

Figure 4.1 Sequence alignment of Chux with homologues from various Gram-negative bacteria. 93

Figure 4.2 Ribbon diagram of the four ChuX molecules in the crystallographic asymmetric unit. 95 Figure 4.3 Structural similarity of ChuX with ChuS and putative heme

binding cleft. 98 Figure 4.4 Spectral analysis of ChuX reconstituted with heme. 102 Figure 4.5 Spectra of ChuX and mutants reconstituted with heme 103

Figure 4.6 ChuX protection against heme toxicity and heme transfer 106 by ChuX. Figure 5.1 Heme degradation by ChuS in the presence of the hHO-1 specific inhibitor (2[2-(4-chlorophenyl)ethyl]-2- [(1H-imidazol-1-yl)methyl]-1,3-dioxolane). 126 Figure 5.2 Modified skematic of the heme assimilation system from

Gram-negative bacteria and organization of the heme utilization operon from Escherichia coli O157:H7. 129

Figure A.1 Sequence alignment of LptA with various Gram-negative bacterial homologues, many of which are human pathogens. 154 Figure A.2 Q-TOF Mass Spectrometry result, highlighting the LptA has been processed prior to crystallization. 155 Figure A.3 Anisotropic diffraction pattern from LptA MAD (left) and two-molecule native (right) crystals. 158

Figure A.4 Ribbon diagram of LptA with two molecules in the asymmetric unit at 2.6 Å resolution. 160 Figure A.5 Ribbon diagram of LptA highlighting residues conserved in sequence alignment with other Gram-negative homologues. 161 Figure A.6 Ribbon diagram of LptA with eight molecules in the asymmetric unit at 3.25 Å resolution. 162

x

ABBREVIATIONS

The Abbreviations used throughout this manuscript are:

ANL Argonne National Laboratory APS Advanced Photon Source ATP adenosine triphosphate BGG bovine globulin G BV biliverdin BR bilirubin BNL Brookhaven National Laboratory BSA bovine serum albumin CCD charge-coupled device CHESS Cornell High Energy Synchrotron Source CN cyanide CO carbon monoxide CPR cytochrome P450 reductase DLS dynamic light scattering DNA deoxyribonucleic acid EHEC enterohemmorhagic Escherichia coli

EPR eletroparamagnetic resonance FUR ferric uptake regulator GST glutathione S transferase HO heme oxygenase HO-1 heme oxygenase-1 HO-2 heme oxygenase-2 HPLC high-performance liquid chromatography HUS haemolytic uraemic syndrome IPTG isopropyl-1-thio-b-D-galactopyranoside IM inner membrane KDO 3-deoxy-D-manno-oct-2-ulosonic acid LB Luria Burtani Broth lipid A a phosphorylated glucosamine disaccharide acylated with fatty acids LPS lipopolysaccharide MacCHESS Macromolecular diffraction facilities at CHESS MAD multiple anomalous dispersion MS/MS tandem mass spectrometry MW molecular weight NADH reduced β-nicotinamide adenine dinucleotide NADPH reduced β-nicotinamide adenine dinucleotide phosphate Ni-NTA nickel nitrilotriacetate agarose NO nitrogen monoxide NMR nuclear magnetic resonance spectroscopy NSLS National Synchrotron Light Source OM outer membrane ORF open reading frame

xi

PAGE polyacrylamide gel electrophoresis PCR polymerase chain reaction PEG polyethylene glycol PDB protein data bank rmsd root mean squared deviation SAD single anomalous dispersion SCOP structural classification of proteins SDS sodium dodecyl sulphate SSM secondary structure matching STEC shiga-toxin producing Escherichia coli TB terrific broth TCA trichloroacetic acid

xii

LIST OF MUTANT PROTEINS Mutants generated as part of this investigation include: ChuS Mutants:

H73A, H193N, N-ChuS (N-terminal residues M1-P171 fused to a C-terminal His6-tag),

C-ChuS (C-terminal residues V172-A342 fused to an N-terminal GST-tag).

ChuX Mutants:

H65L, H98D, BetaX (ChuX triple mutant V69E/L71H/F73S), and DM (ChuX double

mutant H65L/H98D).

Chapter 1

General Introduction

2

1.1. General introduction and relevance

Escherichia coli is a constituent of the normal flora of the large intestine in

healthy vertebrates, and is widely distributed in soil and water. However, certain

pathogenic strains of E. coli such as O157 and CFT073 are known to cause a host of

clinical diseases in humans and animals including: mild, self-limiting diarrhea; severe

invasive gastro-intestinal infections such as dysentery and hemorrhagic colitis; urinary

tract infections; septicemia; and meningitis (131). While many of the patients infected

with these enterohaemorrhagic strains of E. coli (EHEC) (estimated 73,000 cases

annually in the United States of America (108)) respond well to treatment and recover

within a short period of time (70), the World Health Organization estimates that up to

10% of EHEC cases develop into hemolytic uremic syndrome (HUS), with an associated

fatality rate of 3-5 % (3). As a result, O157 is classified as an opportunistic pathogen and

a global public health concern. Other E. coli strains such as O26:H11 and O111:NM

share a similar pathogenic potential, but it is the O157:H7 serotype that has recently been

responsible for a few large EHEC outbreaks as well as a series of smaller cases

worldwide (56, 99). The sporadic cases of O157 outbreaks are often due to exposure to

undercooked ground beef (1), unpasteurized dairy (71), contaminated water and fruits or

vegetables (4, 13, 150), person to person transfer (23), and is especially harmful to

patients under five years of age (108). In Ontario, the highly publicized Walkerton

tragedy (May, 2000) was due to E. coli O157:H7 and Campylobacter jejuni contaminated

drinking water, which resulted in an estimated 2,300 people being infected; seven cases

of which were fatal (62).

3

Despite that E. coli is the model organism for scientific investigation there still

remain a large number of proteins and open reading frames (ORFs) from various strains

with either unknown function or with debatable annotation based on weak sequence

homology. Even in K-12, which is the most common and non-pathogenic strain of E. coli,

an estimated 30% of identified genes have unknown function and many other genes are

poorly characterized (68). A comparison of the O157 and K-12 E. coli genomes using

Sakai oligoDNA microarray and whole genome PCR scanning revealed that the two

chromosomes share 4.1 Mb of genetic information, with 1.4 Mb of O157 sequence-

specific genetic information (56, 99). Much of this O157-specific DNA is thought to have

arisen for horizontally transferred foreign DNA, and contains many bacteriophage

elements that mediate genetic reorganization and exchange, contributing to its

pathogenesis through frequent genetic change (56). Sequence analysis suggested the

O157 specific DNA encodes for 1632 unique ORFs, 20 unique tRNAs, and that at least

131 of the encoded proteins are recognized virulence factors (56). These unique elements

are therefore highly important, as understanding their function and interplay with other

protein partners will contribute to our overall understanding of bacterial pathogenesis as

well as principles and mechanism of infection. One level of protein characterization

involves structural determination of O157 specific proteins via cryo-electron microscopy,

NMR, and X-ray crystallography. As structural information can provide a plethora of

detail into the precise 3-D folding, interaction of proteins with substrates and other

molecules, mechanisms of action, and interaction with other proteins; as a whole,

structural characterization has the potential to improve our ability to prevent disease and

improve treatments in infected patients (74). The focus of this body of work is the

4

structural determination and functional annotation of ChuS and ChuX, two members of

the heme utilization operon in pathogenic E. coli O157:H7.

1.2. Bacterial iron acquisition from heme sources

Iron is the fourth most abundant element in the Earth's crust (6). However, due to

the low solubility of iron(III) salts at physiological pH in the presence of oxygen, iron

uptake, storage, and transport is a serious obstacle for bacteria to overcome for survival

and pathogenesis. A variety of bacterial iron-acquisition and transport systems have

evolved to assist survival and pathogenesis, such as the secretion of low-molecular-

weight (MW) siderophores (FyuA and IreA) capable of binding tightly to free iron and

mediating iron internalization via iron-specific membrane receptors (33, 100).

Reinforcing the importance for bacterial iron acquisition is that the iron assimilation

strategies they employ are highly redundant, for example, at least seven different

mechanisms have been identified in E. coli (130). As part of a coordinated immune

response against invading microorganisms, mammals specifically limit iron availability

by producing the iron-binding proteins transferrin and lactoferrin, which depress free

extracellular iron to a concentration of ~1x10-18 M in serum, insufficient to sustain

bacterial growth (6, 20). Furthermore, invading microorganisms are confronted with the

obstacle that 99.9% of iron within the human body is sequestered as protein bound heme

in hemoglobin, myoglobin, and various other heme enzymes (130).

5

Through the sequential participation of eight different enzymes, partly in the

mitochondria and partly in the cytoplasm, nucleated human cells and many bacteria can

synthesize heme from glycine and succinyl CoA (154). Heme is a propionic substituted

tetrapyrrole porphyrin group complexed to a central iron atom (Figure 1.1A). As a result

of the delocalized π-electron system of the porphyrin ring, the redox properties of the

central iron atom, and the variety of unique interactions possible for heme coordination

(104), heme is a widely used prosthetic group that enables proteins from various

organisms to fulfill many vital biological roles in photosynthesis, cell differentiation and

proliferation, as well as oxygen binding, transport, and utilization (89, 154). Beyond

heme simply acting as a prosthetic group, in HeLa cells, succinyl acetone-induced heme

deficiency was shown to increase the protein levels of the tumor suppressor gene product

p53 and CDK inhibitor p21. The effect was an overall decrease in the protein levels of

Cdk4, Cdc2, and cyclin D2, and diminish the activation/phosphorylation of Raf, MEK1/2,

and ERK1/2-components of the MAP kinase signaling pathway (168). This result

suggests a role for heme as an effector molecule beyond the regulation of proteins

associated with heme and iron metabolism. Furthermore, the regulatory potential effected

by hemoproteins contribute to all levels of gene expression including; DNA splicing,

transcription, mRNA stability, translation, and post-translational modifications (154). The

cumulative effect of heme uptake is therefore a cascade of gene up-regulation, with heme

effectively acting as a signaling molecule for invading microorganisms that they have

colonized a host (83). As a result, understanding the proteins and mechanisms involved

heme metabolism, transfer, and utilization has the potential to contribute greatly to our

understanding of many integral cellular processes as well as pathogen-host interactions.

6

αααα

ββββ

γγγγ

δδδδ

B

A

αααα

ββββ

γγγγ

δδδδ

B

A

Figure 1.1 Structure of Heme (protoporphyrin IX) and mechanism of heme degradation catalyzed by heme oxygenases. A) Carbon (yellow), oxygen (red), nitrogen (blue) form a conjugated planar molecule that coordinates a central iron (orange) atom. Meso-carbon positions are denoted with Greek symbols. The heme group is rotated to show that the molecule porphyrin is planar, and that the propionic side chains and vinyl groups extend out of the plane of the porphyrin ring. B) Heme oxygenase catalyzed degradation of heme to biliverdin requires a total of seven electrons and two equivalents of molecular oxygen. Most heme oxygenases are regioselective for the α-meso carbon position of the heme group.

7

Some microorganisms such as Haemophilus and Bacterioides are unable to

synthesize heme and hence must acquire heme in large amounts from host organisms.

Other bacteria, such as Borrelia and streptococci, are thought to neither synthesize nor

require heme for growth (130). Many other pathogens have been identified which use

host heme sources and have had heme utilization genes identified including: Bordetella

pertussis (18), Bradyrhizobium japonicum (106), C. jejuni (111), Haemophilus influenzae

(29), Neisseria meningitides (129), Porphyromonas gingivalis (124), Serratia marcescens

(50), Vibrio cholera (58), Yersinia enterocolitica (128), Yersinia pestis (63), E. coli

O157:H7 (91), and Shigella dysenteriae (91). Certain strains have evolved the capacity to

actively seek out and utilize host heme sources though the production of host-specific

hemoprotein proteases such as EspC (38), Pic (59), or Hbp (147), which separate heme

from host proteins such as hemoglobin and hepatoglobin. Heme may then be taken up by

bacterially produced hemophores such as HxuA (28) and HasA (50, 85), or transferred

directly to outer membrane receptors that have been identified in at least 18

microorganisms, many of which are human pathogens (48). Another class of receptors

interact with hemoproteins directly, mediating heme delivery without the need for

hemophores (156). However, sequence comparison of human heme oxygenase-1 (HO-1)

and bacterial HOs have failed to reveal homologues in some heme-utilizing bacteria such

as Escherichia, Plesiomonas, Shigella, Vibrio, and Yersinia (130).

Heme has a higher solubility relative to free iron and is therefore a more readily

transportable source of the metal (130). Once internalized, heme can readily be degraded

by HOs, releasing iron (Fe2+) along with the products biliverdin (BV) and carbon

monoxide (CO). In mammals, the downstream products of heme catabolism have

8

recently been found to mediate the antioxidant, anti-apoptotic, anti-proliferative,

vasodilatory, and anti-inflammatory properties of HO-1 in mammalian cells (45). For

bacteria, the liberated iron atom can then be incorporated into various proteins where it

acts as a biocatalyst or electron carrier, enabling organisms to fulfill many vital biological

processes such as N2 fixation, methanogenesis, H2 production and consumption,

respiration, the citric acid cycle, oxygen transport, gene regulation, and DNA

biosynthesis (6, 89). Alternatively, heme cleaved at the α-meso carbon position by HOs

produces BV IXa, which can subsequently be converted to phytobilins, precursors of the

chromophores for the light harvesting phycobiliproteins and photoreceptor phytochromes

(44). The structure and function of HOs will be addressed in Section 1.4.

While heme is one of the most wide spread enzyme cofactors in nature (130), the

properties which contribute to its usefulness in a controlled protein associated form also

make heme potentially cytotoxic in the free form. This is because heme is both a

hydrophobic molecule capable of freely diffusing into membranes where it can alter the

bilayer structure, thereby disrupting cell integrity (119), and can act as a potent

prooxidative agent, promoting the light-dependent formation of reactive oxygen species

that can cause non-enzymatic redox reactions (54). Heme synthesis and metabolism with

its utilization must therefore be tightly regulated; for this reason heme is normally found

tightly associated in various hemoproteins (54). Furthermore, invading microorganisms

must also tightly regulate heme acquisition and metabolism by either degrading heme or

storing it for later use in various processes (155). The regulation of heme uptake and

utilization must also be coordinated with iron requirements in order to avoid iron induced

toxicity. Using a human-based example, the disease associated with chronic iron overload

9

known as hemochromatosis causes rapidly progressing, multisystem disorders (32). Iron

metabolism is therefore tightly regulated in host organisms through the use of iron

scavenging elements such as transferrin and lactoferrin (6, 20), as well as through effects

mediated by transcriptional regulatory proteins such as Fur (48), DtxR (117), or FecI (40),

which bind to iron related elements on DNA known as Fur binding sites.

Enterohemorrhagic pathogens such as E. coli O157:H7, for which isolates from

human patients have been shown to have stimulated growth in the presence of heme and

hemoglobin; may induce hemolytic lesions in order to gain access to serum sources of

nutrients such as heme and iron (83). It may therefore be in part as a result of attempting

to acquire heme which elicits the clinical effects of pathogenic E. coli in HUS. Using a

laboratory strain of E. coli (1017 (ent::Tn5)) that was unable to import heme into the

cytoplasm, it was established that the outer membrane receptor ChuA (E. coli heme-

utilization gene A) is required for heme uptake along with TonB, an energy-transducing

protein associated with ExbB and ExbD that uses the proton motive force of the

cytoplasmic membrane for the passage of ligands into the periplasm (146). Four other

members of the heme uptake operon from Y. enterocolitica were shown to be required to

complete heme uptake, transport, and use as an iron source by E. coli K-12 (127)

including: the periplasmic heme-binding protein HemT, the heme permease protein

HemU, the ATP-binding hydrophilic protein HemV, and the protein HemS, whose role

may be to regulate the amount of non-protein associated heme in the cell (127, 128).

Restriction mapping and sequence analysis of selected regions of E. coli O157:H7

genome have revealed that the organization of the heme transport locus of this strain is

strikingly similar to that of other enteric bacteria such as Y. enterocolitica and

10

S. dysenteriae (127, 128). This genetic organization is especially true with respect to the

region encoding heme outer membrane receptors such as ChuA, ShuA, and HemR,

followed by a short gap, and then the gene encoding ChuS or one of its homologues. In S.

dysenteriae the intergenic region between ShuA and ShuS are separated by a 48-

nucleotide region with sequence homology to the consensus Fur box, but no obvious –10

and –35 promoter elements (128). However, in response to iron limiting conditions, ShuS

was upregulated, permitting heme as an iron source and preventing heme toxicity (165).

A summary of the proteins encoded within the heme utilization operon from E. coli

O157:H7 is summarized in Table 1.1.

11

Table 1.1 Summary of proteins encoded within the heme utilization operon from pathogenic Escherichai coli O157:H7.

Protein Localization Length (amino acids) Function

ChuA Outer

Membrane 660 Heme Receptor

ChuT Periplasm 304 Heme Transport

ChuU Inner

Membrane 330 Permease ABC Transport Protein

ChuV Inner

Membrane 256 Permease ABC Transport Protein

ChuS Cytoplasm 342 Heme/hemoglobin transport

protein ChuW Cytoplasm 445 Coproporphyrinogen III oxidase ChuX Cytoplasm 164 Hypothetical ChuY Cytoplasm 207 Hypothetical

12

1.3. Role of heme and structural diversity of heme-proteins

While traditional heme studies have focused on heme as a prosthetic group within

hemoglobin, myoglobin, and iron processing elements, novel heme-Fcontaining proteins

have recently been discovered. These include the cytochrome c chaperone CcmE (123),

the iron-dependent regulator of heme biosynthesis Irr (107), and the heme-based

aerotactic transducer Hem-AT (64). Other hemoproteins function through conformational

changes induced as a result of modifications to the ferrous heme environment upon the

coordination of small effector ligands such as NO, O2, or CO, as in the case of guanylate

cyclase (172), FixL (52), and CooA (81), respectively. Other newly identified heme-

associated signaling proteins include AxPDEA1, NPAS2, and EcDos whose interaction

between heme and various gases in the sensor domains mediate structural changes in the

effector domains of these proteins (115). While the cellular consequences mediated by

these signaling proteins are diverse, the common feature between them is that small

molecules can trigger major changes to the overall heme-protein environment, which can

distort the heme group itself or the interactions stabilizing the heme group, resulting in a

robust structural and functional transition.

While heme helps to mediate diverse responses, it is the interaction of heme with

various coordinating proteins that ultimately confer molecular function. Therefore,

characterizing the interactions that contribute to heme stabilization will also help to

characterize how heme mediates cell signaling. Furthermore, while heme helps to

mediate diverse responses, it is the interaction of heme with various coordinating proteins

that ultimately confer molecular function. Structural insights provided from vibrational

13

spectroscopy (EPR and Raman), nuclear magnetic resonance spectroscopy (NMR), and

X-ray crystallography have indicated that heme proteins often coordinate the iron center

of the heme group between a histidine side chain and a variety of other possible residues

including tyrosine (8), methionine (22), cysteine (39), proline (81), histidine (103),

arginine (137); while other proteins have been shown to coordinate the heme iron center

without additional assistance from other side chains (52, 121). Other proximal side chains

and regions of the polypeptide backbones have been shown to interact with other parts of

the heme group, contributing to the stabilization of the heme moiety within the protein

environment, and participating in the interactions which ultimately confers hemoprotein

function. These heme-protein interactions are both hydrophobic and hydrophilic,

stabilizing the porphyrin ring or propionic elements of the heme group, respectively, and

are mediated through various interactions.

Given the diversity of functions attributed to heme proteins together with the

various heme polypeptide interactions that have evolved, it is not surprising that heme

proteins feature a broad range of structural motifs and overall architectures. The structure

of sperm whale myoglobin, the first high resolution macro molecular structure solved by

X-ray crystallography, consists of eight α-helices that wrap around the heme group,

seating it within a hydrophobic pocket (72). The overall architecture of myoglobin is

similar to that of HO-1 whose overall structure is also mainly α-helical, and sandwiches

the heme group between two α-helices that are termed the proximal and distal helices

(121). In the smaller hemoprotein cytochrome c, the heme group is associated within a

simple, mainly α-helical fold (86) but resides closer to the surface than either myoglobin

or HO-1. Heme proteins may also assume a variety of other folds. In the heme

14

sequestering protein hemopexin found in serum, heme is coordinated between two β-

propeller domains (103), and in HasA the heme binding site is formed by loop regions

which extend from an αβ-fold (8). The structure of FixL resembles an open pita or clam

which coordinates heme in the center of the open mouth between a series of antiparallel

β-sheets at the protein core and an α-helix (52). A unique example of heme coordination

was recently reported for the iron regulated lipoprotein from C. jejuni, ChaN, whose

function is poorly understood, but seems to associate with the outer membrane receptor

ChuR (25). ChaN forms a homodimer which sandwiches a pair of heme molecules

between a pair of conserved tyrosine residues donated from opposing monomers, that

originate at the base of two α-helices. ChaN does not have a homologue in O157 or K-12,

but homologues have been identified in other E. coli strains.

15

1.4. Bacterial heme oxygenases

Although many reports have focused on the proteins involved in heme transport

across the outer membrane and periplasm, the fate of heme once it has reached the

bacterial cytoplasm has only recently received attention. Investigations of eukaryotes

revealed that release of iron stores requires the heme porphyrin ring to be degraded by

monooxygenases known HOs. HOs are enzymatically unique, as they use heme as both a

substrate and as a cofactor in the binding of oxygen for intramolecular degradation of the

porphyrin macromolecule to BV, CO, and Fe2+ (Figure 1.1B) (174). Structural insight

contributed by EPR and X-ray crystallography studies suggests that heme cleavage

catalyzed by HOs proceeds via similar mechanisms (73, 92). Initially, HO bound heme

accepts an electron from either NADPH-cytochrome P450 reductase (CPR), yielding a

ferrous (Fe2+) heme-HO complex. Molecular oxygen binds with the complex between the

iron atom and the axial histidine residue, forming a metastable oxy-form, that receives

another electron from CPR and a proton from the distal water pocket, and is rapidly

converted to a hydroperoxide intermediate (Fe3+–OOH). The oxygen of the

hydroperoxide intermediate attacks the α-meso carbon (or other regioselective position),

yielding a ferric α-meso-hydroxyheme. This intermediate then receives another oxygen

molecule and an electron and is converted to ferrous-verdoheme. Another oxygen and

reducing equivalent are received, transforming ferrous-verdoheme to ferric-BV and is the

least understood part of the mechanism (44). In the last step ferric-BV is further reduced

to the ferrous state, yielding BV and Fe2+. Overall the reaction requires an input of seven

electrons and two molecules of oxygen (Figure 1.1) (73, 92).

16

Several bacterial and mammalian HOs have been recently identified, including

HemO (173), HmuO (116), HmuQ/D (106), cyanobacterial HO-1 and HO-2 (30), IsdG/I

(125), mammalian HO-1 and HO-2 (121), nm-HO (122), PCC_6803 (132), and

PigA/BphO (46); and crystal structures have been determined for HemO (122), HmuO

(61), nm-HO (122), pa-HO (46), PigA (46), and Syn HO-2 (132), as well as a partial

structural characterization of BphO by NMR (157). Furthermore, extensive structural

work has been conducted on various HOs in complex with effector molecules such as

azide (135), biliverdin (134), CO and cyanide (136), ferric and ferrous forms of iron (79),

imidazole, and nitrogen monoxide (79) in order to effectively capture “snapshots” along

the reaction coordinate of heme degradation.

The overall topology of known bacterial HOs share strong structural similarity to

mammalian HOs in that they are mainly α-helical and sandwich heme between two

helices at the base of the structures, with the propionic side chains oriented away from the

protein (44). In this mode of heme coordination, HOs are regioselective for O2 addition

across the α-meso position of heme, which is oriented toward a series of polar residues

that do not participate directly in the reaction, but likely contribute to regioselectivity

(121). The exception to this mode of coordination is pa-HO from Pseudomonas

aeruginosa which is regioselective for the β- and δ-meso positions of heme in a 30:70

ratio, which has been attributed to heme being rotated by 100º in the active site (46). The

purpose of the generation of these two BV isomers is currently unknown, but is likely

related to the fact that P. aeruginosa has a second HO, BphO, with α-meso carbon

specificity (44). Despite the structural conservation across bacterial HOs identified to

date, it should be noted that many of these were targeted because they shared sequence

17

homology with mammalian HOs. In pa-HO the propionic side chains of the heme moiety

are stabilized in a unique fashion compared to other HOs identified. This highlights how

a variety of protein-heme interactions can mediate heme catalysis beyond those well

characterized in the hHO-1 related system. In many bacteria where sequence analysis

fails to suggest the enzyme responsible for heme processing, other approaches must be

utilized such as systematic gene knockouts and phenotypic characterization to identify

the genes responsible for heme utilization (165). At the initiation of our investigation

presented in this thesis, bioinformatics analysis had failed to identify a HO in any E. coli

strain, although both E. coli O157:H7 and CFTO73 were known to use heme as an iron

source, suggesting that a heme degrading enzyme was necessary.

1.5. Heme utilization protein ChuS

Because of its poorly understood role in modulating heme utilization and its

genetic organization downstream of the outer membrane receptor chuA, we directed our

attention to the gene product of chuS, the E. coli O157:H7 and CFT073 homologue of

shuS, as a candidate for intracellular heme breakdown (128). Sharing 98% identity to the

S. dysenteriae homologue ShuS, the structure and function of these two homologues is

likely highly similar. In the initial report characterizing ShuS in isolation of other

members of the heme utilization operon, ShuS was shown to form a high molecular

weight oligomer (650 kDa) from 37 kDa subunits, to nonspecifically bind double

stranded DNA, and was not observed to exhibit any HO activity in standard NADPH or

ascorbate based assays (158). In contrast, in an S. dysenteriae knockout, the shuS mutant

18

was defective in utilizing heme as an iron source and ShuS was required under

microaerobic and anaerobic conditions for the optimal utilization of heme (165). ShuS

expression also protected cells from heme toxicity at high concentrations, and was

therefore suggested to bind to and potentially transfer heme from the transport proteins in

the membrane to either heme containing or heme degrading proteins, or involve itself as a

heme degrading enzyme (165). Furthermore, expression of the Y. enterocolitica

homologue to ChuS, HemS, protected E. coli cells against heme toxicity (127).

1.6. ChuW, ChuX, ChuY and bacterial heme proteins

In addition to chuS, three other genes: chuW, chuX, and chuY are located within

the heme utilization locus of O157 were poorly characterized at the initiation of our study.

ChuW does not share any homology to any protein identified in heme transport, but does

have weak homology to HemN, the oxygen-independent form of the heme biosynthetic

enzyme co-proporphyrinogen oxygenase III present in various bacteria including Bacillus

ssp. Salmonella typhimurium and E. coli K-12 (164, 167). This homology suggests that

ChuW may play a role in some aspect of heme transport or metabolism. In S. dysenteriae

shuW begins 19 nucleotides downstream of shuT, and the lack of any apparent

transcription termination or promoter sequences suggests that shuT and shuW are located

on the same transcript (127, 128, 164, 165). In S. dysenteriae type 1, a premature stop

codon exists within shuW suggesting that ShuW may be defective, although in E. coli

O157:H7 this stop codon is absent in chuW suggesting that ChuW may be completely

19

functional (164). Furthermore, DNA microarray analysis of CFTO73 strains suggested

that chuW was upregulated by more than 5-fold during urinary tract infection (126).

Twelve nucleotides downstream of shuW is the start codon for shuX, and the

initiation condon of shuY overlaps the stop codon of shuX. This indicate that these two

smaller genes may also be co-transcribed with shuT (164). In a similar sequence analysis

of the heme uptake locus from Y. pestis, homologous to chuX and chuY (orfX and orfY,

respectively) were determined to be located immediately downstream of the heme

receptor gene hmuR and were therefore likely co-expressed (145). In Vibrio anguillarum,

which can use heme and hemoglobin as a source of iron, using a lacZ reporter system it

was shown that a homologue of chuX, huvX, was co-transcribed with huvZ under iron

limiting conditions (95). An E. coli homologue to huvZ was not identified via NCBI

Blastp sequence analysis at (www.ncbi.nlm.nih.gov/BLAST/). Furthermore, an

examination of the -10 and -35 elements downstream of huvX revealed sequence with

similarity to sigma70-promoters, as well as the presence of a Fur element (95).

Characterization of either chuX or chuY gene products or their homologues from various

heme utilizing organisms had not received any attention at the beginning of our

investigation. However, based on Blastp sequence analysis ChuX was suggested to be a

protein associated with heme-iron utilization, and that ChuY is a hypothetical protein or

nucleoside-diphosphate-sugar epimerase (143). A schematic-diagram of the heme

assimilation system of Gram-negative bacteria, as well as the structures and functions of

members of the heme utilization operon at the initiation of our investigation is

summarized in Figure 1.2.

20

OM

PP

IM

HasA

Hbp

ChuA?

TonB

ExbB? ExbD?ExbB?ExbB? ExbD?

ChuT?

ChuU? ChuV?

ChuW?ChuW?

Function

V U Y X W T A S

? ? ? ?

Figure 1.2 Schematic-diagram of the heme assimilation system from Gram-negative bacteria at the initiation of our investigation and organization of the heme utilization operon from Escherichia coli O157:H7. Proteins involved in heme transport from pathogenic E. coli O157:H7 are presented in their locations in the outer membrane (OM), periplasm (PP), and inner membrane (IM). Where the structures of proteins are unknown, structures of homologues have been included and names followed by a question mark. Putative promoter elements within the heme utilization operon are depicted with arrows. The function of ChuS, ChuW, ChuX, and ChuY were unknown at that time. The proteins focused on as part of this body of work are ChuX and ChuS.

21

In light of the general importance of heme as a signaling molecule and cofactor,

this thesis examines the structure and function of two members of the E. coli O157:H7

heme utilization operon: the utilization protein ChuS in its apo and heme bound forms,

and the heme binding protein ChuX. The method of X-ray crystallography, specifically,

the anomalous dispersion method using seleno-methionine labeled proteins has been used

for structural determination. A combination of spectral characterization, site-directed

mutagenesis of conserved residues, gas chromatography, and sensitivity of cultures

challenged with heme provide biochemical evidence for the function of ChuS and ChuX.

The application of structure based functional characterization as well as the implications

of our findings for heme uptake and utilization in E. coli O157:H7 and Gram-negative

bacteria in general is addressed.

22

Chapter 2

Identification of a novel Escherichia coli O157:H7 heme oxygenase with tandem functional repeats

Preface:

This chapter was published in The Proceedings of the National Academy of Sciences:

Michael D. L. Suits, Gour P. Pal, Kanji Nakatsu, Allan Matte, Miroslaw Cygler, and

Zongchao Jia (2005) Identification of a novel Escherichia coli O157:H7 heme oxygenase

with tandem functional repeats. PNAS, 102(47):16955-60.

Michael Suits was responsible for protein expression, purification, crystallization, data

collection, structure solution and refinement, and all biochemical experiments in the

investigation of function. Dr. B. McLaughlin helped with gas chromatography and

interpretation, Dr. B. Hill conducted EPR experimentation and interpretation, and expert

help from Dr. Mike Nesheim in heme reconstitution setup and interpretation of results to

calculate the Kd for ChuS-heme. The chuS gene was cloned by Pietro Iannuzzi and the N-

and C-terminal halves of ChuS by Jim Blonde. The manuscript was written by Michael

Suits with editorial input from Dr. Zongchao Jia and with additional help from Dr. John S.

Elce.

Data deposition footnote: Coordinates and structure factors for ChuS are deposited in the RSCB Protein Data Bank under accession code 1U9T.

23

2.1. Abstract

Heme oxygenases (HOs) catalyze the oxidation of heme to biliverdin, carbon

monoxide (CO), and free iron. Iron acquisition is critical for invading microorganisms to

enable survival and growth. Here we report the crystal structure of ChuS, which displays

a novel fold and is unique compared to other HOs characterized to date. Despite only

19% sequence identity between the N- and C-terminal halves, these segments of ChuS

represent a structural duplication, with a root mean square deviation of 2.1 Å between the

two repeats. ChuS is capable of using either ascorbic acid or cytochrome P450 reductase-

NADPH as electron sources for heme oxygenation. CO detection confirmed that ChuS is

a HO, the first to be identified in any strain of Escherichia coli. Based on sequence

analysis, this novel HO is present in many bacteria, though not in the E. coli K-12 strain.

The N- and C-terminal halves of ChuS are each a functional HO.

24

2.2. Introduction

The incorporation of iron into proteins as a biocatalyst or electron carrier enables

organisms to fulfill many vital biological processes including photosynthesis, N2 fixation,

methanogenesis, H2 production and consumption, respiration, the tricarboxylic acid cycle,

oxygen transport, gene regulation, and DNA biosynthesis (89). The importance of iron

for bacterial survival and pathogenesis is evident by the variety of iron-acquisition and

transport systems that have evolved, such as the secretion of low-molecular-weight (MW)

siderophores capable of binding free iron and transporting it into the cell via specific

membrane receptors (33). As part of a coordinated immune response against invading

microorganisms, mammals specifically limit iron availability by producing the iron-

binding proteins transferrin and lactoferrin, which depress free extracellular iron to levels

insufficient to sustain bacterial growth (3). Alternatively, invading pathogens may

acquire iron directly from host iron sources such as heme, hemoglobin, or hemopexin

either through the production of specific outer membrane receptors that bind directly to

host heme-sequestering proteins or through the secretion of heme-binding proteins

(hemophores) that then deliver the heme-protein complex to bacterial cell surface

receptors (48). Enterohemorrhagic pathogens such as Escherichia coli O157:H7, for

which isolates from human patients have been shown to have stimulated growth in the

presence of heme and hemoglobin, may induce hemolytic lesions in order to gain access

to these serum sources of iron (83).

Using a laboratory strain of E. coli (1017 (ent::Tn5)) that was unable to import

heme into the cytoplasm, it was established that the outer membrane receptor ChuA

25

(E. coli heme-utilization gene A) is required for heme uptake along with TonB, an

energy-transducing protein associated with ExbB and ExbD that uses the proton motive

force of the cytoplasmic membrane for the passage of ligands into the periplasm (146).

Four other members of the heme uptake operon from Yersinia enterocolitica are required

to complete heme uptake, transport, and use as an iron source by E. coli K-12 (127)

including: the periplasmic heme-binding protein HemT, the heme permease protein

HemU, the ATP-binding hydrophilic protein HemV, and the protein HemS, whose role

may be to regulate the amount of non-protein-associated heme in the cell (128).

Restriction mapping and sequence analysis of selected regions of E. coli O157:H7

DNA, the serotype responsible for outbreaks of hemorrhagic colitis and hemolytic uremic

syndrome, has revealed that the organization of the heme transport locus of this strain is

strickingly similar to that of other enteric bacteria such as Y. enterocolitica (164). This

organization is especially true with respect to the region encoding heme outer membrane

receptors such as ChuA, which are followed by a short gap and then by ChuS or one of

its homologues. In Shigella dysenteriae the intergenic region between ShuA and ShuS are

separated by a 48-nucleotide region with sequence homology to the consensus Fur box

but no obvious –10 and –35 promoter elements (127). Furthermore, a promoter was likely

to be present within the intergenic region because a mini=Tn10 insertion in ShuA was not

polar with respect to ShuS expression in minicells (91).

Although many reports have focused on the proteins involved in heme transport

across the outer membrane and periplasm, the fate of heme once it has reached the

bacterial cytoplasm has only recently received attention. Investigations in eukaryotes

revealed that release of iron stores requires the heme porphyrin ring to be degraded by

26

monooxygenases known as heme oxygenases (HOs). HOs are enzymatically unique, as

they use heme as both a substrate and a cofactor in the binding of oxygen for

intramolecular degradation of the porphyrin macromolecule to biliverdin, CO, and free

iron (121). Several bacterial HOs have been identified, including HemO (173), HmuO

(26), IsdG/I (125), cyanobacterial HO-1 and HO-2 (132), and PigA/BphO (110). All of

these were reviewed (44), and the structures of HemO (122), HmuO (61), and PigA (46)

have been determined, as well as a partial structural characterization for BphO (157). The

crystal structures of HemO, HmuO, and PigA all share a strong structural similarity to

mammalian HOs: they are mainly α-helical and sandwich heme between two helices. At

present, no sequence comparison or structural analysis has identified a HO in an E. coli

strain.

Here we report the crystal structure of ChuS, which has a novel fold with two

tandem repeats and does not show structural similarity to known mammalian or bacterial

HOs. ChuS is capable of binding heme and can break down heme using ascorbic acid or

cytochrome P450 reductase-NADPH (CPR) as electron sources for oxygenation. CO

detection by gas chromatography confirmed that ChuS is a HO, the first to be identified

in E. coli. In addition, HO activity occurs independently in the N- and C-terminal halves

of ChuS.

27

2.3. Materials and methods

2.3.1 Sequence analysis, protein expression, and purification

Sequence analysis was carried out using Blast (5). Full-length ChuS and the N-

and C-terminal halves of ChuS were subcloned into pET expression vectors, yielding

fusion proteins to either a N- or C-terminal His tag or an N-terminal GST tag. Based on

the results of small-scale expression trials, we selected the following constructs: ChuS

fused to an N-terminal His6 tag (ChuS), the N-terminal half fused to a C-terminal His6 tag

(N-ChuS), and the C-terminal half fused to an N-terminal GST tag (C-ChuS). In each

case, 1-L cultures of BL21(DE3) cells carrying plasmid for the respective recombinant

protein were grown at 37°C in Terrific Broth (Bioshop, Burlington, Canada)

supplemented with 100 µg/ml ampicillin. Protein expression was induced using 0.4 mM

IPTG for 5 h. ChuS protein substituted with seleno-methionine was expressed in the

metA- E. coli strain DL41 in LeMaster medium (60).

All proteins were purified using standard methods. Briefly, His6-tagged proteins

were purified via Ni-NTA batch purification in phosphate buffer (pH 8.0), followed by

purification by Resource Q and Hi-Trap metal-chelating columns on an ATKA Explorer

FPLC. Collected fractions were monitored by SDS-PAGE, for which those that contained

pure protein were pooled together and dialyzed against 50 mM Tris-HCl buffer (pH 8.0),

150 mM NaCl for subsequent crystallization, or 100 mM sodium phosphate buffer (pH

7.0) for HO spectral and activity measurements. C-ChuS was purified on a GSTrap FF

column.

28

2.3.2 Crystallization of full-length ChuS

Crystals were obtained by the hanging-drop vapor diffusion method by mixing

2.5 µl of 20 mg/ml ChuS with 2.5 µl of reservoir solution consisting of 20% (w/v) PEG

3350 and 100 mM Bis-Tris (pH 6.5). Crystals appeared after 1-3 d. ChuS crystals were

further improved by microseeding using 12–20% (w/v) PEG 3350. SeMet derivative

crystals were obtained in the same way. For cryoprotection, crystals were soaked in the

reservoir solution containing ethylene glycol (concentration increased stepwise to 30%

[v/v]), picked up using a nylon loop, and flash-cooled in the N2 cold stream at 100K. Two

nonisomorphous crystal forms belonging to space group P21 were obtained. Crystals

having the larger unit cell contain three molecules in the asymmetric unit, whereas those

with the smaller unit cell contain only one molecule.

2.2.3 Data collection and structure determination

Native and multiple-wavelength anomalous dispersion (MAD) data sets were

collected, respectively, for the two different crystal forms at beamlines 17-ID and 19-BM

of the Advanced Photon Source, Argonne National Laboratory. The native data were

collected for 180° in total, with 1.0° oscillation. SeMet MAD data were collected at Se

peak and inflection wavelengths for 360° and remote wavelength for 180°, all with 1.0°

oscillation. Data were processed with HKL2000 (102). Initial phases for the three-

molecule form was determined using the MAD data with the program SOLVE (144).

Density modification, including three-fold NCS averaging, was carried out using

29

RESOLVE (144). Automatic chain tracing was performed using RESOLVE (144), and

the remainder of the model was built by manual fitting using XtalView/Xfit (90). When

the model of one of the three molecules was about 80% built, it was used as a search

model in molecular replacement using data obtained from the single-molecule form. This

resulted in an unambiguous solution that was subsequently refined using CNS (19).

2.3.4 Reconstitution with heme and absorbance measurements

Heme complexes of ChuS, N-ChuS, and C-ChuS were prepared by dissolving

heme into 0.5% (v/v) ethanolamine and adding small volumes of this mixture to protein

until a final ratio of 4:1 (mol:mol) was obtained for ChuS and 2:1 for N-ChuS and

C-ChuS. ChuS-heme and N-ChuS-heme were then applied to Resource Q or Superdex

200 Prep Grade columns, respectively, to remove noncoordinated heme. Reconstituted

protein-heme complex concentrations were determined by adapting the reported

extinction coefficient for ShuS (ε410) of 79.5 and 159 mM-1cm-1 for ChuS and N-ChuS,

respectively (11). Pure fractions of C-ChuS were concentrated and used for

spectrophotometric analysis and CO detection in which twice the amount of heme was

added to protein prior to measurements. Absorbance data were then collected using a

microplate spectrophotometer (Bio-Tek Instruments, Inc.). As a control in separate

experiments, 10 µM catalase and 10 µM superoxide dismutase were added to the reaction

wells, followed by ascorbic acid or cytochrome P450 reductase (CPR)-NADPH, and the

difference spectra were recorded. A quadruplicate series of ChuS-heme reconstitution

titrations were also conducted to examine the ChuS-heme interaction.

30

2.3.5 CO activity assay and inhibition

Briefly, in 2-ml vials 250 µl reaction volumes of 20 µM heme and 10 µM ChuS,

N-ChuS, or C-ChuS in 100 mM phosphate buffer (pH 7.0) were incubated with constant

shaking at room temperature for 6 min at which time enzymatic heme degradation was

initiated by adding either CPR and NADPH or ascorbic acid at final concentrations of

17.7 nM and 135 µM, or 50 µM, respectively. Vials were then sealed with screw caps

and blue silicone rubber septa and the headspace above each was immediately purged

with CO-free air, and incubated for 45 min at 24°C with constant shaking. The reaction

was stopped by placing the vials on powdered dry ice (-78°C), where they remained for

approximately 30 min. The amount of liberated CO in the headspace of each vial was

measured using a gas chromatograph equipped with an HgO reduction detector (Trace

Analytical). The amount of CO resulting from ChuS-catalyzed breakdown of heme was

determined by comparing peak area measurements for CO against linear CO standard

curves (n = 20 determinations; average correlation coefficient 0.993). Technical details

are described elsewhere (27, 153). In a parallel experiment, Sn(IV) mesoporphyrin IX

dissolved in 0.5% (v/v) ethanolamine was also added during initial mixing which acted as

a competitive inhibitor.

2.3.6 Dynamic light scattering of C-ChuS and N-ChuS

31

Dynamic light scattering (DLS) was conducted for ChuS, ChuS-heme, and N-

ChuS each at 1 mg/ml concentration in 100 mM phosphate buffer (pH 7.0) using a

DynaPro99 instrument (Protein Solutions). DLS was also conduced on samples of

1 mg/ml ChuS and ChuS-heme in 50 mM Tris and 150 mM NaCl (pH 8.0).

2.3.7 EPR experiments

EPR spectra were obtained on a Bruker EMX spectrometer at X-band fitted with

an HSQ cavity equipped with a liquid helium flow cryostat (Oxford Instruments). The

standard instrument conditions were T = 10K, modulation amplitude = 20 G, power =

2 mW at a gain of 1 × 104.

2.4 Results

2.4.1 Expression and purification

ChuS overexpression resulted in cells exhibiting a blue-green pigment

characteristic of biliverdin accumulation (87, 157, 174). Following Ni-NTA purification

and dialysis in low-ionic-strength Tris buffer, fractions containing ChuS appeared purple

and turned to a straw color over time. EPR of ChuS fractions following Ni-NTA

purification did not reveal the presence of any coordinated metal ion, and spectral data

did not reveal the source of the pigmentation. This pigmentation was not observed in

ChuS grown in LeMaster medium in DL41(DE3) cells or for the N-ChuS or C-ChuS

32

constructs. The purple coloration of ChuS was lost following the subsequent FPLC anion

exchange step. Expression of full-length ChuS and N-ChuS yielded more than 60 mg of

pure protein from 1 L culture. Recombinant C-ChuS was insoluble in pilot expression

trials for N- or C-terminal His-tagged versions, and could only be expressed in soluble

form as a GST-C-ChuS construct. Average yields of 2.6 mg pure GST-C-ChuS/ L culture

were obtained.

2.4.2 Structure determination and analysis

Two different crystal forms were obtained from the same crystallization

conditions, that shared the same overall morphology and space group (P21), but differed

in unit cell dimensions and number of molecules in the asymmetric unit. We collected

MAD data for the three-molecule crystal form and native data for the single-molecule

crystal form, and the crystal structure of ChuS was solved using a combination of MAD

and molecular replacement with the two crystal forms. The higher resolution (2.15 Å)

structure from the single-molecule form is presented here, although the final structures

for ChuS from the two crystal forms are essentially identical, with an average root mean

square deviation (rsmd) of 0.92 Å. The final structure of ChuS contains residues 1–337 of

a total 342 amino acids with no ligand. Due to an absence of electron density, a gap exists

in the structure between residues 12 and 23 (12-QNPGKYARDI-23), and between 165

and 173 (165-KAVDAPV-173). In total, 99.7% of residues are within the allowed region

of the Ramachandran plot. Only one residue, Asp 114, was in the disallowed region and

33

resides within a surface exposed β-turn. Final refinement statistics are summarized in

Table 1.

The overall structure of ChuS comprises a central core of two large pleated

β-sheets, each consisting of nine anti-parallel β-strands sandwiched together and bowing

outward in a saddle motif (Figure 2.1). Each β-sheet is flanked at its N-terminus by one

pair of parallel α-helices and at its C-terminus by a set of three α-helices in an α-loop-α-

loop-α configuration. Two large clefts are found on opposite sides of the central core of

β-pleated sheets, with the third side of the clefts delineated by the flanking sets of three

α-helices. A long stretch of coil connecting the N- and C-terminal halves runs from I156

to V183, in which the gap in density occurs (Figure 2.1). Structural comparison of ChuS

against structures within the SCOP database (97) was performed with the program SSM

(77). This search did not reveal any significant match; hence ChuS represents a novel

fold.

34

Table 2.1 Diffraction data and refinement statistics for ChuS

Native Peak Inflection Remote Space group P21 P21 Cell dimensions a (Å) 41.683 46.072 b (Å) 57.878 194.172 c (Å) 59.575 60.461 β (°) 96.020 100.383 No. of molecules in ASU 1 3 Wavelength (Å) 0.9795 0.9803 0.9804 0.9642 Resolution range (Å) 50–2.15 50–2.60 50–2.60 50–2.40

Total reflections 53,596a 170,485b 176,691b 149,352b

Unique reflections 14,606a 32,694b 32,966b 40,453b

Completeness (%) 96.9 (82.8)# 99.8 (99.6) 99.7 (98.4) 96.3 (76.5)

Rsym(I) (%) 7 9.7 9.1 7.8 I/σI 15.0 (2.4) 17.4 (2.4) 16.6 (2.0) 11.4 (1.0) Refinement (single molecule form) Resolution range (Å) 50 – 2.15 Rwork, (%) 20.1 Rfree, (%) 26.3 No. reflections total/ Rfree 12315/ 744 No. atoms, protein / solvent (H2O)/ total 2428/ 331/ 2759

B factors (Å2) protein/ solvent (H2O)/ total 29.7/ 48.6/ 32.0 RMS, bond length (Å)/ bond angle(o) 0.009/ 1.44

#Values in parentheses are for the outermost shell (2.23–2.15 Å) aNumber of reflections following merging bNumber of reflections without merging of Bijovet pairs

35

Figure 2.1 Ribbon diagram of ChuS. The core of nine anti-parallel β-strands is visible in the front face. The gaps in the structure are shown with a dashed lines.

36

2.4.3 ChuS contains a structural duplication

On close inspection, the two halves of ChuS were found to be structurally similar.

Including side chains the two halves superimpose to an rmsd of 2.1 Å (Figure 2.2).

Sequence alignment of the two halves reveals only 19% identity. Nevertheless, at the

structure based level, many residues are well conserved structurally, including R29 and

R209, F127 and F304, and Y138 and Y315. The two halves of ChuS associate across the

central portion of the β-pleated sheets. At the interface, residues from the C-terminal half

are hydrophobic (V244, V250, F274, L276), point toward the interior of the C-terminal

subunit, and expose regions of the backbone across the interface. In contrast, most

interface residues from the N-terminal half are hydrophilic (N66, Y70, H73, N75, R95)

and point across the interface between the two halves. The side chains of the N-terminal

interacting residues are in contact with the backbone region of the C-terminal half. This

structural similarity suggests that the two halves of ChuS may be functionally

independent or that they have dual functions. Furthermore, the second gap in the structure

is exactly at the midpoint of the ChuS sequence in a flexible/disordered linker connecting

the two halves.

37

Figure 2.2 The two halves of ChuS superimposed, with an overall rsmd of 2.1 Å. The N-terminal portion of ChuS is shown in red, the C-terminal portion in blue.

38

2.4.4 Spectral properties of the heme-ChuS complex

Based on difference spectra comparing free heme to full-length ChuS, N-ChuS,

and C-ChuS, we determined that all three formed complexes when reconstituted with

heme (Figures 2.3A and 2.3B). The spectra of ChuS resemble those of other HOs in that

a Soret maximum is evident at 408 nm, with a smaller set of peaks for the β-band at

545 nm and for the α-band at 580 nm, suggesting that the ChuS-heme complex formed is

ferric hexacoordinate in the high-spin state at neutral pH (26). The spectrum for N-ChuS

has a peak at 402 nm (which is smaller than for ChuS), a smaller peak at 530 nm, and a

shoulder at 625 nm. Although the spectrum for N-ChuS with heme did not show

definitive heme coordination, it remained unchanged following gel filtration, indicating a

specific interaction between the two. C-ChuS with heme showed a broad peak centered at

384 nm, a smaller peak at 540 nm, and two small shoulders at 585 and 625 nm.

Furthermore, EPR spectra analysis suggested the presence of two populations of heme

molecules (Figure 2.4).

Despite structural similarity between the two halves, the spectra for C-ChuS and

N-ChuS differed, perhaps indicating different environments for heme coordination.

Based on absorbance at 408 nm during ChuS-heme reconstitution and assuming two

binding sites, the estimated Kd is 1.0 ± 0.3 µM (n = 3) based on the following expression:

])24}2({2[5.0)( 2/100

20000 HsitesPKHPsitesHKPsites ddfb ⋅−++⋅−++⋅−=∆Α εε ,

39

where P0 is the amount of protein, H0 is the concentration of heme, and (εb – εf) is the

difference in the Soret absorption spectra between the bound and free states (Figure 2.5).

This estimated Kd suggests a slightly weaker association than that characterized for the

mammalian hHO-1 (0.84 ± 0.2 µM) (159, 160) but tighter than that for other bacterial

heme oxygenases: HmuO (2.5 ± 1.0 µM) (162), IsdG (5.0 ± 1.5 µM) and IsdI (3.5 ±

1.4 µM) (125).

40

Figure 2.3 Spectral analysis of ChuS reconstituted with heme. Full-length ChuS and the N- and C-terminal halves of ChuS all form complexes when reconstituted with heme, with peaks at 408, 402, and 384 nm, respectively (back lines). A) Following addition of ascorbic acid as an electron donor, ChuS and its two halves were all capable of heme oxygenation, as evident from the spectral shifts (gray lines). Absorbance values corresponding to wavelengths greater than 500 nm have been multiplied by 10 to amplify absorbance signals corresponding to the β- and α-bands. B) Similar to A), but NADPH and the P450 reductase system were used as the electron source.

41

Figure 2.4 EPR analysis of ChuS reconstituted with heme. Based on spectra comparison with myoglobin reconstituted with heme, ChuS-heme exists in two spin states.

42

µµµµµµµµ

Figure 2.5 ChuS reconstituted with heme. Based on absorbance at 408 nm during ChuS-heme reconstitution and assuming two binding sites, the estimated Kd is 1.0 ± 0.3 µM (n = 3).

43

2.4.5 Heme oxygenation by ChuS

Bacterial and mammalian HOs use the CPR-NADPH system, ferredoxin,

flavodoxin, or ascorbic acid as reducing partners in vitro (160). To effectively monitor

and quantify the HO activity of ChuS, 50 µM ascorbic acid was used. This concentration

is between 100- (21) and 350-fold (26) less than that used for characterizing other HOs.

The use of this lower concentration of ascorbic acid demonstrates highly efficient

catalysis of ChuS with this electron donor. Furthermore, by using small amounts of

ascorbic acid, heme degradation through coupled oxidation was minimized (34). To

examine the heme degrading activity of ChuS, we monitored the reaction by UV-visible

spectroscopy after the addition of ascorbic acid (Figure 2.6). As the reaction progressed,

the Soret peak at 408 nm decreased, the small broad peak at 560 nm increased, and the

absorption at 680 nm initially increased then decreased. Because the addition of catalase

and superoxide dismutase did not affect the trends of the spectral change, this effect can

be attributed to specific heme degradation by ChuS and its variants. These spectral shifts

are similar to those observed for other HOs, which suggests the formation of biliverdin

and free iron rather than the Fe3+-biliverdin complex formed by HO-1. Furthermore,

similar spectral shifts were also observed for heme reconstituted both to N-ChuS and C-

ChuS, indicating that both halves of ChuS are capable of degrading heme independently

(Figure 2.3A). Similar endpoint spectra were also observed for ChuS, N-ChuS, and C-

ChuS when CPR was used as the electron donor (Figure 2.3B).

44

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

320

340

360

380

400

420

440

460

480

500

520

540

560

580

600

620

640

660

680

700

Wavelength (nm)

Ab

so

rban

cy Zero

45 sec

1.5 min

3 min

4.5 min

6 min

7.5min

8.5 min

10 min

16 min

20 min

1 hour

5x

Figure 2.6 Spectral change of ChuS reconstituted with heme, monitored over time following ascorbic acid addition. Absorbance measurements between 500 and 700 nm have been multiplied by five to amplify the change in the β- and α-bands. Arrows follow changes in absorption at 408, 545, and 680 nm during reaction progression. The spectral change corresponds with the formation of the product, biliverdin.

45

2.4.6 Detection of CO release

All three ChuS constructs were capable of releasing CO (Figure 2.7). When

ascorbic acid was used as the electron source, C-ChuS was 1.5 times more effective at

CO production than ChuS and 10 times more effective then N-ChuS, whose release was

only slightly above the background level of heme breakdown. When both N-ChuS and

C-ChuS were added to the same reaction vessel, the amount of CO evolved was greater

than the levels observed for ChuS alone, but not as great as those observed for C-ChuS.

When CPR was used as the electron source, the amount of CO produced by ChuS and

C-ChuS was five times less than when ascorbic acid was used. In contrast, N-ChuS

produced more than four times as much CO as when ascorbic acid was used and was

slightly more effective at producing CO than either the C-ChuS or full-length ChuS.

46

0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0

Hem

e

ChuS

N-C

huS

C-C

huS

N +

C C

huS

Av

era

ge

CO

Re

lea

se

d (

pm

ol/

min

) Ascorbic Acid

Cytochrome P450Reductase-NADPH

Figure 2.7 Compiled average CO detected for ChuS, N-ChuS, and C-ChuS. Electron donors were either ascorbic acid (red) or cytochrome P450 reductase-NADPH (blue) as an electron source for heme oxygenation. Error bars represent standard errors.

47

2.4.7 Organization of the heme uptake operon

Given the similarity between the heme uptake operon of Y. enterocolitica and

E. coli O157:H7, it is likely that homologues of HemR, HemT, HemU, HemV, and HemS

would be required for heme assimilation and use in E. coli. Sequence comparison

between the heme uptake loci of these organisms revealed the homologous genes ChuA,

ChuT, ChuU, various putative ATP-binding components of the heme transport system,

and ChuS. Sequence identities ranged between 38% and 68%, and similarities were

between 58% and 80%. The highest identity was conserved between HemS and ChuS,

suggesting conservation of function in iron acquisition and prevention of heme toxicity.

The lower identity with the other proteins may reflect differences between the

periplasmic and membrane environments of the two organisms. Furthermore, sequence

comparison revealed the following ChuS orthologs with strong sequence similarity: ChuS

(CFT073), ShuS (Shigella), EhuS (Enterobacter, Erwinia), HemS (Yersinia), plu2633

(Photorhabdus), HmuS (Rhizobium, Sinorhizobium), and BhuS (Bordetella), suggesting

that this HO is present in many other bacteria (Figure 2.8). Furthermore, conserved

histidine residues at positions 73, 87, 193 and 277 may be important for heme

coordination.

2.4.8 Dynamic light scattering

The derived molecular radii of N-ChuS, ChuS, and ChuS-heme in phosphate

buffer and ChuS and ChuS-heme in Tris buffers were all approximately 2.90 nm,

48

corresponding to a MW of 41 kDa for an elongated protein. This is in agreement with

that expected for an N-ChuS dimer or ChuS monomer. The longest dimension of ChuS

from the crystal structure corresponds to a radius of 2.88 nm, indicating that the N-ChuS

homodimer is similar in shape to full-length ChuS. The monomeric organization of ChuS

was confirmed using a gel filtration column calibrated with a set of MW standards. This

finding is contradictory to a previous report showing that a homologue of ChuS, ShuS

(98% identity), forms an oligomer with an estimated MW of 650 kDa (11). When ChuS

or ChuS reconstituted with heme were examined in the Tris buffer system following 4 d

of incubation at 4°C, DLS results suggested the formation of various high-MW

aggregates without any detection of a ChuS monomer. In contrast, the protein remained

monomeric after 1 week in phosphate buffer.

2.4.9 Inhibition by Sn(IV) mesoporphyrin

In the presence of Sn(IV) mesoporphyrin IX, a competitive inhibitor for heme

binding (152), a dose-dependent decrease in HO activity was observed from 2- to 7-fold.

This level of inhibition is similar to that characterized for other HOs in the presence of

competitive inhibitors (152) (Figure 2.9).

49

10 20 30 40 50 60

ChuS:O157:H7 MNHYTR WLELKEQNPG KYARDIAGLM NIREAELAFA RVT-HDAWRM HGDIREILAA LESVGETKCI

ShuS:Shigella MNHYTR WLELKEQNPG KYVRDIAGLM NIREAELAFA RVT-HDAWRM RGDIREILAA LESVGETKCI

ChuS:CFT07 MNHYTR WLELKEQNPG KYARDIAGLM NISEAELAFA RVT-HDAWRM RGDIREILAA LESVGETKCI

EhuS:Enterobac MVIPFGATIQ KDKIMGHYTR WLELKEEHPG KYARDIAGLM HISEAELAFA RVG-HDAWRL RGEIREILAA LESVGETKCI

HemS:Yersinia MSKSIYEQ YLQAKADNPG KYARDLATLM GISEAELTHS RVS-HDAKRL KGEARALLAA LEAVGEVKAI

EhuS:Erwinia MQTSIYQQ YLQAKVEHPR KYARDLATLL NISEAELTHA RVG-HDAQRL QGEAKVWLSA LEQVGNTKSI

plu2633:Photor MNSSLYER YLQAKSDNKA KYARDLAAYL NVSEGELLHS RVG-IDAKRL NVEAPTLLEE LAVVGEIKAI

HmuS:Rhizobium M TEQTRPAPAE IRTFRAENPK MRERDIAAQL KISEAALVAA ETG-ISVTRI DGSALKLLER VAGLGEVMAL

HmuS:Sinorhizo MTM TESIRPTPSE IRAYRAENPK LRERDIAAKL GISEAALVAA ECG-LTAVRI ESSAGRFLER VEELGEVLAL

BhuS:Bordetell MTDINF ETRAADLRAR HDAYAASQPG QRARNAAQAL EVSEAEWVAA GCGGARVTAL RGEPQTIFRE LGTLGEVMAL

Consensus ...... .......y.r .l..ka.np. kyaR#.A..$ .!sEael..a rvg..da.rl .g.a...l.. le.vG#vkai

70 80 90 100 110 120 130 140

ChuS:O157:H7 CRNEYAVHEQ VGTFTNQHLN GHAGLILNPR ALDLRLFLNQ WASVFHIKEN TARGE-RQSI QFFDHQGDAL LKVYATDNTD

ShuS:Shigella CRNEYAVHEQ VGAFTNQHLN GHAGLILNPR ALDLRLFLNQ WASVFHIKEN TARGE-RQSI QFFDHQGDAL LKVYATDNTD

ChuS:CFT073 CRNEYAVHEQ VGAFTNQHLN GHAGLILNPR ALDLRLFLNQ WASVFHIKEN TARGE-RQSI QFFDHQGDAL LKVYATDNTD

EhuS:Enterobac CRNEYAVHEQ VGAFTNQHLG GHAGLVLNPR ALDLRLFLNQ WASAFHISET TSRGE-RQSI QFFDHQGDAL LKVYTTDHTD

HemS:Yersinia TRNTYAVHE. MGRYENQHLN GHAGLILNPR NLDLRLFLNQ WASAFTLTEE TRHGV-RQSI QFFDHQGDAL HKVYVTEQTD

EhuS:Erwinia TRNEYAVHEH TGYYQNQTLD GHAGLILNPR ELDLRLFLSQ WASAFALREN TARGE-RQSI QFFDHQGDAI LKVYTTDTTD

plu2633:Photor TRNDFAVHEH IGRYVNTHFS SHAGLVLNPR ELDLRIFFGH WSSAFYLVEP AKNGM-RHSI QFFDHQGDAL HKVYTTGNTD

HmuS:Rhizobium SRNESAVHEK IGVFENIKSG AQAAIVLGE- NIDLRIFPSR WEHGFAVSKK DGDQE-RLSL QYFDKTGNAV HKVHLRPSSN

HmuS:Sinorhizo TRNESAVHEK VGVYENIKQG HAAALVLGS- EIDLRIFPGA WAHGFAVTKT D.KGEVRRSL QFFDKWGHAV HKVHLRPQSN

BhuS:Bordetell SRNEWCVHER HGRYEDIQAQ GHVGLVLGP- DIDLRVFFGH WKSAWAV--- --EQDGRHSL QFFDGAGVAV HKVYRTEATD

Consensus .RNe.aVHE. .G.%e#.... ghagl!Lnpr .lDLR.F... Was.fa..e. ...ge.R.Si Q%FDhqGdA. hKVy.t..t#

150 160 170 180 190 200 210 220

ChuS:O157:H7 MAAWSELLAR FITDE--NTP LELKAVDA-P -VVQTRADAT VVEQEWRAMT DVHQFFTLLK RHNLTRQQAF NLVADDLACK

ShuS:Shigella MAAWSELLAR FITDE--NMP LELKAVDA-P -VVQTRADAT VVEQEWRAMT DVHQFFTLLK RHNLTRQQAF NLVADDLACK

ChuS:CFT073 MAAWSELLAR FITDE--NTP LELKAVDA-P -VVQTRADAS VVEQEWRAMT DVHQFFTLLK RHNLTRQQAF NLVADDLACK

EhuS:Enterobac VAAWGDVLTR FIIAD--NPT LALKAVDA-P -AHSDGADAG SVEKEWRAMT DVHQFFSLLK RHNLSRQQAF RLVSDDLACK

HemS:Yersinia MPAWEALLAQ FITTE--IPE LQLEPLSA-P EVTEPTATDE AVDAEWRAMT DVHQFFQLLK RNNLTRQQAF RAVGNDLAYQ

EhuS:Erwinia MAAWQLFVEK FTSSE--NPA LTRRPASN-T AVEQPV.HNP QFEADWRAMT DVHDFFILLR RYNLTRQQAF NSVADDLAYR

plu2633:Photor MTAWEALIEK YQTED--NPA LVIKPIEEIT YSTLTNELKA QLDEEWRAMT NVHQFFTLLK NHNLSRQQVF NAVHDDLAYK

HmuS:Rhizobium VEAYHAIVAE LKLEDQSQEF VEAETSNAAD DTADVSRD-- ELRDNWSKLT DTHQFFGMLK RLKIGRQAAV RTVGDDYAWR

HmuS:Sinorhizo LAAYDKMVED LRL.DQSQDF IADAGAPATD DAVDAAVDTA ELRDRWSKLT DTHQFPGMLR KLNIGRRRAL HAIGDDFAWR

BhuS:Bordetell EAAWNALIEK .A----GPPV WPLPTAIAAA EESTEVTDPQ ALREAWLAMQ DTHDFFPLLR KFKVSRLAAL AAVGSDLAQA

Consensus .aAw.al.e. f.......p. l......a.. .......d.. .l...Wra$t #vH#Ff.$Lk r.nl.Rqqaf .a!gdDlACK

230 240 250 260 270 280 290

ChuS:O157:H7 VSNSALAQIL ESAQQDGNEI MVFVGNRGCV QIFTGVVEKV V---PM---K GWLNIFNPTF TLHLLEESIA EAWVTRKPTS

ShuS:Shigella VSNSALAQIL ESAQQDGNEI MVFVGNRGCV QIFTGVVEKV V---PM---K GWLNIFNPTF TLHLLEESIA EAWVTRKPTS

ChuS:CFT073 VSNSALAQIL ESAQQDGNEI MVFVGNRGCV QIFTGVVEKV V---PM---K GWLNIFNPTF TLHLLEESIA EAWVTRKPTS

EhuS:Enterobac VDNSALAQLL ETARQDGNEI MIFVGNRGCV QIFTGVVEKL T---PM---K GWLNIFNPTF TLHLLEETIA ETWVTRKPTA

HemS:Yersinia VDNSSLTQLL NIARQEQNEI MIFVGNRGCV QIFTGMIEKV T---PH---Q DWINVFNQRF TLHLIETTIA ESWITRKPTK

EhuS:Erwinia VENDALSQIL FAAREDQNEI MVFVGNRGCV QIFTGTLERV ERMEQH---A EWINIFNKDF TLHLIEGAIA ESWITRKPTA

plu2633:Photor VENSALNELI NTAYNDQNEI MIFVSNHGCV QIFTGKLERL MPHQEENSDQ KWINIFNRNF TLHLIESAIA ESWVTRKPTE

HmuS:Rhizobium LDNSATAEMM HASVKSGLPI MCFVANDGIV QIHSGPIFNV QSMGP----- -WINIMDPTF HLHLRQDHIA ETWAVRKPTT

HmuS:Sinorhizo LDTACIEMMM RSAAETALPI MCFVGNGGVI QIHSGPVVKI GTMGP----- -WLNVMDETF HLHLRTDHIA ELWAVRKPTA

BhuS:Bordetell VPLDSAERML TQAAACGLPI MCFVGNRGMI QIHTGPVESL RRTGP----- -WFNVLDARF NLHLNT.AIA QSWVVNKPTA

Consensus v.nsal...l ..a...gneI M.FVgNrGc! QIftG.ve.. ....p..... .W.N!f#..F tLHL.e..IA #sWvtrKPT.

300 310 320 330 340 Identity Similarity

ChuS:O157:H7 DGYVTSLELF AHDGTQIAQL YGQRTEGEQE QAQWRKQIAS LI-PEGVAA - -

ShuS:Shigella DGYVTSLELF AHDGTQIAQL YGQRTEGEQE QAQWRKQIAS LI-PEGVAA 98% 98%

ChuS:CFT073 DGHVTSLELF AHDGTQIAQL YGQRTEGEQE QAQWRKQIAS LI-PEGVTA 98% 98%

EhuS:Enterobac DGHVTSLELF AADGTQIAQL YGQRTEGEPE QSQWRRQIDA LT-PEGLAA 82% 90%

HemS:Yersinia DGFVTSLELF AADGTQIAQL YGQRTEGQPE QTQWRDEIAR LN-NKDIAA 66% 77%

EhuS:Erwinia DGIVTSLELY AAEGSQIAQL FGQRTEGTPE QQQWRNQIDA LKLRKDLAA 64% 76%

plu2633:Photor DGFVTSLELF DANGDQIAQL FGQRTEGTPE QTQWREQVAA LPKLQSIQEE RVA 55% 70%

HmuS:Rhizobium DGHVTSLEAY NAEGEMIIQF FGKRQEGSDE RAAWREIMEN LPRAASVAA 36% 55%

HmuS:Sinorhizo DGHVTSLEGL DAKGEMIIQF FGKRKEGSSE RAEWRSLVEE LPRLEAVVAA 37% 54%

BhuS:Bordetell DGWVTSLEIY AANGDLIVQF FGERKPGKPE LPVWRELMLS LCEQPLAA 37% 54%

Consensus DG.VTSLEl. aa.G.qIaQl %GqRteG.pE q.qWR..... L......aa. ...

Figure 2.8 Sequence alignment of ChuS with homologues from various Gram-negative bacteria. Highly conserved residues are highlighted in red, residues with sequence similarity are presented in blue. Numbering corresponds to ChuS. Conserved histidines across the organisms examined at positions 73, 87, 193 and 277 may be important for heme coordination by ChuS.

50

0

1

2

3

4

5

6

7

8

9

Ratio Sn to Heme

pm

ol

CO

/min

100 µµµµM

Heme

1:1 2:1 4:1 20:1 100 µµµµM

SnHeme

No ChuS

0

1

2

3

4

5

6

7

8

9

Ratio Sn to Heme

pm

ol

CO

/min

100 µµµµM

Heme

1:1 2:1 4:1 20:1 100 µµµµM

SnHeme

No ChuS

Figure 2.9 Inhibition of ChuS heme degradation by Sn(IV) mesoporphyrin IX. ChuS activity was inhibited in a dose dependent manner and was similar to that observed for other heme oxygenases. Error bars represent standard error.

51

2.5. Discussion

ChuS is the first HO identified in an E. coli strain. For pathogenic organisms that

cause hemolytic lesions, such as the E. coli O157:H7 and CFT073 strains, heme

degradation by oxygenation may be a significant mechanism for iron acquisition and/or

prevention of heme toxicity. Monitoring the spectral changes of ChuS reconstituted with

heme when ascorbic acid or NADPH-CPR were used as electron donors provided direct

evidence that ChuS catalyzes the breakdown of heme to biliverdin, CO, and free iron. In

vivo, E. coli may use reductants such as ferredoxins and flavodoxin as electron sources

for heme oxygenation (157). The liberation of CO and inhibition of CO release by

addition of the prototypical HO inhibitor Sn(IV) mesoporphyrin IX support ChuS as a

HO. While the spectra of N-ChuS and C-ChuS do not definitively suggest HO activity,

unequivocal detection of CO release indicates that both halves of ChuS possess the

ability to break down heme. Based on the significant sequence similarity of proteins

between Enterobacter, Erwinia, Shigella, Yersinia, to E. coli ChuS, we can extend our

identification of ChuS as an HO to these other bacterial genera. Further work will

characterize heme breakdown products resulting from the action of N-ChuS and C-ChuS

and the underlying mechanisms.

Although the overall structure of all bacterial and mammalian HOs characterized

to date share the same mainly α-helical fold, the structure of ChuS is unique in that it is

comprised of two central sets of anti-parallel β-sheets, each flanked by two pairs of

α-helices. The absorption spectra of the two halves of ChuS following reconstitution with

heme are notably different from each other and the full-length protein. This is especially

52

true for the C-terminal half, whose broad peak centered at 384 nm is unique compared to

the Soret range of 402 to 412 nm observed for other HOs identified to date (44).

The structure of ChuS consists of a structural duplication, despite the fact that the

two halves share only 19% sequence identity. This unique structural feature may be

explained in part by the expression profiles of recombinant full-length ChuS, N-ChuS,

and C-ChuS. Recombinant forms of full-length ChuS and N-ChuS were highly expressed,

whereas the expression of C-ChuS was at least 30 times lower and required a large GST

tag to increase protein solubility. Considering its structure, this decreased solubility of C-

ChuS is not surprising. In contrast to the N-terminal half, the interface of the C-terminal

half consists of hydrophobic residues (V244, V250, F274, L276), exposure of which

would certainly lead to poor solubility. CO release experiments provide further insights

as to why an enzyme having a structural repeat should exist when the N-terminal half

alone would suffice. In the presence of ascorbic acid as an electron donor, C-ChuS

produced over 18 times more CO than the weakly acting N-ChuS. When CPR was used

as the electron source, N-ChuS was almost twice as reactive as C-ChuS. Together, the

two halves of ChuS would enable pathogenic E. coli to use different electron donors

while stabilizing the less-soluble C-terminal half of the molecule. This, in turn, would

increase the chance of survival of E. coli strains in the host, thereby maximizing their

virulence.

In addition to having an essential role in the acquisition of iron from host sources,

ChuS may also protect E. coli against heme toxicity. There is growing evidence that

heme accumulation can result in cell damage and tissue injury, as heme catalyzes the

formation of reactive oxygen species, resulting in oxidative stress (9, 12). Furthermore,

53

because heme has a low MW and is lipophilic, it can easily intercalate into the membrane

and impair lipid bilayers and organelles, such as mitochondria and nuclei, and destabilize

the cytoskeleton (113, 149). Therefore, the HO activity of ChuS may be required to

convert heme to biliverdin, a precursor to the less-reactive species bilirubin. This role for

ChuS was also suggested through the appearance of a blue-green pigment during ChuS

expression, characteristic of biliverdin accumulation. The ORF YhhX has been identified

in E. coli K12, O157:H7, and CFT073. YhhX is downstream of two Fur binding sites

(149), shares 19% sequence identity to human BVR chain A, and may be the oxido-

reductase that converts biliverdin to bilirubin. The structure of a homologue to YhhX

(41% identity) has recently been deposited in the PDB, and publication of its structural

and functional analysis could confirm the identity of the second enzyme in the heme

degradation pathway. Finally, our results from DLS demonstrated that ChuS and ChuS-

heme both form high-molecular-weight aggregates after 4 d of incubation in Tris buffer

at 4°C. This could be the cause of the absence of HO activity previously reported for

ShuS, a homologue of ChuS (11).

In summary, we determined the structure of ChuS, which represents a novel fold

and is different from that of known mammalian and bacterial HOs. The structure contains

tandem structural repeats, which are both functional in terms of heme oxygenation

activity, albeit with somewhat different properties. Through heme spectral analysis and

CO quantification, we identified ChuS as an HO, the first to be identified in E. coli.

54

2.6. Acknowledgments

We thank Dr. Gerald Marks for his support, Dr. B. McLaughlin for help with gas

chromatography, Dr. B. Hill for EPR experimentation and interpretation, other members

of the Jia lab (especially M. Adams) for help with synchrotron data collection, and Jim

Blonde for help with cloning. The expert help of Dr. Mike Nesheim is greatly appreciated.

We are also grateful to the Advanced Photon Source beamlines 17-ID and 19-BM, where

the diffraction data were collected. This work was supported by the CIHR. MS was

supported by an E.G. Bauman Fellowship, and ZJ is a Canada Research Chair in

Structural Biology and an NSERC Steacie Fellow.

55

Chapter 3

Structure of the Escherichia coli O157:H7 heme oxygenase ChuS in complex with heme and enzymatic inactivation by mutation of the heme coordinating residue

His-193

Preface:

This chapter was published in The Journal of Biological Chemistry: Michael D. L. Suits,

Neilin Jaffer, and Zongchao Jia (2006). Structure of the Escherichia coli O157:H7 heme

oxygenase ChuS in complex with heme and enzymatic inactivation by mutation of the

heme coordinating residue His-193. J Biol Chem. 281(48): 36776-82.

Michael Suits was responsible for protein expression, purification, crystallization, data

collection, structure solution and refinement, and all biochemical experiments in the

investigation of function. The chuS gene was cloned by Pietro Iannuzzi and Neilin Jaffer

was responsible for generation of the two ChuS mutants: H73A and H193N. The

manuscript was written by Michael Suits with editorial input from Dr. Zongchao Jia and

with additional help from Dr. John S. Elce and Jocelyne Suits.

Coordinates for ChuS-heme are deposited in the RSCB Protein Data Bank under accession code 2HQ2.

56

3.1 Abstract

Heme oxygenases catalyze the oxidation of heme to biliverdin, CO, and free iron.

For pathogenic microorganisms, heme uptake and degradation is a critical mechanism for

iron acquisition that enables multiplication and survival within hosts they invade. Here

we report the first crystal structure of the pathogenic Escherichia coli O157:H7 heme

oxygenase ChuS in complex with heme at 1.45 Å resolution. Compared to other heme

oxygenases, ChuS has a unique fold, including structural repeats and a beta sheet core.

Not surprisingly, the mode of heme coordination by ChuS is also distinct, whereby heme

is largely stabilized by residues from the C-terminal domain, assisted by a distant

arginine from the N-terminal domain. Upon heme binding, there is no large

conformational change beyond fine tuning of a key histidine (H193) residue. Most

intriguingly, in contrast to other heme oxygenases, the propionic side chains of heme are

orientated towards the protein core, exposing the α-meso carbon position where O2 is

added during heme degradation. This unique orientation may facilitate presentation to an

electron donor, explaining the significantly reduced concentration of ascorbic acid

needed for the reaction. Based on the ChuS-heme structure, we converted the histidine

residue responsible for axial coordination of the heme group to an asparagine residue

(H193N), as well as a second histidine to an alanine residue (H73A) for comparison

purpose. We employed spectral analysis and CO measurement by gas chromatography to

analyze catalysis by ChuS, H193N and H73A, demonstrating that H193 is the key residue

for the heme degrading activity of ChuS.

57

3.2 Introduction

Iron is an essential cofactor whose role as a biocatalyst or electron carrier enables

organisms to fulfill many vital biological processes including respiration, the citric acid

cycle, oxygen transport and utilization, gene regulation, and DNA biosynthesis (89). For

bacteria invading mammalian hosts, iron acquisition, transport, and utilization is essential

for survival and pathogenesis. However, due to the low solubility of iron(III) salts at

physiological pH in the presence of oxygen, iron uptake, storage, and transport are

considerable obstacles for invading bacteria to overcome. Enterohemorrhagic pathogens

such as Escherichia coli O157:H7 may induce hemolytic lesions in order to gain access

to host nutrient stores, such as heme; the most abundant source of invertebrate host iron

(83). Furthermore, growth of E. coli O157 isolated from human patients is stimulated in

the presence of heme and hemoglobin, suggesting that heme may be an important

signaling molecule for the up-regulation of factors contributing to host colonization and

virulence (83).

The proteins involved in heme internalization by enterohemorrhagic pathogens

have been shown to involve homologues to members of the heme utilization locus (127).

Using a laboratory strain of E. coli (1017 (ent::Tn5)) that was unable to import heme into

the cytoplasm, it was established that the outer membrane receptor ChuA was required

for heme uptake along with TonB, an energy-transducing protein associated with ExbB

and ExbD, forming a complex that uses the proton motive force of the cytoplasmic

membrane for the transport of ligands into the periplasm (146). Four other members of

the heme uptake operon from Yersinia enterocolitica were shown to be required to

58

complete heme uptake, transport, and use as an iron source by E. coli K-12 (127)

including: the periplasmic heme-binding protein HemT, the heme permease protein

HemU, the ATP-binding hydrophilic protein HemV, and the protein HemS, whose role

was supposed to have been involved in regulating the amount of non-protein-associated

heme in the cell (128). These Y. enterocolitica proteins are homologous to members of

the E. coli O157:H7 heme utilization locus chuT, chuU, chuV, and chuS, respectively.

We recently reported the crystal structure of apo-ChuS and confirmed via spectral

analysis and CO measurement that ChuS was a HO, the first to be identified in any strain

of E. coli (41). A Shigella dysenteriae chromosomal knockout of shuS, a homologue to

chuS, was shown to be defective in utilizing heme as an iron source and that expression

of ShuS was upregulated when challenged with iron-limiting conditions (22).

Furthermore, shuS was implicated to have a cytoprotective role against heme toxicity in

that growth of the shuS knockout mutant was impaired at intermediate concentrations of

heme (15 µM) and that shuS was absolutely required for colony growth at high heme

concentrations (40 µM) (22). Together, these studies suggest that ChuS and its

homologues are up-regulated under iron-limiting conditions, and that ChuS activity

enables pathogenic E. coli O157:H7 to use heme as an iron source, while conferring a

cytoprotective role in the prevention of heme toxicity. It appears very probable that ChuS

is an essential protein in heme utilization by various pathogenic, enterohemorrhagic

bacteria.

The crystal structures of HOs from other Gram-negative bacteria such as HemO

(174), HmuO (61), and PigA (46) all share a high structural similarity with human HO-1

(122); they are mainly α-helical and the heme is sandwiched between a histidine residue

59

and a glycine residue. However, the structure of ChuS revealed a unique architecture

compared to other HOs in that it contains a central set of antiparallel β-sheets and that the

N- and C-terminal halves are structural repeats. In light of the identification of ChuS as a

novel HO and its unique overall structure, it was of interest to elucidate the mechanism of

its heme binding and degradation. To this end, we have determined the X-ray crystal

structure of ChuS in complex with heme at 1.45 Å resolution. We have further performed

site-directed mutagenesis of the histidine residue (H193) (shown in the crystal structure

to be responsible for axial coordination of the heme group within ChuS) to an asparagine

(H193N), and of another conserved histidine residue (H73) to an alanine (H73A) as a

control. Using spectral analysis and CO measurement by gas chromatography to analyze

heme degradation catalyzed by ChuS, H193N, and H73A, we have demonstrated that

H193 is a key residue for the heme degrading activity of ChuS.

3.3 Experimental Procedures

3.3.1 Protein expression and purification, site-directed mutagenesis

ChuS fused to an N-terminal His6-tag was subcloned into a pET expression

vector. Based on the structure of the heme complex and sequence conservation analysis,

the mutations of H73A and H193N of ChuS were selected (41). Using the Quickchange®

site-directed mutagenesis kit (Stratagene, CA), wild-type ChuS was mutated at these two

histidine positions using the following primers. Base pair changes have been underlined.

H73A: sense 5'-CGTAATGAATATGCAGTCGCGGAGCAAGTTGGTACG-3',

60

antisense 5'-CGTACCAACTTGCTCCGCGACTGCATATTCATTACG-3', and H93N:

sense 5'-GGGCGATGACCGACGTTAATCAGTTTTTTACGTTGCTCAAG-3',

antisense 5'-GAGCAACGTAAAAAACTGATTAACGTCGGTCATCGCCC-3'.

Briefly, polymerase chain reaction conditions and cycling parameters recommended by

Stratagene (123) were followed using PfuTurbo polymerase for thirty cycles between 30

sec 95°C melt, 60 sec 55°C anneal, and 9 min 68°C extend. Methylated (non-PCR

amplified) DNA was then digested with Dpn1 restriction enzyme for 1 h and the digested

DNA was heat shock transformed into MC1061 cells. Transformed cells were

subsequently plated onto LB-agar plates supplemented with 100 µg/ml ampicillin. Single

colonies were grown overnight at 37°C in 5 ml Terrific Broth (Bioshop, Burlington, CAN)

supplemented with 100 µg/ml ampicillin (TB-amp). Plasmid DNA was purified by means

of a QIAprep Spin Miniprep Kit (Qiagen, Mississauga, CAN) was transformed into

BL21(DE3), and the desired mutations confirmed by sequence analysis (Robarts

Research Inst., London, CAN).

Cultures (1 L) of BL21(DE3) cells carrying plasmids for the respective

recombinant protein were grown at 37°C in TB-amp. Protein expression was induced at a

culture OD600 between 0.6-1.0 using 0.4 mM IPTG and grown for 5 h. ChuS, H73A, and

H193N were then purified via nickel nitrilotriacetate agarose (Ni-NTA) batch purification

in phosphate buffer (pH 8.0) and dialyzed overnight in 30 mM Tris, 5 mM NaCl, pH 8.0.

Over-expression of ChuS and H73A resulted in cells exhibiting a blue-green pigment

which is characteristic of biliverdin accumulation (87, 156, 174). ChuS, H73A and

H193N all expressed at high levels (>20 mg/L culture depending on batch) and were

soluble in Tris (pH 8.0) or phosphate (pH 7.0) buffers following Ni-NTA purification.

61

Purification of each protein was monitored by SDS PAGE, and each was estimated to be

>95% pure prior to spectral analysis, CO measurement, or heme reconstitution. H193N

and H73A were subsequently dialyzed twice more to eliminate any residual imidazole.

Protein concentrations were determined by the BioRad protein assay (Biorad,

Mississauga, CAN) in 96 well format, using Bovine Globulin G as a protein standard.

3.3.2 ChuS-heme reconstitution

The ChuS-heme complex was prepared by dissolving heme (ferriprotoporphyrin

IX chloride) (Sigma-Aldrich, Oakville, CAN) into 0.5% (v/v) ethanolamine for 30 min,

then into 30 mM Tris, and the pH was adjusted to 8.0. Small volumes of this mixture

were then added to dialyzed ChuS until a final ratio of 4:1 heme to protein was reached.

The ChuS-heme mixture was then incubated for 15 min, centrifuged at 16,000 rpm for 15

min, in a JA-20 Beckman Coulter rotor (Beckman Coulter, Mississauga, CAN), and

purified using a Resource Q column on an ATKA Explorer FPLC with 30 mM Tris, 1 M

NaCl (pH 8.0) as the elution buffer. Collected fractions were examined by SDS-PAGE,

and those that were evaluated to contain pure protein were pooled. Pure ChuS was

concentrated using an Amicon Ultra 15 kDa cutoff centrifugal filtration device (Fisher,

Mississauga, CAN) and used immediately for crystallization. Reconstituted ChuS-heme

complex concentrations were determined by using the extinction coefficient reported for

ShuS (εεεε410) of 159 mM-1cm-1 for ChuS (11).

62

3.3.3 Crystallization of ChuS-heme

ChuS-heme crystals were obtained only by the sitting drop vapor diffusion

method, in 96-well plates (Greiner Bio-One, NC, USA), using freshly prepared protein

and reagents. Crystallization was achieved by mixing 1.75 µl of 20 mg/ml ChuS-heme

with 1.75 µl of chilled reservoir solution consisting of 12-15% (w/v) PEG 3350, 0.15-

0.20 M magnesium formate, and 10-20 mM NAD. Crystallization plates were incubated

in the dark at 4°C, with both hexagonal and triangular prism crystals appearing after 1 to

3 d. For cryoprotection, crystals were soaked in the reservoir solution containing ethylene

glycol (concentration increased stepwise to 30% [v/v]), picked up using a nylon loop, and

flash-cooled in the N2 cold stream at 100K. Two crystal forms belonging to space groups

P65 and R3 were obtained whose morphology were hexagonal and triangular prism,

respectively.

3.3.4 Data collection and structure determination

Diffraction data of the P65 and R3 crystals were collected at beam line X29

equipped with an ADSC Quantum-315, nine quadrant, CCD detector at the National

Synchrotron Light Source, Brookhaven National Laboratory. For both the P65 and R3

crystals, data were collected for 180° (0.5° and 1.0° oscillations, respectively) at the

wavelength 1.100 Å. Data were processed with HKL2000 (102) and the structure of the

R3 crystal was solved via molecular replacement using the program Phaser (88) with

apo-ChuS (PDB ID code 1U9T) as the initial probe structure. Although the P65 crystals

63

diffracted beyond 2.0 Å resolution, due to the long unit cell dimensions and resulting spot

proximity, diffraction was only collected to 2.7 Å. The heme moiety was manually placed

in the ChuS structure according to the difference density in both space groups. The

ChuS-heme complex structure in R3 space group was completed by iterative cycles of

manual fitting using XtalView/Xfit (84) and refinement using Refmac5 (96, 148) at a

final resolution of 1.45 Å. Coordinates and structure factors for ChuS-heme are deposited

in the RSCB Protein Data Bank under accession code 2HQ2. Due to lower resolution and

poor data quality in regions along the long axes, the P65 complex structure was not further

refined beyond confirming heme the position.

3.3.5 Spectral analysis and CO measurements

Reconstitution of ChuS, H73A, and H193N with heme was achieved by

dissolving heme into 0.5% (v/v) ethanolamine and diluted into phosphate buffer just prior

to mixing with 10 µM protein samples to a final ratio of 1:1 protein to heme. Absorbance

data were collected after 15 min incubation at room temperature in 100 mM phosphate

buffer (pH 7.0) using a microplate spectrophotometer (Bio-Tek Instruments, Inc. NY,

USA.). Heme degradation was initiated in parallel by adding ascorbic acid to a final

concentration of 50 µM, and spectra were recorded between 300 and 700 nm at 5 nm

intervals.

For CO measurement, in 2-ml amber vials, 250 µl reaction volumes of 10 µM

heme and 10 µM ChuS, H73A, or H193N in 100 mM phosphate buffer (pH 7.0) were

incubated with constant shaking at room temperature for 15 min at which time enzymatic

64

heme degradation was initiated by adding ascorbic acid to a final concentration of 50 µM.

The amber vials were then sealed with screw caps and blue silicone rubber septa, the

headspace above each was immediately purged with CO-free air, and the vials were

incubated for 25 min at 37°C with constant shaking. The reaction was stopped by placing

the vials on powdered dry ice (-78°C), where they remained for approximately 30 min.

The amount of liberated CO in the headspace of each vial was measured using a gas

chromatograph equipped with an HgO reduction detector (Trace Analytical). The amount

of CO resulting from ChuS-catalyzed breakdown of heme was determined by comparing

n=3 peak area measurements for CO against linear CO standard curves (n=2

determinations; average correlation coefficient 0.9859). Technical details of CO detection

are described elsewhere (27, 153).

3.4 Results

3.4.1 Structure determination and analysis of ChuS-heme

Crystallization of ChuS reconstituted with heme using the hanging drop vapor

diffusion method turned from red to blue-green following room temperature incubation

periods of over one week. In the sitting drop vapor diffusion setup, two different crystal

forms of ChuS-heme were obtained from the same set of crystallization conditions over

the same time period, but differed dramatically in their morphology. In the case of the

P65 crystals, which appeared more frequently at room temperature, the crystals appeared

as broad (>1mm), flat, hexagonal plates. Crystals belonging to the R3 space group

appeared as clusters of long (>1 mm) triangular prisms and diffracted at 1.45 Å resolution.

65

Crystals from the P65 space group contained two molecules in the asymmetric unit

whereas those belonging to the R3 space group contained only one.

The final structure of ChuS-heme complex contains residues 0-337 of a total 354

amino acids including the His6-affinity tag for the R3 space group (Figure 3.1). In total,

100% of residues are within the allowed region of the Ramachandran plot. Only one

residue, A96, was in the generously allowed region and resides just prior to a surface-

exposed β-turn. Final refinement statistics are summarized in Table 3.1.

The ChuS-heme complex structure is virtually identical to that of apo-ChuS. The

overall structure comprises a central core of two large pleated β-sheets, each consisting of

nine anti-parallel β-strands sandwiched together and bowing outward in a saddle motif

(Figure 3.1). Each β-sheet is flanked at its N-terminus by one pair of parallel α-helices

and at the C-terminus by three α-helices in an α-loop-α-loop-α motif, forming a

structural repeat. When ChuS-heme is superimposed with apo-ChuS, the only notable

structural change is that the two α-helices N-terminal of H193 have shifted toward the

core of the protein (H193 by 3.58 Å), resulting in the appropriate orientation of this key

histidine residue for heme coordination (see Section 3.4).

66

Table 3.1: Diffraction data and refinement statistics for ChuS-heme

Space Group P65 R3 Cell Dimensions a = b (Å) 58.4 106.5 c (Å) 390.0 90.2 α = β (°) 90 90

γ (°) 120 120 No. molecules in ASU 2 1

λ (Å) 1.100 1.100 Resolution Range 50 - 2.7 50 - 1.45 Total Reflectionsa 92612 361180 Unique Reflectionsa 20724 67663 Completeness (%)b 85.2 (97.0) 98.7 (98.4) Rsym (I) (%) 10.0 6.7 I/σIb 23.9 (14.8) 38 (1.8)

Refinement Statistics Resolution range (Å) 50 – 1.45

Rwork, (%) 17.05

Rfree, (%) 19.30 No. reflections unique/ Rfree 63392 / 3371 No. atoms, protein / Heme/ solvent (H2O) / total 2606 / 45 / 354 / 3005

B factors (Å2) protein / heme/ solvent / total 21.4 / 24.3 / 32.2 / 22.7 RMS, bond length (Å) / bond angle(o) 0.0199 / 1.894 PDB Accession Code 2HQ2

aNumber of reflections following merging bValues in parentheses are for the outermost shell (P65: 2.81-2.70 Å, and R3: 1.50-1.45 Å)

67

N-terminal Domain

C-terminal Domain

H193

R100

H73

N-terminal Domain

C-terminal Domain

H193

R100

H73

Figure 3.1 Ribbon diagram of ChuS in complex with heme. ChuS binds to heme in a cleft region delineated by the C-terminal half (green) and N-terminal half (slate), between H193 at the base of an α-helix and R100 via two water molecules (blue) from central set of β-sheets at the core of ChuS. The residues important for heme coordination as well as the mutant control position H73 are depicted in red. Fig. 3.1 and 3.2 were generated using PyMOL (DeLano, 2006, The PyMOL Molecular Graphics System, www.pymol.org).

68

3.4.2 Heme binding

The heme group is coordinated in a cleft region delineated between H193 at the

base of a C-terminal α-helix and R100 from the central set of β-sheets at the core of

ChuS (Figure 3.1). A network of hydrogen bonds is formed between R100 to the Fe atom

of the heme group via two water molecules with distances of 2.93, 2.17, and 1.93 Å

respectively. In this cleft the heme group is stabilized by the non-polar series L90, L92,

F102, V192, and F243, and the polar series R100, H193, R206, M241, K291, Q313,

Y315 and R318 (Figure 3.2). In contrast to other HOs, the propionic side chains of the

heme moiety point toward the interior of one of the structural lobes formed between the

two halves of ChuS and are stabilized by the positively charged side chains R206, K291,

and R318 (Figure 3.2). These interactions are probably important in maintaining the

correct orientation of heme within ChuS. The burying of the propionic side chains

exposes the α-meso carbon of the heme group which is where O2 is selectively added

during heme degradation in HO reactions. Although the P65 data were of lower quality,

the identical heme position in this structure was unmistakable. The heme orientation in

ChuS is unique compared to other HOs, such as HO-1, in which the propionic side chains

point outward (117), although the heme degrading enzymes IsdG/IsdI may bind heme

with the propionic side chains pointing inward (125).

Essential to the heme coordination is His193, which makes axial interaction with

the central iron atom of the heme group. In ChuS, the interaction distance is 2.03 Å

between the NE2 atom of H193 and the center of the iron atom, allowing H193 to clamp

the heme firmly into place. R100, from the N-terminal domain, interacts with heme via

69

two water molecules and originates from the protein core beneath the heme group on one

of the central β-sheets (Figure 3.2). These water molecules were clearly visible even at

early refinement using multiple data sets, and are well defined in the final high-resolution

structure.

70

H193

F102

L90

L92

R100

M241

R318

Y315

Q313

V192

R206

K291

F243

αααα-mesocarbon

H193

F102

L90

L92

R100

M241

R318

Y315

Q313

V192

R206

K291

F243

αααα-mesocarbon

Figure 3.2 Difference omit map for the heme moiety and two water molecules (blue) within the active site pocket of ChuS contoured to 2σ σ σ σ in the R3 space group. Heme and waters were omitted from refinement prior to map calculations. Residues contributing to heme stabilization include the non-polar series L90, L92, F102, V192, and F243, and the polar series R100, H193, R206, M241, K291, Q313, Y315 and R318. Heme is therefore mainly coordinated by the C-terminal half but also by a distant residue, R100 from the N-terminal half. A network of hydrogen bonds is formed between R100 to the Fe atom of the heme group via two water molecules with distances of 2.93, 2.17, 1.93 Å respectively. In this orientation the propionic side chains of the heme group point toward the protein interior, exposing the α-meso carbon edge (black arrow). This presentation of the α-meso edge may facilitate electron attack during heme degradation.

71

3.4.3 Spectroscopic properties of ChuS, H73A, and H193N in complex with heme

H193 resides at the end of an α-helix, pointing into a cleft delineated by the

α-helix and the core set of β-sheets, interacting directly with iron in the heme group.

Thus, the mutation of H193 was expected to inactivate the protein. On the other hand,

H73 is within the central β-sheets, pointing between the N- and C-terminal halves and

does not make contact with the heme group. Since H73 is fully conserved, but does not

interact with heme, it provides a good control for H193. As previously reported, the

absorbance spectrum of native ChuS exhibits a clear Soret peak at 410 nm, together with

the common characteristic α/β-bands of HOs around 550 and 580 nm (11, 41, 118). The

H73A mutation was not observed to instigate any spectral changes compared to native

ChuS. However, the spectrum of the H193N mutant broadened, showed a decrease in the

magnitude of absorbance, and shifted the peak absorbance to 390 nm compared to the

Soret absorption observed for ChuS and H73A. A similar trend of spectral flattening and

peak shift to 625 nm was also observed for H193N in the α/β-band region.

3.4.4 Heme degradation by ChuS

Bacterial and mammalian HOs use the Cytochrome P450 Reductase-NADPH

system, ferredoxin, flavodoxin, and/or ascorbic acid as reducing partners in vitro (26, 160,

174). We have previously demonstrated the ability of ChuS to use either CPR-NADPH or

ascorbic acid as an electron donor for heme degradation (16, 41). It is important to note

that in our analysis we employed only 50 µM ascorbic acid, a concentration between 100-

72

(77) and 350-fold (26) less than that used for characterizing other HOs, thereby

minimizing the non-enzymatic degradation of heme through coupled oxidation.

Furthermore, calatase and superoxide dismutase were used for additional controls. To

examine the heme-degrading activity of ChuS, H73A, and H193N, we monitored the

reaction by UV-visible spectroscopy for 1 h after the addition of ascorbic acid. As the

reaction proceeded, the Soret peak for ChuS and H73A at 410 nm decreased, the small

broad peak at 560 nm increased, and the absorption at 680 nm initially increased, then

decreased (Figure 3.3). When ascorbic acid was added to H193N reconstituted with heme,

the profile of the spectrum did not change, but was observed to increase slightly in

absorbance. In the case of native ChuS, addition of catalase and superoxide dismutase did

not affect the observed spectral change, demonstrating specific enzyme activity.

When HOs degrade heme to biliverdin, CO is released as a byproduct. Using gas

chromatography, we measured CO release for wild-type ChuS, as well as the two mutants.

As shown in Figure 3.4, whereas wild-type ChuS and H73A were able to generate CO as

expected, the H193N mutation inactivated the enzyme. However, based on the

appearance of a broad peak for H193N near the Soret region of native ChuS, this mutant

seems to retain some interaction with heme.

73

Figure 3.3 Spectral change of ChuS reconstituted with heme, monitored over time following ascorbic acid addition. Absorbance measurements between 500 and 700 nm have been amplified five fold to emphasize the change in the β- and α-bands. Coloured arrows corresponding to the spectra of each protein follow changes in absorption at 408, 545, and 680 nm during reaction progression.

74

0

20

40

60

80

100

120

Heme H193N H73A ChuS

% A

ctiv

ity

- C

O M

easu

rem

ent

Figure 3.4 CO measurement via gas chromatography of ChuS, H73A, H193 and heme following ascorbic acid addition and 25 min incubation (n=3). Mutation at conserved H193 abolished heme degradation activity by ChuS.

75

3.5 Discussion

The X-ray crystal structure of the E. coli O157:H7 HO ChuS-heme complex has

revealed that the ChuS employs a unique mode of heme coordination. Whereas the

structures of other HOs characterized to date coordinate heme between a set of α-helices,

ChuS coordinates the central iron atom of the heme group between His193 at the end of

an α-helix and R100 (via two water molecules) originating from one of the central set of

β-sheets. In this fashion, the heme binding site of ChuS occurs between two distinct

interfaces, delineated by the repeating structural elements contributed mainly from the

C-terminal repeat and, to a lesser extent, the N-terminal half. While this heme

coordination is unique compared to other HOs, it is analogous to the mode of heme

coordination assumed for the heme degrading enzymes IsdG and IsdI from

Staphylococcus aureus (125) and is also similar to the heme binding PAS domain of

BjFixL (21) and EcDos (16).

Due to an absence of electron density in the data derived from the ChuS-heme

complex, a gap exists in the structure between residues 168 and 176 (V168-DAPVVQT-

R176), dividing the structural repeats of ChuS almost at the mid point. A similar gap was

also observed for apo-ChuS in the region of K166 to V173. Interestingly, the key heme

coordination residue H193 is located at the end of an α-helix towards the C-terminus of

the gap. In fact, H193 is in the only region that exhibits considerable conformational

change in the entire ChuS structure. It is thus conceivable that the flexible linker between

the N- and C-halves provides the flexibility for the critical iron-coordinating H193

residue of the C-terminal half. In the N-terminal half, R100 is also important for

76

coordination of the iron atom from the β-sheet core via a hydrogen bonding network

created through two water molecules. Therefore, the flexibility observed in the linker

between the two halves of ChuS may be required for heme binding. Communication

between the two lobes may also result in increased control over heme processing, similar

to that of the hemopexin dimer (40).

In spite of the evidence that genes homologous to chuS have been shown to be up-

regulated under iron limiting conditions, that these homologues prevented heme toxicity

in S. dysenteriae (22), and that no other heme degrading enzyme has been identified in

any E. coli strain, the precise role for ChuS and its homologues has been downplayed in

recent reports to be that of a cytoplasmic trafficker of heme, or a non-specific oligomeric

complex that binds to DNA (11, 118). By carefully examining the spectral progression

and CO measurement of ChuS catalyzed degradation of heme in the presence of minimal

amounts of ascorbic acid, we conclude that ChuS specifically degrades heme and that

His193 is the essential amino acid responsible. The confirmation of H193 as an integral

catalytic residue is affirmed through its sequence conservation, its role in coordination of

the heme moiety in the crystal structure, as well as the comparison of the H73A control

mutant. The spectral shifts observed for ChuS and H73A are similar to those observed for

other HOs, which suggest the formation of biliverdin and free iron rather than the

Fe3+-biliverdin complex formed by HO-1. Because the addition of catalase and

superoxide dismutase did not affect the trends of the spectral change, the observed

activity of ChuS can be attributed only to specific heme degradation (41).

It is difficult to extrapolate the patho-physiological significance of the heme

degrading ability of ChuS to its homologues in other bacteria due to the fact that other

77

redundant systems have been identified for iron acquisition and utilization from heme

sources including proteins from Pseudomonas aeruginosa (PigA and BphP) and

Staphylococcus aureus (IsdG and IsdI). Furthermore, redox enzymes are notoriously non-

specific with respect to their reaction with oxygen species (52). This has led to a lack of

straightforward characterization of other bacterial cytoplasmic heme binding proteins

outside those with sequence homology to HO-1. Therefore, understanding the interplay

between sequence and structural motifs in heme binding and utilization are of particular

interest in order to clarify our understanding of heme utilization by pathogenic bacteria.

As the propionic side chains of the heme group point toward the β-sheet core of ChuS,

the orientation of the heme group within ChuS may allow for specific targeting of heme

for degradation. This degradation may occur through the initial activation of the two

water molecules by R100, which would activate O2 in a similar fashion to the mechanism

hypothesized for HO-1 and its homologues (46). The distinct orientation of heme may

further explain the minimal amount of ascorbic acid needed for the degradation reaction,

because the α-meso carbon atom is exposed to facilitate interaction with an electron

donor.

During our investigation, a variety of practical obstacles were overcome which

could help in future studies of ChuS and its homologues. As previously reported, ChuS

and its homologues form high MW aggregates following incubation over a few days at

4oC (11, 16, 118). In order to overcome aggregation, fresh, unfrozen preparations of pure

ChuS were absolutely necessary for crystallization and biophysical studies. Furthermore,

even residual amounts of imidazole were observed to inhibit crystallization with heme

78

and the heme degradation reaction. Purification of ChuS must therefore ensure complete

removal of imidazole from reaction buffers.

In conclusion, we have provided a first view of the heme-bound structure of ChuS;

a novel HO from a pathogenic bacterium with a unique structure. The heme binding site

in ChuS is very different from those of other known HO-heme complexes. H193 plays a

key role in anchoring the heme group by tightly interacting with the central iron ion. The

majority of the heme stabilization is contributed by the C-terminal half, along with R100

from the N-terminal half that also interacts with iron via two fully conserved water

molecules. The exposed α-meso region, which is unique among known HO-heme

complexes, may facilitate presentation to an electron donor and thus provide an

explanation for the much lower concentration of ascorbic acid needed for the reaction.

The H193N mutation, examined by spectral and CO measurements, clearly demonstrates

the critical role of H193.

3.6 Footnotes

We thank Dr. B. McLaughlin and Dr. K. Nakatsu for help with gas

chromatography. We would like to express our gratitude to Howard Robinson and

Brookhaven National Laboratory beamline X-29, where the diffraction data were

collected for both the P65 and R3 crystals of ChuS-heme. Excellent equipment

management was also contributed by Hans Metz. This work was supported by the CIHR.

MS was supported by an E.G. Bauman Fellowship and an Ontario Graduate Scholarship.

ZJ is a Canada Research Chair in Structural Biology and an NSERC Steacie Fellow.

79

Coordinates and structure factors for ChuS-heme are deposited in the RSCB Protein Data

Bank under accession code 2HQ2.

80

Chapter 4

Structure of the Escherichia coli O157:H7 heme binding protein ChuX and its role in regulating heme uptake

Preface: Michael Suits was responsible for protein expression, purification, mutagenesis

crystallization, data collection, structure solution and refinement, and all biochemical

experiments in the investigation of function. Dr. Gour P. Pal conducted the initial

expression and crystallization which was refined and expanded by Michael Suits. Dr.

Mike Nesheim helped in heme reconstitution setup and interpretation of results to

calculate the Kd for ChuX-heme. The chuX gene was cloned by Pietro Iannuzzi. The

manuscript was written by Michael Suits with editorial input from Dr. Zongchao Jia and

with additional help from Brent Wathen and Jocelyne Suits.

81

Coordinates and structure factors for ChuX are deposited in the RCSB Protein Data Bank (www.rcsb.org) under accession code 2OVI. 4.1 Abstract

The genes encoded within the heme utilization operon enable various Gram-

negative bacteria to effectively take up and utilize heme as an iron source. For many

pathogenic microorganisms, iron acquisition from heme sources stimulates growth and

enables successful multiplication and survival within hosts they invade. While the

proteins involved in heme internalization are well characterized functionally, the fate of

heme once it has reached the cytoplasm has only recently begun to be resolved. Here we

report the first crystal structure of the conserved heme utilization protein ChuX from

pathogenic Escherichia coli O157:H7 determined at 2.05 Å resolution. ChuX forms a

dimer which, surprisingly, displays a very similar fold to the monomer structure of two

other heme utilization operon proteins ChuS and HemS from E. coli and Yersinia

enterocolitica, despite sharing less than 19% sequence identity. Absorption spectral

analysis of heme reconstituted ChuX demonstrates that ChuX binds to heme in a 1:1

manner implying that each homodimer coordinates two heme molecules, in contrast to

ChuS where only one heme molecule is bound. Based on sequence and structural

comparisons, we designed a number of site-directed mutations in ChuX to probe the

heme binding sites and dimer interface. Mutations of both H65L and H98D were required

to abolish heme binding capacity of ChuX. Further, mutations within the dimer interface

resulted in compromised heme binding, supporting the notion that varying ChuX

juxtapositions observed in the structure may serve to modulate heme binding. Taken

82

together, it appears that ChuX acts upstream of ChuS and regulates heme uptake through

ChuX-mediated heme binding and release.

4.2 Introduction

Iron is the fourth most abundant element in the Earth's crust (6). However, due to

the low solubility of iron(III) salts at physiological pH in the presence of oxygen, iron

uptake, storage, and transport is a serious obstacle for bacteria to overcome for survival

and pathogenesis. Furthermore, invading microorganisms are confronted with the

obstacle that 99.9% of remaining body iron is sequestered as protein bound heme in

hemoglobin, myoglobin, and various other host heme coordinating proteins (127).

Through mechanisms of heme detection and liberation from host hemoproteins, bacteria

have evolved the ability to effectively uptake heme through translocation machinery and

then to utilize heme as a prosthetic group or as a valuable pool of iron. Additionally, the

products of heme degradation, biliverdin and bilirubin, may serve a cytoprotective role in

bacteria where they can act as antioxidants (73).

Enterohemorrhagic pathogens such as Escherichia coli O157:H7, for which

isolates from human patients have been shown to have stimulated growth in the presence

of heme and hemoglobin, may induce hemolytic lesions in order to gain access to this

potential sources of heme and iron (83). Using a laboratory strain of E. coli (1017

(ent::Tn5)) that is unable to import heme into the cytoplasm, it was established that the

outer membrane receptor ChuA (E. coli heme-utilization protein A) was required for

heme uptake along with TonB, an energy-transducing protein associated with ExbB and

ExbD that uses the proton motive force of the cytoplasmic membrane for the passage of

83

ligands into the periplasm (146). Three other members of the heme uptake operon from

Y. enterocolitica were shown to be required to complete heme uptake and transport by

E. coli K-12 (127) including: the periplasmic heme-binding protein HemT, the heme

permease protein HemU, and the ATP-binding hydrophilic protein HemV (128).

Once internalized, heme would be rapidly degraded by bacterial proteins such as

ChuS (E. coli O157:H7) (139), Cj1613c (Campylobacter jejuni) (111), HmuQ

(Bradyrhizobium japonicum) (106), IsdG (Bacillus anthracis) (125), and ShuS (Shigella

dysenteriae) (165), all of which could use the abundance of potential electron donors to

liberate iron that can then be incorporated into various proteins where it acts as a

biocatalyst or electron carrier. Alternatively, internalized heme could be stored or used in

various vital biological processes including photosynthesis, gene regulation, oxygen

utilization, N2 fixation, methanogenesis, H2 production and consumption, the

tricarboxylic acid (TCA) cycle, or DNA biosynthesis (89, 91). A tightly regulated

balance between heme uptake and degradation is of critical importance.

Restriction mapping and sequence analysis of selected regions of E. coli O157:H7

DNA, the serotype responsible for outbreaks of hemorrhagic colitis and hemolytic uremic

syndrome, has revealed that the organization of the heme transport locus is strikingly

similar to that of other enteric bacteria such as Y. enterocolitica and S. dysenteriae (127,

128). These include the three other genes clustered within the heme utilization locus,

chuW, chuX, and chuY which are poorly characterized, but whose homologues appear to

be co-transcribed with the periplasmic binding protein shuT in S. dysenteriae (127, 128).

Twelve nucleotides downstream of shuW is the start codon for shuX, and the initiation

codon of shuY overlaps the stop codon of shuX which may indicate that these two smaller

84

genes are co-transcribed (164). In a similar sequence analysis of the heme uptake locus

from Y. pestis, genes homologues to chuX and chuY (orfX and orfY, respectively) were

evaluated to be located immediately downstream of the heme receptor gene hmuR and

were therefore likely co-expressed (145). In Vibrio anguillarum, which can use heme and

hemoglobin as a source of iron, using a lacZ reporter system it was shown that a

homologue of chuX, huvX, was co-transcribed with huvZ under iron limiting conditions

(95). An E. coli homologue to huvZ was not identified via NCBI Blastp sequence analysis

at (www.ncbi.nlm.nih.gov/BLAST/). Furthermore, an examination of the -10 and -35

elements downstream of huvX revealed sequence with similarity to sigma70-promoters, as

well as the presence of a Fur element (95). Characterization of either chuX or chuY gene

product or their homologues from various heme utilizing organisms had not received any

attention at the beginning of our investigation. However, based on Blastp sequence

analysis ChuX was suggested to be either a hypothetical protein or protein associated

with heme-iron utilization, and that ChuY is a hypothetical protein or nucleoside-

diphosphate-sugar epimerase (143).

While traditional heme studies have focused on heme as a prosthetic group within

hemoglobin and myoglobin, or general heme involvement in iron processing, many novel

heme associating proteins have been discovered recently. These include the cytochrome c

chaperone CcmE (123), the iron-dependent regulator of heme biosynthesis Irr (107), and

the heme-based aerotactic transducer Hem-AT (64). Other heme proteins function

through conformational changes induced by a modification to the ferrous heme

environment upon the coordination of an effector ligand such as NO, O2, or CO, as in the

case of guanylate cyclase (172), FixL (52), and CooA (81), respectively. Furthermore,

85

heme has recently been shown to have functionality beyond simply acting as a prosthetic

group or a source of iron. In HeLa cells, succinyl acetone-induced heme deficiency was

shown to increase the protein levels of the tumor suppressor gene product p53 and CDK

inhibitor p21, decrease the protein levels of Cdk4, Cdc2, and cyclin D2, and diminish the

activation/phosphorylation of Raf, MEK1/2, and ERK1/2-components of the MAP kinase

signaling pathway (168). This set of results highlights a role for heme as an effector

molecule beyond the regulation of proteins associated with heme and iron metabolism.

Despite the emerging knowledge concerning hemoproteins in many Gram-

negative bacteria, very little has been published concerning the three lesser characterized

members of the heme utilization operon, chuW, chuX, or chuY. We have determined the

crystal structure of ChuX at 2.05 Å resolution. Through site-directed mutagenesis of

conserved histidine residues, spectral analysis, and structural comparison of ChuX with

ChuS and HemS we have identified heme coordination sites within ChuX and the

structural basis of heme binding and release. Furthermore, we have shown that ChuX

plays a protective role against heme toxicity.

4.3 Experimental Procedures

4.3.1 Sequence analysis, protein expression and purification, site-directed

mutagenesis

Sequence analysis was carried out using Blastp (www.ncbi.nlm.nih.gov) (5) and

Multialign (http://bioinfo.genopole-toulouse.prd.fr/multalin/) (31). ChuX fused to an

86

N-terminal His6-tag was subcloned into a pET 15b expression vector. Based on the

apo-ChuX structure and sequence conservation analysis, the two single mutants H65L

and H98D, a double mutant (H65L/H98D) (termed DM) and a triple mutant

(V69E/L71H/F73S) (termed BetaX) of ChuX were created. The BetaX triple mutant

alters three nonpolar residues shown in the crystal structure to interact across the central

β-sheets in ChuX homo-dimers, changing them to polar ones in an attempt to disrupt

homo-dimerization. Using the QuickChange® site-directed mutagenesis kit (Stratagene,

CA), wild-type ChuX was mutated at these three positions using the following sense

primers. Base pair changes have been underlined. H65L: sense

5'-CGTCACCACGTTAGTACTTACTGCCGATGTAATCC-3', H98D: sense

5'-GCACGGCATGTCCGGGGATATCAAAGCAGAAAACTG-3', and BetaX:

sense 5'-GTACATACTGCCGATGAAATCCACGAATCTAGCGGCGAACTGCCTTC

GGG-3', Antisense primers were reverse complements to sense primers in each case.

Briefly, polymerase chain reaction conditions and cycling parameters recommended by

Stratagene (123) were followed using PfuTurbo polymerase for thirty cycles between

30 sec 95°C melt, 60 sec 55°C anneal, and 9 min 68°C extend. Methylated (non-PCR

amplified) DNA was then digested with the restriction enzyme Dpn1 for 1 h and the

digested DNA was heat shock transformed into MC1061 cells. Transformed cells were

subsequently plated onto Lauria Broth (LB)-agar plates supplemented with 100 µg/ml

ampicillin. Single colonies were grown overnight at 37°C in 5 ml Terrific Broth (Bioshop,

Burlington, CAN) supplemented with ampicillin. Plasmid DNA purified by means of a

QIAprep Spin Miniprep Kit (Qiagen, Mississauga, CAN) was transformed into

BL21(DE3), and the desired mutations confirmed by sequence analysis. 1 L cultures of

87

BL21(DE3) cells carrying plasmids for the respective recombinant protein were grown at

37°C in Terrific Broth-ampicillin. Protein expression was induced at a culture OD600

between 0.6-1.0 using 0.4 mM isopropyl-1-thio-β-D-galactopyranoside and grown for 5 h.

ChuX protein substituted with seleno-methionine was expressed in the metA- E. coli

strain DL41 in LeMaster medium supplemented with ampicillin (51). Over-expression of

ChuX resulted in cells exhibiting a slight yellow colour compared to H65L, H98D, DM,

or BetaX cultures. ChuX and mutants were all purified via nickel nitrilotriacetate agarose

batch purification in phosphate buffer (pH 8.0) and dialyzed overnight in 30 mM Tris, pH

8.0. Proteins were then purified further via anion exchange, using a 6 ml Resource Q

column on an ATKA Explorer FPLC in 30 mM Tris, pH 8.0 with 30 mM Tris, 500 mM

NaCl, pH 8.0 as the elution buffer. Purification was monitored by SDS PAGE, and each

protein was estimated to be >95% pure prior to concentrated using an Amicon Ultra

15 kDa cutoff centrifugal filtration device (Fisher, Mississauga, CAN). Protein

concentrations were determined by BioRad protein assay (Biorad, Mississauga, CAN)

using Bovine Globulin G as a protein standard and used immediately for crystallization.

ChuX, H65L, H98D, and DM all expressed at high levels (>20 mg/L culture depending

on batch) and were soluble in Tris (pH 8.0) or phosphate (pH 7.0) buffers following

affinity purification. BetaX expression levels (2.4 mg/L culture) were much lower than

those for native ChuX or the other mutants. ChuX used in reconstitution experiments was

stored in liquid nitrogen, the concentrations of which were equalized prior to spectral

analysis to ensure that spectral comparisons were the result of differences in heme

coordination and not due to small changes in relative protein concentrations.

88

4.3.2 Crystallization of ChuX

ChuX crystals were obtained using the hanging drop vapor diffusion method,

using freshly prepared protein and reagents. Crystallization was achieved by mixing

equal parts 40 mg/ml ChuX with reservoir solution consisting of 1.50 M ammonium

sulfate, 50 mM Hepes, pH 7.5, 2% (v/v) PEG 400. Large rectangular crystals (3.0 x 1.5 x

0.75 mm) appeared after 2-4 d. For cryoprotection, crystals were looped through

Paratone-N oil, and flash-cooled in the N2 cold stream at 100K.

4.3.3 Data collection and structure determination

ChuX crystals were determined to belong to the space group P41 with four

molecules in the asymmetric unit. Native and single-wavelength anomalous dispersion

(SAD) data sets were collected, respectively, at beamlines X8C and X25, Brookhaven

National Laboratory. SeMet SAD data were collected at a peak wavelength determined

by fluorescence scan to be 0.97950 Å, for 360° with 1.0° oscillations. The native data

were collected for 180° in total, using 1.0° oscillations. Both datasets were processed

with HKL2000 (102). Heavy atom positions were initially determined using HySS (64,

168) refined, and phases determined with SOLVE (144). Density modification, solvent

flattening, and automatic chain tracing were performed using RESOLVE (144). The

remainder of the model was built by manual fitting using XtalView/Xfit (84). This

89

structure was refined using native data with the program CNS (19). A structural

homology search using the SCOP database (97) was performed with the program SSM

(77) and protein structures were superimposed using the CCP4 program LSQKAB

(Superpose) (2, 67, 77, 78). Modeling of heme binding to ChuX was done using

Autodock 3.0 (93) and crystal contacts were evaluated with CryCo (42).

4.3.4 Heme reconstitution

ChuX-heme was prepared by dissolving heme (ferriprotoporphyrin IX chloride)

(Sigma, Oakville, CAN) in 0.5% (v/v) ethanolamine for 30 min, and transferring the

solution into 50 mM NaH2PO4. The pH was adjusted to 7.0 just prior to incubation with

ChuX in all reconstitution experiments. Small volumes of the heme mixture were then

added until a final ratio of 2:1 heme to ChuX was reached. The ChuX-heme mixture was

then incubated for 15 min, centrifuged at 16,000 rpm for 15 min at 4oC in a JA-20

Beckman Coulter rotor (Beckman Coulter, Mississauga, CAN), and purified using a

Sepharose-G200 column on an ATKA Explorer FPLC in 50 mM NaH2PO4. The

spectrum of concentrated, ChuX-heme purified by size exclusion was then recorded.

ChuX binding kinetics were determined by the addition of 15 µM ChuX to

various amounts of heme (0-30 µM), followed by a 15 min incubation at room

temperature, and absorbance measured at 404 nm. Spectral characterization of ChuX

mutants was achieved by adding 5 µM heme to a molar equivalent of protein. In heme

transfer experiments, protein heme complexes of ChuX or ChuS were setup as in the

mutant spectral characterization experiments. To a set of ChuX-heme mixtures an

90

equivalent amount of ChuS was added, and 50 µM of ascorbic acid was added to each.

Changes in absorption at 408 nm, the peak wavelength for ChuS-heme, were then

monitored over time. All absorbance data were collected in duplicate using a microplate

spectrophotometer (Bio-Tek Instruments, Inc. NY, USA.) for spectra between 300 and

700 nm at 1 nm intervals, 250 µl volumes, and the mean values plotted.

4.3.5 Growth of ChuX and DM in heme

Quadruplicates of BL21(DE3) cells were freshly transformed with pET15b

plasmids that contained ChuX, DM, or C1528 (a 12.2 kDa hypothetical protein from

E. coli CFTO73 unrelated to heme processing). Individual colonies were selected and

grown overnight in 100 ml LB supplemented with 100 µg/ml ampicillin. Cultures were

then diluted in 50 ml pre-warmed LB ampicillin to an OD600 reading of 0.05 and grown

for 20 min at 37oC with aeration. Protein expression was induced using 0.2 mM

isopropyl-1-thio-β-D-galactopyranoside, and cultures were grown for another 20 min.

Sterile filtered heme was then added to bring the solution to a concentration of 2 µM and

the OD650 readings were then measured over time. This wavelength was selected to

minimize the amount of additional absorbance detected due to heme coordination.

4.4 Results

4.4.1 Sequence Analysis

91

NCBI Blastp sequence analysis suggested that ChuX belongs to a family of

conserved putative heme-iron utilization proteins with the conserved domain DUF1008

(143). ChuX has high sequence similarity with proteins from human pathogens such as

S. dysenteriae (ShuX) and various Yersinia strains (COG3721), as well as

Photobacterium damselae (HutX), and Listonella anguillarum (HuvX). ChuX

homologues also appear in photoactive bacteria such as Halorhodospira halophila and

Rhodopseudomonas palustris which use either iron or the products of heme degradation

in the production of light harvesting photoreceptors (47). Sequence alignment of ChuX

and its homologues revealed many conserved residues (Figure 4.1), especially in the C-

terminal half. Notably residues H65, H98, and the segment of residues K134 to L145 are

highly conserved across ChuX homologues.

4.4.2 ChuX protein fold and structural analysis

The final structure of ChuX contains four molecules in the asymmetric unit, with

two sets of monomers interacting to form two homodimers (Figure 4.2). Due to the

absence of clear electron density for a beta turn, a gap in the structure exists in molecule

B (residues 90-93). The center of each dimer is composed of two sets of nine antiparallel

β-sheets, which bow outward to form an overall saddle motif. Each set of β-sheets are

made up of six strands contributed by one of the ChuX monomers and three others

contributed via a domain swap from the partnering ChuX molecule. Two α-helices flank

the small gap that is formed as a result of the bowing β-core, and the N-terminal 36

residues of each ChuX monomer flank each end of the central β-core in an α-loop-α-

92

loop-α motif. Each monomer couple interacts across equivalent polar residues from the

core sets of β-sheets, forming two pairs of closely associated dimers. Molecular weight

evaluation by gel filtration and dynamic light scattering suggests that ChuX forms a

homodimer in solution (data not shown). Ramachandran analysis of the final ChuX

structure shows 488 residues (88.1%) within the most favored region, 55 (9.9%) in

additional allowed, 10 (1.8%) in generously allowed, and only 1 residue (0.2%) in the

disallowed region (K11 from molecule C). Crystallographic statistics for diffraction data

and final structural refinement are presented in Table 4.1.

93

1 10 20 30 40 50 60 65

70 80 90 98 110 120 130 140 150 160

1 10 20 30 40 50 60 65

70 80 90 98 110 120 130 140 150 160

Figure 4.1 Sequence alignment of ChuX with homologues from various Gram-negative bacteria. Highly conserved residues are highlighted in red, residues with sequence similarity are presented in blue. Numbering corresponds to ChuX. Conserved histidines across the organisms examined at positions 65 and 98 are underlined.

94

Table 4.1. Diffraction data and refinement statistics for ChuX

Data Collection Native Se-Met

Space group P41 P41

Cell Dimensions a and b (Å) 76.56 76.83 c (Å) 140.86 141.64 α, β, and γ (°) 90 90 No. of molecules in ASU 4 4 Wavelength (Å) 1.1000 0.9795 Resolution range (Å) 50 – 2.05 50 – 2.20

Total reflections 359,110a 529,767a

Unique reflections 50,833a 82,255b

Completeness (%) 100.0 (100.0)# 98.6 (94.7)#

Rsym(I) (%) 4.3 4.0

I/σI 28.3 (1.7) # 45.3 (6.0) #

Refinement Statistics Resolution Range (Å) 50 – 2.05

Rwork, (%) 20.5

Rfree, (%) 27.4

No. reflections total/ Rfree 47915/ 3137

No. atoms, protein / solvent (H2O)/ total 4833/ 360/ 5193

B-factors (Å2) protein/ solvent (H2O)/ total 49.3/ 53.8/ 49.6

RMS, bond length (Å)/ bond angle(o) 0.020/1.87 #Values in parentheses are for the outermost shells (Native: 2.12–2.05 Å, SeMet: 2.27-2.20 Å) aNumber of reflections following merging bNumber of reflections without merging of Bijovet pairs

95

C

A

B

C

D

C

C

NN

N

C

A

B

C

D

C

C

NN

N

Figure 4.2 Ribbon diagram of the four ChuX molecules in the crystallographic asymmetric unit. Each of the four ChuX monomers that make up two pairs of ChuX homo-dimers are depicted with different colours. ChuX homo-dimerization occurs via interactions between equivalent residues of red (A) and blue (B), yellow (C) and green (D) monomers, respectively. Interactions between ChuX monomers occur across the central sets of β-sheets, involving a domain swap between associated ChuX couples. The gap in the structure is shown with a dotted line.

96

Structural comparison revealed that the ChuX monomer shares structural

similarity with the unpublished hypothetical protein AGR_C_4470 (2HQV, 1.3 Å) from

Agrobacterium tumefaciens and overall topology with the E. coli metal chelating enzyme

glyoxalase I (57) (1FA8, 3.8 Å). Furthermore, an excellent structural alignment exists

between the ChuX dimer and the monomeric structures of two other members of the

heme utilization operon, ChuS and HemS. ChuX aligns with these protein structures with

rmsd values between 2.4 and 2.8 Å, depending on which ChuX dimer is used and

whether the apo- or holoenzyme structure of ChuS or HemS is used (Figure 4.3A). This

structural similarity was surprising given that ChuX shares less than 18.3% sequence

identity with either the N- or C-terminal sequences of ChuS or HemS. Within each ChuX

dimer, a pair of α-helices together with several segments of the β-core, delineate two sets

of broad clefts. These clefts are equivalent to the heme binding positions demonstrated

for ChuS (137) and HemS (120) (Figure 4.3B).

In both ChuS and HemS the histidine side chain responsible for heme

coordination originates from one of the flanking α-helices. An equivalent residue is,

however, absent in the structure of ChuX. Instead, the two conserved histidine residues,

H65 and H98, originating from elements of the central β-core, are proximal to these

structurally conserved heme binding clefts (Figure 4.3B). These histidines are contributed

from opposite ChuX monomers, suggesting that domain swapping is involved in heme

coordination similar to that characterized for HasA (35). The first three residues of the

conserved sequence K134 to L145 reside on the medial of the three β-sheets contributed

in the domain swap to the β-sheet core and may therefore be important for dimerization.

This region of ChuX is also at the base of the conserved heme binding cleft, with K134

97

extending to only 3.76 Å from H65, potentially contributing to the stabilization of the

imidazole group for iron coordination. The other residues of this conserved segment

reside on a loop that delineates one edge of the heme binding cleft and may therefore be

involved in guiding heme insertion, or stabilizing the heme moiety once it is coordinated

by ChuX.

98

Figure 4.3 Structural similarity of ChuX with ChuS and putative heme binding cleft. A) Superimposition of ChuX (blue) and ChuS-heme (red) giving an rmsd of 2.8 Å. Despite low sequence similarity, the topology formed by the ChuX dimer is akin to the structural repeat of the ChuS monomer. Notably, the α-helix responsible in ChuS for heme coordination via H198 has shifted away from the heme binding cleft (*). B) Ribbon diagram of the ChuX homo dimer oriented to highlight the differences between the open cleft (top) and closed cleft (bottom) conformations. The putative heme binding clefts are emphasized with arrows. The conserved histidines shown to be responsible for heme coordination through mutagenesis and subsequent spectral analysis are highlighted (H65 in green, H98 in purple). The α-carbon positions of the central set of β-sheets used to generate the beta mutant form of ChuX in the inhibited dimer form are highlighted in green.

99

4.4.3 The ChuX structure may represent an open and closed conformation

Structural alignment of the two homodimers in the ChuX asymmetric unit

revealed an interesting difference in that one of the four putative heme binding clefts is

significantly more open compared to the other three (Figure 4.3B). In the three closed

cleft conformations, the gap between α-carbons of H84 and M115 are 13.3, 15.3 and

15.5 Å across, whereas the equivalent gap distance of the open cleft is 23.3 Å. This

difference is largely due to a 3.2 Å shift of the α-helix delineated by residues 16 to 26

that define the lateral edge of the cleft. In ChuX, this helix is structurally equivalent to the

helix responsible for claming heme into place in ChuS and HemS, and which shifts upon

heme binding by as much as 3.8 and 4.0 Å, respectively (120, 137). Crystal contact

analysis (42) revealed relatively few contacts among symmetry-related molecules. Those

that were observed interacted in a similar fashion with either ChuX homodimer,

suggesting that the structural difference was not due to crystal packing.

The overall effect of the gap differential is that H65 is effectively buried at the base

of the putative heme binding cleft in the closed cleft conformations, while H98 remains

oriented into the cleft space. When heme was modeled into each of the four clefts, the

docking positions were equivalent to those characterized for ChuS and HemS, with the

central iron atom proximal to an axial H98 in each of the closed conformations, but was

axial to H65 in the open cleft example. In either model, the docking energy predicted for

heme binding by ChuX were favorable (93), with corresponding mean values for ∆Gpred

of -9.9 and -11.4 kcal/mol for the open and closed clefts respectively. Furthermore, this

docking potential could reflect how the difference in dimer juxtapositions between the

100

open and closed cleft conformations of ChuX could help mediate heme uptake in that

H98 is important for heme coordination in the open cleft conformations while H65 is

similarly important in closed conformations. Other residues located within the heme

binding cleft which could contribute to heme stabilization include T17, E19, T61, L63,

I70, E72, F114, K134, and F136. All of which, except T17, are conserved across

sequence homologues of ChuX.

4.4.4 ChuX-heme association

Heme proteins have characteristic visible spectra that convey information about

the structure and coordination state of their heme binding sites. The spectra of ChuX

resemble these of other heme proteins in that a Soret maximum is evident at 404 nm

(Figure 4.4A) depending on buffer pH, and accompanied by a smaller peak for the

α-band at 620 nm. The spectra of ChuX-heme is consistent with the formation of a ferric

hexacoordinate histidine-imidazole complex (169). Monitoring the appearance of this

peak during heme reconstitution to ChuX suggests that the association occurs rapidly,

reaching its maximum within five minutes. Furthermore, differential spectroscopy of

heme versus ChuX-heme suggests that the association is 1:1, implying that the ChuX

homodimer binds two molecules of heme (Figure 4.4B).

Spectral analysis of the ChuX mutants indicate that H65L and H98D both affect

heme binding as the Soret peak for each has shifted to 414 nm, with the absorbance

intensity at 360 nm decreased, while the shoulders in the β-band region intensified

compared to those of ChuX (Figure 4.5). The spectrum of the BetaX mutant is similar to

101

ChuX with the exception of a decrease in the magnitude of the Soret peak, suggesting a

reduction in heme coordination. Notably, while the independent mutations of the

conserved histidines did not significantly affect heme coordination by ChuX, the

spectrum of DM ChuX reflects a reduction in heme coordination due to the elimination of

a clear Soret peak, as well as a reduction in absorption of the α- and β-bands.

102

y = 0.6977x

y = 0.3823x

y = 0.3808x

0

0.2

0.4

0.6

0.8

1

1.2

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2

Ratio of Heme to ChuX

Ab

so

rban

ce 4

04n

m

ChuX plus Heme <1:1

Heme Standards

ChuX plus Heme >1:1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

320 330 340 350 360 370 380 390 400 410 420 430 440

Wavelength (nm)

Ab

so

rban

ce

ChuX - Heme

Heme

A B

y = 0.6977x

y = 0.3823x

y = 0.3808x

0

0.2

0.4

0.6

0.8

1

1.2

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2

Ratio of Heme to ChuX

Ab

so

rban

ce 4

04n

m

ChuX plus Heme <1:1

Heme Standards

ChuX plus Heme >1:1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

320 330 340 350 360 370 380 390 400 410 420 430 440

Wavelength (nm)

Ab

so

rban

ce

ChuX - Heme

Heme

y = 0.6977x

y = 0.3823x

y = 0.3808x

0

0.2

0.4

0.6

0.8

1

1.2

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2

Ratio of Heme to ChuX

Ab

so

rban

ce 4

04n

m

ChuX plus Heme <1:1

Heme Standards

ChuX plus Heme >1:1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

320 330 340 350 360 370 380 390 400 410 420 430 440

Wavelength (nm)

Ab

so

rban

ce

ChuX - Heme

Heme

A B

Figure 4.4 Spectral analysis of ChuX reconstituted with heme. A) A clear Soret peak is visible at 404 nm for reconstituted ChuX-heme purified by size exclusion chromatography, suggesting that the ChuS-heme complex formed is ferric hexacoordinate histidine-imidazole in the high-spin state at neutral pH (52). B) Differential spectroscopy of heme versus ChuX-heme at 404 nm. A transition in the slope of the line of fit for ChuX-heme inflection point is evident for ChuX to heme at a ratio of 1:1. At higher heme concentrations the slope of the line changes to one similar to that observed for free heme, suggesting that ChuX has become saturated.

103

4x4x

Figure 4.5 Spectra of ChuX and mutants reconstituted with heme. The region between 500 and 700 nm has been amplified four-fold to highlight peak differences. The spectrum of ChuX-heme resembles that of other heme proteins in that a Soret maximum is evident at 404 and an α-band at 620 nm, consistent with the formation of a ferric hexacoordinate histidine-imidazole complex (169). Mutations H65L and H98D both affect heme binding, evident in the shift in the wavelength of the Soret peak to 414 nm. The spectrum of the BetaX mutant has a decrease in the magnitude of the Soret peak, suggesting reduced heme coordination. Notably, while the H65L and H98D independent mutations did not significantly affect heme coordination by ChuX, the spectrum of DM ChuX reflects a reduction in heme coordination due to the elimination of a clear Soret peak, as well as a reduction in absorption of the α- and β-bands.

104

4.4.5 ChuX heme binding affinity

Based on absorbance at 404 nm during ChuX-heme reconstitution, the estimated

Kd is 1.99 ± 0.02 µM (n = 2) as calculated from the following expression:

∆A = (εb – εf)0.5[Po + Kd + Ho – ({Po + Ho+Kd}2 -4·Po·Ho)

1/2]

where P0 is the amount of protein, H0 is the concentration of heme, and (εb – εf) is the

difference in the Soret absorption spectra between the bound and free states. The

estimated Kd for ChuX-heme complex formation suggests a weaker heme association for

ChuX than that characterized for other heme binding proteins such as the mammalian

hHO-1 (0.84 ± 0.2 µM) (159, 161), HasA of 5.0 ± 3.1 nM (85), and ChuS (1.0 ± 0.3 µM)

(139), but tighter than that reported for HmuO (2.5 ± 1.0 µM) (15), IsdG (5.0 ± 1.5 µM)

and IsdI (3.5 ± 1.4 µM) (125).

4.4.6 ChuX protection from heme toxicity

Heme is a potentially toxic molecule due to its ability to disrupt normal redox

processes and catalyze free radical formation (54, 101). As a result, uncommitted heme

concentrations are normally maintained below 10-9 M (141). When challenged with low

concentrations of heme (2 µM), LB-ampicillin cultures expressing ChuX were protected

against heme toxicity and grew at an enhanced rate relative to those expressing the DM

ChuX mutant or the C1528 control (Figure 4.6A).

105

When an equal amount of ChuS was added to ChuX reconstituted with heme, a

small spectral shift in the Soret peak was evident from 404 to 408 nm (Figure 4.6B). This

Soret peak shift indicates that heme was being released from ChuX and taken up by ChuS.

Furthermore, when ascorbic acid, an electron donor for heme degradation was added, a

decrease in the spectra for heme coordination was observed compared to either ChuX-

heme or free heme, indicating that heme was degraded as previously characterized

(Figure 4.6B) (139). However, because the ChuX homodimer is similar in shape, size,

and molecular weight to the ChuS monomer, as well as having similar Soret peaks, it was

difficult to evaluate the specific amount of heme transfer (14, 80).

106

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 30 60 90 120 150 180 210 240

Time (minutes)

Rela

tive A

bso

rban

ce O

D 6

50n

m

ChuX-Uninduced

ChuX-No Heme

ChuX

DM

C1528

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0 10 20 30 40 50 60 70 80 90 100 110 120 130

Time (minutes)

Ab

so

rban

ce

408

nm

Buffer

Heme

ChuX only

ChuS only

ChuX and ChuS

BA

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 30 60 90 120 150 180 210 240

Time (minutes)

Rela

tive A

bso

rban

ce O

D 6

50n

m

ChuX-Uninduced

ChuX-No Heme

ChuX

DM

C1528

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0 10 20 30 40 50 60 70 80 90 100 110 120 130

Time (minutes)

Ab

so

rban

ce

408

nm

Buffer

Heme

ChuX only

ChuS only

ChuX and ChuS

BA

Figure 4.6 ChuX protection against heme toxicity and heme transfer by ChuX. A) Progression of BL21(DE3) cells expressing ChuX, DM ChuX, and C1528. ChuX, DM, C1528 were each challenged with 2 µM heme. Cells expressing ChuX were protected relative to those expressing the ChuX double mutant or C1528. B) Transfer of heme from ChuX to the heme degrading enzyme ChuS. An initial increase in the peak for ChuX and ChuS shows that ChuS is becoming reconstituted with heme. Following ascorbic acid addition ChuS catalyzed degradation of heme was then tracked by monitoring the decrease in the Soret peak at 408 nm.

107

4.5 Discussion

Here we present the first crystal structure of ChuX, a member of the heme

utilization operon from pathogenic E. coli O157:H7, and propose that ChuX may help to

regulate heme utilization through structure-mediated heme binding and release. Heme is a

hydrophobic molecule capable of freely diffusing into membranes where it can alter the

bilayer structure and disrupt cell integrity (119). Furthermore, heme is a potent

prooxidative, promoting the light-dependent formation of reactive oxygen species that

can cause non-enzymatic redox reactions (54). Organisms must therefore tightly regulate

heme acquisition and metabolism by either degrading heme or storing it for later use in

various processes (155). Largely through gene knockouts and phenotypic characterization

the transport machinery responsible for heme uptake in E. coli O157:H7 has been

elucidated (127, 128). However, the regulation of internalized heme metabolism has only

recently received attention. Furthermore, in order to avoid iron induced toxicity, heme

degradation and iron storage must also be balanced. This has been shown to occur

through the regulation of iron scavenging elements by transcriptional regulator proteins

such as Fur (168), DtxR (116), and FecI (40). Alternatively, iron toxicity can also be

avoided by reducing the amount of heme degraded within the cytoplasm through heme

storage or short term sequestering until the organism is experiencing iron limiting

conditions (55). A heme storage role has recently been suggested for the gene products

ght from Neisseria meningitides (109) and hmsT from Yersinia pestis (66), as well as the

protein HutZ from Vibrio cholerae (166). Our results suggest a cytoprotective role for

ChuX based on bacterial cultures challenged with low levels of heme relative to ChuX

108

mutants or non-heme processing controls. Furthermore, we demonstrated that ChuX is

capable of transferring heme to other heme utilizing proteins such as ChuS, although we

cannot discount at this time that heme transfer may have occurred through passive

diffusion.

Structural insights provided from vibrational spectroscopy (EPR and Raman),

NMR, and X-ray crystallography have indicated that heme proteins often coordinate the

iron center of the heme group between a histidine side chain and a variety of other

possible residues including tyrosine (8), methionine (22), cysteine (39), proline (81),

histidine (103), arginine (137); some heme proteins can even coordinate their heme iron

centers via a sole histidine sidechain without additional sidechain assistance (52, 121).

Other proximal side chains and polypeptide backbones have been shown to interact with

the propionic and vinyl parts of the heme group, contributing to the stabilization of the

heme moiety within the protein environment which ultimately confers heme protein

function. The structure of the dimeric oxygen sensing kinase FixL resembles an open pita

or clam which coordinates heme in the center of the open mouth between a series of

antiparallel β-strands arranged in a β-barrel, and an α-helix (121). This mode of heme

coordination is also similar to the topology utilized by ChuS and HemS in which the

central iron atom of heme is sandwiched between a histidine side chain and an arginine

side chain via two water molecules (120, 137). The structural similarity shared between

ChuX and ChuS/HemS suggests a similar mechanism of heme coordination. However,

the axial histidine side chain responsible for heme coordination in ChuS/HemS was

notably absent when the structure of ChuX was superimposed with either ChuS or HemS.

109

Spectral analysis of ChuX reconstituted with heme suggested that a histidine

residue was involved in ChuX heme coordination and that heme coordination occurred in

a 1:1 fashion. When the ChuX sequence was aligned with homologues from various

Gram-negative bacteria, the conserved residues H65 and H98 were identified as potential

heme coordinating residues. In the crystal structure of ChuX, both of these residues are

within the putative ChuX heme binding cleft, but single residue mutants that changed

these histidine residues did not abolish heme binding. However, spectral analysis of the

ChuX double H65/H98 mutants demonstrated that (i) either both residues may be

involved in heme coordination; (ii) that alternate histidine residues may be involved; or

(iii) that heme binding sites within ChuX may be overlapping. This variability in protein-

heme coordination has been detailed for the extracellular hemophore HasA from Serratia

marcescens, which is also a domain swapping dimer (85). Lastly, the spectra of heme

reconstituted with the BetaX mutant, suggested that under conditions where ChuX

dimerization was impaired, heme coordination was compromised. Collectively, based on

the fact that the putative heme binding sites of ChuX are contributed by two ChuX

molecules, the identification of conserved residues and sequence segments through

sequence homology comparison, and the evaluation of mutant ChuX variants, we

conclude that ChuX dimerization is absolutely obligatory for coordinating heme

molecules and that the two conserved residues (H65 and H98) located within the

conserved heme binding cleft are both essential for heme coordination.

Two different ChuX homodimer conformations were observed within the ChuX

asymmetric unit. One ChuX dimer assumed a structure with two similarly spaced gaps

across the conserved heme binding cleft, while the second dimer had one gap similar in

110

size to those found in the first dimer conformation as well as a noticeably broader one.

We propose that this structural plasticity may reflect an open, ligand accepting form and

a closed, ligand bound form of ChuX. Similar protein flexibility has been shown to be

important for hemoglobin to reversibly regulate heme coordination (53). We found

ChuX-heme coordination to be conformation-dependent: when heme was docked to the

closed conformation of ChuX, heme coordination was observed to cluster around H65

with favorable energy predictions, whereas when heme was docked to the open

conformation of ChuX, heme coordination was found to cluster around H98. Furthermore,

many conserved residues in the ChuX sequence are clustered within and proximal to the

heme binding cleft. Together, the variation of ChuX juxtapositions in homodimers as

well as the docking of heme to ChuX at the two conserved histidine residues H65 and

H98 located within the conserved heme binding pocket, suggest a potential structural

mechanism for heme binding and sequestering.

In conclusion, based on the crystal structure of ChuX we have proposed that

dimer juxtaposition may serve as a structural mechanism in regulating heme binding and

that ChuX may function as a cytoplasmic heme shuttling protein, thereby regulating

heme processing through sequestering or storage.

111

4.6 Footnotes

Thank you to Dr. Michael Nesheim for his expertise in designing and interpreting

heme binding studies and John Wagner for the generation of the H98D mutant. We thank

Dr. Mirek Cygler and Dr. Allan Matte for their support. We are also grateful to

Brookhaven National Laboratory beamline staff Leonid Flaks and Martin McMillan

(X8C) as well as Anand Saxena (X12C), where the SAD and native data for ChuX were

collected respectively. This work has been supported by Bracken E.G. Bauman and R.

Samuel McLaughlin Queen’s Graduate Fellowships as well as CIHR and OGS. ZJ is a

Canada Research Chair in Structural Biology. Coordinates and structure factors for ChuX

are deposited in the RCSB Protein Data Bank (www.rcsb.org) under accession code

2OVI.

112

Chapter 5

General discussion and summary

113

5.1 Gaining functional information from protein structures

With the completion of sequencing of many microbial and eukaryotic genomes,

the essential next steps toward understanding the molecular mechanisms and interactions

that ultimately facilitate cellular function. To do this, one must first identify the genomic

complement of ORFs, and then to determine the structures and functions of the proteins

encoded. Historically, the first stage of ORF characterization has involved primary

sequence homology searches to infer functionality from previously annotated proteins or

segments of proteins. However, it has become obvious that current sequence homology-

based functional annotation alone is not sufficient for complete genome functional

coverage. During translation, protein sequences fold into conformations that serve as

scaffolds for residues whose interactions establish the microenvironments for directing

localization as well as protein-protein and protein-ligand interfaces that ultimately

determine a protein’s function. Consequently, the global interactions present in the final

fold of the protein can be better conserved than its sequence alone. Protein structure

comparisons may therefore have the potential to reveal evolutionarily distant

relationships that may be used for functional annotation through the identification of

active sites and integral functional residues, suggesting potential ligands, and may also

contribute information as to possible mechanisms of action.

Whereas traditional structural studies have often involved targeting proteins of

known function in order to completely understand their mechanism of action, our efforts

simultaneously combined structure determination and functional characterization of two

proteins, ChuS and ChuX, which had little functional information attributed to them at

the initiation of our investigation. To successfully establish this concurrent structural

114

determination and functional annotation methodology, a set of criteria must be

established to maximize the allocation of resources. The most important of these is target

selection; to highlight proteins from organisms of medical or functional interest that may

be functionally significant in cell operation. Initially, the evaluation of putative genes

relies on sequence comparisons and homology searches as well as investigations of the

partial or complete functions attributed to genes organized into operons. This process was

facilitated for ChuS and ChuX through the use of sequence analysis tools such as the

Clusters of Orthologous Groups (142). Protein selection may also enrich for proteins

present in pathogenic organisms which do not have homologues in more benign strains.

ChuS and ChuX are genes encoded within the heme utilization operon from E. coli

O157:H7 and CFT073, but not in the benign K-12 strain. Depending on project

methodology, protein selection could also encompass well characterized proteins in order

to better understand their mechanism of action or the nature of their interactions with

ligands, inhibitors, and/or binding partners.

Lastly, protein selection must also apply bioinformatics-based analysis to address

the inherent challenges of X-ray crystallography. As a large number of potential targets

may be initially selected, a systematic ranking of targeted protein priority must also be

assigned (17). This may include an evaluation of protein solubility and cellular

localization. Furthermore, protein analysis may consider whether full length proteins,

truncations, or co-expression with other proteins should be implemented. Other important

considerations such as whether inhibitors or supplements to culture media may be

necessary, as well as special considerations into growth conditions such as whether

special expression systems may be better suited for certain proteins of interest. This latter

115

analysis may also be applied retroactively once difficulties are encountered during

various phases of the investigation.

Once the structures of ChuS and ChuX had been determined, in silico functional

prediction proceeded via three structure-based methods simultaneously due to the

inherent strengths and weaknesses attributed to each (114). These levels of structural

analysis were: (i) structure homology and overall topology searches, (ii) putative cleft

identification, and (iii) cleft and surface charge distributions. Together these

computational methods for probing structure have the potential to provide functional

information at phenotypic, cellular and molecular levels (114). Conveniently, many of the

tools used for fold and sequence alignment, structure modeling, model evaluation, and

META servers are implemented as web servers, and are readily available to the public.

In structure homology searches new structures may be compared against a

database of structures with or without functional annotation, or a sample of structures

from the protein data bank. As the number of experimentally determined structures

increases, the application of homology based or comparative modeling will likely become

more frequent in structural evaluation, as it provides speed, accuracy, and detail that other

modeling programs lack (94). Evidence suggests that there are a limited number of

different classifications of protein folds (49). It may therefore be possible to assign

specific fold and function classifications, patterns of phylogenetic occurrence, expression

data, and suggest protein-protein interactions from architecture homology searches (49).

Several current search engines such as CATH, CE, DALI, DEJAVU, MATRAS, PALI,

SCOP, SUPFAM, and VAST provide the opportunity for this type of analysis, and have

the added benefit of increasing their value as more structures have their function

116

annotated (49, 98, 140). Other comparisons of closely related structure families can be

performed using the servers CASP, CRABP, and DDMP, which can suggest substrate

specificity or protein-protein interactions. A recent evaluation of these class of databases

suggested that CE, DALI, MATRAS, and VAST showed the best performance, but

understandably none of the servers achieved a 100% success rate (98). For both ChuX

and ChuS, a DALI search was the most informative, but only when the homologous

structures were subsequently evaluated in greater detail; including sequence and

structural investigation.

The second method for predicting function would be to use software such as

pvSOAR, DOCK, GRASP, LigPROF, PROFUNC, PROSITE, SLIDE, SUMO and

SURFNET that compare highlighted regions against catalogues of chemical structures

and functional mechanisms (49). Akin to randomly fitting molecular pieces into a protein

jigsaw puzzle, this method involves docking libraries of molecules to regions on the

“virtual screening” structure. When the program DOCK is applied, factors such as

rigidity, energy, contact, and solvation scorings are considered to determine how well a

molecule fits sites on the structure. For both ChuS and ChuX, pvSOAR correctly

predicted that both were heme binding proteins, and even suggested putative residues

responsible for heme coordination. However, in both cases the correct pvSOAR

predictions were not the top results, or even in the top 10. This experience highlights that

where some functional information has already been acquired, it can facilitate the

processing of the vast amount of information output during in silico characterization. An

advantage of PROFUNC is that it combines multiple analysis of a protein structure in

parallel including: a prediction of cellular localization, sequence and structure alignments,

117

and cleft identification. PROFUNC therefore provides the advantage of a complete

summary of sequence and structural based analysis in a single output, which facilitates

comparisons and may highlight more distant relationships that would otherwise be

overlooked.

In the final technique, a system such as principal component analysis would be

used to compare subtle surface and cleft characteristic differences across orthologous

groups of structures (24), such as our analysis of apo ChuS with heme-ChuS, and ChuX

with ChuS and HemS. Key regions of conformational changes were highlighted, in

addition to consensus pockets and surface regions for interactions. These analyses were

particularly informative in suggesting the position of heme coordination in ChuX and

potentially suggested a mechanism of heme uptake. More broadly applied, this method

could provide information concerning substrate specificity and inhibitors that could be

tested in binding studies, applied in co-crystallization experiments, or in catalytic

investigations with the potential to postulate molecular mechanisms for action (24).

However, this method is often more computationally expensive than the previous two;

requiring time and computing power, but has the potential to yield a more complete

characterization of a class of structures with careful analysis.

With the initiation of structural genomics projects, many examples of structure

based functional annotation have emerged. The crystal structure of a previously

uncharacterized H. influenzae protein YbaK was reported and through the

implementation of combination of sequence and structure homology searches to

demonstrate that YbaK was a tRNA synthetase involved in nucleotide/oligonucleotide

binding (171). Alternatively, structures of previously uncharacterized proteins may

118

provide insight into their function through the shape and location of the active site

pockets or through the presence of unexpected cofactors in the solved structures. The

structure of MJ0577 was solved with an ATP molecule bound and shared the

corresponding binding motif with a family of ATP binding structures (170). These

attributes strongly suggested that the protein was an ATPase or ATP-mediated molecular

switch, which was later confirmed through biochemical experimentation. Furthermore,

obtaining protein structures that share sequence homology to families of proteins with

annotated function may complete our understanding of their molecular mechanisms,

highlight key regions important for function and substrate specificity, or reveal novel

mechanisms of action. An example of this last class of structures was demonstrated for

the G-domain of Rnd3/RhoE, a protein member within a subfamily of Rho GTPases (43).

The structure of Rnd3/RhoE shared a similar fold to other Rho GTPases but revealed

striking differences with respect to charge distribution, GTPase center, ligand binding

sites, and interfaces for known regulators and effectors. Together these examples, as well

as the plethora of positive cases across a variety of prokaryotes, eukaryotes, and

archaebacteria, demonstrate the feasibility, pitfalls, and potential impact for functional

assignment following structural solution of ORFs and poorly annotated proteins (10).

5.2 ChuS and ChuX: structural and functional implications for heme utilization

Due to their localization within the heme binding operon, and phenotypic

evaluation of gene knockouts, a putative function had been ascribed for ChuS and

suggested for ChuX prior to the initiation of our studies. It was therefore important for

119

our investigation to develop testable hypotheses as to particular mechanistic and in vitro

functions for these two proteins. Primarily, information was gathered based on literature

evaluation and through the incorporation of as much bioinformatics-based information as

possible into the study. In order to process the large amount of computational based data

generated, a combined approach of bioinformatic analysis, experimental observations,

and structural characterization improved our overall analysis. In the case of ChuS,

biochemical clues were revealed prior to a successful structure determination. During

ChuS over-expression, a blue-green pigment developed and persisted in fractions

containing ChuS following affinity purification (139). As ChuS is organized into an

operon responsible for heme uptake and utilization, the formation of this pigment was

suggested to be the result of the production of biliverdin (BV) or bilirubin (Figure 1.1B).

Furthermore, the colour causing pigment was eliminated following subsequent

purification indicating that the pigment producing agent was not tightly coordinated or

covalently bound by ChuS. These observations suggested that while bilirubin or

biliverdin may be responsible for the observed pigment produced during ChuS

over-expression, neither was the preferred substrate or ligand for ChuS.

Concurrent with efforts to obtain the structure of ChuS, we initiated experiments

to reconstitute ChuS with heme and conducted spectral analysis of the protein. When

hemoproteins coordinate heme, a characteristic spectrum results that is dependent on the

nature of the protein-heme interactions. The spectra of ChuS reconstituted with heme has

a Soret maximum at 408 nm, with a smaller set of peaks at 545 and 580 nm, suggesting

that the ChuS-heme complex formed was ferric hexacoordinate and that a histidine was

responsible for this coordination (26). A comparison of the ChuS-heme spectra with other

120

hemoproteins revealed similarities with HOs (26, 158). HOs use molecular oxygen

together with an electron donor such as cytochrome P450 reductase-NADPH or ascorbic

acid to degrade heme. When either electron donor was added, the spectra of ChuS-heme

changed over time in a similar fashion to that characterized for other HOs, producing BV,

CO, and free iron. Furthermore, we used sensitive gas chromatography to measure CO

production during heme degradation, confirming that ChuS was a HO. In order to

confirm that this change was not due to non-enzymatic coupled oxidation, catalase and

superoxide dismutase were also added prior to the addition of either electron donor and

similar trends in spectral analysis and CO measurement were observed, demonstrating

specific HO activity.

Concurrent with ChuS-heme interaction study, the structure of the apo-ChuS had

initially been solved by our group prior to experimental characterization of ChuS to be a

HO. However, as ChuS represented a novel fold at that time, structural information alone

failed to suggest any immediate functional information. Based on the appearance of a

Soret peak in heme coordination studies, we postulated that an axial histidine side chain

may be responsible for heme coordination. An examination of the structure of ChuS in

light of highly conserved residues from this protein family, revealed four candidate

histidine residues (H73, H87, H198, H277), all in proximity to two broad clefts. These

clefts in ChuS are delineated on opposite sides of the central core of β-pleated sheets,

with the third side of the clefts delineated by the flanking sets of three α-helices (Figure

2.1). The exception was H73 which was proximal to the putative cleft, but was oriented

toward the interior of the structure. A careful examination of the ChuS structure revealed

that ChuS is a structural repeat (Figure 2.2). This would not have been predicted from

121

sequence analysis, as the N- and C-terminal halves of ChuS share low sequence identity

(139). Notably, the structure of ChuS does not share any structural homology to

characterized heme degrading enzymes, which are mainly α-helical and coordinate heme

by sandwiching it between a histidine residue and a glycine, such as observed in hHOs

(structures reviewed in (105)), but has since been determined to share homology with

ChuX, HemS, AGR_C_4470, and glyoxalase I.

In an attempt to identify the key heme coordinating residues, we initiated

mutagenesis of these conserved histidine residues. Simultaneously, we obtained two

isoforms of ChuS-heme co-crystals, which led to structure determination of ChuS in

complex with heme and provided a detailed view of the heme environment. Spectral

analysis and CO detection by gas chromatography of native ChuS together with H73A

and H193N mutants was conducted. Together, the mutagenesis study and the structural

solution of ChuS-heme indicated that H193 was important for axial ligation of heme, and

that R100 further stabilized the heme group from the medial side of the protein via two

water molecules. Our initial hypothesis was that ChuS bound to two equivalents of heme.

Despite that the conserved histidine 277 was in proximity to the second putative heme

binding cleft it originates from the central set of β-sheets, and not the α-helix shown to be

responsible for heme coordination by ChuS. This difference between the two clefts may

explain why ChuS and HemS have been shown through crystal structure analysis to only

bind one equivalent of heme (120, 137). Furthermore, we must amend our evaluation of

the Kd for ChuS-heme to 0.4 ± 0.1 µM from our previously estimated Kd of 1.0 ± 0.3 µM.

In summary, ChuS represented a novel fold; it was the first HO to be identified in

any strain of E. coli, and is therefore an important factor in heme utilization as an iron

122

source. Furthermore, the mode of heme coordination was unique compared to other heme

degrading enzymes which potentially could be related to the increased activity observed

for ChuS relative to other bacterial enzymes.

The axiom of sequence based structure and function prediction is that similar

sequences have similar folds. The corollary of this statement, that dissimilar sequences

will fold into slightly different structures, has been shown to be untrue in the cases of

AGR_C_4470, ChuS (137, 139), ChuX (138), and HemS (120). Despite sharing low

sequence similarity, all four structures have very similar topology. Furthermore, as the

monomeric structures of ChuS and HemS resemble the homodimer formed by ChuX and

AGR_C_4470, this could imply an evolutionary relationship where ChuS has evolved

from a chuX gene duplication event. Absorption spectral analysis of ChuX reconstituted

heme demonstrates that ChuX binds to heme in a 1:1 manner, implying that each

homodimer coordinates two heme molecules, in contrast to ChuS where only one heme

molecule is bound. Due to the structural similarity observed between the two proteins, we

extended the mechanism of ChuS heme binding to ChuX, and suggested the putative

location in the protein responsible for heme coordination. Based on sequence

comparisons of ChuX and its homologues, as well as structural comparisons with ChuS

and HemS, we designed a number of site-directed mutations to probe heme binding sites

as well as the dimer interface. In ChuS a single His mutation abolished the capacity to

coordinate heme, but in ChuX, two mutations (H65L and H98D) were required to

achieve the same result. Further, dimer interface mutations resulted in compromised

heme binding, supporting the notion that varying ChuX juxtapositions observed in the

structure may serve to modulate heme binding and release. We also demonstrated that

123

ChuX can transfer heme to ChuS for degradation although currently we cannot exclude

the possibility that this transfer is occurring through diffusion. Furthermore, we

demonstrated that ChuX expression had a cytoprotective role against heme toxicity. In

conclusion, ChuX appears to act upstream of ChuS and may regulate proper heme uptake

and utilization through ChuX-mediated heme binding and release.

5.3 Inhibition of heme utilization as a potential mode for treatment

HOs catalyze the degradation of heme to BV, CO, and free iron. In bacteria, this

liberated iron can be incorporated into various proteins where it can act as a cofactor in

many reactions that contribute to pathogenesis. Beyond a role in iron metabolism, it has

also been revealed that the two other products of oxidative heme cleavage, BV and CO,

are important signaling molecules in mammalian systems. BV is subsequently reduced by

BV reductase, forming BR, a potent cellular antioxidant (12, 113). While excessive levels

of BR cause toxicity, adequate levels play an important role in cytoprotection form

oxidative stress (69). Furthermore, approximately 85% of the physiologically derived CO

is produced through HO-1 mediated heme degradation (112). CO has been shown to act

as an important signaling molecule that plays a role in modulating vasodilation,

neurotransmission in the central or peripheral nervous systems, immune responses, and

anti-inflammatory, anti-apoptotic, and/or anti-proliferative potentials (7, 36, 75). In

addition, HO-1 activity has been shown to be important for tumor growth, implicating a

role for HO-1 inhibitors as anticancer agents (37).

124

Because of the broad range of effects mediated by the products of heme

degradation, HO-1 specific inhibitors have the potential to modulate the function of many

physiological systems. Traditional HO-1 inhibitors have involved the use of various

metalloporphyrins, where the central iron atom of heme is replaced with various other

metals including chromium, tin, and zinc (7). However, these heme derivatives will also

interact with other important hemoproteins such as nitric oxide synthase and soluble

guanylyl cyclase (76). New classes of HO-1 inhibitors have emerged that are based on

azalanstat and azole (76, 151), as well as targeted delivery of HO-1 inhibitors (65).

Investigation into these class of compounds include the structure and biochemical

properties of a ternary complex of rat HO-1, heme, and an imidazole-dioxolane

compound (2[2-(4-chlorophenyl)ethyl]-2-[(1H-imidazol-1-yl)methyl]-1,3-dioxolane) was

recently described (133). The usefulness of these classes of chemicals in HO inhibition

stems from their similarity to the natural imidazole group of the preferred coordinating

histidine residue. Thereby contributing additional stabilization of the heme moiety by

interacting with the iron atom and shielding the heme group from the oxygen and electron

addition required for heme degradation.

We have shown that in the presence of Sn(IV) mesoporphyrin IX, a competitive

inhibitor of heme binding and degradation (152), the activity of ChuS decreased between

2- to 7-fold in a dose-dependent manner (139) (Figure 2.7). This level of inhibition was

similar to that characterized for HO-1 (152). However, when we added the HO-1 specific

inhibitor (2[2-(4-chlorophenyl)ethyl]-2-[(1H-imidazol-1-yl)methyl]-1,3-dioxolane) to

ChuS reconstituted with heme, we did not detect any inhibition during spectral analysis

(Figure 5.1). The inability of this compound to inhibit the activity of ChuS, is likely due

125

to the structural differences between ChuS and HO-1, and their different modes of heme

coordination and overall stabilization. From a therapeutic perspective these differences

may potentially be exploited to specifically target the mechanism of iron acquisition

mediated by ChuS during E. coli O157:H7 infections, while not disrupting normal HO-1

function and signaling. As the structure of ChuS has been solved in complex with heme,

it may be used for inhibitor design and in silico screening. This could involve parallel

screening of the existing catalogue of azalanstat and azole inhibitors of HO-1, or could

involve the synthesis of unique inhibitors that potentially would interact with ChuS-heme.

126

Figure 5.1 Heme degradation by ChuS in the presence of the hHO-1 specific inhibitor (2[2-(4-chlorophenyl)ethyl]-2-[(1H-imidazol-1-yl)methyl]-1,3-dioxolane). Reaction progression by monitoring the decrease in the Soret peak (408 nm) for ChuS were similar in the presence or absence of the hHO-1 inhibitor.

127

5.4. Conclusion

At the initiation of our experiments, the proteins involved in heme uptake through

the OM, periplasm, and IM had been identified largely through genetic knockouts (91,

127, 128). However, the fate of heme once it reached the cytoplasm was lacking in detail.

It was recognized that heme was a signaling molecule that would simultaneously

stimulate increases in DNA splicing, transcription, mRNA stability, translation, and post

translational modifications (154) and, therefore, the cumulative effect of heme uptake

was effectively a signal for successful colonization of a host (83). While many bacterial

heme degrading enzymes had been identified (130), and it had been hypothesized that a

heme binding or storage protein also existed in E. coli (158, 165), the genes responsible

had not been identified. We have therefore contributed both structural and functional

insight for two proteins involved in cytoplasmic heme processing in pathogenic E. coli

O157:H7, and can extend our findings for ChuS catalyzed degradation of heme for the

production of antioxidants and as a source of iron. Based on conclusions drawn from this

structural and function insight, we can therefore augment the structures and mechanism

of heme uptake in pathogenic E. coli O157:H7 (Figure 5.2). Having demonstrated that

homologues to ChuS and ChuX are present in many Gram-negative bacteria, including

Enterobacter, Erwinia, Shigella, and Yersinia, we can extend our findings for these two

proteins across various organisms with respect to heme utilization. Due to the

implications of heme degradation for iron acquisition, and heme binding as it pertains to

prevention against heme toxicity and proper utilization of heme, our characterization of

these two proteins has contributed to our overall understanding of the pathogenesis of

E. coli O157:H7.

128

Future work will involve characterization of the interactions of the cytoplasmic

proteins ChuS, ChuW, ChuX, and ChuY and their roles in heme processing and

utilization in E. coli O157:H7 or its homologues. Through improved characterization of

the heme utilization proteins and the interplay in mediating heme utilization, we will

likely be able to extend our findings in bacterial systems to fungal, protozoan, and

mammalian heme transports systems, which at this point are poorly understood.

Follow-up investigations of E. coli O157:H7 system may involve tracking heme

internalization (82, 163) as well as a characterization of protein expression profiles

during heme limiting and heme surplus conditions. In order to characterize the protein-

protein interactions involved a synthetic heme derivative may be used where two heme

molecules are covalently attached by a linker. This may be used in a type of tethered pull

down experiment to explore possible protein-protein interactions. Conversely,

understanding the structural interplay and modes of heme coordination in infectious

bacteria may also be exploited to mediate medical treatments through screening of

chemical compounds that may interact with these proteins. As the structures of ChuS and

ChuX revealed unique modes of coordination compared to mammalian hHO-1, structure

based drug design may be viable to specifically inhibit Gram-negative bacteria. At this

time these azole based inhibitors seem to be the most suitable candidates for inhibiting

heme uptake.

129

C

NN

C

23.3Å

15.5Å

H65

H65

H98

H98

C

NN

C

23.3Å

15.5Å

H65

H65

H98

H98

OM

PP

IM

HasA

Hbp

ChuA?

TonB

ExbB? ExbD?ExbB?ExbB? ExbD?

ChuT?

ChuU? ChuV?

ChuW?

ChuX

ChuS

V U Y X W T A S

? ? Function

V U Y X W T A S

? ? Function

Figure 5.2 Modified schematic of the heme assimilation system from Gram-negative bacteria and organization of the heme utilization operon from Escherichia coli O157:H7. Proteins involved in heme transport from pathogenic E. coli O157:H7 are presented in their locations in the outer membrane (OM), periplasm (PP), and inner membrane (IM). Where the structures of proteins are unknown, structures of homologues have been included and names followed by a question mark. Putative promoter elements within the heme utilization operon are depicted with arrows. The structure and function of ChuS, ChuS-heme, and ChuX have been included, while the function of ChuW and ChuY is still unknown at this time.

130

Reference List

1. 1993. Update: multistate outbreak of Escherichia coli O157:H7 infections from hamburgers--western United States, 1992-1993. MMWR Morb. Mortal. Wkly. Rep. 42:258-263.

2. 1994. The CCP4 suite: programs for protein crystallography. Acta Crystallogr. D. Biol. Crystallogr. 50:760-763.

3. 1999. New frontiers in the development of vaccines against enterotoxinogenic (ETEC) and enterohaemorrhagic (EHEC) E. coli infections. Part I. Wkly. Epidemiol. Rec. 74:98-101.

4. Ackers, M. L., B. E. Mahon, E. Leahy, B. Goode, T. Damrow, P. S. Hayes, W. F. Bibb, D. H. Rice, T. J. Barrett, L. Hutwagner, P. M. Griffin, and L. Slutsker. 1998. An outbreak of Escherichia coli O157:H7 infections associated with leaf lettuce consumption. J Infect. Dis. 177:1588-1593.

5. Altschul, S. F., T. L. Madden, A. A. Schaffer, J. Zhang, Z. Zhang, W. Miller, and D. J. Lipman. 1997. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res 25:3389-3402.

6. Andrews, S. C., A. K. Robinson, and F. Rodriguez-Quinones. 2003. Bacterial iron homeostasis. FEMS Microbiol. Rev. 27:215-237.

7. Appleton, S. D., M. L. Chretien, B. E. McLaughlin, H. J. Vreman, D. K. Stevenson, J. F. Brien, K. Nakatsu, D. H. Maurice, and G. S. Marks. 1999. Selective inhibition of heme oxygenase, without inhibition of nitric oxide synthase or soluble guanylyl cyclase, by metalloporphyrins at low concentrations. Drug Metab Dispos. 27:1214-1219.

8. Arnoux, P., R. Haser, N. Izadi, A. Lecroisey, M. Delepierre, C. Wandersman, and M. Czjzek. 1999. The crystal structure of HasA, a hemophore secreted by Serratia marcescens. Nat. Struct. Biol 6:516-520.

9. Balla, G., H. S. Jacob, J. W. Eaton, J. D. Belcher, and G. M. Vercellotti. 1991. Hemin: a possible physiological mediator of low density lipoprotein oxidation and endothelial injury. Arterioscler. Thromb. 11:1700-1711.

10. Banfield, M. J., J. J. Barker, A. C. Perry, and R. L. Brady. 1998. Function from structure? The crystal structure of human phosphatidylethanolamine-binding protein suggests a role in membrane signal transduction. Structure. 6:1245-1254.

11. Barker, P. D., E. P. Nerou, S. M. Freund, and I. M. Fearnley. 1995. Conversion of cytochrome b562 to c-type cytochromes. Biochemistry 34:15191-15203.

131

12. Beri, R. and R. Chandra. 1993. Chemistry and biology of heme. Effect of metal salts, organometals, and metalloporphyrins on heme synthesis and catabolism, with special reference to clinical implications and interactions with cytochrome P-450. Drug Metab Rev. 25:49-152.

13. Besser, R. E., S. M. Lett, J. T. Weber, M. P. Doyle, T. J. Barrett, J. G. Wells, and P. M. Griffin. 1993. An outbreak of diarrhea and hemolytic uremic syndrome from Escherichia coli O157:H7 in fresh-pressed apple cider. JAMA. 269:2217-2220.

14. Bhakta, M. N. and A. Wilks. 2006. The mechanism of heme transfer from the cytoplasmic heme binding protein PhuS to the delta-regioselective heme oxygenase of Pseudomonas aeruginosa. Biochemistry. 45:11642-11649.

15. Bibb, L. A., N. D. King, C. A. Kunkle, and M. P. Schmitt. 2005. Analysis of a heme-dependent signal transduction system in Corynebacterium diphtheriae: deletion of the chrAS genes results in heme sensitivity and diminished heme-dependent activation of the hmuO promoter. Infect. Immun. 73:7406-7412.

16. Bonnett, R. and A. F. McDonagh. 1973. The meso-reactivity of porphyrins and related compounds. VI. Oxidative cleavage of the haem system. The four isomeric biliverdins of the IX series. J Chem Soc [Perkin 1] 9:881-888.

17. Bray, J. E., R. L. Marsden, S. C. Rison, A. Savchenko, A. M. Edwards, J. M. Thornton, and C. A. Orengo. 2004. A practical and robust sequence search strategy for structural genomics target selection. Bioinformatics. 20:2288-2295.

18. Brickman, T. J., C. K. Vanderpool, and S. K. Armstrong. 2006. Heme transport contributes to in vivo fitness of Bordetella pertussis during primary infection in mice. Infect. Immun. 74:1741-1744.

19. Brunger, A. T., P. D. Adams, G. M. Clore, W. L. DeLano, P. Gros, R. W. Grosse-Kunstleve, J. S. Jiang, J. Kuszewski, M. Nilges, N. S. Pannu, R. J. Read, L. M. Rice, T. Simonson, and G. L. Warren. 1998. Crystallography & NMR system: A new software suite for macromolecular structure determination. Acta Crystallogr. D. Biol. Crystallogr. 54:905-921.

20. Bullen, J. J., H. J. Rogers, and E. Griffiths. 1978. Role of iron in bacterial infection. Curr. Top. Microbiol. Immunol. 80:1-35.

21. Caignan, G. A., R. Deshmukh, A. Wilks, Y. Zeng, H. W. Huang, P. Moenne-Loccoz, R. A. Bunce, M. A. Eastman, and M. Rivera. 2002. Oxidation of heme to beta- and delta-biliverdin by Pseudomonas aeruginosa heme oxygenase as a consequence of an unusual seating of the heme. J Am Chem Soc 124:14879-14892.

22. Campos, A. P., A. P. Aguiar, M. Hervas, M. Regalla, J. A. Navarro, J. M. Ortega, A. V. Xavier, M. A. De La Rosa, and M. Teixeira. 1993. Cytochrome

132

c6 from Monoraphidium braunii. A cytochrome with an unusual heme axial coordination. Eur. J Biochem 216:329-341.

23. Carter, A. O., A. A. Borczyk, J. A. Carlson, B. Harvey, J. C. Hockin, M. A. Karmali, C. Krishnan, D. A. Korn, and H. Lior. 1987. A severe outbreak of Escherichia coli O157:H7--associated hemorrhagic colitis in a nursing home. N. Engl. J Med. 317:1496-1500.

24. Casari, G., C. Sander, and A. Valencia. 1995. A method to predict functional residues in proteins. Nat. Struct. Biol. 2:171-178.

25. Chan, A. C., B. Lelj-Garolla, I. Rosell, K. A. Pedersen, A. G. Mauk, and M. E. Murphy. 2006. Cofacial heme binding is linked to dimerization by a bacterial heme transport protein. J Mol. Biol. 362:1108-1119.

26. Chu, G. C., K. Katakura, X. Zhang, T. Yoshida, and M. Ikeda-Saito. 1999. Heme degradation as catalyzed by a recombinant bacterial heme oxygenase (HmuO) from Corynebacterium diphtheriae. J. Biol. Chem. 274:21319-21325.

27. Cook, M. N., K. Nakatsu, G. S. Marks, B. E. McLaughlin, H. J. Vreman, D. K. Stevenson, and J. F. Brien. 1995. Heme oxygenase activity in the adult rat aorta and liver as measured by carbon monoxide formation. Can. J. Physiol Pharmacol. 73:515-518.

28. Cope, L. D., S. E. Thomas, J. L. Latimer, C. A. Slaughter, U. Muller-Eberhard, and E. J. Hansen. 1994. The 100 kDa haem:haemopexin-binding protein of Haemophilus influenzae: structure and localization. Mol. Microbiol. 13:863-873.

29. Cope, L. D., R. Yogev, U. Muller-Eberhard, and E. J. Hansen. 1995. A gene cluster involved in the utilization of both free heme and heme:hemopexin by Haemophilus influenzae type b. J Bacteriol. 177:2644-2653.

30. Cornejo, J., R. D. Willows, and S. I. Beale. 1998. Phytobilin biosynthesis: cloning and expression of a gene encoding soluble ferredoxin-dependent heme oxygenase from Synechocystis sp. PCC 6803. Plant J 15:99-107.

31. Corpet, F. 1988. Multiple sequence alignment with hierarchical clustering. Nucleic Acids Res 16:10881-10890.

32. Cox, T. M. and D. J. Halsall. 2002. Hemochromatosis-neonatal and young subjects. Blood Cells Mol. Dis. 29:411-417.

33. Crosa, J. H. 1989. Genetics and molecular biology of siderophore-mediated iron transport in bacteria. Microbiol. Rev. 53:517-530.

34. Crusats, J., A. Suzuki, T. Mizutani, and H. Ogoshi. 1998. Regioselective Porphyrin Bridge Cleavage Controlled by Electronic Effects. Coupled Oxidation

133

of 3-Demethyl-3-(trifluoromethyl)mesohemin IX and Identification of Its Four Biliverdin Derivatives. J Org. Chem 63:602-607.

35. Czjzek, M., S. Letoffe, C. Wandersman, M. Delepierre, A. Lecroisey, and N. Izadi-Pruneyre. 2007. The Crystal Structure of the Secreted Dimeric Form of the Hemophore HasA Reveals a Domain Swapping with an Exchanged Heme Ligand. J Mol. Biol. 365:1176-1186.

36. Devadas, K. and S. Dhawan. 2006. Hemin activation ameliorates HIV-1 infection via heme oxygenase-1 induction. J Immunol. 176:4252-4257.

37. Doi, K., T. Akaike, S. Fujii, S. Tanaka, N. Ikebe, T. Beppu, S. Shibahara, M. Ogawa, and H. Maeda. 1999. Induction of haem oxygenase-1 nitric oxide and ischaemia in experimental solid tumours and implications for tumour growth. Br. J Cancer. 80:1945-1954.

38. Drago-Serrano, M. E., S. G. Parra, and H. A. Manjarrez-Hernandez. 2006. EspC, an autotransporter protein secreted by enteropathogenic Escherichia coli (EPEC), displays protease activity on human hemoglobin. FEMS Microbiol. Lett. 265:35-40.

39. Dugad, L. B., X. Wang, C. C. Wang, G. S. Lukat, and H. M. Goff. 1992. Proton nuclear Overhauser effect study of the heme active site structure of chloroperoxidase. Biochemistry 31:1651-1655.

40. Enz, S., V. Braun, and J. H. Crosa. 1995. Transcription of the region encoding the ferric dicitrate-transport system in Escherichia coli: similarity between promoters for fecA and for extracytoplasmic function sigma factors. Gene 163:13-18.

41. Escolar, L., L. de, V, and J. Perez-Martin. 1997. Metalloregulation in vitro of the aerobactin promoter of Escherichia coli by the Fur (ferric uptake regulation) protein. Mol. Microbiol. 26:799-808.

42. Eyal, E., S. Gerzon, V. Potapov, M. Edelman, and V. Sobolev. 2005. The limit of accuracy of protein modeling: influence of crystal packing on protein structure. J Mol. Biol. 351:431-442.

43. Fiegen, D., L. Blumenstein, P. Stege, I. R. Vetter, and M. R. Ahmadian. 2002. Crystal structure of Rnd3/RhoE: functional implications. FEBS Lett. 525:100-104.

44. Frankenberg-Dinkel, N. 2004. Bacterial heme oxygenases. Antioxid. Redox. Signal. 6:825-834.

45. Fredenburgh, L. E., M. A. Perrella, and S. A. Mitsialis. 2007. The role of heme oxygenase-1 in pulmonary disease. Am. J Respir. Cell Mol. Biol. 36:158-165.

134

46. Friedman, J., L. Lad, H. Li, A. Wilks, and T. L. Poulos. 2004. Structural basis for novel delta-regioselective heme oxygenation in the opportunistic pathogen

Pseudomonas aeruginosa. Biochemistry 43:5239-5245.

47. Gambetta, G. A. and J. C. Lagarias. 2001. Genetic engineering of phytochrome biosynthesis in bacteria. Proc. Natl. Acad. Sci. U. S. A. 98:10566-10571.

48. Genco, C. A. and D. W. Dixon. 2001. Emerging strategies in microbial haem capture. Mol. Microbiol. 39:1-11.

49. Gerstein, M. 2000. Integrative database analysis in structural genomics. Nat. Struct. Biol. 7 Suppl:960-3.:960-963.

50. Ghigo, J. M., S. Letoffe, and C. Wandersman. 1997. A new type of hemophore-dependent heme acquisition system of Serratia marcescens reconstituted in Escherichia coli. J Bacteriol. 179:3572-3579.

51. Gilles-Gonzalez, M. A. and G. Gonzalez. 2005. Heme-based sensors: defining characteristics, recent developments, and regulatory hypotheses. J Inorg. Biochem 99:1-22.

52. Gong, W., B. Hao, S. S. Mansy, G. Gonzalez, M. A. Gilles-Gonzalez, and M. K. Chan. 1998. Structure of a biological oxygen sensor: a new mechanism for heme-driven signal transduction. Proc Natl Acad Sci U S A 95:15177-15182.

53. Halder, P., J. T. Trent, III, and M. S. Hargrove. 2007. Influence of the protein matrix on intramolecular histidine ligation in ferric and ferrous hexacoordinate hemoglobins. Proteins. 66:172-182.

54. Hamza, I., S. Chauhan, R. Hassett, and M. R. O'Brian. 1998. The bacterial irr protein is required for coordination of heme biosynthesis with iron availability. J Biol. Chem. 273:21669-21674.

55. Hamza, I., Z. Qi, N. D. King, and M. R. O'Brian. 2000. Fur-independent regulation of iron metabolism by Irr in Bradyrhizobium japonicum. Microbiology. 146:669-676.

56. Hayashi, T., K. Makino, M. Ohnishi, K. Kurokawa, K. Ishii, K. Yokoyama, C. G. Han, E. Ohtsubo, K. Nakayama, T. Murata, M. Tanaka, T. Tobe, T. Iida, H. Takami, T. Honda, C. Sasakawa, N. Ogasawara, T. Yasunaga, S. Kuhara, T. Shiba, M. Hattori, and H. Shinagawa. 2001. Complete genome sequence of enterohemorrhagic Escherichia coli O157:H7 and genomic comparison with a laboratory strain K-12. DNA Res. 8:11-22.

57. He, M. M., S. L. Clugston, J. F. Honek, and B. W. Matthews. 2000. Determination of the structure of Escherichia coli glyoxalase I suggests a structural basis for differential metal activation. Biochemistry. 39:8719-8727.

135

58. Henderson, D. P. and S. M. Payne. 1993. Cloning and characterization of the Vibrio cholerae genes encoding the utilization of iron from haemin and haemoglobin. Mol. Microbiol. 7:461-469.

59. Henderson, I. R., J. Czeczulin, C. Eslava, F. Noriega, and J. P. Nataro. 1999. Characterization of pic, a secreted protease of Shigella flexneri and enteroaggregative Escherichia coli. Infect. Immun. 67:5587-5596.

60. Hendrickson, W. A., J. R. Horton, and D. M. LeMaster. 1990. Selenomethionyl proteins produced for analysis by multiwavelength anomalous diffraction (MAD): a vehicle for direct determination of three-dimensional structure. EMBO J 9:1665-1672.

61. Hirotsu, S., G. C. Chu, M. Unno, D. S. Lee, T. Yoshida, S. Y. Park, Y. Shiro, and M. Ikeda-Saito. 2004. The crystal structures of the ferric and ferrous forms of the heme complex of HmuO, a heme oxygenase of Corynebacterium

diphtheriae. J. Biol. Chem. 279:11937-11947.

62. Holme, R. 2003. Drinking water contamination in Walkerton, Ontario: positive resolutions from a tragic event. Water Sci. Technol. 47:1-6.

63. Hornung, J. M., H. A. Jones, and R. D. Perry. 1996. The hmu locus of Yersinia pestis is essential for utilization of free haemin and haem--protein complexes as iron sources. Mol. Microbiol. 20:725-739.

64. Hou, S., R. W. Larsen, D. Boudko, C. W. Riley, E. Karatan, M. Zimmer, G. W. Ordal, and M. Alam. 2000. Myoglobin-like aerotaxis transducers in Archaea and Bacteria. Nature 403:540-544.

65. Iyer, A. K., K. Greish, J. Fang, R. Murakami, and H. Maeda. 2007. High-loading nanosized micelles of copoly(styrene-maleic acid)-zinc protoporphyrin for targeted delivery of a potent heme oxygenase inhibitor. Biomaterials. 28:1871-1881.

66. Jones, H. A., J. W. Lillard, Jr., and R. D. Perry. 1999. HmsT, a protein essential for expression of the haemin storage (Hms+) phenotype of Yersinia

pestis. Microbiology. 145:2117-2128.

67. Kabsch, W. A solution for the best rotation to relate two sets of vectors.

68. Kang, Y., T. Durfee, J. D. Glasner, Y. Qiu, D. Frisch, K. M. Winterberg, and F. R. Blattner. 2004. Systematic mutagenesis of the Escherichia coli genome. J Bacteriol. 186:4921-4930.

69. Kapitulnik, J. 2004. Bilirubin: an endogenous product of heme degradation with both cytotoxic and cytoprotective properties. Mol. Pharmacol. 66:773-779.

136

70. Karch, H., P. I. Tarr, and M. Bielaszewska. 2005. Enterohaemorrhagic Escherichia coli in human medicine. Int. J Med. Microbiol. 295:405-418.

71. Keene, W. E., K. Hedberg, D. E. Herriott, D. D. Hancock, R. W. McKay, T. J. Barrett, and D. W. Fleming. 1997. A prolonged outbreak of Escherichia coli O157:H7 infections caused by commercially distributed raw milk. J Infect. Dis. 176:815-818.

72. Kendrew, J. C. 1959. Structure and function in myoglobin and other proteins. Fed. Proc. 18:740-751.

73. Kikuchi, G., T. Yoshida, and M. Noguchi. 2005. Heme oxygenase and heme degradation. Biochem. Biophys. Res. Commun. 338:558-567.

74. Kim, B. Y., J. Kang, and K. S. Kim. 2005. Invasion processes of pathogenic Escherichia coli. Int. J Med. Microbiol. 295:463-470.

75. Kim, H. P., S. W. Ryter, and A. M. Choi. 2006. CO as a cellular signaling molecule. Annu. Rev. Pharmacol. Toxicol. 46:411-49.:411-449.

76. Kinobe, R. T., R. A. Dercho, J. Z. Vlahakis, J. F. Brien, W. A. Szarek, and K. Nakatsu. 2006. Inhibition of the enzymatic activity of heme oxygenases by azole-based antifungal drugs. J Pharmacol. Exp. Ther. 319:277-284.

77. Krissinel, E. and K. Henrick. 2004. Secondary-structure matching (SSM), a new tool for fast protein structure alignment in three dimensions. Acta Crystallogr D Biol Crystallogr 60:2256-2268.

78. Krissinel, E. and K. Henrick. 2004. Protein structure comparison in 3D based on secondary structure matching (SSM) followed by Cα alignment, scored by a new structural similarity function., p. 88. In Andreas J.Kungl & Penelope J.Kungl (Eds.) (ed.).

79. Lad, L., J. Wang, H. Li, J. Friedman, B. Bhaskar, P. R. Ortiz de Montellano, and T. L. Poulos. 2003. Crystal structures of the ferric, ferrous, and ferrous-NO forms of the Asp140Ala mutant of human heme oxygenase-1: catalytic implications. J Mol. Biol. 330:527-538.

80. Lansky, I. B., G. S. Lukat-Rodgers, D. Block, K. R. Rodgers, M. Ratliff, and A. Wilks. 2006. The Cytoplasmic Heme-binding Protein (PhuS) from the Heme Uptake System of Pseudomonas aeruginosa Is an Intracellular Heme-trafficking Protein to the delta-Regioselective Heme Oxygenase. J Biol. Chem. 281:13652-13662.

81. Lanzilotta, W. N., D. J. Schuller, M. V. Thorsteinsson, R. L. Kerby, G. P. Roberts, and T. L. Poulos. 2000. Structure of the CO sensing transcription activator CooA. Nat. Struct. Biol 7:876-880.

137

82. Lara, F. A., U. Lins, G. H. Bechara, and P. L. Oliveira. 2005. Tracing heme in a living cell: hemoglobin degradation and heme traffic in digest cells of the cattle tick Boophilus microplus. J Exp. Biol. 208:3093-3101.

83. Law, D. and J. Kelly. 1995. Use of heme and hemoglobin by Escherichia coli O157 and other Shiga-like-toxin-producing E. coli serogroups. Infect. Immun. 63:700-702.

84. Layer, G., J. Moser, D. W. Heinz, D. Jahn, and W. D. Schubert. 2003. Crystal structure of coproporphyrinogen III oxidase reveals cofactor geometry of Radical SAM enzymes. EMBO J 22:6214-6224.

85. Letoffe, S., C. Deniau, N. Wolff, E. Dassa, P. Delepelaire, A. Lecroisey, and C. Wandersman. 2001. Haemophore-mediated bacterial haem transport: evidence for a common or overlapping site for haem-free and haem-loaded haemophore on its specific outer membrane receptor. Mol. Microbiol. 41:439-450.

86. Louie, G. V. and G. D. Brayer. 1990. High-resolution refinement of yeast iso-1-cytochrome c and comparisons with other eukaryotic cytochromes c. J Mol. Biol. 20;214:527-555.

87. Maki, H., K. Matsuura, K. Shimada, and K. V. Nagashima. 2003. Chimeric photosynthetic reaction center complex of purple bacteria composed of the core subunits of Rubrivivax gelatinosus and the cytochrome subunit of Blastochloris

viridis. J Biol. Chem 278:3921-3928.

88. McCoy, A. J. 2007. Solving structures of protein complexes by molecular replacement with Phaser. Acta Crystallogr. D. Biol. Crystallogr. 63:32-41.

89. McHugh, J. P., F. Rodriguez-Quinones, H. Abdul-Tehrani, D. A. Svistunenko, R. K. Poole, C. E. Cooper, and S. C. Andrews. 2003. Global iron-dependent gene regulation in Escherichia coli. A new mechanism for iron homeostasis. J. Biol. Chem. 278:29478-29486.

90. McRee, D. E. 1999. XtalView/Xfit--A versatile program for manipulating atomic coordinates and electron density. J Struct. Biol 125:156-165.

91. Mills, M. and S. M. Payne. 1995. Genetics and regulation of heme iron transport in Shigella dysenteriae and detection of an analogous system in Escherichia coli

O157:H7. J Bacteriol. 177:3004-3009.

92. Montellano, P. R. 2000. The mechanism of heme oxygenase. Curr. Opin. Chem Biol. 4:221-227.

93. Morris, G. M., D. S. Goodsell, R. S. Halliday, R. Huey, W. E. Hart, R. K. Belew, and A. J. Olson. 1998. Automated docking using a Lamarckian genetic algorithm and an empirical binding free energy function. J Comput Chem 19:1639-1662.

138

94. Moult, J. and E. Melamud. 2000. From fold to function. Curr. Opin. Struct. Biol. 10:384-389.

95. Mourino, S., C. R. Osorio, M. L. Lemos, and J. H. Crosa. 2006. Transcriptional organization and regulation of the Vibrio anguillarum heme uptake gene cluster. Gene. 374:68-76.:68-76.

96. Murshudov, G. N., A. A. Vagin, and E. J. Dodson. 1997. Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallogr. D. Biol. Crystallogr. 53:240-255.

97. Murzin, A. G., S. E. Brenner, T. Hubbard, and C. Chothia. 1995. SCOP: a structural classification of proteins database for the investigation of sequences and structures. J Mol. Biol. 247:536-540.

98. Novotny, M., D. Madsen, and G. J. Kleywegt. 2004. Evaluation of protein fold comparison servers. Proteins. 54:260-270.

99. Ogura, Y., K. Kurokawa, T. Ooka, K. Tashiro, T. Tobe, M. Ohnishi, K. Nakayama, T. Morimoto, J. Terajima, H. Watanabe, S. Kuhara, and T. Hayashi. 2006. Complexity of the genomic diversity in enterohemorrhagic Escherichia coli O157 revealed by the combinational use of the O157 Sakai OligoDNA microarray and the Whole Genome PCR scanning. DNA Res. 13:3-14.

100. Okeke, I. N., I. C. Scaletsky, E. H. Soars, L. R. Macfarlane, and A. G. Torres. 2004. Molecular epidemiology of the iron utilization genes of enteroaggregative

Escherichia coli. J Clin. Microbiol. 42:36-44.

101. Olczak, T., K. Siudeja, and M. Olczak. 2006. Purification and initial characterization of a novel Porphyromonas gingivalis HmuY protein expressed in Escherichia coli and insect cells. Protein Expr. Purif. 49:299-306.

102. Otwinowski, Z. and W. Minor. 1997. [20] Processing of X-ray diffraction data collected in oscillation mode, p. 307-326. In Methods in Enzymology. Academic Press.

103. Paoli, M., B. F. Anderson, H. M. Baker, W. T. Morgan, A. Smith, and E. N. Baker. 1999. Crystal structure of hemopexin reveals a novel high-affinity heme site formed between two beta-propeller domains. Nat. Struct. Biol 6:926-931.

104. Paoli, M., J. Marles-Wright, and A. Smith. 2002. Structure-function relationships in heme-proteins. DNA Cell Biol. 21:271-280.

105. Poulos, T. L. 2005. Structural and functional diversity in heme monooxygenases. Drug Metab Dispos. 33:10-18.

139

106. Puri, S. and M. R. O'Brian. 2006. The hmuQ and hmuD genes from

Bradyrhizobium japonicum encode heme-degrading enzymes. J Bacteriol. 188:6476-6482.

107. Qi, Z., I. Hamza, and M. R. O'Brian. 1999. Heme is an effector molecule for iron-dependent degradation of the bacterial iron response regulator (Irr) protein. Proc Natl Acad Sci U S A 96:13056-13061.

108. Rangel, J. M., P. H. Sparling, C. Crowe, P. M. Griffin, and D. L. Swerdlow. 2005. Epidemiology of Escherichia coli O157:H7 outbreaks, United States, 1982-2002. Emerg. Infect. Dis. 11:603-609.

109. Rasmussen, A. W., H. L. Alexander, D. Perkins-Balding, W. M. Shafer, and I. Stojiljkovic. 2005. Resistance of Neisseria meningitidis to the toxic effects of heme iron and other hydrophobic agents requires expression of ght. J Bacteriol. 187:5214-5223.

110. Ratliff, M., W. Zhu, R. Deshmukh, A. Wilks, and I. Stojiljkovic. 2001. Homologues of neisserial heme oxygenase in gram-negative bacteria: degradation of heme by the product of the pigA gene of Pseudomonas aeruginosa. J Bacteriol. 183:6394-6403.

111. Ridley, K. A., J. D. Rock, Y. Li, and J. M. Ketley. 2006. Heme utilization in Campylobacter jejuni. J Bacteriol. 188:7862-7875.

112. Rodgers, P. A., H. J. Vreman, P. A. Dennery, and D. K. Stevenson. 1994. Sources of carbon monoxide (CO) in biological systems and applications of CO detection technologies. Semin. Perinatol. 18:2-10.

113. Ryter, S. W. and R. M. Tyrrell. 2000. The heme synthesis and degradation pathways: role in oxidant sensitivity. Heme oxygenase has both pro- and antioxidant properties. Free Radic. Biol. Med. 28:289-309.

114. Sanchez, R., U. Pieper, F. Melo, N. Eswar, M. A. Marti-Renom, M. S. Madhusudhan, N. Mirkovic, and A. Sali. 2000. Protein structure modeling for structural genomics. Nat. Struct. Biol. 7 Suppl:986-90.:986-990.

115. Sasakura, Y., T. Yoshimura-Suzuki, H. Kurokawa, and T. Shimizu. 2006. Structure-function relationships of EcDOS, a heme-regulated phosphodiesterase from Escherichia coli. Acc. Chem Res. 39:37-43.

116. Schmitt, M. P. 1997. Utilization of host iron sources by Corynebacterium

diphtheriae: identification of a gene whose product is homologous to eukaryotic heme oxygenases and is required for acquisition of iron from heme and hemoglobin. J Bacteriol. 179:838-845.

140

117. Schmitt, M. P. and R. K. Holmes. 1991. Iron-dependent regulation of diphtheria toxin and siderophore expression by the cloned Corynebacterium diphtheriae

repressor gene dtxR in C. diphtheriae C7 strains. Infect. Immun. 59:1899-1904.

118. Schmitt, M. P., E. M. Twiddy, and R. K. Holmes. 1992. Purification and characterization of the diphtheria toxin repressor. Proc Natl Acad Sci U S A 89:7576-7580.

119. Schmitt, T. H., W. A. Frezzatti, Jr., and S. Schreier. 1993. Hemin-induced lipid membrane disorder and increased permeability: a molecular model for the mechanism of cell lysis. Arch. Biochem. Biophys. 307:96-103.

120. Schneider, S., K. H. Sharp, P. D. Barker, and M. Paoli. 2006. An induced fit conformational change underlies the binding mechanism of the heme transport proteobacteria-protein HemS. J Biol. Chem. 281:32606-32610.

121. Schuller, D. J., A. Wilks, P. R. Ortiz de Montellano, and T. L. Poulos. 1999. Crystal structure of human heme oxygenase-1. Nat. Struct. Biol 6:860-867.

122. Schuller, D. J., W. Zhu, I. Stojiljkovic, A. Wilks, and T. L. Poulos. 2001. Crystal structure of heme oxygenase from the gram-negative pathogen Neisseria

meningitidis and a comparison with mammalian heme oxygenase-1. Biochemistry 40:11552-11558.

123. Schulz, H., H. Hennecke, and L. Thony-Meyer. 1998. Prototype of a heme chaperone essential for cytochrome c maturation. Science 281:1197-1200.

124. Simpson, W., T. Olczak, and C. A. Genco. 2000. Characterization and expression of HmuR, a TonB-dependent hemoglobin receptor of Porphyromonas

gingivalis. J Bacteriol. 182:5737-5748.

125. Skaar, E. P., A. H. Gaspar, and O. Schneewind. 2004. IsdG and IsdI, heme-degrading enzymes in the cytoplasm of Staphylococcus aureus. J. Biol. Chem. 279:436-443.

126. Snyder, J. A., B. J. Haugen, E. L. Buckles, C. V. Lockatell, D. E. Johnson, M. S. Donnenberg, R. A. Welch, and H. L. Mobley. 2004. Transcriptome of uropathogenic Escherichia coli during urinary tract infection. Infect. Immun. 72:6373-6381.

127. Stojiljkovic, I. and K. Hantke. 1994. Transport of haemin across the cytoplasmic membrane through a haemin-specific periplasmic binding-protein-dependent transport system in Yersinia enterocolitica. Mol. Microbiol. 13:719-732.

128. Stojiljkovic, I. and K. Hantke. 1992. Hemin uptake system of Yersinia

enterocolitica: similarities with other TonB-dependent systems in Gram-negative bacteria. EMBO J. 11:4359-4367.

141

129. Stojiljkovic, I., V. Hwa, M. L. de Saint, P. O'Gaora, X. Nassif, F. Heffron, and M. So. 1995. The Neisseria meningitidis haemoglobin receptor: its role in iron utilization and virulence. Mol. Microbiol. 15:531-541.

130. Stojiljkovic, I. and D. Perkins-Balding. 2002. Processing of heme and heme-containing proteins by bacteria. DNA Cell Biol 21:281-295.

131. Su, C. and L. J. Brandt. 1995. Escherichia coli O157:H7 infection in humans. Ann. Intern. Med. 123:698-714.

132. Sugishima, M., Y. Hagiwara, X. Zhang, T. Yoshida, C. T. Migita, and K. Fukuyama. 2005. Crystal structure of dimeric heme oxygenase-2 from Synechocystis sp. PCC 6803 in complex with heme. Biochemistry. 44:4257-4266.

133. Sugishima, M., Y. Higashimoto, T. Oishi, H. Takahashi, H. Sakamoto, M. Noguchi, and K. Fukuyama. 2007. X-ray Crystallographic and Biochemical Characterization of the Inhibitory Action of an Imidazole-Dioxolane Compound on Heme Oxygenase. Biochemistry. 46:1860-1867.

134. Sugishima, M., H. Sakamoto, Y. Higashimoto, M. Noguchi, and K. Fukuyama. 2003. Crystal structure of rat heme oxygenase-1 in complex with biliverdin-iron chelate. Conformational change of the distal helix during the heme cleavage reaction. J Biol. Chem. 278:32352-32358.

135. Sugishima, M., H. Sakamoto, Y. Higashimoto, Y. Omata, S. Hayashi, M. Noguchi, and K. Fukuyama. 2002. Crystal structure of rat heme oxygenase-1 in complex with heme bound to azide. Implication for regiospecific hydroxylation of heme at the alpha-meso carbon. J Biol. Chem. 277:45086-45090.

136. Sugishima, M., H. Sakamoto, M. Noguchi, and K. Fukuyama. 2003. Crystal structures of ferrous and CO-, CN(-)-, and NO-bound forms of rat heme oxygenase-1 (HO-1) in complex with heme: structural implications for discrimination between CO and O2 in HO-1. Biochemistry. 42:9898-9905.

137. Suits, M. D., N. Jaffer, and Z. Jia. 2006. Structure of the Escherichia coli O157:H7 heme oxygenase ChuS in complex with heme and enzymatic inactivation by mutation of the heme coordinating residue His-193. J Biol. Chem. 281:36776-36782.

138. Suits, M. D., G. P. Pal, and Z. Jia. 2007. Structure of the Escherichia coli O157:H7 heme binding protein ChuX and its mechanism of heme uptake. Submitted to J. Biol. Chem.

139. Suits, M. D., G. P. Pal, K. Nakatsu, A. Matte, M. Cygler, and Z. Jia. 2005. Identification of an Escherichia coli O157:H7 heme oxygenase with tandem functional repeats. Proc. Natl. Acad. Sci. U. S. A. 102:16955-16960.

142

140. Szustakowski, J. D. and Z. Weng. 2000. Protein structure alignment using a genetic algorithm. Proteins. 38:428-440.

141. Taketani, S. 2005. Aquisition, mobilization and utilization of cellular iron and heme: endless findings and growing evidence of tight regulation. Tohoku J Exp. Med. 205:297-318.

142. Tatusov, R. L., E. V. Koonin, and D. J. Lipman. 1997. A genomic perspective on protein families. Science. 278:631-637.

143. Tatusova, T. A. and T. L. Madden. 1999. BLAST 2 Sequences, a new tool for comparing protein and nucleotide sequences. FEMS Microbiol. Lett. 174:247-250.

144. Terwilliger, T. C. and J. Berendzen. 1999. Automated MAD and MIR structure solution. Acta Crystallogr. D. Biol. Crystallogr. 55:849-861.

145. Thompson, J. M., H. A. Jones, and R. D. Perry. 1999. Molecular characterization of the hemin uptake locus (hmu) from Yersinia pestis and analysis of hmu mutants for hemin and hemoprotein utilization. Infect. Immun. 67:3879-3892.

146. Torres, A. G. and S. M. Payne. 1997. Haem iron-transport system in enterohaemorrhagic Escherichia coli O157:H7. Mol. Microbiol. 23:825-833.

147. Tsutsui, K. and G. C. Mueller. 1982. A protein with multiple heme-binding sites from rabbit serum. J Biol. Chem. 257:3925-3931.

148. Vagin, A. A., R. A. Steiner, A. A. Lebedev, L. Potterton, S. McNicholas, F. Long, and G. N. Murshudov. 2004. REFMAC5 dictionary: organization of prior chemical knowledge and guidelines for its use. Acta Crystallogr. D. Biol. Crystallogr. 60:2184-2195.

149. Vassinova, N. and D. Kozyrev. 2000. A method for direct cloning of fur-regulated genes: identification of seven new fur-regulated loci in Escherichia coli. Microbiology 146 Pt 12:3171-3182.

150. Verma, A., F. J. Bolton, D. Fiefield, P. Lamb, E. Woloschin, N. Smith, and R. McCann. 2007. An outbreak of E. coli O157 associated with a swimming pool: an unusual vehicle of transmission. Epidemiol. Infect. :1-4.:1-4.

151. Vlahakis, J. Z., R. T. Kinobe, R. J. Bowers, J. F. Brien, K. Nakatsu, and W. A. Szarek. 2005. Synthesis and evaluation of azalanstat analogues as heme oxygenase inhibitors. Bioorg. Med. Chem Lett. 15:1457-1461.

152. Vreman, H. J., D. A. Cipkala, and D. K. Stevenson. 1996. Characterization of porphyrin heme oxygenase inhibitors. Can. J Physiol Pharmacol. 74:278-285.

143

153. Vreman, H. J., L. K. Kwong, and D. K. Stevenson. 1984. Carbon monoxide in blood: an improved microliter blood-sample collection system, with rapid analysis by gas chromatography. Clin. Chem. 30:1382-1386.

154. Wagener, F. A., H. D. Volk, D. Willis, N. G. Abraham, M. P. Soares, G. J. Adema, and C. G. Figdor. 2003. Different faces of the heme-heme oxygenase system in inflammation. Pharmacol. Rev. 55:551-571.

155. Wandersman, C. and P. Delepelaire. 2004. Bacterial iron sources: from siderophores to hemophores. Annu. Rev. Microbiol. 58:611-47.:611-647.

156. Wandersman, C. and I. Stojiljkovic. 2000. Bacterial heme sources: the role of heme, hemoprotein receptors and hemophores. Curr. Opin. Microbiol. 3:215-220.

157. Wegele, R., R. Tasler, Y. Zeng, M. Rivera, and N. Frankenberg-Dinkel. 2004. The Heme Oxygenase(s)-Phytochrome System of Pseudomonas aeruginosa. J. Biol. Chem. 279:45791-45802.

158. Wilks, A. 2001. The ShuS protein of Shigella dysenteriae is a heme-sequestering protein that also binds DNA. Arch Biochem Biophys 387:137-142.

159. Wilks, A., S. M. Black, W. L. Miller, and P. R. Ortiz de Montellano. 1995. Expression and characterization of truncated human heme oxygenase (hHO-1) and a fusion protein of hHO-1 with human cytochrome P450 reductase. Biochemistry 34:4421-4427.

160. Wilks, A. and P. R. Ortiz de Montellano. 1993. Rat liver heme oxygenase. High level expression of a truncated soluble form and nature of the meso-hydroxylating species. J Biol. Chem 268:22357-22362.

161. Wilks, A., P. R. Ortiz de Montellano, J. Sun, and T. M. Loehr. 1996. Heme oxygenase (HO-1): His-132 stabilizes a distal water ligand and assists catalysis. Biochemistry 35:930-936.

162. Wilks, A. and M. P. Schmitt. 1998. Expression and characterization of a heme oxygenase (Hmu O) from Corynebacterium diphtheriae. Iron acquisition requires oxidative cleavage of the heme macrocycle. J Biol Chem 273:837-841.

163. Worthington, M. T., S. M. Cohn, S. K. Miller, R. Q. Luo, and C. L. Berg. 2001. Characterization of a human plasma membrane heme transporter in intestinal and hepatocyte cell lines. Am. J Physiol Gastrointest. Liver Physiol. 280:G1172-G1177.

164. Wyckoff, E. E., D. Duncan, A. G. Torres, M. Mills, K. Maase, and S. M. Payne. 1998. Structure of the Shigella dysenteriae haem transport locus and its phylogenetic distribution in enteric bacteria. Mol. Microbiol. 28:1139-1152.

144

165. Wyckoff, E. E., G. F. Lopreato, K. A. Tipton, and S. M. Payne. 2005. Shigella

dysenteriae ShuS promotes utilization of heme as an iron source and protects against heme toxicity. J Bacteriol. 187:5658-5664.

166. Wyckoff, E. E., M. Schmitt, A. Wilks, and S. M. Payne. 2004. HutZ is required for efficient heme utilization in Vibrio cholerae. J Bacteriol. 186:4142-4151.

167. Xu, K., J. Delling, and T. Elliott. 1992. The genes required for heme synthesis in Salmonella typhimurium include those encoding alternative functions for aerobic and anaerobic coproporphyrinogen oxidation. J Bacteriol. 174:3953-3963.

168. Ye, W. and L. Zhang. 2004. Heme controls the expression of cell cycle regulators and cell growth in HeLa cells. Biochem Biophys Res Commun. 315:546-554.

169. Yi, X., M. Mroczko, K. M. Manoj, X. Wang, and L. P. Hager. 1999. Replacement of the proximal heme thiolate ligand in chloroperoxidase with a histidine residue. Proc. Natl. Acad. Sci. U. S. A. 96:12412-12417.

170. Zarembinski, T. I., L. W. Hung, H. J. Mueller-Dieckmann, K. K. Kim, H. Yokota, R. Kim, and S. H. Kim. 1998. Structure-based assignment of the biochemical function of a hypothetical protein: a test case of structural genomics. Proc. Natl. Acad. Sci. U. S. A. 95:15189-15193.

171. Zhang, H., K. Huang, Z. Li, L. Banerjei, K. E. Fisher, N. V. Grishin, E. Eisenstein, and O. Herzberg. 2000. Crystal structure of YbaK protein from Haemophilus influenzae (HI1434) at 1.8 A resolution: functional implications. Proteins. 40:86-97.

172. Zhao, Y., C. Hoganson, G. T. Babcock, and M. A. Marletta. 1998. Structural changes in the heme proximal pocket induced by nitric oxide binding to soluble guanylate cyclase. Biochemistry 37:12458-12464.

173. Zhu, W., D. J. Hunt, A. R. Richardson, and I. Stojiljkovic. 2000. Use of heme compounds as iron sources by pathogenic neisseriae requires the product of the hemO gene. J Bacteriol. 182:439-447.

174. Zhu, W., A. Wilks, and I. Stojiljkovic. 2000. Degradation of heme in Gram-negative bacteria: the product of the hemO gene of Neisseriae is a heme oxygenase. J. Bacteriol. 182:6783-6790.

145

Appendix I

Crystal structure of the conserved Gram-negative lipopolysaccharide transport

protein LptA

Michael D. L. Suits and Zongchao Jia

Preface: structure and functional information are currently in development for this

chapter: Michael D. L. Suits and Zongchao Jia (2007) Crystal structure of the conserved

Gram-negative lipopolysaccharide transport protein LptA.

Michael Suits was responsible for protein expression, purification, crystallization, data

collection, structure solution and refinement. Dr. Christian M. Udell cloned LptA into the

expression vector. Dr. Howard Robinson collected the LptA MAD diffraction data. Dr.

Clemens Vonrhein helped with the application of autoSHARP that ultimately led to a

structural solution. Dr. Michael Sawaya provided useful discussions on how to overcome

anisotropy and developed the anisotropic correction server employed herein. This chapter

was written by Michael Suits with editorial input from Dr. Zongchao Jia and with

additional help from Jocelyne Suits.

146

A1.1 Abstract

LptA is an essential lipopolysaccharide (LPS) transport protein that is conserved

across various Gram-negative bacteria, many of which are human and plant pathogens.

LptA has been implicated in LPS transport from the inner membrane to the outer

membrane, thereby contributing to maintaining cell envelope integrity and outer

membrane assembly. Here we present the first crystal structure of the essential LPS

periplasmic transport protein LptA in two crystal forms. This is the processed form of

LptA which has been exported to the periplasm and has concurrently had the N-terminal

27 residues removed during the export process. The structure of LptA represents a novel

fold which comprises a set of eight antiparallel β-sheets, bent in the middle to resemble a

ribcage or β-jellyroll. The structure of LptA is not completely symmetrical along the

length but opens slightly at both the N and C-terminal ends, creating two putative clefts.

The N-terminal cleft is slightly larger than the C-terminal cleft, and is surrounded by a

conserved segment of residues from P35-K83. Other conserved residues align along the

length of the β-sheets, implicating a structural contribution. Severe anisotropy in

diffraction data was corrected in order to obtain the final refined structure.

147

A1.2 Introduction

Gram-negative bacteria such as Escherichia coli are compartmentalized via an

inner membrane (IM) and outer membrane (OM) into the cytoplasm, periplasm, and extra

cellular environment. This compartmentalization facilitates efficient cellular operation by

concentrating functional components, while avoiding potentially disruptive cross talk,

similar to the system of organelles in eukaryotic cells. Focusing on the outermost

components, the IM, periplasm, and OM make up the cell envelope, a barrier which

separates bacteria from their environments, while affording efficient chemical transport,

detoxification, and cell protection. Between the two membranes is the highly viscous,

oxidizing environment of the periplasm. This compartment houses the peptidoglycan or

cell wall, consisting of a polymer mesh of glycan chains cross-linked by oligopeptides

that helps to maintain structural integrity (9, 25). Proteins are targeted to the periplasm

through type II secretion, a process whereby N-terminal signal sequences direct translated

peptides to IM transport proteins such as the Sec class of proteins (8). In the process of

translocation, the targeting signal sequences are removed from periplasmic proteins,

resulting in a mature form of the protein that can then participate in various functions

such as preventing cellular toxicity through small molecule transport, chemical

modifications, as well as participating in polymer synthesis and degradation (31).

Furthermore, periplasmic proteins interact with the IM and OM, maintaining structural

integrity and proper membrane activity.

While both the IM and OM are highly dynamic lipid bilayers with embedded

proteins and other factors, they have significant differences in their compositions which

148

contribute to their specialized roles as cytoplasmic barrier and semipermeable cell surface

layer, respectively. In the IM, both leaflets are composed of phospholipids such as

phosphatidylethanolamine, with embedded proteins that are typically α-helical bundles. It

has been estimated that as many as 20% of the open reading frames identified in E. coli

encode for IM proteins; the roles of these proteins are diverse, but relate to cellular

metabolic process such energy generation and conversion in the respiratory chain, cell

division, signal transduction, and transport processes (18). The OM however is an

asymmetric bilayer, with an inner leaflet composed of phospholipids and an outer leaflet

composed primarily of lipopolysaccharide (LPS) (33). Proteins embedded in the OM are

primarily β-barrels, and although they are encoded by only 2-3% of the genome in Gram-

negative bacteria (40), they makeup about 90% of all cellular lipoproteins (38).

Furthermore, lipoproteins are attached to the inner leaflet of the OM through post

transcriptional modifications such as acylation (19). This asymmetry of the OM

contributes to its role as a protective, semi-permeable barrier, which effectively interacts

with its environment. However, the asymmetry of the OM also poses a unique challenge

in the actual assembly of the lipid bilayer, in protein transport and insertion, as well as

effective extracellular communication and pathogenesis.

The unique feature of the OM is the essential component LPS, a distal O-antigen

polysaccharide attached to an endotoxin lipid A core (a phosphorylated glucosamine

disaccharide acylated with fatty acids). The O-antigen can be divided further into an outer

core oligosaccharide of various lengths, and an inner core sugar KDO (3-deoxy-D-

manno-oct-2-ulosonic acid) (30). While various E. coli mutants have been identified

which lack O-antigens, Lipid A and KDO have been found to be essential for proper

149

growth (31). Largely through characterization of gene knockouts, more than 50 genes

have been identified to be required for synthesis and assembly of LPS at the cell surface

(42), many of which are validated drug targets including the LpxC inhibitor L-161,240

(11, 22). Furthermore, two genes, kdsD and kdsC (formerly yrbH and yrbI, respectively)

have recently been implicated in KDO assembly (23, 41), and that their action is

genetically redundant (35).

While the genes that contribute to proper LPS assembly and transport have been

highlighted, the mechanisms of how these factors interact to actually facilitate the process

are poorly understood. The exceptions are the proteins MsbA which is responsible for

flipping LPS across the IM (27, 29), and the Imp/RlpB complex involved in LPS

targeting to the OM (42). However, the mechanism and factors involved in LPS transport

across the periplasm, through the peptidoglycan remain largely uncharacterized. Two of

the four other genes in the same operon as kdsD and kdsC, are the lipopolysaccharide

transport genes A and B (lptA and lptB, formerly yhbN and yhbG) have been implicated

in maintaining cell envelope integrity (35) and LPS transport to the OM (34). lptB is

immediately downstream of the sequence for rpoN encoding for sigma factor σ54. The

other two genes in this operon are yrbG and yrbK, which have been suggested to be

homologous to cation/calcium exchanger and IM anchored, respectively.

The lptA gene encodes for a 185 amino-acid protein that has an N-terminal 27

residue periplasmic targeting sequence and is co-transcribed with lptB (35). LptA has

been shown to be essential for growth in E. coli (32, 34), and is upregulated during

swarming in Salmonella, contributing to pathogenesis in Gram-negative bacteria (13).

Conditional gene knockouts of lptA were shown to be sensitive to hydrophobic chemicals,

150

suggesting that these mutants had defective outer membranes (35). Furthermore, mutants

depleted in LptA produce an anomalous form of LPS, resulted in GlcNAc accumulation

in a novel membrane fraction, and were defective in LPS transport to the OM (34). In

order to characterize the essential LPS assembly protein LptA, we have determined its

crystal structures in two crystal forms: one with two molecules in the asymmetric unit at

2.15 Å resolution and another at 3.25 Å resolution with eight molecules in the

asymmetric unit.

A1.3 Materials and methods

A1.3.1 Sequence analysis, protein expression and purification

Sequence analysis was carried out using Blastp (www.ncbi.nlm.nih.gov) (2) and

Multialign (http://bioinfo.genopole-toulouse.prd.fr/multalin/) (7). GeneContext was used

to highlight other proteins in the same operon as LptA (6). LptA fused to a C-terminal

His6-tag (LptA-SGRVEHHHHHH) was subcloned into a pET 21b expression vector and

transformed into BL21(DE3) cells. Cultures (1L) carrying plasmids for the recombinant

protein were grown at 37°C in Terrific Broth-ampicillin (100 µg/ml), with five times the

suggested amount of glycerol. Protein expression was induced at a culture OD600 between

1.0-1.2 using 0.1 mM isopropyl-1-thio-β-D-galactopyranoside and grown for 16 h at

16°C. LptA protein substituted with seleno-methionine was expressed in a similar fashion,

in the metA- E. coli strain DL41, in LeMaster medium supplemented with ampicillin (10).

LptA was purified via nickel nitrilotriacetate agarose batch purification in phosphate

151

buffer (pH 7.0) supplemented with 2% glycerol and dialyzed overnight in 20 mM sodium

acetate (pH 5.5), 2% (v/v) glycerol. LptA was then purified further via anion exchange,

using a 6ml Resource Q column on an ATKA Explorer FPLC in dialysis buffer, with 20

mM sodium acetate (pH 5.5), 2% (v/v) glycerol, 500 mM NaCl as the elution buffer.

During each stage, purification was monitored by SDS PAGE, and LptA was estimated to

be >95% prior to dialysis in 20 mM sodium acetate (pH 5.5), 50 mM NaCl, 2% glycerol

pure prior to concentrated using an Amicon Ultra 5 kDa cutoff centrifugal filtration

device (Fisher, Mississauga, CAN). Protein concentration was determined by BioRad

protein assay (Biorad, Mississauga, CAN) using BGG as a protein standard and used

immediately for crystallization. LptA expressed at low levels (~2.5 mg/L culture

depending on batch). A Micromass Q-TOF mass spectrometer (Mississauga, CAN) was

used to evaluate the processed form of purified LptA.

A1.3.2 Crystallization of LptA

Two different LptA crystal forms were obtained; one with two molecules in the

asymmetric unit and another with eight molecules in the asymmetric unit (see Section

A1.3.3). Crystallization of both crystal forms were obtained by hanging drop vapour

diffusion method using freshly prepared protein and reagents. Crystallization of the two-

molecule form of LptA was achieved by mixing equal parts 15 mg/ml LptA with

reservoir solution consisting of 15% (v/v) PEG 3500, 17% (v/v) glycerol, 25 mM MgSO4,

60 mM BisTris (pH 6.5), 2% Methanol. Rectangular prism crystals (2.2 x 1.0 x 1.0 mm)

appeared after 14-21 d. The eight-molecule form of LptA crystallized in similar

conditions to the two molecule form, except that reservoir solution was supplemented

152

with 5 mg/ml Ra-LPS or lauric acid. Crystals corresponding to the eight-molecule form

had a morphology that resembled an American football (1.6 x 0.9 mm) and developed

after 14-21 d. Where these crystals nucleated from, or grew into the cover-slip, the cross

section of the football shaped crystals were hexagonal in their appearance. For both two-

and eight-molecule forms, crystals were briefly soaked in reservoir solution

supplemented with 30% glycerol as a cryoprotectant, and flash-cooled in the N2 cold

stream at 100K prior to diffraction data collection.

A1.3.3 Data collection and structure determination

LptA crystals were determined to belong to the P212121 and P3221 space groups

with two and eight molecules in the asymmetric unit, respectively. The multiple-

wavelength anomalous dispersion (MAD) data sets were collected on the P212121 form at

beamline X29, Brookhaven National Laboratory. Data collection strategy involved data

collection at the peak λ first, followed by the remote λ, and then at the inflection λ , each

with 1.0° oscillations for 360°. Native two- and eight-molecule data was collected at

beamlines F1 and A1, respectively, Cornell High Energy Synchrotron source. In both

cases diffraction data was collected for 180°, using 1.0° oscillations. For both native and

MAD datasets diffraction data were processed with HKL2000 and scaled with Scalepack

(28). Heavy atom positions were determined and refined with autoSHARP, including

density modification with SOLOMON, and solvent flattening (39). Electron density and

structural solutions were attempted with many other software packages, but failed due to

a lack of suitable anisotropic correction to correctly identify and refine heavy atom

153

positions. Automatic chain tracing was performed using RESOLVE (37). The remainder

of the model was built by manual fitting using XtalView/Xfit (16). Due to the absence of

any tryptophan residues in LptA which would facilitate assigning protein sequence, the

heavy atom positions of selenomethionine were relied upon, and were later evaluated to

be quite accurately estimated by autoSHARP (39). Once a suitable model of LptA was

built using the peak wavelength dataset in the P212121 form, it was used as a molecular

replacement model to place the second molecule in the asymmetric unit prior refinement

of the final structure (21). A single LptA molecule from the high resolution refined LptA

structure was used for locate 8 molecules by molecular replace in the P3221 form.

For both native datasets, an anisotropic correction was applied via Michael

Sawaya’s Anisotropy Correction Server (www.doe-mbi.ucla.edu/~sawaya/anisoscale/)

(36), that facilitated both molecular replacement and improved the quality of the electron

density obtained. In the eight-molecule form, Phaser successfully found six LptA

molecules, and the other two were revealed by RESOLVE (37) through automatically

chain tracing. The extra two molecules revealed by RESOLVE were then used together

with the program LSQKAB (Superpose)(15) to correctly place two more LptA molecules

in the asymmetric unit. The structure of LptA was refined based on both two- and eight-

molecule datasets using CNS (5). Structural homology search using the SCOP database

(24) was performed with the program SSM (14) and protein structures were

superimposed using the CCP4 program LSQKAB (Superpose) (1, 12, 14, 15).

154

A1.4 Results and Discussion

A1.4.1 Sequence analysis

NCBI Blastp sequence analysis suggested that LptA contains sequence similarity

with the Imp and OstA domains that have been implicated in tolerance to organic

solvents (3, 42). Furthermore, LptA is highly conserved across Gram-negative bacteria,

with homologues detected in various strains of E. coli, Erwinia, Haemophilus,

Salmonella, Shigella, and Yersinia, many of which are human pathogens (Figure A1).

While LptA mutants have been shown to have defective OMs (34), LptA overexpression

was also an important obstacle to overcome. Sequence alignment suggested residues with

high homology across the species examined are present throughout the sequence of LptA

with a clustered stretch of residues extending from L45 to P77. Including residues which

have less similarity throughout the sequences examined, we can extend sequence

conservation throughout LptA, with the exception of the N-terminal 27 residues and the

C-terminal 15 residues. Furthermore, LptA was evaluated to contain an N-terminal signal

sequence (M1-A27) (20, 26), that is removed when LptA is exported into the periplasm

(17). This result was corroborated by our own mass spectrometry analysis in which

purified LptA was evaluated to have a molecular mass of 18603 Da, corresponding to a

processed recombinant LptA-His6 protein which has had the N-terminal 27 residues (M1-

A27) removed (Predicted MW= 18645 (4)) prior to crystallization (Figure A2). LptA

processing was also corroborated through size exclusion chromatography which further

suggested that LptA is a monomer in solution at pH 5.5.

155

%ID %SIM91.9 96.8

96.4 96.469.5 81.1

67.6 81.956.1 73

48.9 67.251.1 69.449.5 67.247 64.935.3 54.7

36.4 47.7

%ID %SIM91.9 96.8

96.4 96.469.5 81.1

67.6 81.956.1 73

48.9 67.251.1 69.449.5 67.247 64.935.3 54.7

36.4 47.7

Figure A.1 Sequence alignment of LptA with various Gram-negative bacterial homologues, many of which are human pathogens. Residues highlighted with red represent high homology, and those highlighted with blue represent sequence similarity. While regions with high homology are sporadic throughout protein homologous to LptA, a notable cluster of conserved residues corresponding to residues P35-K83 are evident. Numbering corresponds to LptA, with gaps. Alignment was performed with Multalign.

156

-1.00E+04

0.00E+00

1.00E+04

2.00E+04

3.00E+04

4.00E+04

5.00E+04

6.00E+04

7.00E+04

8.00E+04

18500 18550 18600 18650 18700 18750 18800 18850 18900 18950 19000

Mass (Da)

Rel

ativ

e In

tens

ity

18603

18625

18642

1866418682

18703

Figure A.2 Q-TOF Mass Spectrometry result, highlighting the LptA has been processed prior to crystallization. The predominant protein present in the sample is 18603 Da, corresponding to a processed recombinant LptA that has had its N-terminal 27 residues (M1-A27) cleaved. Four other peaks are ~20 Da apart.

157

A1.4.2 Overcoming severe anisotropy in diffraction data

In all three LptA crystals (Table 1), diffraction data was severely anisotropic,

which has the effect of varying the maximum resolution of diffraction data along

different directions (a*, b*, and c*) (Figure A3). For the MAD dataset the diffraction

limits were 4.2, 3.9, 3.2 Å along a*, b*, and c*, respectively, and in the native two-

molecule form the diffraction limits were 2.7, 2.9, and 2.1 Å along a*, b*, and c*,

respectively. In the eight-molecule form, the data was still severely anisotropy, with

diffraction limits of 3.3, 3.3, 3.9 Å along a*, b*, and c*, respectively. The effect of this

inherent anisotropy was discontinuous electron density especially for loop regions, poor

side chain density for longer residues such as lysine and arginine, and the absence of

clear density for solvent atoms. In order to overcome the problem of severe anisotropy

and correct the structure factors associated with the LptA diffraction data, an anisotropy

correction was applied (36). Effectively, this correction truncated any data outside of a

defined ellipse instead of a sphere which is normally used, the data was then scaled into

the defined ellipse, and a negative b-factor correction was applied in order to restore the

magnitude of the high resolution reflections to their original values. The structural

improvement following these corrections improved Rwork and Rfree for the two-molecule

native dataset from 36% and 40% for the uncorrected data, to 30% and 37%, respectively,

prior to solvent addition. Ramachandran analysis of LptA was also improved. A similar

improvement in structural refinement was observed for the eight-molecule form. Data

processing and refinement statistics are summarized for MAD and native crystals, as well

as structural refinement in Table 1.

158

Table A.1 Diffraction data and refinement statistics for LptA

aNumber of reflections following merging #Values in parentheses are for the outermost shell

159

Figure A.3 Anisotropic diffraction pattern from LptA MAD (left) and two-molecule native (right) crystals. Red circle highlights spherical area for scaling data and blue enclosed areas highlight diffraction which would normally not have been included. Anisotropic correction allowed for the inclusion of the higher resolution data.

160

A1.4.3 LptA protein fold and structural analysis

The overall refined structure of LptA resembles a series of eight antiparallel

β-sheets that wind back along the path of the previous peptide stretch, throughout the

length of the protein. These antiparallel sheets are folded approximately in the middle

along their length, creating an overall topology where LptA resembles a ribcage or

β-jellyroll (Figure A4), where each rib is approximately seven residues in length.

Furthermore, the structure of LptA is not completely symmetrical along the length but

opens slightly at both the N- and C-terminal ends, creating two clefts. The N-terminal

cleft is slightly larger than the C-terminal cleft, and is surrounded by the conserved

segment of residues from P35-K83 (Figure A5). Other conserved residues align along the

length of the β-sheets, implicating a structural contribution. An evaluation of the surface

electrostatics of LptA suggested a mainly negatively charged molecule, with a

pronounced cleft delineated by the N-terminal cleft. In the eight-molecule form LptA is

organized in a head-to-tail fashion, effectively forming fibers of LptA that intertwine

(Figure A6). In this case, there was an absence of density for LPS or lauric acid in the

partially refined structure of LptA. At this time it is therefore unclear whether the change

observed for LptA has been influenced due to an interaction with LPS or lauric acid.

Structural comparison of LptA against structures within the SCOP database (24) did not

reveal any significant structural matches; hence, LptA represents a previously

uncharacterized fold, although the structure of LptA does resemble a left-handed β-helix

or β-jellyroll in general.

161

N

C

C

N

N

C

C

N

C

N

N

C

C

N

N

C

90o

N

C

C

N

N

C

C

N

C

N

N

C

C

N

N

C

90o90o

Figure A.4 Ribbon diagram of LptA with two molecules in the asymmetric unit at 2.15 Å resolution. Each of the LptA molecules is presented with different colours and gaps with dotted lines. LptA has also been rotated by 90o to highlight the gap along the length of LptA.

162

90o

N

N

C

90o

N

N

C

Figure A.5 Ribbon diagram of LptA highlighting residues conserved in sequence alignment with other Gram-negative homologues. Two clefts are present at the ends of LptA due to opening of the folded β-sheets that run along its length. A notable cluster of conserved residues flank the N-terminal cleft delineated at the end of the LptA structure, corresponding to residues P35-K83.

163

Figure A.6 Ribbon diagram of LptA with eight molecules in the asymmetric unit at 3.25 Å resolution. LptA molecules are organized head to tail fashion, forming a set of intertwined fibres. There was no clear density that resembled LPS or lauric acid in the partially refined model of LptA.

164

A1.4.4 Conclusions

Here we present the crystal structure of the processed form of the periplasmic

localized, LPS transport protein LptA at 2.6 Å resolution with two molecules in the

asymmetric unit and at 3.25 Å resolution with eight molecules in the asymmetric unit in

the presence of LPS or lauric acid. The structure of LptA forms a unique set of eight

antiparallel β-sheets which open at their ends. Conserved residues are not clustered to a

particular structural location in LptA, with the exception of the N-terminal residues P35-

K83 which are clustered proximal to one of the terminal clefts in LptA. This could

potentially be a position that contributes to ligand binding or to mediate protein-protein

contact. Based on sequence homology we can extend the structural information for LptA

to various Gram-negative bacteria. Furthermore, due to the implication of LptA function

in maintaining membrane integrity and LPS transport we can extend our findings to

multiple organisms, with implications for an LptA contribution for pathogenesis.

A1.4.5 Footnotes

Thank you to Dr. Christian M. Udell for cloning LptA into the expression vector.

Thank you to Dr. Robert Sweet for organizing the Brookhaven National Laboratory

(BNL), Rapid Data collection course 2006 that ultimately led to a structural solution for

LptA. LptA has not been deposited yet in the Protein Data Bank, pending further

refinement. We are also grateful to Howard Robinson X-29 where the MAD dataset was

collected, and the staff at MacCHESS where native data for both isoforms was collected.

165

Thank you to Dr. Clemens Vonrhein for expert assistance in implementing autoSHARP,

and Dr. Michael Sawaya for useful discussions and anisotropy correction server to

improve the map quality obtained from LptA data. This work has been supported CIHR.

MS is supported by Bracken E.G. Bauman and R. Samuel McLaughlin Queen’s Graduate

Fellowships as well as OGS. ZJ is a Canada Research Chair in Structural Biology.

A1.4.6 Appendix References 1. 1994. The CCP4 suite: programs for protein crystallography. Acta Crystallogr. D.

Biol. Crystallogr. 50:760-763.

2. Altschul, S. F., T. L. Madden, A. A. Schaffer, J. Zhang, Z. Zhang, W. Miller, and D. J. Lipman. 1997. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res 25:3389-3402.

3. Aono, R., T. Negishi, K. Aibe, A. Inoue, and K. Horikoshi. 1994. Mapping of organic solvent tolerance gene ostA in Escherichia coli K-12. Biosci. Biotechnol. Biochem. 58:1231-1235.

4. Appel, R. D., A. Bairoch, and D. F. Hochstrasser. 1994. A new generation of information retrieval tools for biologists: the example of the ExPASy WWW server. Trends Biochem. Sci. 19:258-260.

5. Brunger, A. T., P. D. Adams, G. M. Clore, W. L. DeLano, P. Gros, R. W. Grosse-Kunstleve, J. S. Jiang, J. Kuszewski, M. Nilges, N. S. Pannu, R. J. Read, L. M. Rice, T. Simonson, and G. L. Warren. 1998. Crystallography & NMR system: A new software suite for macromolecular structure determination. Acta Crystallogr. D. Biol. Crystallogr. 54:905-921.

6. Ciria, R., C. breu-Goodger, E. Morett, and E. Merino. 2004. GeConT: gene context analysis. Bioinformatics. 20:2307-2308.

7. Corpet, F. 1988. Multiple sequence alignment with hierarchical clustering. Nucleic Acids Res 16:10881-10890.

8. Danese, P. N. and T. J. Silhavy. 2003. Targeting and assembly of periplasmic and outer-membrane proteins in Escherichia coli. Annual. Review of Genetics 32:59.

166

9. Fujiki, K., A. Fukuda, and Y. Okada. 1976. Amino acid composition of peptidoglycan in Caulobacter crescentus. J Biochem. (Tokyo). 80:1453-1455.

10. Gilles-Gonzalez, M. A. and G. Gonzalez. 2005. Heme-based sensors: defining characteristics, recent developments, and regulatory hypotheses. J Inorg. Biochem 99:1-22.

11. Jackman, J. E., C. A. Fierke, L. N. Tumey, M. Pirrung, T. Uchiyama, S. H. Tahir, O. Hindsgaul, and C. R. Raetz. 2000. Antibacterial agents that target lipid A biosynthesis in Gram-negative bacteria. Inhibition of diverse UDP-3-O-(r-3-hydroxymyristoyl)-n-acetylglucosamine deacetylases by substrate analogs containing zinc binding motifs. J Biol. Chem. 275:11002-11009.

12. Kabsch, W. A solution for the best rotation to relate two sets of vectors.

13. Kim, W. and M. G. Surette. 2004. Metabolic differentiation in actively swarming Salmonella. Mol. Microbiol. 54:702-714.

14. Krissinel, E. and K. Henrick. 2004. Secondary-structure matching (SSM), a new tool for fast protein structure alignment in three dimensions. Acta Crystallogr D Biol Crystallogr 60:2256-2268.

15. Krissinel, E. and K. Henrick. 2004. Protein structure comparison in 3D based on secondary structure matching (SSM) followed by Cα alignment, scored by a new structural similarity function., p. 88. In Andreas J.Kungl & Penelope J.Kungl (Eds.) (ed.).

16. Layer, G., J. Moser, D. W. Heinz, D. Jahn, and W. D. Schubert. 2003. Crystal structure of coproporphyrinogen III oxidase reveals cofactor geometry of Radical SAM enzymes. EMBO J 22:6214-6224.

17. Lopez-Campistrous, A., P. Semchuk, L. Burke, T. Palmer-Stone, S. J. Brokx, G. Broderick, D. Bottorff, S. Bolch, J. H. Weiner, and M. J. Ellison. 2005. Localization, annotation, and comparison of the Escherichia coli K-12 proteome under two states of growth. Mol. Cell Proteomics. 4:1205-1209.

18. Luirink, J., H. G. von, E. Houben, and J. W. de Gier. 2005. Biogenesis of inner membrane proteins in Escherichia coli. Annu. Rev. Microbiol. 59:329-55.:329-355.

19. Magee, T. and M. C. Seabra. 2005. Fatty acylation and prenylation of proteins: what's hot in fat. Curr. Opin. Cell Biol. 17:190-196.

20. Marino-Ramirez, L., J. L. Minor, N. Reading, and J. C. Hu. 2004. Identification and mapping of self-assembling protein domains encoded by the Escherichia coli K-12 genome by use of lambda repressor fusions. J Bacteriol. 186:1311-1319.

167

21. McCoy, A. J. 2007. Solving structures of protein complexes by molecular replacement with Phaser. Acta Crystallogr. D. Biol. Crystallogr. 63:32-41.

22. Mdluli, K. E., P. R. Witte, T. Kline, A. W. Barb, A. L. Erwin, B. E. Mansfield, A. L. McClerren, M. C. Pirrung, L. N. Tumey, P. Warrener, C. R. Raetz, and C. K. Stover. 2006. Molecular validation of LpxC as an antibacterial drug target in Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 50:2178-2184.

23. Meredith, T. C. and R. W. Woodard. 2003. Escherichia coli YrbH is a D-arabinose 5-phosphate isomerase. J Biol. Chem. 278:32771-32777.

24. Murzin, A. G., S. E. Brenner, T. Hubbard, and C. Chothia. 1995. SCOP: a structural classification of proteins database for the investigation of sequences and structures. J Mol. Biol. 247:536-540.

25. Nanninga, N. 1991. Cell division and peptidoglycan assembly in Escherichia coli. Mol. Microbiol. 5:791-795.

26. Nielsen, H., J. Engelbrecht, S. Brunak, and H. G. von. 1997. Identification of prokaryotic and eukaryotic signal peptides and prediction of their cleavage sites. Protein Eng. 10:1-6.

27. Oliver, D. B. and J. Beckwith. 1982. Regulation of a membrane component required for protein secretion in Escherichia coli. Cell. 30:311-319.

28. Otwinowski, Z. and W. Minor. 1997. [20] Processing of X-ray diffraction data collected in oscillation mode, p. 307-326. In Methods in Enzymology. Academic Press.

29. Polissi, A. and C. Georgopoulos. 1996. Mutational analysis and properties of the msbA gene of Escherichia coli, coding for an essential ABC family transporter. Mol. Microbiol. 20:1221-1233.

30. Raetz, C. R. and C. Whitfield. 2002. Lipopolysaccharide endotoxins. Annu. Rev. Biochem. 71:635-700. Epub;2001 Nov 9.:635-700.

31. Ruiz, N., D. Kahne, and T. J. Silhavy. 2006. Advances in understanding bacterial outer-membrane biogenesis. Nat. Rev. Microbiol. 4:57-66.

32. Serina, S., F. Nozza, G. Nicastro, F. Faggioni, H. Mottl, G. Deho, and A. Polissi. 2004. Scanning the Escherichia coli chromosome by random transposon mutagenesis and multiple phenotypic screening. Res. Microbiol. 155:692-701.

33. Smit, J., Y. Kamio, and H. Nikaido. 1975. Outer membrane of Salmonella

typhimurium: chemical analysis and freeze-fracture studies with lipopolysaccharide mutants. J Bacteriol. 124:942-958.

168

34. Sperandeo, P., R. Cescutti, R. Villa, B. C. Di, D. Candia, G. Deho, and A. Polissi. 2007. Characterization of lptA and lptB, two essential genes implicated in lipopolysaccharide transport to the outer membrane of Escherichia coli. J Bacteriol. 189:244-253.

35. Sperandeo, P., C. Pozzi, G. Deho, and A. Polissi. 2006. Non-essential KDO biosynthesis and new essential cell envelope biogenesis genes in the Escherichia

coli yrbG-yhbG locus. Res. Microbiol. 157:547-558.

36. Strong, M., M. R. Sawaya, S. Wang, M. Phillips, D. Cascio, and D. Eisenberg. 2006. Toward the structural genomics of complexes: crystal structure of a PE/PPE protein complex from Mycobacterium tuberculosis. Proc. Natl. Acad. Sci. U. S. A. 103:8060-8065.

37. Terwilliger, T. C. and J. Berendzen. 1999. Automated MAD and MIR structure solution. Acta Crystallogr. D. Biol. Crystallogr. 55:849-861.

38. Tokuda, H. and S. Matsuyama. 2004. Sorting of lipoproteins to the outer membrane in E. coli. Biochim. Biophys. Acta. 1694:IN1-IN9.

39. Vonrhein, C., E. Blanc, P. Roversi, and G. Bricogne. 2006. Automated Structure Solution With autoSHARP. Methods Mol. Biol. 364:215-30.:215-230.

40. Wimley, W. C. 2003. The versatile beta-barrel membrane protein. Curr. Opin. Struct. Biol. 13:404-411.

41. Wu, J. and R. W. Woodard. 2003. Escherichia coli YrbI is 3-deoxy-D-manno-octulosonate 8-phosphate phosphatase. J Biol. Chem. 278:18117-18123.

42. Wu, T., A. C. McCandlish, L. S. Gronenberg, S. S. Chng, T. J. Silhavy, and D. Kahne. 2006. Identification of a protein complex that assembles lipopolysaccharide in the outer membrane of Escherichia coli. Proc. Natl. Acad. Sci. U. S. A. 103:11754-11759.