Coupled Divalent Metal-Ion Transporter 1, DMT1

The functional properties and intestinal role of the H+-

coupled divalent metal-ion transporter 1, DMT1

A dissertation submitted to the

Graduate School

of the University of Cincinnati

in partial fulfillment of the

requirements for the degree of

Doctor of Philosophy

in the Department of Molecular and Cellular Physiology

of the College of Medicine

Ali Shawki

B.S. University of Cincinnati

October 2015

Committee Chair: Bryan Mackenzie, Ph.D.

Abstract:

Iron is an essential micronutrient in higher organisms. Despite its abundance on Earth, its bioavailability to mammals is poor. Thus, mammals have evolved specialized systems to acquire and conserve iron; however, too much iron is toxic. Iron deficiency is the most prevalent micronutrient deficiency worldwide. Deficiency of this metal results in iron-deficiency anemia, and neurological and developmental disorders in infants and children. But too much iron—such as occurs in hemochromatosis or thalassemia—results in liver cirrhosis, hepatocellular cancer, cardiomyopathy, endocrine disorders (e.g. diabetes), skin disorders, and joint disorders. This thesis examines the role of DMT1 in iron homeostasis. I tested the hypothesis that intestinal DMT1 is required for the absorption of iron, copper, manganese, and zinc by examining metal metabolism and directly measuring intestinal absorption of radiotracer metals in the DMT1int/int mouse. I demonstrate here that DMT1 serves as the primary or only gateway for the absorption of nonheme iron from the diet, but that DMT1 is not required for the absorption of copper, manganese, or zinc. DMT1 is driven by the proton electrochemical gradient in vitro; however, the provenance of the protons in vivo is unknown. I explored the role of the brush-border Na+/H+ exchangers in iron absorption and metabolism. I demonstrate that the activity of DMT1 relies on the acidic microclimate at the brush border of the small intestine generated by Na+/H+ exchanger-3. The critical roles of DMT1 are evident in animal models and human probands bearing DMT1 mutations that lead to severe iron-deficiency anemia. I demonstrate that human mutations in DMT1, in general, result in impaired iron-transport activity of DMT1. My analysis of the impact of DMT1 mutations helps to explain the anemia phenotype of the probands but also uncovers novel aspects of the molecular physiology of DMT1. Dysregulation of iron sensing can lead to upregulation of DMT1 and consequently iron overload (e.g. hereditary hemochromatosis, thalassemia, sickle-cell disease). DMT1 is a validated target in the treatment of iron overload. The outcome of this work will help drive the development of therapies for improved iron nutrition and the treatment or prevention of iron overload.

iii

Preface:

I am greatly indebted to the training, support, and guidance of my mentor, Bryan Mackenzie. Without his trust, patience, and belief in me, none of the work presented in this thesis would have been possible. Although Bryan was my mentor, we had a strong friendship that I believe is an important quality in a successful PhD tenure. I am also grateful to my committee members, Hamid Eghbalnia, PhD, Tomas Ganz, PhD, MD, Rohit Kohli, MBBS, MS, Anil Menon, PhD, and Gary E. Shull, PhD for their guidance, support, and feedback throughout my study.

I am greatly indebted to the guidance from my father (Araz) who was always willing to share with me his wisdom and also the unconditional support and encouragement from my mother (Amel), sister (Sarwah), and my brother (Amin).

I would like to thank my fiancé, Gwendolyne Rodriguez for providing me with ample nutrition, love, patience and unbounded support. I am incredibly grateful to her family for their compassion and generosity.

I would like to thank my friends for remaining as my friends during the times I spent in solitude.

Last but not least, I would like to express my extreme gratitude to all the professors who provided me with an enriched educational experience and guided me to become the successful scientist that I am today.

Acknowledgements:

I thank the following colleagues for helpful discussions:

Nancy C. Andrews, MD, PhD (Duke University)

François Canonne-Hergaux, PhD (INSERM, Toulouse, France)

Supak Jenkitkasemwong, PhD (University of Florida)

Theodosia A. Kalfa, MD, PhD (University of Cincinnati and Cincinnati Children’s Hospital)

Mitchell D. Knutson, PhD (University of Florida)

John N. Lorenz, PhD (University of Cincinnati)

Jack Rubenstein, MD (University of Cincinnati)

Roger T. Worrell, PhD (University of Cincinnati)

Brian K. Sparkman, MD (University of Cincinnati)

I thank the following colleagues for their generous gifts as noted:

Nancy C. Andrews, MD, PhD (Duke University) for providing us with the floxed DMT1 mouse line

Robert E. Fleming, PhD (St. Louis University School of Medicine) for providing us with the HFE-null mouse line

Elizabeta Nemeth, PhD and Bo Qiao, MD (University of California Los Angeles) for sharing with us the EGFP labeled DMT1 construct

National Cancer Institute Mouse Repository for providing the villin–Cre transgenic mouse line

I thank the following colleagues for their technical contributions and other assistance in the laboratory:

John H. Alexander, MD Robert S. Kim, MD

Avni A. Amratia, BS Patrick B. Knight, MD

Sarah R. Anthony, BS Colin J. Mitchell, BS

Rusty A. Baik, MD Eric J. Niespodzany, MD

Tomasa Barrientos, PhD Yasuhiro Nose, PhD

Emily M. Bradford, PhD Helena Öhrvik, PhD

Melinda A. Engevik, PhD Yogi Patel, BS

Anthony C. Illing, MD Shahana Prakash

Gianna Katsaros, MS Theodore A. Ruwe, BS

Specific contributions of coauthors are noted where appropriate.

The financial support of the following is greatly appreciated:

PHS Grant R01-DK080047 (to B. Mackenzie) from the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK)

University of Cincinnati Department of Molecular and Cellular Physiology

Albert J. Ryan Foundation – Predoctoral Fellowship to A. Shawki

PHS Grant P30-DK078392 (to Digestive Health Center, University of Cincinnati and Cincinnati Children’s Hospital Medical Center) from the NIDDK

The production of copper-64 at Washington University–St Louis School of Medicine is supported by the U.S. Department of Energy (DoE), Nuclear Physics Isotope Program

The content of this work is solely the responsibility of the author and does not necessarily represent the official views of the NIDDK, National Institutes of Health, or Department of Energy.

Table of Contents:

1 General Introduction ...... 1

1.1 Iron ...... 1

1.1.1 History ...... 1

1.1.2 Elemental iron ...... 1

1.1.3 Application of iron ...... 1

1.2 Roles of iron in mammals ...... 2

1.2.1 Redox state of iron ...... 2

1.2.2 The erythron—the largest iron-containing compartment ...... 3

1.2.3 Enzymes and other iron-containing proteins ...... 4

1.3 Systemic iron homeostasis ...... 6

1.3.1 Overview of systemic iron homeostasis ...... 6

1.3.2 Regulation of systemic iron homeostasis ...... 10

1.4 Divalent metal-ion transporter-1 ...... 12

1.4.1 History of DMT1 ...... 12

1.4.2 DMT1 isoforms ...... 13

1.4.3 Tissue and cellular distribution of DMT1 ...... 16

1.4.4 Functional properties of alternative splices of DMT1 ...... 16

1.4.5 Substrate profile and physiological roles of DMT1 ...... 16

1.4.6 Regulation of DMT1 expression ...... 19

1.5 Iron disorders ...... 22

1.5.1 Iron-deficiency anemia ...... 22

1.5.2 Iron overload...... 23

1.5.3 Targeting DMT1 in iron overload disorders ...... 25

2 Role of intestinal DMT1 in metal homeostasis ...... 27

2.1 Abstract ...... 27

2.2 Introduction ...... 28

vii

2.3 Materials and methods ...... 29

2.4 Results ...... 33

2.4.1 Intestine-specific DMT1 knockout mouse model ...... 33

2.4.2 Intestine-specific knockout of DMT1 produces a severe hypochromic–microcytic anemia resulting from iron deficiency...... 35

2.4.3 Effect of DMT1 intestinal knockout on metal homeostasis in mice ...... 41

2.4.4 Effect of intestinal DMT1 ablation on the regulation of genes involved in iron absorption and homeostasis ...... 43

2.4.5 Role of intestinal DMT1 in the absorption of iron, copper, and manganese ...... 45

2.4.6 Haplosufficiency of intestinal DMT1 in mice ...... 47

2.4.7 Role of intestinal DMT1 in the neonate ...... 48

2.5 Discussion ...... 50

2.5.1 Intestinal DMT1 is required for mammalian iron absorption and homeostasis ...... 50

2.5.2 Intestinal DMT1 is not required for copper, manganese, or zinc absorption ...... 52

2.5.3 Model of severe iron deficiency and iron-deficiency anemia ...... 53

2.5.4 How is the DMT1int/int mouse able to survive for up to 5 months? ...... 55

2.6 Conclusions ...... 56

3 Roles of intestinal H+-transporters in driving iron absorption ...... 57

3.1 Abstract ...... 57

3.2 Introduction ...... 58

3.3 Materials and methods ...... 59

3.4 Results ...... 63

3.4.1 DMT1-mediated iron transport in vitro is activated at low pH ...... 63

3.4.2 Iron metabolism and homeostasis in NHE2-null (NHE2–/–) mice ...... 63

3.4.3 Iron metabolism and homeostasis in NHE3-null (NHE3–/–) mice ...... 65

3.4.4 Ablation of NHE3 but not NHE2 severely impairs intestinal iron absorption ...... 68

3.4.5 Rescue of the iron-overload phenotype of the hemochromatosis mouse ...... 69

3.5 Discussion ...... 70

viii

3.5.1 Provenance of the acidic microclimate ...... 71

3.5.2 Prevention or treatment of iron overload ...... 71

4 Exploring the molecular physiology and structure-function of DMT1 ...... 73

4.1 Abstract ...... 73

4.2 Introduction ...... 74

4.3 Materials and Methods ...... 76

4.4 Results ...... 79

4.4.1 N-linked glycosylation sites...... 79

4.4.2 Chemical modification of critical residues in DMT1 ...... 80

4.4.3 Functional analysis of human mutations in DMT1 ...... 85

4.5 Discussion ...... 93

4.5.1 Mapping critical structural elements in DMT1 ...... 94

4.5.2 Functional analysis of human mutations in DMT1 ...... 96

5 General Discussion ...... 98

5.1 Physiological roles of DMT1 ...... 98

5.1.1 Is DMT1 required for intestinal iron absorption in the neonate? ...... 98

5.1.2 Can we target NHE3 in the prevention of iron overload? ...... 99

5.1.3 What is the role of DMT1 in heme absorption? ...... 99

5.1.4 How does DMT1 contribute to Cd absorption? ...... 100

5.1.5 What is the role of DMT1 in the kidney in health and disease? ...... 100

5.1.6 What are the roles of DMT1 in the brain in health and disease? ...... 101

5.2 Molecular physiology of DMT1 ...... 102

5.2.1 Site-directed mutagenesis and functional analysis ...... 102

5.2.2 Structure-function analysis of DMT1 ...... 103

List of tables, figures and equations:

Figure 1.1 ...... 2

Equation 1.1...... 2

Figure 1.2 ...... 4

Figure 1.3 ...... 7

Figure 1.4 ...... 8

Figure 1.5 ...... 10

Figure 1.6 ...... 11

Figure 1.7 ...... 14

Table 1.1 ...... 15

Figure 1.8 ...... 18

Figure 1.9 ...... 19

Figure 1.10 ...... 20

Figure 1.11 ...... 23

Equation 2.1 ...... 31

Table 2.1 ...... 34

Figure 2.1 ...... 35

Figure 2.2 ...... 37

Figure 2.3 ...... 38

Figure 2.4 ...... 39

Figure 2.5 ...... 40

Figure 2.6 ...... 42

Figure 2.7 ...... 43

Figure 2.8 ...... 44

Figure 2.9 ...... 46

Figure 2.10 ...... 48

Figure 2.11 ...... 49

Figure 2.12 ...... 50

Figure 2.13 ...... 56

Equation 3.1 ...... 60

Equation 3.2 ...... 61

Figure 4.1 ...... 75

Equation 4.1 ...... 77

Equation 4.2 ...... 78

Equation 4.3 ...... 78

Equation 4.4 ...... 78

Figure 4.2 ...... 80

Table 4.1 ...... 81

Figure 4.3 ...... 82

Figure 4.4 ...... 84

Figure 4.5 ...... 85

Table 4.2 ...... 86

Figure 4.6 ...... 87

Figure 4.8 ...... 90

Figure 4.9 ...... 91

Figure 4.10 ...... 93

Table 4.3 ...... 95

Table A.1 ...... 129

Equation A.1 ...... 129

Abbreviations:

ANOVA Analysis of variance

BBMV Brush border membrane vesicles

CBC Complete blood count

Crp Ceruloplasmin

Ctr1 Copper transporter-1 (SLC31A1)

Cybrd1 Cytochrome-B reductase-1 (DcytB)

DEPP N′,N′-diethylpiperazine

DMT1 Divalent metal-ion transporter-1 (DCT1, NRAMP2, SLC11A2)

FDR False discovery rate

Fpn Ferroportin (IREG-1, SLC40A1)

Hepc-Fpn axis Hepcidin-ferroportin axis

GFP Green fluorescent protein

Hepc Hepcidin (Hamp1)

HFE Hemochromatosis gene

Hgb Hemoglobin

HH Hereditary hemochromatosis

Description Comments Mutation OMIM*

Hemochromatosis type 1 ‘Classic’ HFE 235200

Hemochromatosis type 2A ‘Juvenile’ Hemojuvelin (HJV) 602390

Hemochromatosis type 2B ‘Juvenile’ Hepcidin (Hamp) 606464

Hemochromatosis type 3 TfR2 604250

Hemochromatosis type 4 Autosomal dominant Ferroportin 604653

Atransferrinemia Transferrin 209300

Neonatal hemochromatosis Unknown

*www.omim.org

Hif2α Hypoxia inducible factor 2 alpha

HO Heme oxygenase

xii

ICP-MS Inductively coupled plasma mass spectrometry

I.P. Intraperitoneal

IRE Iron responsive element

IRP Iron regulatory protein

S . Half-maximal concentration for substrate S MCH Mean corpuscular (cell) hemoglobin

MCHC Mean corpuscular (cell) hemoglobin concentration

MCV Mean corpuscular (cell) volume

MES 2-(N-morpholino)ethanesulfonic acid

NHE2 Na+/H+ exchanger isoform 2 (SLC9A2)

NHE3 Na+/H+ exchanger isoform 3 (SLC9A3)

NRAMP2 See DMT1

NTA Nitrilotriacetic acid

NTBI Nontransferrin bound iron qPCR Quantitative real-time polymerase chain reaction

PCR Polymerase chain reaction

RBC count Red blood cell count

ScaDMT Staphylococcus capitis divalent metal ion transporter

SI Serum iron

SLC9A2 See NHE2

SLC9A3 See NHE3

SLC11A2 See DMT1

SLC31A1 See Ctr1

SLC40A1 See Fpn

SOE Splicing by overlap extension

Steap2 Six transmembrane epithelial antigen of prostate 2

Tf Transferrin

Tfsat Transferrin saturation, i.e. that portion of serum transferrin bound to iron

xiii

TfR1 Transferrin receptor 1

TM Transmembrane domain

UIBC Unsaturated iron binding capacity

UTR Untranslated region

xiv

1 General Introduction

1.1 Iron

1.1.1 History

Iron is the sixth most abundant metal in the universe, and the most abundant metal on Earth, making up approximately 5% of Earth’s crust and 35% of the Earth’s core almost entirely as iron-nickel alloys (source: Wikipedia). Our words for iron come from the Proto-Germanic precursor ‘isarnan’, Greek ‘sideros’, and Latin ‘ferrum’. The use of iron by man dates back to ancient times with its use as weapons and tools; however, the interaction of pure iron with oxygen and moisture causes the metal to rust and erode. Today, iron—mixed with other metals to prevent it from rust and degradation—is used in almost every metallurgic reaction because of its malleable yet strong chemical nature. One such example is steel which contains iron and is useful for reinforcing concrete, making automobiles, and stainless steel kitchenware.

1.1.2 Elemental iron

Iron in its elemental form is a bright white metal that is strong yet malleable. Iron (Fe), a transition metal with atomic number 26 and atomic weight 55.845 consists of 4 stable isotopes. The major of the four stable isotopes found on earth is iron-56. The electron configuration of iron allows for the metal to exhibit many characteristics (e.g. high electrical conductivity and magnetism) and take part in many chemical reactions (e.g. electron transfer). Although complex, and not the focus of this thesis, one important property of iron is termed ‘magnetism’. The partial oxidation of ferrous oxide results in magnetite and this substance exhibits magnetic properties. That molten iron accounts for much of Earth’s core explains the magnetic field around earth that serves as a protective shield from cosmic rays of the sun.

1.1.3 Application of iron

Iron can readily accept or donate an electron making it highly versatile in industrial applications as well as advantageous for humans; however, this same chemical property of iron can lead to corrosion and harmful oxidative damage, and the mammalian cell is no exception.

Iron is used in the treatment of cesium-137 poisoning; chelation of cesium-137 poisoning by Radiogardase—also known as Prussian blue or potassium ferrous ferricyanide used to stain for iron deposits in tissues—relies on its strong binding affinity to cesium. Iron is critical for many processes and its role in mammals will be discussed below.

1.2 Roles of iron in mammals

1.2.1 Redox state of iron

Iron is the most abundant transition metal on Earth; however, only trace amounts of this metal are found in animal cells. Iron is an invaluable trace element because of its ability to easily accept or donate an electron depending on its oxidation state (Figure 1.1).

2+ 3+ Fe Fe

– Figure 1.1. e Iron can readily undergo what is known as the Fenton reaction generating highly reactive and damaging oxygen species under uncontrolled conditions. Ferrous ion (Fe2+) and ferric ion (Fe3+) serve as

catalysts for the Haber-Weiss reaction (Equation 1.1). The reduced metal reacts with H2O2, a byproduct of aerobic respiration, to form HO– and the highly reactive hydroxyl radical OH• to form ferric (Fe3+) ion. The • reverse reaction reducing ferric ion to ferrous ion results in the formation of O (Equation 1.1). The net result of the forward and reverse reaction is described by the Fenton reaction (Equation 1.1).

Haber-Weiss reactions:

• Fe +HO →Fe +OH + OH

• O +Fe →O +Fe Fenton reaction:

• • O +HO →O +OH + OH Equation 1.1.

Iron is naturally found in its oxidized form; however, the solubility of the free form of ferric ion in aqueous solution at physiological pH is low, thereby reducing its bioavailability, and absorption from plant based foods is insufficient (Ganz, 2013; Andrews & Schmidt, 2007). Conversely, iron found in hemoglobin and myoglobin is more readily bioavailable but its dietary availability is socioeconomically limited (Ganz, 2013). That iron is an abundant metal in nature but not readily bioavailable for higher organisms highlights the requirement for a tightly controlled homeostatic system. Life has evolved systems for obtaining iron, transporting iron, storing iron, and conserving iron. Hereditary hemochromatosis (HH)—most commonly resulting from mutations in the hemochromatosis gene (HFE)—may have flourished at a time when iron was extremely low in the diet so it may have conferred an advantage; however, in individuals on a present- day western diet, the gene results in iron overload and toxicity.

Sections 1.2.2, 1.3.1.1–1.4.6, 1.5.1.1, 1.5.2, and 1.5.3 contain both my original work and excerpts from

the following co-authored publication Shawki A, Knight, PB, Maliken, BD, Niespodzany EJ, Mackenzie

B (2012) H+-coupled divalent metal-ion transporter-1: Functional properties, physiological roles and

therapeutics. Curr Top Membr. 70, 169-214.

1.2.2 The erythron—the largest iron-containing compartment

Iron is a critical micronutrient in higher organisms. The total amount of iron found in most healthy humans is 3–4 g (Ganz, 2013). In healthy individuals, the bulk of the iron (2–3 g) cycles within the erythron, which is made up of red blood cells, splenic macrophages, and erythroid marrow. Most of the iron within the erythron is bound by heme, the prosthetic groups of hemoglobin in red blood cells. Blood plasma contains only one thousandth of the total body iron and most of that is bound by transferrin (Tf), the major iron carrier in the blood. Iron can enter the erythroid precursor via the transferrin receptor-mediated endocytic pathway (Harding et al., 1983) (Figure 1.2). Transferrin receptor-1 (TfR1) binds monoferric or diferric transferrin and the complex is internalized by clathrin-coated pits into recycling endosomes (Harding et al., 1983). The V-type H+-ATPase (V-ATPase) is responsible for acidifying the endosome, an event that permits the dissociation of Fe3+ from the transferrin–transferrin receptor complex (Saroussi & Nelson, 2009). Fe3+ is reduced to Fe2+ by the ferrireductase Steap3 (Ohgami et al., 2005; Knutson, 2007), and is then transported out of the endosome via the divalent metal-ion transporter-1, DMT1 (Figure 1.2). Iron is then transported into mitochondria by poorly-understood mechanisms that may include mitoferrin (Mfrn) (Shaw et al., 2006) (Figure 1.2) and an isoform of DMT1 (Wolff et al., 2014). Within the mitochondrion, iron is incorporated into heme for the production of hemoglobin. TfR1 is then recycled to the plasma membrane. This same Tf cycle also accounts for Fe uptake into most other cell types under physiological conditions. The average lifespan of a red cell is approximately 120 days (Bessis & Weed, 1973; Shemin & Rittenberg, 1946). Senescent red blood cells are taken up by macrophages of the spleen to be digested. Hemoglobin recycled from red blood cells can be further processed and digested liberating the iron from heme. The recycling of iron from red blood cells provides the body with a means to maintain iron levels without the need to absorb large amounts of the metal. Since there exists no regulated mechanism for

the excretion of iron, iron homeostasis has to be achieved by tightly regulating dietary iron absorption and thereby controlling whole-body iron status.

Figure 1.2. Transferrin-receptor mediated endocytosis in the erythrocyte. Iron transport in erythroid

2+ precursors comprises endocytosis of diferric transferrin (Fe2Tf ) and DMT1-mediated Fe export from acidified endosomes into the cytosol (TfR transferrin receptor, V-ATPase vacuolar H+-ATPase). Reproduced from (Shawki et al., 2012), redrawn from (Mackenzie & Hediger, 2004).

1.2.3 Enzymes and other iron-containing proteins

Iron is essential for scores of critical metabolic enzymes and other cellular processes which I discuss in this section.

1.2.3.1 Iron-storage and iron-carrier proteins

Iron storage and carrier proteins maintain iron in a relatively unreactive form both in the blood plasma and at the cellular level. Ferritin—The cytoplasmic ferroxidase, ferritin is the primary and principal cellular iron storage protein converting cytoplasmic Fe2+ into an oxidized Fe(III) yet unreactive mineral form. Serum ferritin plays a very minor role in the storing or shuttling of iron; however, serum ferritin levels correlate closely with iron stores (Munro & Linder, 1978) and is therefore a useful clinical indicator of iron stores avoiding the need for more invasive measurements.

Transferrin—The major iron carrier in blood plasma is transferrin, a 75-kDa protein that binds 1 or 2 Fe(III) molecules and is responsible for iron delivery to most tissues via TfR1-mediated uptake (Kozyraki et al., 2001).

Metalloenzymes—Non-heme iron can be found in metallofactors such as iron/sulfur (Fe/S) clusters and flavoproteins. Heme iron is found in cytochromes and the prosthetic group of heme, iron protoporphyrin IX. Fe/S clusters and heme are primarily and exclusively made in mitochondria (Sheftel et al., 2009). Iron containing metalloproteins from both classes are found in different locations of the electron-transport chain of the mitochondria. Aconitase, another Fe/S cluster protein, is critical to the catalysis of citrate to isocitrate in the tricarboxylic acid (TCA) cycle (Ruud, 1954). Failure to properly assemble cytochromes and/or Fe/S proteins is associated with severe and frequently fatal respiratory chain defects and neurodegenerative, metabolic, or hematological diseases, respectively (Barrientos et al., 2002; Lill et al., 2012).

Globins

The final step of heme synthesis occurs in the mitochondria where ferrochelatase catalyzes the reaction incorporating iron into the protoporphyrin IX ring (Sheftel et al., 2009). Heme containing proteins (i.e. nitric oxide synthases, NADPH oxidase and myeloperoxidase, and cytochromes) participate in cellular signaling, host defense, and electron transport. The oxygen-binding globins hemoglobin and myoglobin are absolutely required for normal oxygen delivery to all the tissues of the body and storage of oxygen in skeletal muscle.

Myoglobin—The primary role of myoglobin in skeletal muscle and cardiac tissue is the storage and delivery of oxygen to generate ATP from oxidative phosphorylation. Myoglobin serves as the oxygen storage protein in muscle (Ordway & Garry, 2004). A single polypeptide chain of 154 amino acids, myoglobin is a cytoplasmic hemoprotein expressed strictly in cardiac myocytes and oxidative skeletal muscle (Ordway & Garry, 2004). Myoglobin binds oxygen in a reversible fashion; however, oxygen binding to myoglobin is limited by the existence of only one heme group per molecule.

Myoglobin knockout mouse models exhibit no phenotype even when stressed with exercise or hypoxic conditions (Garry et al., 1998). That the myoglobin knockout mouse functions normally under the tested conditions does not explain the conservation of this protein throughout species. Survival of the myoglobin knockout mouse is made possible by (i) the existence of some robust system in which oxygen delivery to muscle is maintained even under severe stress, and (ii) adaptation by compensatory mechanisms such as nitric oxide induced vasodilation and the ability to utilize dissolved oxygen in the

blood via passive diffusion (Garry et al., 1998). Muscle iron is preserved even in cases of iron deficiency and only under chronic severe cases of iron deficiency anemia will muscle iron content become depleted (Beutler, 2010; Shawki et al., 2015) consistent with an important role for iron in muscle myoglobin.

Hemoglobin—Hemoglobin is a multi-subunit globular protein that is composed of four heme groups. Each heme group contains one iron molecule in the center. Like myoglobin, the final step in the synthesis of heme requires the incorporation of iron into the protoporphyrin ring of heme. Hemoglobin participates

in the exchange of CO2 and oxygen. The iron-containing heme molecules in hemoglobin bind oxygen and

carbon dioxide in a reversible manner, allowing for the exchange of CO2 for O2 in the lungs and vice versa in the tissues. That iron in myoglobin and hemoglobin is a critical component in the delivery of oxygen to maintain living tissue in higher organisms is a prime example of its indispensable role in biology.

1.3 Systemic iron homeostasis

The French physiologist Claude Bernard was a pioneer of the study of homeostasis, a central concept in the discipline of physiology. The constancy of the environment presupposes a perfection of the organism such that external variations are at every instant compensated and brought into balance. In consequence, far from being indifferent to the external world, the higher animal is on the contrary in a close and wise relation with it, so that its equilibrium results from a continuous and delicate compensation established as if the most sensitive of balances (Bernard et al., 1974).

Bernard described a phenomenon vital to the function of all organisms. Homeostasis is the balance of nature, chemical reactions, and the response to changes in the external and/or internal environments of the human body. The aim of this thesis is to focus on the systemic homeostasis of one of the most critical micronutrients in humans—iron.

1.3.1 Overview of systemic iron homeostasis

Iron homeostasis in higher organisms is achieved at the systemic level and at the cellular level depending on iron status. There are three major cell types that participate in systemic iron homeostasis: enterocytes, hepatocytes, and macrophages (both splenic macrophages and Kupffer cells of the liver).

1.3.1.1 Enterocytes

Dietary iron is present in two main forms, heme iron and nonheme iron. Dietary heme comes from the proteolytic digestion of hemoglobin and myoglobin in the lumen, and heme is thought to be taken up intact by enterocytes (Conrad et al., 1967; Andrews, 2005) (Figure 1.3). It is believed that iron is liberated from heme by the endosomal heme oxygenase (Raffin et al., 1974). How iron is transported from the heme-containing endo–/lysosome to the cytoplasm is not known (see Discussion section 5.1.3). Whereas bioavailability of heme iron exceeds that of nonheme iron, an individual on a typical western, mixed diet will absorb about 10% of the iron consumed.

Figure 1.3. Model of intestinal iron transport. Fe3+ is reduced to Fe2+ by luminal ascorbic acid or by apical- membrane ferrireductases (including Cybrd1, also known as DcytB). I postulate that an acidic microclimate generated by Na+/H+ exchanger-3 (NHE3) provides the proton-motive force for DMT1-mediated Fe2+ transport into the enterocyte (see Chapter 3). Redrawn from (Mackenzie & Garrick, 2005).

Nonheme iron is comprised of iron salts and organic complexes but is poorly bioavailable. Those on a vegetarian diet may absorb as little as 2% of the iron they consume (Ganz, 2013). Gastric acid plays some modest role in iron absorption by promoting the solubility of Fe(III) complexes (Golubov et al., 1991; Mackenzie & Garrick, 2005); however, we have evidence that gastric acid is not required for iron homeostasis (Engevik MA, Shawki A, Anthony SR, Baik RA, Kim RS, Worrell RT, Shull GE, Mackenzie B, manuscript in preparation). Ascorbic acid, secreted in the stomach, promotes the formation of soluble complexes and promotes the reduction of Fe(III) to Fe(II) (Davidsson, 2003). Iron that is not reduced by luminal ascorbic acid must be reduced by the surface ferrireductase cytochrome b reductase-1 (Cybrd1,

also known as duodenal cytochrome b, DcytB) (Latunde-Dada et al., 2002) prior to its uptake by enterocytes, chiefly in the duodenum (Figure 1.3).

Figure 1.4. Systemic iron homeostasis. In healthy individuals, the bulk of the iron cycles through the erythron and recycled mainly by splenic macrophages. Under normal conditions, there is less than 1 gram of iron found in tissues such as the liver, which is the primary site of iron storage, spleen, and other tissues. There is no regulated mechanism for the excretion of iron, so iron homeostasis has to be achieved by tightly regulating intestinal iron absorption. Reproduced from (Ganz, 2013).

Approximately 1–2mg of iron is absorbed in the small intestine per day (Ganz, 2013) (Figure 1.4). Once iron is absorbed into the enterocyte, the metal can be stored in the cytoplasmic iron storage protein, ferritin (Harrison & Arosio, 1996) (Figure 1.3). Iron can then exit the cell via the basolateral iron export protein, ferroportin (Fpn) (Ganz, 2013; Abboud & Haile, 2000; McKie et al., 2000; Mackenzie & Garrick, 2005) (Figure 1.3). It is thought that the ferroxidase hephaestin is an electron acceptor associated with the basolateral export of Fe2+ by Fpn before handing off Fe3+ to Tf (Figure 1.3). Accumulation of iron in the

8 enterocyte in the copper-deficient sla mouse that harbors defective hephaestin demonstrates the requirement of hephaestin for the exit of iron from the enterocyte (Vulpe et al., 1999). It is interesting to note that most of the hephaestin protein, however, localizes to perinuclear vesicles rather than the basolateral membrane. This may simply reflect a regulatory redistribution, but it seems worth exploring whether hephaestin could also be involved in shuttling iron to Fpn. Ceruloplasmin, a homologue of hephaestin, is another membrane-bound or plasma ferroxidase involved in iron export from the liver and other tissues (Wessling-Resnick, 2006). Patients and animal models with defective ceruloplasmin develop iron overload in liver and other tissues (Cherukuri et al., 2005) indicating that ceruloplasmin is required for proper iron export similarly to the role of hephaestin in the intestine.

Transferrin, the major iron carrier in the blood, readily accepts iron delivered to the serosal side (Figure 1.3). Dysregulation of iron sensing, or conversely, iron absorption leads to iron overload or iron deficiency and iron deficiency anemia, respectively. That the bioavailability of dietary non-heme iron is limited suggests that higher organisms must tightly regulate the absorption and recycling of the metal.

1.3.1.2 Hepatocytes

Newly absorbed iron enters the Tf-bound iron pool in the blood where it is directed to the liver through the hepatic portal vein. Iron can enter the liver by two mechanisms: (i) Tf transferrin-receptor mediated endocytic pathway (Zhao et al., 2010), and (ii) non-Tf bound iron uptake via the Zrt-/Irt-like transporter- 14 (ZIP14) (Liuzzi et al., 2006). Iron in the hepatocyte is stored in the intracellular iron storage protein, ferritin. Hepatocytes are also able to sense iron levels and produce the systemic iron-regulatory hormone hepcidin (Hepc), the major effector of iron homeostasis (Ganz, 2013) (See section 1.3.2.1) (Figure 1.4). Hepatocytes and hepatic macrophages, also known as Kupffer cells, are able to store iron in ferritin as needed. Iron exit from the hepatocyte or Kupffer cell into the plasma occurs via the extracellular iron export protein Fpn, in a similar fashion to export from enterocytes and macrophages (Ganz, 2013; Ramey et al., 2010) (Figure 1.4).

1.3.1.3 Macrophages (splenic and Kupffer cells)

The blood Tf bound iron pool consists of newly absorbed iron and recycled iron from senescent red blood cells filtered through the spleen to provide approximately 3–4 grams of iron to the erythron for the generation of new red blood cells (Figure 1.4). Hemoglobin from digested red blood cells within the

9 phagosome/early endosome of the macrophage (principally splenic) is the source of heme iron (Figure 1.5). Iron is liberated from heme and transported into the cytoplasm by DMT1 where the metal can either be stored in ferritin or transported out of the cell into the blood circulation by Fpn (Ganz, 2013) (Figure 1.5). Iron export from enterocytes, hepatocytes, and macrophages via Fpn is regulated by the liver produced hormone Hepc (Nemeth et al., 2004b; Ramey et al., 2010) (Figure 1.3).

Figure 1.5. Recycling of iron in the phagosome (macrophage). Macrophages digest senescent erythrocytes and recycle the iron contained in them. The heme regulatory gene-1 (HRG1) is capable of transporting heme out of the phagolysosome. It is unknown if heme oxygenase (HO), which may be cytoplasmic or endosomal, then liberates iron from heme in the cytoplasm or at the plasma membrane of the endosome. The V-ATPase acidifies the heme-containing endosome before DMT1 can transport iron out of the endo– /lysosome. Ferroportin exports recycled iron to the blood plasma from the macrophage. Ceruloplasmin (Crp) then oxidizes the iron before the iron is bound to transferrin (Tf). Redrawn from (Mackenzie & Hediger, 2004).

1.3.2 Regulation of systemic iron homeostasis

1.3.2.1 Hepcidin

Hamp1 is the gene that encodes for the liver produced hormone hepcidin (Hepc) (Park et al., 2001). Hepc, initially translated into an 84 amino acid propeptide before amino-terminal processing (Park et al., 2001),

10 is a 25-AA peptide hormone produced in the liver. The N-terminal of Hepc, known to interact with its receptor Fpn, and the four disulfide bonds are highly conserved sequences (Ganz, 2013) (Figure 1.6).

Figure 1.6. Amino acid and protein structure of full length hepcidin. Hepcidin is a 25-amino acid propeptide with four disulfide bonds between the following amino acids (also labeled yellow): C11 with C19, C10 with C13, C22 with C14, and C7 with C23. The pink shaded region is known to bind to Fpn (Preza et al., 2011). Reproduced from (Ganz, 2013).

Synthesis of Hepc is controlled at the transcriptional level by several factors. Inflammation and acute infection upregulate Hepc synthesis, primarily an effect of interleukin-6 (IL6R–JAK–STAT3 pathway) (Lee et al., 2005; Pietrangelo et al., 2007; Wrighting & Andrews, 2006; Nemeth et al., 2004a). This largely explains the anemia of inflammation. Regulation of Hepc by liver iron stores is not well understood but involves the bone morphogenic protein receptor BMPR (and the SMAD pathway) (Andriopoulos et al., 2009). Regulation by plasma iron involves transferrin receptor-2 (TfR2) in a sort of ‘sensing’ role, and also the hemochromatosis-related protein HFE (Goswami & Andrews, 2006), hemojuvelin and TMPRSS6 (matriptase-2) (Rapisarda et al., 2010). Hepc has two major targets—macrophages recycling iron (20–25 mg/day) and enterocytes absorbing iron (1–2 mg/day) (Ganz, 2013) (Figure 1.4).

1.3.2.2 Ferroportin

Ferroportin (Fpn) is the only known cellular iron-export protein. It is critical for the release of iron from enterocytes, hepatocytes, and macrophages (Ganz, 2013). Ferroportin binds hepcidin, a step that triggers the internalization and degradation of Fpn (Nemeth et al., 2004b). Hepcidin binding to ferroportin leads to lysine phosphorylation of Fpn, internalization of Fpn in clathrin-coated pits, and its subsequent dephosphorylation (Ross et al., 2012). Ferroportin is then ubiquitinated and targeted for degradation in the late endo–/lysosome (Qiao et al., 2012). Loss of Fpn from the plasma membrane leads to decreased iron recycling by macrophages or decreased intestinal iron absorption and, subsequently, a decrease in

the plasma Fe2Tf (holotransferrin) pool (Ganz, 2013). Mutations in Fpn or Hepc that disrupt the hepcidin- ferroportin axis (Hepc-Fpn) result in disorders in iron metabolism (Ganz, 2013) (see section 1.5.2).

1.4 Divalent metal-ion transporter-1

Since there exists no regulated mechanism for the excretion of iron, systemic iron homeostasis must be achieved by regulating absorption of the metal at the intestinal level.

1.4.1 History of DMT1

DMT1 (NRAMP2, DCT1) was cloned by Hediger’s group in 1997 by the functional screening of a complementary DNA library prepared with duodenal mRNA isolated from rats fed a low-iron diet (Gunshin et al., 1997). Those same investigators found that expression of the rat DMT1 in RNA-injected Xenopus oocytes stimulated 55Fe2+ uptake and Fe2+-evoked currents and that this widely expressed transporter was reactive with several other divalent metals ions (Gunshin et al., 1997). Andrews’ group around the same time had used a positional cloning strategy to identify the defective gene responsible for the microcytic anemia phenotype of the mk inbred mouse strain (Fleming et al., 1997). These two findings established DMT1 as a divalent metal-ion transporter with a central role in iron metabolism.

During the intervening 15 years, DMT1 has been the focus of intense research effort—mainly because DMT1 plays nonredundant roles in iron homeostasis. Its widespread expression—present in at least some cell types in every organ tested (Gunshin et al., 1997; Hubert & Hentze, 2002; Wang et al., 2012)—places DMT1 at the nexus of multiple pathways of epithelial iron transport and cellular iron acquisition in health and disease, and the transport of other metals (e.g. cadmium, manganese) particularly in chronic exposure (for discussion see section 5.1.4-5.1.6). DMT1 is a mechanistically complex transporter with multiple isoforms that function in varied environments.

1.4.2 DMT1 isoforms

There exists at least four isoforms of the human DMT1 protein (Mackenzie et al., 2007) (Figure 1.7). The bulk of the protein (531 or more amino acid residues) is conserved among all isoforms; however, they differ in both their N- and C-termini as a result of variant transcription of the human SLC11A2 gene (Table 1.1) at locus 12q13.

5′– regulation (N-terminus)—Transcript variants arise from use of at least two, possibly three, discrete transcription initiation sites. Initiation from exon 1A, upstream of a second initiation site in exon 1B, adds 29 N-terminal amino acids not present in the 1B isoform (Hubert & Hentze, 2002) (Table 1.1). Both variants are spliced to exon 2 and, because exon 1B contains no start codon, translation of the 1B isoform begins with a start codon in exon 2 (Table 1.1). Orthologous transcript variants were also found in rat and mouse (Hubert & Hentze, 2002). Whereas the 1A isoform is found predominantly in intestine and kidney, 1B isoforms are widely expressed (Hubert & Hentze, 2002). In the NCBI Reference Sequence record for the human SLC11A2 gene (NG_021139.1, version 27 June 2012), these exons—originally described as 1A, 1B, and 2—are annotated exons 1, 2A, and 4, respectively. Initiation of transcription from exon 5 codes a putative, shorter isoform with a distinct N-terminus (MSTVDYL) (Table 1.1). All three 5′ variants are spliced to exon 6A, and transcription proceeds in common for most of the remainder of the coding region.

3′– regulation (3′-UTR)—Alternative splicing at the 3′ end of the SLC11A2 gene results in mRNA transcripts that differ in both their 3′-translated and 3′-untranslated regions (Lee et al., 1998; Canonne-Hergaux et al., 1999). One variant, IRE(+), contains in its 3′-untranslated region (3′-UTR) a stem-loop structure featuring a consensus iron-responsive element (IRE) that is important for post-transcriptional mRNA stabilization according to cellular iron status (Figure 1.10). Thus, the IRE(+) and IRE(−) isoforms are expected to differ mainly in their regulation by iron (Hubert & Hentze, 2002). The IRE(+) and IRE(−) transcripts also give rise to variability in the C-terminal amino acid sequences (Table 1.1). Investigators have found a total of 7 ‘coding’ transcript variants for human DMT1, and these code for four isoforms (Table 1.1). A fifth isoform, 1A/IRE(−), is presumpve because a corresponding human mRNA that would code for it has not been reported in the database. No mouse tissue tested expressed exclusively 1A and IRE(−) transcripts (Hubert & Hentze, 2002), although a 1A-IRE(−) cDNA has been generated from rat kidney cortex RNA (Abouhamed et al., 2006). An engineered 1A/IRE(−) human construct has been expressed in RNA-injected Xenopus oocytes, and the expressed protein was functional (Mackenzie et al., 2007).

Figure 1.7. The multiple isoforms of DMT1. (A) Variant transcripts of the human SLC11A2 gene. There are at least four DMT1 mRNA transcripts (illustrated) that differ in their UTRs (green) and in their coding regions (Hubert & Hentze, 2002). The 1A exon contains a start codon upstream of that in 1B, adding a 5′- coding region (white) that introduces 29 additional amino acids at the N-terminus of the protein in 1A isoforms. DMT1 transcripts also vary at the 3′-end (Lee et al., 1998; Canonne-Hergaux et al., 1999; Herrmann et al., 2004): those variants containing an IRE in the 3′-UTR, i.e. IRE(+) isoforms, also have an isoform-specific 3′ coding region (yellow) in place of the 3′ coding region (red) in the isoforms lacking the IRE, i.e. IRE(−). The C-terminus of the IRE(+) forms contains 18 amino acid residues that substitute for the final 25 of the IRE(−) forms. (B) Coupled transcription–translation of the multiple isoforms of human DMT1 (hDMT1) in a cell-free system, in the absence (X) or presence (M) of canine pancreatic microsomes. L- [35S]methionine-labelled products were separated by SDS/PAGE and the autoradiograph exposed for 3 days (3d exp, upper panel) or 35 days (35d exp, lower panel). A+, 1A/IRE(+)-hDMT1 isoform; A−, 1A/IRE(−); B+, 1B/IRE(+); B−, 1B/IRE(−). Molecular mass of translation products was estimated using Bio-Rad standards of 31.8, 41.3, 89, 131 and 210 kDa. (C) DMT1 immunofluorescence in control oocytes and oocytes expressing DMT1 isoforms. Reproduced from (Mackenzie et al., 2007).

Table 1.1. Multiple isoforms of human DMT1 arise from variant transcription of the SLC11A2 gene. The SLC11A2 gene coding for DMT1 is described at NCBI Reference Sequence NG_021139.1 (accessed on June 27, 2012). The 1A/IRE(−) isoform is presumptive since a full-length transcript variant that would code for this isoform has not been recorded so far. Reproduced from (Shawki et al., 2012).

Transcript variant Protein isoform

Transcript NCBI reference Isoform NCBI reference Peptide Name N- and C- terminal sequences Predominant localization number sequence number sequence length N– Apical membrane, polarized 1 NM_001174125.1 1 1A/IRE(+) NP_001167596.1 590 MRKKQLKTEAAPHCELKSYSKNSATQVSTMVLGPEQK epithelia MSDD… / …VSISKGLLTEEATRGYVK–C

2 NM_001174126.1 NP_001167597.1 N–MVLGPEQKMSDD… / Early (recycling) endosomes, 2 1B/IRE(−) 568 3 NM_001174127.1 NP_001167598.1 …CHLGLTAQPELYLLNTMDADSLVSR–C erythroid precursors

4 NM_000617.2 NP_000608.1

5 NM_001174128.1 3 1B/IRE(+) NP_001167599.1 561 N–MVLGPEQKMSDD… / …VSISKGLLTEEATRGYVK–C Late endosomes and 6 NM_001174129.1 NP_001167600.1 lysosomes, widespread cellular distribution Short- 7 NM_001174130.1 4 form NP_001167601.1 557 N–MSTVDYL… / …VSISKGLLTEEATRGYVK–C IRE(+)

8 NR_033421.1 Noncoding

9 NR_033422.1 Noncoding

N– 1A/IRE(−) 597 MRKKQLKTEAAPHCELKSYSKNSATQVSTMVLGPEQK Recycling endosomesa MSDD … / …CHLGLTAQPELYLLNTMDADSLVSR–C a Based on expression of a GFP-tagged construct in cultured cells (Yanatori et al., 2010).

1.4.3 Tissue and cellular distribution of DMT1

The DMT1 isoforms differ also in their tissue distribution and subcellular targeting (Table 1.1). Epithelial cell lines predominantly express the IRE(+) isoforms whereas blood cell lines, the IRE(−) isoforms (Tabuchi et al., 2002). The N- and C-termini of the DMT1 isoforms may contain signal sequences that direct the subcellular targeting of the protein. Mutational analyses in transfected epithelial cell lines (HEp-2 and

MDCK) have revealed that disruption of the N-terminal region H13CELKS (critically, Leu-16) prevents the

insertion of the 1A isoform into the plasma membrane (Yanatori et al., 2010), and that the Y555XLXX region in the C-terminal cytoplasmic tail of the IRE(−) isoform is required for its targeng to early endosomes (Tabuchi et al., 2002). In contrast, IRE(+) was localized to the apical plasma membrane, late endosomes, and lysosomes of polarized cells, and its apical-membrane insertion was dependent upon N-glycosylation of DMT1 (Tabuchi et al., 2002).

1.4.4 Functional properties of alternative splices of DMT1

The N- and C-terminal sequence variations between the isoforms do not however alter the functional properties of the isoforms. Mackenzie et al have expressed 1A/IRE(+), 1A/IRE(−), 1B/IRE(+), and 1B/IRE(−) isoforms of human DMT1 in Xenopus oocytes and found (i) that the cellular uptake of 55Fe2+ correlated with the levels of expression at the plasma membrane, and (ii) that the transport cycle turnover rate did not differ among the four isoforms, i.e. all four isoforms transport Fe2+ with equal efficiency (Mackenzie et al., 2007). We and others also found, after normalization for expression levels, that the 1A/IRE(+) and 1B/IRE(+) isoforms did not differ in their apparent affinity for Fe2+, pH dependence, or relative ability to transport a range of metal ions (Mackenzie et al., 2007; Mackenzie et al., 2010). Likewise, doxycycline- induced expression of the rodent 1A/IRE(+) and 1B/IRE(+) isoforms in stably transfected human embryonic kidney HEK293 cells revealed no differences in their ability to transport Fe2+ or Mn2+ (Mackenzie et al., 2010; Garrick et al., 2006).

1.4.5 Substrate profile and physiological roles of DMT1

Which metal ions are transported by DMT1? That the question has been addressed in nearly as many reviews as there have been original research articles on the topic, probably reflects both the high level of interest in the answer and the lack of complete consensus. A broad range of metal ions evoked currents in voltage-clamped Xenopus oocytes expressing rat or human DMT1 (Gunshin et al., 1997; Mackenzie et al., 2007). Many of these metal ions have been shown to inhibit radiotracer iron uptake in mammalian

16 cell preparations in which DMT1 is constitutively or heterologously expressed (Garrick & Dolan, 2002; Conrad et al., 2000; Picard et al., 2000; Bannon et al., 2003; Linder et al., 2006; Zhang et al., 2008). Nevertheless, currents and inhibition only demonstrate reactivity with the transporter, but do not demonstrate per se that the metal ion is transported by DMT1. We performed a comprehensive substrate-profile analysis of DMT1 by using two methods that would directly test for metal-ion transport in the oocyte system—radiotracer assay and a fluorescence- based assay using the metal-sensitive fluorophore PhenGreen SK (Illing et al., 2012). We found that Cd2+, Co2+, Fe2+, Mn2+, Ni2+, VO2+, and Zn2+ are transported substrates of DMT1 (Figure 1.8A); however, we found no evidence of DMT1-mediated transport of Ca2+, Cr2+, Cr3+, Cu1+, Cu2+, Fe3+, Ga3+, Hg2+, or VO1+ (Illing et al., 2012; Shawki & Mackenzie, 2010). Our findings extended the catalogue of DMT1 substrates and nonsubstrates and, to the extent that a few of these metals had been tested previously (see (Illing et al., 2012) for references), were consistent with previous studies with the notable exception of copper. Other investigators have measured copper transport in the human intestinal cell model Caco-2 cells and concluded that DMT1 can transport cuprous ion (Cu1+) (Arredondo et al., 2003; Espinoza et al., 2012), or have measured copper uptake in intestinal brush border membrane vesicles (BBMV) and concluded that DMT1 can transport cupric ion (Cu2+) (Knöpfel et al., 2005a). We found instead that DMT1 did not transport Cu1+ or Cu2+ in the oocyte preparation (Figure 1.8B). One advantage of the oocyte system is its superior sensitivity—expression of DMT1 stimulated iron transport more than 700-fold over control (Illing et al., 2012)—whereas assigning transport activity to DMT1 is less clear in the mammalian systems. Still, it is possible that a subunit normally expressed in mammalian cells and required for copper transport is not expressed in the Xenopus oocyte, or that an isoform other than the 1A/IRE(+) expressed in the study just cited is capable of transporting copper. Clearly, further studies are needed if we are to conclude that copper is a transported substrate of DMT1 and, if so, to determine the valence of the transported species.

Figure 1.8. Radiotracer metal-ion (*Mv+) transport in Xenopus oocytes expressing DMT1. (A) uptake of 2 μM radiotracer metal ions (109Cd2+, 55Fe2+, 54Mn2+, and 65Zn2+) at pH 5.5 in control oocytes (gray bars) and oocytes expressing DMT1 (black bars) (n = 7–12). Two-way ANOVA revealed an interaction (P < 0.001); Holm-Šidák all-pairwise multiple comparisons showed that DMT1 differed from control for each metal ion (aP < 0.001; bP = 0.028) and that, within DMT1, all metal ions differed from one another (P < 0.001). (B) uptake of 2 μM radiotracer metal ions (64Cu2+, 64Cu+, and 55Fe2+) at pH 6.0 in control oocytes (gray bars) and oocytes expressing DMT1 (black bars) (n = 16–23). (We added 1 mM L-histidine to solutions containing copper, and 1 mM L-ascorbic acid to solutions containing Cu+ or Fe2+.) Two-way ANOVA revealed an interaction (P < 0.001); DMT1 differed from control only for 55Fe2+ (P < 0.001) but not for 64Cu2+ (P = 0.71) or 64Cu+ (P = 0.72). Data from A and B are derived from independent preparations. Reproduced from (Illing et al., 2012).

Assessing whether DMT1 plays important roles in the absorption or cellular transport of each of its substrates requires that we first take into account the relative selectivity with which DMT1 reacts with and transports each of its substrates. We found that DMT1 exhibited a strong preference (selectivity) for

2+ Fe Fe (. ≈ 1 µM) over any of its other physiological substrates (Illing et al., 2012). We ranked DMT1 substrates by using the “specificity constant” as follows: Cd2+ > Fe2+ > Co2+, Mn2+ >> Zn2+, Ni2+, VO2+ (Figure 1.9). With some knowledge of prevailing metal-ion concentrations in vivo, these data help us predict DMT1’s roles in transport of metals aside from iron, and such predictions can be tested in tissue preparations or animal models. Intestinal absorption of manganese and its uptake into reticulocytes is disrupted in the b rat indicating that DMT1 plays some role in manganese transport in those tissues (Chua & Morgan, 1997). Nevertheless, the involvement of DMT1 in the transport of manganese (moderate Mn selectivity, . ≈ 4 μM; (Illing et al., 2012)) may be of greater pathophysiological significance than physiological, e.g. nasopulmonary absorption of manganese and its uptake into the brain in chronic

Zn manganese exposure (see section 5.1.6). We found zinc to be a very poor substrate (. ≈ 19 μM)—and weak competitive inhibitor—of DMT1 (Illing et al., 2012). Because also the prevailing iron concentrations generally exceed those of zinc, we would expect iron to outcompete zinc under most conditions and DMT1 contribute little if at all to the absorption or cellular uptake of zinc. If DMT1 participates at all in zinc absorption, then it is not sufficient to compensate for the loss of the Zrt-/Irt-like transporter-4 (ZIP4), because hereditary defects in ZIP4 result in acrodermatitis enteropathica (Wang et al., 2002a). We found that human DMT1 transports the toxic heavy metal Cd2+ more efficiently than it does Fe2+ (Illing et al., 2012), suggesting that DMT1 is a likely route of intestinal cadmium absorption and its cellular uptake (see section 5.1.4).

Figure 1.9. Metal-ion substrate selectivity. Metal-ion substrate selectivity is expressed as the ratio M M max⁄. (i.e. the specificity constant). Data are means, SE from n independent trials for each metal ion (M). For each oocyte, we measured the metal-ion-evoked currents at −70 mV and pH 5.5 over a range of at least seven concentrations and fit the data by Equation 4.1. Fe2+ was superfused in the presence of 100

M 2+ μM L-ascorbic acid. max was normalized by the current evoked by 50 μM Mn in each individual oocyte. One-way ANOVA (P < 0.001) followed by Holm-Šidák all-pairwise comparisons; all differed from one another (P ≤ 0.004) except among those marked with following: aP = 0.64; bP ≥ 0.61. Reproduced from (Illing et al., 2012).

1.4.6 Regulation of DMT1 expression

Regulation of DMT1 occurs primarily via two pathways: (i) directly by cellular iron status, and (ii) indirectly by systemic iron status.

Direct regulation (IRE/IRP)

At the purely cellular level, iron homeostasis involves mRNA iron-responsive elements (IRE) and iron- regulatory proteins (IRP). For example, ferritin mRNA contains a 5’-mRNA stem-loop structure that can

19 associate with iron. When cellular iron levels are low, the IRP can bind to this loop and this prevents the initiation of translation (Figure 1.10). When cellular iron levels are high, the IRP cannot bind, and this permits translation of ferritin, ultimately storing the increased cellular iron. TfR1 mRNA is an example of one containing 3’-IREs. When cellular iron levels are low, IRPs can bind to these 3’ structures and have the effect of stabilizing the mRNA by slowing its degradation (Figure 1.10). Increased translation of TfR1 then serves to increase the supply of iron to the cell.

Figure 1.10. IRP-mediated cellular regulation of steady-state mRNA levels coding for iron-related proteins. Whereas low iron in the cell results in the binding of IRP1/2 to 5’–IRE containing mRNAs (translational repression) and 3’–IRE containing mRNAs (mRNA stabilization), high iron results in loss of IRP affinity to IREs and permits translational activation and mRNA degradation by RNases. Reproduced from (Wallander et al., 2006).

The dominant mechanism by which DMT1 is regulated at the cellular level is thought to be via post-transcriptional mRNA stabilization of the IRE(+) isoform in response to cellular iron levels. Binding of the 3′-UTR mRNA IRE by iron-responsive binding protein-1 (IRP1) confers RNA stability under conditions of low iron (Gunshin et al., 2001). In contrast, the IRE(−) form is not subject to iron-dependent regulation (Zoller et al., 2001; Rolfs et al., 2002). Less is known about IRE/IRP interactions in DMT1 than for other IRE-containing transcripts, but control of expression of the DMT1 IRE(+) isoforms appears analogous to that of the transferrin receptor (Pantopoulos, 2004; Wallander et al., 2006) (Figure 1.10). Other metals,

20 such as Mn2+, are also bound by iron-responsive binding proteins (Kwik-Uribe et al., 2003) and may exert some control over DMT1 expression.

Hypoxia-inducible factor

DMT1 is also regulated at the transcriptional level by hypoxia and stress-related signaling. Exposure of P19 embryonic carcinoma cells to nitric oxide decreased expression of the 1B but not 1A isoform (representing both IRE(+) and IRE(−) isoforms). This eﬀect appears to have resulted from decreased binding of the nuclear factor NF-κB (p65 subunit) to the 1B promoter region of DMT1 (Paradkar & Roth, 2006). In the intestine, hypoxia regulates the expression of DMT1 via stabilization of the hypoxia-inducible transcription factor HIF2α, which can bind the 1A promoter region of DMT1, and intestine-specific ablation of HIF2α blunted the increased expression of DMT1 in response to iron deficiency (Shah et al., 2009; Mastrogiannaki et al., 2009).

Indirect regulation (hepcidin)

The liver hormone hepcidin (Hepc) regulates iron metabolism predominantly via its interaction with the iron-export protein ferroportin (Fpn) (Ganz & Nemeth, 2006). Ferroportin binds hepcidin, triggering the internalization and degradation of this transporter. The resulting transient increase in intracellular iron is thought to act via the IRE/IRP system to downregulate expression of DMT1 in enterocytes. Nevertheless, novel mechanisms are emerging by which intestinal DMT1 may be regulated independently of Fpn or independently of Hepc. Two studies have proposed that Hepc directly controls DMT1 expression. Hepcidin inhibited apical 55Fe uptake in both Caco-2 cells and rat duodenal segments apparently by decreasing DMT1 transcription with no change in Fpn mRNA or protein (Mena et al., 2008). Other investigators using similar preparations also observed decreased DMT1 mRNA but attributed the loss of DMT1 transport activity mainly to its ubiquitin-dependent proteasomal degradation (Brasse-Lagnel et al., 2011). Such degradation may be facilitated by the adapter proteins Ndfip1 and Ndfip2 (Foot et al., 2008). Indeed, DMT1 levels in enterocytes were elevated in Ndfip1-null mice fed a low-Fe diet more strongly than in wildtype mice (Foot et al., 2011). Ubiquitination and proteasomal degradation of DMT1 is reviewed elsewhere (Garrick et al., 2012), and data provided there indicate that ubiquitination by Parkin is specific to the 1B/IRE(−) isoform and not 1A/IRE(+). DMT1 is the only known mechanism for the absorption of nonheme iron. DMT1’s vital role in the intestine and in erythroid tissue is illustrated by the impaired intestinal iron absorption and severe hypochromic–microcytic anemia (characteristic of iron deficiency) exhibited by (i) the Belgrade (b) rat and mk mouse inbred rodent strains bearing an identical (G185R) mutation in DMT1 (Su et al., 1998; Fleming

21 et al., 1998), (ii) the intestinal specific DMT1 knockout mouse (DMT1int/int) (see Chapter 2), and (ii) mutations in human DMT1 associated with a disease phenotype (Mims et al., 2005; Iolascon et al., 2006; Beaumont et al., 2006; Blanco et al., 2009; Bardou-Jacquet et al., 2011) (see Chapter 4). Mutations in DMT1, tissue specific ablation, and dysregulation of DMT1 have strengthened the view that DMT1 is essential for iron homeostasis.

1.5 Iron disorders

1.5.1 Iron-deficiency anemia

1.5.1.1 Dietary iron deficiency: Interaction of calcium with DMT1

Iron deficiency remains the most prevalent micronutrient deficiency worldwide and results in iron- deficiency anemia, as well as neurological and developmental defects in children. Iron fortification of infant milk formulas and cereals, and better diet, has decreased the incidence of iron deficiency over recent decades; nevertheless, a more precise understanding of the iron-absorptive machinery will lead to new strategies for improved iron nutrition.

Dietary calcium is known to reduce iron bioavailability, but the molecular basis of this interaction has been poorly understood. Although Ca2+ is not a transported substrate of DMT1, we have found that Ca2+ inhibited 55Fe2+ transport activity (Figure 1.11A) and currents (Figure 1.11B) associated with DMT1 expression in Xenopus oocytes with inhibition constants (Ki) of 1–20 mM (Shawki & Mackenzie, 2010). Because intestinal luminal calcium concentrations periodically reach the millimolar range, dietary calcium (from sources such as milk) could substantially inhibit DMT1-mediated Fe2+ absorption. Calcium may also decrease intestinal brush-border iron uptake by stimulating the derecruitment of DMT1 from the apical membrane. Treatment of Caco-2 cells with 2.5 mM CaCl2 for 4 h was sufficient to cause an 80% reduction in DMT1 immunoreactivity of the membrane fraction (Thompson et al., 2010). Interactions of calcium with DMT1 therefore should be taken into consideration when developing strategies for improved iron nutrition, milk formulation, and iron fortification.

Figure 1.11. Ca2+ is a noncompetitive inhibitor of DMT1-mediated Fe2+ transport. (A) Uptake of 2 µM 55Fe2+ (pH 5.5, 1 mM Mg2+) in the absence or presence of 20 mM Ca2+ in control oocytes and oocytes expressing DMT1 (n = 10-12). Two-way ANOVA revealed an interaction (P < 0.001). (B) Saturation kinetics of 55Fe2+ uptake (pH 5.5, 1 mM Mg2+) in the absence (circles) or presence (triangles) of 20 mM Ca2+ (n = 10-14 per group at each concentration). We obtained the specific DMT1-mediated 55Fe2+ uptake by subtracting from the 55Fe2+ uptake in oocytes expressing DMT1 (filled symbols) the endogenous 55Fe2+ uptake derived from linear fits (dashed lines) of the 55Fe2+ uptake in control oocytes (empty symbols). Specific DMT1-mediated 55Fe2+ uptake data were fit (solid lines) by Equation A.1 (Appendix), yielding for [Ca2+] = 0 the parameters

−1 2 Vmax = 5.0 ± 0.3 pmol.min , nH = 1.2 ± 0.1, and K0.5 = 8.4 ± 1.0 µM (adjusted r = 0.96, P <0.001) and for

2+ −1 [Ca ] = 20 mM the parameters Vmax = 2.6 ± 0.1 pmol.min , nH = 1.2 ± 0.1, and K0.5 = 10.2 ± 1.2 µM (adjusted r2 = 0.97, P <0.001). Reproduced from (Shawki & Mackenzie, 2010).

1.5.1.2 Iron losses (e.g. bleeding) exceeding iron absorption

Chronic loss of iron due to phlebotomy or hemorrhage can exceed the amount of iron that can be delivered, resulting in iron deficiency and iron deficiency anemia. Chronic gastrointestinal blood loss is a common cause of iron deficiency. Whereas acute gastrointestinal bleeding results in adequate upregulation of the iron absorptive machinery (Johnson-Wimbley & Graham, 2011), the body is unable to compensate for chronic loss of iron exceeding 5mg/day, leading to depletion of iron stores and iron deficiency anemia (Rockey, 1999). Iron deficiency due to diet or bleeding results in a physiological response to correct for the loss of iron in an attempt to maintain systemic iron homeostasis.

1.5.2 Iron overload

Defects in the hepcidin–ferroportin axis

Mutations in genes comprising the Hepc-Fpn axis result in hereditary iron disorders. Hereditary defects that either (i) increase Hepc production or (ii) impair Fpn-mediated iron-transport activity lead to abnormally low iron absorption and less iron recycling by macrophages. Consequently, these mutations can result in an iron-restricted anemia with macrophage iron retention. TMPRSS6 (matriptase-2, MT2) normally acts as a negative regulator that decreases Hepc production (Du et al., 2008). Hereditary defects in TMPRSS6 lead to overproduction of Hepc relative to the iron load, and this leads to an iron-restricted anemia (Folgueras et al., 2008). It is referred to as an “iron-refractory iron-deficiency anemia” (or IRIDA) since oral iron does not correct the condition. In fact, parenteral iron is only a short-term solution since iron is retained in the macrophages. Extremely rare mutations in DMT1 cause other forms of IRIDA. IRIDA may also result from phlebotomy in patients bearing any of the several mutations in Fpn that destroy its iron-transport capacity. They are not initially anemic, but do not recover normally following phlebotomy. Mutations in other components of the Hepc-Fpn axis produce iron-overload disorders. Hereditary defects that (i) impair iron sensing or Hepc production in the liver, or (ii) render Fpn insensitive to Hepc, will increase iron absorption and recycling of iron to the plasma. This leads to saturation of the plasma Tf—and that is the critical turning point clinically speaking. For as long as plasma iron exists primarily bound by Tf, then its cellular uptake is regulated or kept in check by the transferrin receptor cycle. When plasma iron levels exceed what can be bound by Tf, as in hereditary hemochromatosis, we observe the characteristic appearance in plasma of nontransferrin-bound iron (NTBI). NTBI will typically exist as complexes of Fe(II) and Fe(III) with citrate, acetate, albumin, and ascorbate. It is NTBI that is the effector of the problems associated with iron-overload conditions since NTBI gains access to cells in liver, heart, pancreas and elsewhere in an uncontrolled fashion, and results in organ damage.

Iron sensing disorders

Excess iron (unless it is lost by cell desquamation or bleeding) is retained in the body. Accumulation in the liver, heart, and other vital organs can result in cirrhosis, hepatocellular carcinoma, cardiomyopathy, endocrine disorders, and arthritis (Fleming & Ponka, 2012; Anderson et al., 2004b; Darshan et al., 2010). Hereditary hemochromatosis (HH) is a disorder characterized by tissue iron overload secondary to excessive dietary-iron absorption, and generally results from a disruption of the Hepc-Fpn axis regulating iron metabolism (Fleming & Ponka, 2012; Darshan et al., 2010). The most common form, HH type 1, is an autosomal recessive disorder resulting from mutations in the hemochromatosis gene, HFE (Harrison & Bacon, 2003). Intestinal expression of DMT1 is elevated in human patients and animal models of HH type 1 in spite of increased serum iron or ferritin levels (Fleming et al., 1999; Byrnes et al., 2002). Less

24 commonly, HH can result also from mutations in the genes coding for hemojuvelin (type 2A), Hepc (type 2B), TfR2 (type 3) or Fpn (type 4); the last of these being autosomal dominant. The mainstay of treatment for HH is phlebotomy; however, the effectiveness of this approach is limited by poor patient compliance and the fact that phlebotomy itself can lead to upregulation of DMT1 mRNA in HH patients (Kelleher et al., 2004; Dostalikova-Cimburova et al., 2011), presumably as a result of signaling from increased erythropoiesis post-phlebotomy and the loss of the stimulatory effect of increased iron stores in hepcidin production.

1.5.3 Targeting DMT1 in iron overload disorders

That DMT1 is the front line in iron absorption in the intestine serves a target for the treatment of iron disorders. Considering the function of DMT1 or targeting the transporter in the intestine may lead to superior therapies for anemia and the prevention of iron overload in HH patients. Validation of intestinal DMT1 as a therapeutic target derives from at least the following three observations: (i) intestinal expression of DMT1 is upregulated in HH patients; (ii) the severity of the anemia phenotype of human probands carrying mutations in DMT1, demonstrating the critical role of DMT1 in human iron metabolism; and (iii) the severe anemia in the intestine-specific DMT1-null mouse (see Chapter 2), demonstrating that DMT1 is the principal or only transporter serving intestinal uptake of nonheme iron, at least in the mouse post-weaning. In thalassemia, an autosomal recessive disorder characterized by defective hemoglobin production, iron overload results from blood transfusions given to treat the primary lesion. Iron-chelation therapy is the main approach used to treat iron overload in these patients. Notably, increased intestinal iron absorption (presumably driven by the ineffective erythropoiesis) is also evident in thalassemia (Hershko, 2010; Darshan et al., 2010), so blockade of intestinal DMT1 may be of benefit in treating this disorder.

Small molecule inhibitors of DMT1

Screening approaches have identified a number of compounds that can inhibit DMT1-mediated iron uptake in mammalian cells. It is likely that some of these do not act directly upon DMT1. For example, two unrelated antioxidants, ebselen and pyrrolidine dithiobarbamate, inhibited DMT1-mediated Fe2+ uptake apparently by altering intracellular redox status because (i) both agents increased intracellular glutathione, and (ii) ebselen did not inhibit transport of Mn2+ (the free metal is not redox active in physiological environments) (Wetli et al., 2006). A screen identified 10 compounds that could inhibit

25 cellular uptake of NTBI in nontransfected HeLa cells (Brown et al., 2004). Of these, six were cytotoxic and of low apparent affinity (i.e. the concentration at which inhibition was half-maximal, IC50 ≥ 20 µM), and may therefore be of limited utility. Two of the 10 also inhibited Tf-dependent iron uptake, so this group of compounds may serve as useful experimental tools with which to discriminate NTBI and Tf-dependent transport pathways. When one of these compounds, a polysulfonated dye (NSC306711), was tested in HEK293T cells stably transfected with DMT1, it reversibly and competitively inhibited DMT1-mediated iron uptake with Ki ≈ 7 µM (Buckett & Wessling-Resnick, 2009); however, the action of NSC306711 is complex because this same compound also induced the internalization and degradation of TfR (Horonchik & Wessling-Resnick, 2008). High-throughput screening has identified the triazinone pyrazolone as an inhibitor of iron

transport in CHO cells transfected with DMT1 (IC50 ≈ 3 µM) (Cadieux et al., 2012). Optimization efforts

generated substituted pyrazoles with lower IC50 values (higher apparent affinity) but which did not completely inhibit iron transport. The same group has also identified a class of diaryl and tricyclic benzylisothiourea compounds that more potently inhibited iron transport with submicromolar IC50 values in CHO cells expressing DMT1 (Zhang et al., 2012). These compounds also boasted low cytotoxicity and low cell permeability—the latter property being significant in that it may be more desirable to block intestinal DMT1 using an agent that is not absorbed, thereby minimizing secondary effects. The mechanism of action of these blockers and the nature of inhibition has not been described; however, inhibition by the substituted pyrazoles did not appear to be mediated via redox effects or metal chelation (Cadieux et al., 2012). The identification of DMT1 blockers offers both novel experimental tools that will help us more clearly define the roles of DMT1 in metal metabolism, and the prospect of improved therapies in iron- overload disorders.

A broad range of disease phenotypes in animals and humans justify my focus on DMT1. My overall goal is to better our understanding of the physiological role of DMT1 and the molecular mechanisms of iron transport in order to (i) drive the development of therapies that will improve metal-ion nutrition and (ii) identify novel therapeutic targets in the treatment of iron overload.

2 Role of intestinal DMT1 in metal homeostasis

Chapter 2 comprises in its entirety the published work Shawki A, Anthony SR, Nose Y, Engevik MA,

Niespodzany EJ, Barrientos T, Öhrvik H, Worrell RT, Thiele DJ, Mackenzie B (2015) Intestinal DMT1 is

critical for iron absorption in the mouse but is not required for the absorption of copper or

manganese. Am J Physiol Gastrointest Liver Physiol, 309, G635-G647, with the addition of unpublished

work in sections 2.4.7 (Results) and 2.5.4 (Discussion).

2.1 Abstract

Divalent metal-ion transporter-1 (DMT1) is a widely expressed iron-preferring membrane-transport protein that serves a critical role in erythroid iron utilization. We have investigated its role in intestinal metal absorption by studying a mouse model lacking intestinal DMT1 (i.e. DMT1int/int). DMT1int/int mice exhibited a profound hypochromic–microcytic anemia, splenomegaly and cardiomegaly. That the anemia was due to iron deficiency was demonstrated by the following observations in DMT1int/int mice: (i) blood iron and tissue nonheme-iron stores were depleted; (ii) mRNA expression of liver hepcidin (Hamp1) was depressed; and (iii) intraperitoneal iron injection corrected the anemia, and reversed the changes in blood iron, nonheme-iron stores, and hepcidin expression levels. We observed decreased total iron content in multiple tissues from DMT1int/int mice compared with DMT1+/+ mice but no meaningful change in copper, manganese, or zinc. DMT1int/int mice absorbed 64Cu and 54Mn from an intragastric dose to the same extent as did DMT1+/+ mice but the absorption of 59Fe was virtually abolished in DMT1int/int mice. This study reveals a critical function for DMT1 in intestinal nonheme-iron absorption for normal growth and development. Further, this work demonstrates that intestinal DMT1 is not required for the intestinal transport of copper, manganese, or zinc.

2.2 Introduction

Iron deficiency is the most prevalent micronutrient deficiency worldwide (Ramakrishnan, 2002; Beutler, 2010). Its deficiency leads to iron-deficiency anemia, and to neurological and developmental disorders in children (Beard, 2008; Beutler, 2010). Since there exists no regulated mechanism for the excretion of iron, regulation of the whole-body iron economy is achieved by tightly controlling absorption of the metal (Ganz, 2013). Failure to regulate iron absorption in a manner appropriate to iron status is characteristic of several hereditary iron-overload disorders and iron-refractive iron-deficiency anemia (Ganz, 2013; Darshan et al., 2010; Fleming & Ponka, 2012).

Divalent metal-ion transporter-1 [DMT1; reviewed, (Shawki et al., 2012)] is a widely expressed mammalian proton-coupled iron transporter (Gunshin et al., 1997; Mackenzie et al., 2006). Mice in which the SLC11A2 gene coding for DMT1 was globally-inactivated (i.e. SLC11A2−/−) exhibited a severe hypochromic–microcytic anemia and did not survive more than 7 days (Gunshin et al., 2005a). A critical role for DMT1 in erythroid iron acquisition was confirmed by the following observations: (i) transfusion of red cells, but not parenteral iron injections, improved survival of SLC11A2−/− mice, and (ii) lethal-dose- irradiated wildtype mice into which the investigators transplanted hematopoietic stem cells from SLC11A2−/− mice exhibited defective erythropoiesis (Gunshin et al., 2005a).

The microcytic (mk) mouse and Belgrade (b) rat models, inbred strains that harbor a Gly185→Arg mutation in DMT1 (Fleming et al., 1998; Fleming et al., 1997), also exhibited an anemia phenotype. Parenteral iron injections partially improved the condition, and tissue or vesicle preparations from the mk mouse and b/b rat revealed deficiencies in intestinal iron transport (Edwards & Hoke, 1972; Knöpfel et al., 2005b). The mk mouse and b/b rat anemia phenotypes were, however, less severe than that of the SLC11A2−/− knockout model, a finding that may be explained by the observation that the G185R mutant retains a fraction of wildtype activity (Su et al., 1998).

Rare mutations in human DMT1 result in severe hypochromic–microcytic anemia (Shawki et al., 2012); however, that four out of five probands also exhibited hepatic iron overload has raised questions as to the importance of DMT1 in intestinal iron absorption and iron homeostasis. Nevertheless, hepatic iron loading is also apparent in the b/b rat model (Thompson et al., 2006). Targeted deletion of DMT1 in the mouse intestine has provided a superior model in which to test the role of DMT1 in intestinal iron absorption. Initial characterization of the intestine-specific DMT1 knockout mouse (i.e. DMT1int/int) revealed a progressive anemia that was prevented by parenteral iron injection (Gunshin et al., 2005a),

28 consistent with a primary defect in iron absorption; however, iron absorption has not been directly examined in that model.

In the present study, we have further examined the role of DMT1 in intestinal metal absorption by studying the DMT1int/int mouse. Our initial goal was to test a critical role for DMT1 in intestinal iron absorption and provide a detailed analysis of the DMT1int/int phenotype in male and female mice. We examined the effects of loss of intestinal DMT1, with and without parenteral iron, on iron homeostasis and the expression of iron-related genes. Secondly, we directly tested the contribution of DMT1 to intestinal nonheme-iron absorption by measuring 59Fe absorption from an oral dose, and localized the site of the primary lesion. Gunshin et al found that the SLC11A2+/− (global) heterozygote exhibited lower iron stores than did wildtype mice, and suggested that haploinsufficiency of intestinal absorption could explain this finding (Gunshin et al., 2005a). We have more directly tested this possibility here in heterozygotes (DMT1+/int) of the intestinal knockout mouse model.

Whereas DMT1 is reactive with a broad range of divalent metal ions in vitro, it exhibits a strong preference for Fe2+ over the other physiologically relevant metal ions it can transport (Illing et al., 2012). We ranked these according to the specificity constant as follows: Fe2+ > Co2+, Mn2+ >> Zn2+. A central objective in the present study was to determine which of these metals relies upon DMT1 for their absorption. We have tested the hypothesis that intestinal DMT1 is required for the absorption of copper, manganese, and zinc by examining metal metabolism and directly measuring intestinal absorption of radiotracer metals in the DMT1int/int mouse. We did not test cobalt because the nutritional requirements for free cobalt (i.e. aside from cobalamin) are very minor. We included copper for the following reasons: (i) Previous studies demonstrated that Ctr1 is a primary high-affinity Cu+ transporter in mammals, particularly within the intestine; however, a Ctr1-independent transport activity has been detected in Ctr1−/− mouse embryonic fibroblasts (Lee et al., 2002b). (ii) Whereas we previously found no evidence that copper is a transported substrate of DMT1 in vitro (Illing et al., 2012), others have measured copper transport in vitro and concluded that DMT1 can transport Cu1+ or Cu2+ (Arredondo et al., 2003; Arredondo et al., 2014; Espinoza et al., 2012; Jiang et al., 2013; Lin et al., 2015). Conversely, in two of the studies just cited, investigators proposed that Ctr1 is also capable of transporting Fe2+ (Espinoza et al., 2012; Lin et al., 2015), so we also examined the expression of Ctr1 in the intestinal DMT1int/int mouse.

2.3 Materials and methods

Reagents—Reagents were obtained from Sigma–Aldrich (St Louis, MO) unless otherwise indicated.

Animals—We examined iron metabolism in adult mice with intestine-specific knockout of DMT1, heterozygotes, and their wildtype littermates, following a protocol approved by the University of Cincinnati Institutional Animal Care and Use Committee. We generated the DMT1int/int model by crossing a floxed DMT1 mouse line (DMT1fl/fl) (Gunshin et al., 2005a) on the 129S6 strain with the villin–Cre transgenic line (Vil–Cre) (el Marjou et al., 2004) on the C57BL/6 strain. Experimental mice (F2) were of a mixed background, 75% 129S6, 25% C57BL/6 except as follows. The Vil–Cre mouse line was also backcrossed over 9 generations onto the 129S6 strain and crossed with the DMT1fl/fl line to produce (i) DMT1int/int mice on a homogeneous 129S6 background used in experiments described in Figures 2.9I–L, 2.10, and (ii) wildtype and Cre-positive mice on a homogeneous 129S6 background used in the experiment described in Table 2.1. All Cre-positive mice were hemizygous for the Vil–Cre recombinase transgene. Animals were genotyped by obtaining DNA from tail clips and analyzed as described previously (el Marjou et al., 2004; Gunshin et al., 2005a). Mice were fed a standard chow containing 350 ppm Fe (7922 NIH-07, Harlan Laboratories, Indianapolis, IN).

Characterization of the DMT1int/int mouse model—We sampled peripheral blood via cardiac puncture and isolated tissues from mice of approximately 120 days of age subjected to isoflurane anesthesia (by inhalation, to effect) in the following set of experiments: (i) male DMT1 wildtype lacking (Cre−) or bearing (Cre+) the Cre recombinase transgene to examine any effect of introducing Cre (Table 2.1); (ii) male experimental control DMT1fl/fl | Cre− (hereaer DMT1+/+) compared with DMT1int/int (Figures 2.1A, 2.2– 2.4, 2.6, 2.8, and 2.9); (iii) female experimental control (DMT1+/+) compared with DMT1int/int (Figures 2.5 and 2.7); (iv) male DMT1fl/+ with and without the Cre transgene, i.e. control mice compared with mice heterozygous for intestinal DMT1 (Figure 2.10). Blood and tissue were isolated from male and female suckling mice at age 7, 14, and 21 days subjected to isoflurane anesthesia (by inhalation, to effect) to examine iron metabolism in the neonate (Figure 2.11).

Blood and tissue analyses—Automated complete blood count (CBC) was performed by Antech Diagnostics (Oak Brook, IL). Serum iron (SI) and unsaturated iron-binding capacity (UIBC) were assayed by using the Iron-SL and UIBC kits (Sekisui Diagnostics, San Diego, CA) according to manufacturer’s protocols, and

transferrin saturation (Tfsat, %) according to Equation 2.1. SI Tf = × 100 sat SI + UIBC 30

Equation 2.1.

Nonheme iron content of liver, heart, spleen, kidney, and skeletal muscle (gastrocnemius) was determined by using a standard acid-digestion, chromogen-based colorimetric assay as described (Torrance & Bothwell, 1980), and normalized by wet tissue weight. Total metal (iron, copper, manganese, and zinc) content was measured by using high-resolution inductively coupled plasma mass spectrometry (ICP–MS) of nitric acid-digested wet tissues from male DMT1+/+ and DMT1int/int mice. We collected an enterocyte-enriched preparation from female DMT1+/+ and DMT1int/int mice by lightly scraping the luminal surface of the proximal 1–5 cm of small intestine (i.e. duodenum) with the edge of a glass microscope slide. Total metal content was normalized by tissue wet weight, except in the case of enterocytes (dry weight).

qPCR analyses—Freshly isolated small intestine was flushed with ice-cold saline, and duodenal enterocytes sampled as described above. Enterocytes and liver tissue were collected into TRIzol reagent (Life Technologies, Carlsbad, CA), homogenized, and frozen at −80 °C prior to their use in quantave real- time PCR (qPCR) analysis. We performed reverse transcription by using 50 µg.mL−1 oligo(dT)-20 primer and reverse transcriptase (Qiagen, Valencia, CA) according to the manufacturer’s instructions. Sample cDNA concentrations were determined by measuring absorbance at 260 nm by using the NanoDrop 2000 Spectrophotometer (Thermo Fisher Scientific, Waltham, MA). We amplified 200 ng sample cDNA with Fast SYBR Green Real-Time PCR master mix (Life Technologies) in a final volume of 20 µL by using the ABI Step One Machine. Gene expression was determined by using the delta delta CT method (Livak & Schmittgen, 2001), with glyceraldehyde 3-phosphate dehydrogenase (GAPDH) mRNA in each tissue sampled as reference, by using primers described in Table A.1 (Appendix) to determine gene expression. qPCR data were normalized by the mean −∆∆CT in untreated control (DMT1+/+) mice and expressed as 2−∆∆CT, i.e. relative fold expression.

Histology—Liver and spleen tissue samples were rinsed in phosphate-buffered saline, fixed by using 4% (w/v) paraformaldehyde (Thermo Fisher Scientific), and embedded in paraffin. Sections of thickness 4 μm were stained with hematoxylin and eosin. Images were acquired using the Olympus BH2 microscope and Olympus Magnafire digital image-capture system with the aid of AxioVision version 4.8.1 software (Zeiss).

Absorption of iron, copper, and manganese in the DMT1int/int mouse—We measured radiotracer metal absorption in female control mice and in DMT1int/int mice made iron replete by two I.P. injections (28 and 56 days) of iron dextran (12.5 mg elemental Fe per kg body weight) in phosphate-buffered saline solution (pH 7.4). 59Fe (specific activity 61.6 mCi.mg−1) and 54Mn (specific activity 196 mCi.mg−1) were obtained from Perkin–Elmer (Boston, MA), and 64Cu (specific activity 0.081–0.25 mCi.mg−1) was obtained from Washington University–St Louis. Conscious mice were administered 59Fe (0.1 µCi per gram body weight as

64 FeSO4 in 1 mM NaCl, 1 mM L-ascorbic acid, 1 mM L-glutamine), Cu (1 µCi per gram body weight as CuCl2

54 in 10 mM NaCl, 1 mM L-ascorbic acid, 1 mM L-histidine), or Mn (0.1 µCi per gram body weight as MnCl2 in 1 mM NaCl, 1 mM L-ascorbic acid, 1 mM L-glutamine) via oral−intragastric gavage following an overnight fast.

Blood was collected into heparinized hematocrit tubes by tail-nick incision at 15, 45 min, 2 and 4 h. Blood samples were centrifuged for measurement of hematocrit and then expelled into 20-mL scintillation vials containing 1 mL Solvable solution (Perkin–Elmer). Mice were euthanized by CO2 asphyxiation at 4 h after which we collected ≈100 mg liver into 1 mL Solvable. Freshly isolated small intestine was flushed with ice-cold wash solution containing 130 mM NaCl, 5 mM KCl, 1 mM CaCl2, 1 mM

MgCl2, 1 mM nitrilotriacetic acid (NTA), 5 mM 2-(N-morpholino)ethanesulfonic acid, and buffered to pH 7.4 by using N′,N′-diethylpiperazine, except that 1 mM L-ascorbate and 1 mM L-histidine were used in place of NTA for the 64Cu study. Duodenal enterocytes were collected (as described above) into 1.5-mL microcentrifuge tubes containing 0.99 mL Solvable plus 10 µL protease inhibitor cocktail I (EMD Biochemicals). 0.5 mL of this sample was transferred to a 20-mL scintillation vial with 1 mL Solvable and the remainder was used to quantify protein concentration by using the Pierce BCA assay (Thermo Fisher Scientific). Tissue samples in Solvable were further processed according to the manufacturer’s protocol. We added Scintisafe−30% cocktail (Thermo Fisher Scientific) and measured sample radioactivity by using liquid-scintillation counting.

Echocardiography of hearts in the DMT1int/int mouse—We measured left-ventricular size and function by echocardiography in female DMT1+/+ and DMT1int/int mice. We obtained systolic and diastolic volume measurements (µL), ejection fraction (%), fractional shortening (%), and cardiac output (mL.min–1) from long-axis view and M-mode images in DMT1+/+ and DMT1int/int mice.

Statistics—Statistical analyses were performed using SigmaPlot version 13 (Systat Software) with critical significance level α = 0.05 and β = 0.20. Data are presented as arithmetic mean and standard deviation (SD) for n independent observations, with the following exceptions: (i) qPCR data (relative fold expression) are expressed as geometric mean and SD, and (ii) enterocyte and liver radiotracer content (Figure 2.9) are presented as box plots (median, interquartile range, and whiskers representing 10th and 90th percentiles). Except where noted, between-group comparisons were made by using (i) two-way analysis of variance (ANOVA) controlled for unequal sample size by general linear model (GLM) or Kaplan–Meier log-rank test (survival analysis) followed by pairwise multiple comparisons by using the Holm–Šidák test when appropriate, or (ii) Student’s t tests (or Welch’s test to control for unequal sample size or variance) controlled by a false-discovery rate (FDR) procedure (Curran-Everett, 2000) in which we declare that a significant effect exists only when the individual P value (Pi) is less than the individual critical significance ∗ level ( ) computed for each comparison tested. Red-cell morphology scores were tested by using one- way ANOVA on ranks; blood radiotracer content and hematocrit as a function of time post intragastric dose were analyzed by using repeated-measures two-way ANOVA; and tissue radiotracer content (data for which were not normally distributed) was analyzed by using the Mann–Whitney rank–sum test.

Because there was a lack of any observed effect in the Cre control study (experiment i, described above in the Characterization of the DMT1int/int mouse model section) and the Het study (experiment iv), I have determined the power to detect a specific effect (δ) on [Hgb] in those studies with α = 0.05. I took data from untreated DMT1+/+ mice in experiment ii as an independent control to provide estimates of the population standard deviation (γ) and mean (µ), and selected < 12.9 g.dL–1 (a) as an indicator of anemia (WHO, 2011). The meaningful specified effect was therefore δ = µ – a. I considered the study sufficiently powered if power ≥ 0.8 (i.e. 1 – β).

2.4 Results

2.4.1 Intestine-specific DMT1 knockout mouse model

We generated an intestine-specific DMT1 knockout mouse model by crossing the floxed DMT1 (DMT1fl/fl) (Gunshin et al., 2005a) and villin–Cre transgenic (Vil–Cre) (el Marjou et al., 2004) mouse lines. In postnatal Vil–Cre mice, Cre/loxP recombination is observed exclusively and homogeneously in intestinal epithelial cells (el Marjou et al., 2004). Introduction of the Cre transgene (Cre+) alone in male 129S6 mice had no effect on body weight, hematological variables, blood-iron variables, spleen or heart weights, or iron

33 stores (Table 2.1). Gunshin et al previously found that introduction of the loxP sites into the DMT1 (SLC11A2) gene (i.e. DMT1fl/fl) did not result in any observable phenotype (Gunshin et al., 2005a).

Table 2.1. Effect of introducing the Cre transgene in 129S6 mice.

P , Significant Cre– Cre+ i by FDR?

Body weight (g) 24.9 ± 5.8 (6) 24.7 ± 2.1 (8) 0.91, No

Hematocrit (%) 45.8 ± 2.2 (6) 45.9 ± 2.2 (8) 0.97, No

Red blood cell count (× 106 / μL) 9.1 ± 0.5 (5) 9.3 ± 0.5 (7) 0.55, No

Mean corpuscular volume (fL) 53.0 ± 1.7 (5) 53.3 ± 0.5 (6) 0.66, No

[Hemoglobin] (g.dL−1) 14.7 ± 0.8 (5) 14.6 ± 0.6 (7) 0.81, No

Mean corpuscular hemoglobin (pg) 16.2 ± 0.3 (5) 15.7 ± 0.4 (7) 0.042, No

Mean corpuscular [hemoglobin] (g.dL−1) 30.6 ± 1.3 (5) 30.1 ± 0.7 (7) 0.45, No

Serum Fe (μM) 39.9 ± 15.5 (6) 20.8 ± 7.9 (8) 0.011, No

Transferrin saturation (%) 67.9 ± 14.5 (5) 49.9 ± 19.3 (8) 0.10, No

Spleen weight (mg/g body weight) 2.1 ± 0.7 (6) 2.2 ± 0.2 (8) 0.86, No

Heart weight (mg/g body weight) 5.1 ± 0.6 (6) 4.9 ± 0.1 (7) 0.43, No

Liver nonheme Fe (μmol.g−1 wet tissue) 5.3 ± 0.8 (6) 4.1 ± 0.3 (8) 0.004, No

Heart nonheme Fe (μmol.g−1 wet tissue) 1.0 ± 0.3 (6) 1.3 ± 0.1 (8) 0.030, No

Spleen nonheme Fe (μmol.g−1 wet tissue) 17.2 ± 4.1 (5) 20.2 ± 3.7 (8) 0.20, No Data are mean ± SD (n). Body and tissue weights, hematological and blood-iron variables, and tissue nonheme iron were analyzed in male mice (wildtype for DMT1) lacking (Cre−) or hemizygous (Cre+) for the villin-Cre transgene. Mean (SD) age of Cre− mice, 106 (35) d (n = 6), did not differ from that of Cre+ mice, 112 (7) d (n = 8) (P = 0.62). Student’s t tests and the FDR procedure revealed that no variable differed between Cre− and Cre+. Power analysis of the study to find a specified effect on [Hgb] (see methods) was 0.99.

In subsequent experiments, we used DMT1fl/fl mice without the Cre transgene as control mice (for simplicity, hereafter described as DMT1+/+) and DMT1fl/fl | Cre+ mice as our intestine-specific DMT1 knockout model (hereafter described as DMT1int/int). We confirmed efficient knockout of intestinal DMT1

34 by using qPCR analysis of an enriched enterocyte preparation with the use of primers targeting the floxed region of DMT1 (Table A.1, Appendix). Intestinal expression of full-length DMT1 in the untreated DMT1int/int mouse was approximately 5% that in DMT1+/+ (Figure 2.1A).

Previously, immunoblot analysis indicated that DMT1 was absent from duodenal epithelial cells of DMT1int/int mice, and Southern blot analysis of several tissues isolated from DMT1int/int mice revealed that the deletion of the floxed region was confined to intestinal tissues (Gunshin et al., 2005a).

Ablation of intestinal DMT1 resulted in early mortality (Figure 2.1B). Mean (± SE) survival time in DMT1int/int mice was 186 ± 7 d (n = 125) compared with 565 ± 14 d (n = 215) in DMT1+/+ mice (P < 0.001).

Figure 2.1. Intestinal DMT1 knockout mouse model. (A) qPCR analysis of the expression of full-length DMT1 in male DMT1-floxed mice lacking (DMT1+/+, n = 3) or bearing the villin-Cre transgene (DMT1int/int, n = 4) at approximately 120 d of age. We used a primer targeting the floxed region of DMT1 (exon 7–9) (Table A.1, Appendix). Data are geometric mean, SD relative fold expression and are unitless; P = 0.008 by Student’s t test. (B) Effect of ablation of intestinal DMT1 on survival in the mouse. Survival analysis (over the study period 1–600 d) of untreated DMT1+/+ (black circles, n = 215), untreated DMT1int/int (black diamonds, n = 125), as well as DMT1+/+ (gray circles, n = 11) and DMT1int/int (gray diamonds, n = 56) given Fe (I.P.) at age approximately 28 and 56 d. Steps denote events, symbols denote censored observations. Log-rank statistic, P < 0.001; Holm–Šidák pairwise comparisons revealed that untreated DMT1int/int differed from untreated DMT1+/+ and Fe-treated DMT1int/int (P < 0.001) but that no other groups differed from one another (0.071 ≤ P ≤ 0.97).

2.4.2 Intestine-specific knockout of DMT1 produces a severe hypochromic–microcytic anemia resulting from iron deficiency

Intestine-specific knockout of DMT1 in the mouse resulted in a hypochromic–microcytic anemia resulting from iron deficiency accompanied by splenomegaly and cardiomegaly.

Peripheral blood smears of male DMT1int/int mice (aged approximately 120 d) revealed marked reticulocytosis and a hypochromic–microcytic anemia, judged from the reduction in erythrocyte size and the increased area of central pallor in erythrocytes compared with DMT1+/+ mice (Figure 2.2A). DMT1int/int mice exhibited marked polychromasia (P < 0.001), marked anisocytosis (P < 0.001), and moderate poikilocytosis (P = 0.005) (severity scores 0–4 from visual inspection, data not shown). Automated CBC revealed significant reductions in hematocrit, red blood cell count, mean corpuscular volume (MCV), hemoglobin concentration, and mean corpuscular hemoglobin (MCH) in DMT1int/int mice compared with DMT1+/+ mice (Figure 2.3A–E). We observed no effect on mean corpuscular hemoglobin concentration (MCHC) in DMT1int/int mice compared with DMT1+/+ mice (Figure 2.3F), a result that may be explained by the coupled effects of reductions in MCV and MCH.

Serum iron and transferrin saturation were decreased in DMT1int/int mice compared with DMT1+/+ mice (Figure 2.3G,H). Parenteral administration of iron dextran (12.5 mg Fe I.P. at age approx. 28 and 56 d) was sufficient to substantially increase blood-iron variables (Figure 2.3G,H), and prevented or reversed the changes in hematological variables and early mortality observed in untreated DMT1int/int mice (Figures 2.1B, 2.3A–E).

Mean body weight of untreated DMT1int/int mice was 19.3 g (SD 3.1 g) (n = 12), significantly lower than that of their DMT1+/+ littermates, 29.8 (2.7) g (n = 12), whereas body weight did not differ between DMT1int/int and DMT1+/+ mice that had received parenteral Fe (interaction, P < 0.001).

The DMT1int/int mouse exhibited marked splenomegaly and cardiomegaly, and a modest increase in kidney weight compared with DMT1+/+ mice (Figure 2.4A–C). Histological examination of the spleen in the DMT1int/int mouse revealed a relative contraction and darkening of the white pulp, and an expansion of red pulp, in contrast to a greater relative volume of white pulp in the spleen of DMT1+/+ mice (Figure 2.2B). Splenomegaly and red-pulp expansion are consistent with an anemia and may indicate increased extramedullary hematopoiesis (Cesta, 2006) in the DMT1int/int mouse.

Figure 2.2. Blood and tissues of control (DMT1+/+) and intestinal DMT1 knockout (DMT1int/int) mice. (A) Peripheral blood smears prepared by using Wright–Giemsa stain (scale bars, 10 µm). (B) Histology of spleen prepared by using hematoxylin and eosin stain (scale bars, 100 μm). (C) Histology of liver prepared by using hematoxylin and eosin stain (scale bars, 40 μm).

Figure 2.3. Hematological and blood-iron variables in male control and intestinal DMT1-null mice. Male control (DMT1+/+) and intestine-specific DMT1 knockout (DMT1int/int) mice (approximately 120 d of age) were untreated (solid bars) or given Fe (I.P.) at age approximately 28 and 56 d (hatched bars). (A) Hematocrit. (B) Red blood cell count. (C) Mean corpuscular volume. (D) Hemoglobin (Hgb) concentration. (E) Mean corpuscular hemoglobin. (F) Mean corpuscular hemoglobin concentration. (G) Serum iron concentration. (H) Transferrin saturation. Statistical analyses (untreated, n = 5–12 per group; Fe-injected, n = 3–4 per group): with the exception of MCHC (P = 0.063), Two-way ANOVA revealed interactions for all variables (serum iron, P = 0.030; all other comparisons, P < 0.001). In A–E, G, H: within DMT1int/int, Fe (I.P.) injection produced a significant difference (P ≤ 0.003) and, within Fe (I.P.)-injected, DMT1int/int did not differ from DMT1+/+ (P ≥ 0.35).

Liver nonheme iron was depleted in DMT1int/int mice compared with DMT1+/+ mice (Figure 2.4D), an effect not associated with any observed morphological change in the liver (Figure 2.2C). Nonheme iron was also depleted in the spleen of DMT1int/int mice compared with DMT1+/+ mice (Figure 2.4E). We found no significant changes in nonheme iron content in heart, kidney, and skeletal muscle of untreated mice (Figure 2.4F–H); however, modest reductions in heart and kidney nonheme iron content of untreated DMT1int/int mice may have been masked by the variability associated with large increases in tissue nonheme iron following parenteral iron treatment.

Figure 2.4. Tissue weights and tissue nonheme iron content in male control and intestinal DMT1-null mice. Male control (DMT1+/+) and intestine-specific DMT1 knockout (DMT1int/int) mice (approximately 120 d of age) were untreated (n = 9–12, solid bars) or given Fe (I.P.) at age approximately 28 and 56 d (n = 3–4, hatched bars). (A–C) Spleen, heart and kidney weight normalized by body weight: Two-way ANOVA revealed interactions (P ≤ 0.029) for each organ; within Fe (I.P.)-injected, DMT1int/int did not differ from DMT1+/+ (P ≥ 0.65). (D) Liver nonheme iron content: Two-way ANOVA revealed an interaction (P < 0.001); within untreated, DMT1int/int differed from DMT1+/+ (P < 0.001). (E) Spleen nonheme iron content: Two- way ANOVA revealed main effects of genotype (P < 0.001) and Fe (I.P.) injection (P < 0.001) but no interaction (P = 0.12). (F–H) Tissue nonheme iron content: Two-way ANOVA revealed no effects of genotype for heart (P = 0.098), kidney (P = 0.57) or skeletal muscle (gastrocnemius) (P = 0.57).

We found similar results in female mice (3–4 mo of age) lacking intestinal DMT1. Female DMT1int/int mice were growth limited (Figure 2.5A) and exhibited a hypochromic–microcytic anemia characterized by reductions in hematocrit, RBC count, MCV, and Hgb concentration (Figure 2.5B–E); however, we observed no significant difference in MCH (Figure 2.5F). We attribute this hypochromic– microcytic anemia in female DMT1int/int mice to severe iron deficiency since it was associated with substantially lower serum iron and transferrin saturation compared with DMT1+/+ mice, splenomegaly, cardiomegaly, and depleted nonheme iron stores (liver, spleen) (Figure 2.5G–L).

Figure 2.5. Effects of intestinal ablation of DMT1 in female mice. Body and tissue weights, hematological and blood-iron variables, and tissue nonheme iron were analyzed in female control (DMT1+/+, n = 8–13) and intestine-specific DMT1 knockout (DMT1int/int, n = 3–6). Mean (SD) age of DMT1+/+ mice was 170 (53) d (n = 13) and that of DMT1int/int mice, 103 (18) d (n = 6). (A) Body weight. (B) Hematocrit. (C) Red blood cell count. (D) Mean corpuscular volume. (E) Hemoglobin (Hgb) concentration. (F) Mean corpuscular hemoglobin (MCH). (G) Serum iron concentration. (H) Serum transferrin saturation. (I–J) Spleen and heart weight normalized by body weight. (K) Liver nonheme iron content. (L) Spleen nonheme iron content. Student’s t tests revealed that DMT1+/+ and DMT1int/int differed for all variables (P ≤ 0.006) except MCH (P = 0.078).

Regardless of sex, intestine-specific ablation of DMT1 produces a hypochromic–microcytic anemia due to severe iron deficiency. The anemia phenotype is associated with growth limitation, splenomegaly, and cardiomegaly. That parenteral iron treatment prevented or corrected anemia phenotype of the

DMT1int/int mouse (i) confirms that the anemia resulted solely from iron deficiency and (ii) provides further evidence that the primary or only lesion is confined to the intestine.

Although our statistical analyses did not directly compare variables between male and female mice, some differences in iron distribution are apparent between male and female control but not DMT1int/int mice (see Figure 2.4D,E cf. Figure 2.5K,L), possibly an effect arising from the use here of a mixed background. Nevertheless, our study reveals no sexual divergence in the critical functional role of intestinal DMT1 in iron homeostasis.

2.4.3 Effect of DMT1 intestinal knockout on metal homeostasis in mice

Tissue levels of total iron (i.e. heme iron and nonheme iron) were markedly decreased in the liver, heart, spleen, kidney, and skeletal muscle of male DMT1int/int mice compared with DMT1+/+ mice (Figure 2.6A–E). The effects were more profound in liver, spleen, and kidney than in heart and skeletal muscle, consistent with the depletion of iron stores; nevertheless total iron (which also includes myoglobin iron) was markedly depleted even in the critical muscle mass (cardiac muscle and gastrocnemius).

Figure 2.6. Tissue metal content in male control and intestinal DMT1-null mice. Tissue metal content (in units ng per mg of wet tissue) measured by inductively coupled plasma mass spectrometry (ICP–MS) in male control (DMT1+/+, n = 7) and intestine-specific DMT1 knockout (DMT1int/int, n = 8) mice (approximately 120 d of age). (A–E) Iron content in liver, heart, spleen, kidney, and muscle, respectively. (F–J) Copper content in liver, heart, spleen, kidney, and muscle, respectively. (K–O) Manganese content in liver, heart, spleen, kidney, and muscle, respectively. (P–T) Zinc content in liver, heart, spleen, kidney, and muscle, respectively. Data were analyzed using Student’s t tests controlled by the FDR procedure for 20

a int/int +/+ b comparisons: DMT1 differed from DMT1 (P ≤ 0.011). No other significant differences: (Pi = 0.041)

c int/int and (Pi = 0.029) were found not significant by the FDR procedure (i.e. DMT1 did not differ from DMT+/+); all other comparisons, 0.057 ≤ P ≤ 0. 57.

We found no meaningful changes in the tissue levels of copper, manganese, or zinc (Figure 2.6F– T). A modest decrease in skeletal muscle copper content of male DMT1int/int mice compared with DMT1+/+ mice was accompanied by an increase in copper content of the spleen (Figure 2.6H,J), and kidney manganese content was only modestly decreased in DMT1int/int mice (Figure 2.6N). Enterocyte total iron content was decreased in female DMT1int/int mice compared with DMT1+/+ mice but there were no differences in the levels of copper, manganese, or zinc (Figure 2.7).

Figure 2.7. Enterocyte metal content in female control and intestinal DMT1-null mice. Enterocyte metal content measured by inductively coupled plasma mass spectrometry (ICP–MS) in female control (DMT1+/+, n = 4) and intestine-specific DMT1 knockout (DMT1int/int, n = 3) mice (approximately 120 d of age). (A–D) Enterocyte total iron, copper, manganese, and zinc content, respectively. Data were analyzed using Student’s t tests: aDMT1int/int differed from DMT1+/+ (P = 0.043); no other effects (0.089 ≤ P ≤ 0.50).

2.4.4 Effect of intestinal DMT1 ablation on the regulation of genes involved in iron absorption and homeostasis

We examined the effect of intestine-specific ablation of DMT1 on the expression of genes involved in iron and copper homeostasis. We used a DMT1 primer complementary to a region upstream of the first loxP site (and downstream of the transcription start site) such that it would serve as a reporter of the regulatory signal independent of expression of functional DMT1. We found that the DMT1 gene was upregulated (30-fold) and the intestinal ferrireductase Cybrd1 (DcytB) upregulated 48-fold in enterocytes of DMT1int/int mice compared with DMT1+/+ mice (Figure 2.8A,B); however, we found no change in the expression of the Steap2 gene coding a second candidate brush-border ferrireductase (Figure 2.8C). mRNA expression of

43 the basolateral iron exporter ferroportin was unchanged in DMT1int/int mice (Figure 2.8D), as was the expression of the Ctr1 gene coding the apical copper transporter (Figure 2.8E). Ablation of intestinal DMT1 downregulated the expression of the iron-regulatory hormone liver Hamp1 gene (hepcidin) to ≈1% that of DMT1+/+ mice (Figure 2.8F). The changes in expression levels of DMT1, Cybrd1, and Hamp1 in DMT1int/int mice were corrected by parenteral Fe injection. Iron treatment increased the expression of Ctr1 in both DMT1int/int and DMT1+/+ mice, i.e. the effect of iron on Ctr1 was independent of DMT1.

Figure 2.8. Gene expression in male control and intestinal DMT1-null mice. Gene expression of intestinal ferrireductases and transporters (A–E) and the hepatic iron-regulatory hormone hepcidin (Hamp1, F) in male control (DMT1+/+) and intestine-specific DMT1 knockout (DMT1int/int) mice (approximately 120 d of age) that were untreated (n = 6–10, dark gray bars) or given Fe (I.P.) at age approximately 28 and 56 d (n = 3–4, light gray bars). Data are expressed as geometric mean (and SD) relative fold expression and are unitless. (A) Intestinal divalent metal-ion transporter-1 (DMT1), primer upstream of the first loxP site. Two-way ANOVA revealed an interaction (P = 0.028); iron-treated groups did not differ by genotype (P = 0.14). (B) Intestinal Cybrd1 (DcytB), interaction (P = 0.047); iron-treated groups did not differ by genotype (P = 0.25). (C) Intestinal Steap2, no effects (0.26 ≤ P ≤ 0.99). (D) Intestinal ferroportin (Fpn), no effects (0.24 ≤ P ≤ 0.28). (E) Intestinal Ctr1, main effect of iron treatment (P = 0.002) but not of genotype (P = 0.27), and no interaction (P = 0.52). (F) Liver Hamp1 (hepcidin), interaction (P < 0.001); iron-treated groups did not differ by genotype (P = 0.61).

2.4.5 Role of intestinal DMT1 in the absorption of iron, copper, and manganese

To examine directly the role of intestinal DMT1 in the absorption of iron, copper, and manganese, we measured metal absorption in control (DMT1+/+) and DMT1int/int mice fed a radiotracer metal dose via oral– intragastric gavage. Although serial blood collections were small, we measured hematocrit over the 4-h time course of these experiments to ensure against large changes in hematocrit or the onset of hypovolemic shock. We observed in the 59Fe and 64Cu studies only very modest reductions in hematocrit over time and no time-dependent differences between DMT1+/+ and DMT1int/int (Figure 2.9D,H); we observed no changes in hematocrit in the 54Mn study (Figure 2.9L).

In DMT1+/+ mice, 59Fe appeared in the blood within 15 min after the oral dose and peaked around 2 h. In contrast, 59Fe appearance in the blood of DMT1int/int mice was substantially blunted, 12% that of DMT1+/+ mice (area under the curve, 0–4 h) (Figure 2.9A). We obtained for DMT1+/+ mice at 4 h robust 59Fe signals in enterocytes and liver (the latter tissue being expected to rapidly clear 59Fe from the portal and peripheral circulation); however, 59Fe content of enterocytes and liver of DMT1int/int mice was only a small fraction of that observed for DMT1+/+ mice (Figure 2.9B,C). These data indicate that intestinal DMT1 is required for iron absorption. The lack of 59Fe within enterocytes 4 h after an oral dose provides the first demonstration that the primary lesion in the DMT1int/int mouse is abolished iron uptake at the intestinal brush border (apical membrane). (We assume that we had bypassed the need for luminal reduction of

59 iron by providing the Fe as FeSO4 in ascorbic acid.)

Blood 64Cu content in DMT1+/+ mice peaked later than did 59Fe (the 64Cu dose was ten-fold greater than the 59Fe given) and was no different in DMT1int/int mice (Figure 2.9E). 64Cu content of enterocytes and liver did not differ between DMT1+/+ and DMT1int/int mice (Figure 2.9F,G). 54Mn appeared in blood within 15 min and did not differ between DMT1int/int and DMT1+/+ mice (Figure 2.9I). Tissue 54Mn content in enterocytes and liver of DMT1int/int mice were no different from DMT1+/+ mice (Figure 2.9J,K). These data indicate that copper and manganese absorption do not rely upon intestinal DMT1.

Figure 2.9. Intestinal iron, copper, and manganese absorption in iron-replete male control (DMT1+/+) and intestinal DMT1 knockout (DMT1int/int) mice. Conscious mice were dosed with 0.1 µCi 59Fe or 54Mn or 1 µCi 64Cu per gram body weight via oral–intragastric gavage. Blood radiotracer data (DMT1+/+, blue symbols, lines; and DMT1int/int, green symbols, lines) were analyzed by Two-way ANOVA with repeated measures over time but, for clarity, are displayed as mean, SD. (A–C) 59Fe absorption in DMT1+/+ (n = 3) and DMT1int/int (n = 7). (A) Blood 59Fe content as a function of time after dosing, interaction (P < 0.001) (B) Enterocyte 59Fe content at 4 h, analyzed by using the rank–sum test (P = 0.017). (C) Liver 59Fe content at 4 h (P = 0.033). (D) Serial hematocrit measurements for DMT1+/+ (blue symbols and lines) and DMT1int/int (green symbols and lines) mice as a function of time after the oral radiotracer dose. Each line-and-scatter plot depicts the

46 data for an individual mouse. Two-way ANOVA with repeated measures (time) revealed a main effect of time (P < 0.001) but not genotype (P = 0.54); no interaction (P = 0.26). (E–H) 64Cu absorption in DMT1+/+ (n = 7) and DMT1int/int (n = 5) mice. (E) Blood 64Cu content as a function of time after dosing, no interaction (P < 0.10) (F) Enterocyte 64Cu content at 4 h (P = 0.88). (G) Liver 64Cu content at 4 h (P = 0.94). (H) Serial hematocrit measurements for DMT1+/+ (blue) and DMT1int/int (green) mice as a function of time after the oral radiotracer dose. Each plot depicts the data for an individual mouse. Two-way ANOVA with repeated measures revealed main effects of time (P < 0.001) and genotype (P = 0.008) but no interaction (P = 0.31). (I–L) 54Mn absorption in DMT1+/+ (n = 3) and DMT1int/int (n = 5) mice. (I) Blood 54Mn content as a function of time after dosing, no interaction (P = 0.34) (J) Enterocyte 54Mn content at 4 h (P = 0.56). (K) Liver 54Mn content at 4 h (P ≥ 0.99). (L) Serial hematocrit measurements for DMT1+/+ (blue) and DMT1int/int (green) mice as a function of time after the oral radiotracer dose. Each plot depicts the data for an individual mouse. Two-way ANOVA with repeated measures revealed no effects (P ≤ 0.48).

2.4.6 Haplosufficiency of intestinal DMT1 in mice

To determine the effect of omitting a single copy of the DMT1 gene in the intestine, we examined heterozygous mice (i.e. DMT1+/int). Mean body weight in DMT1+/int mice (≈150 d of age) was 27.8 g (SD 4.6 g), no different from control (DMT1+/+), 31.7 (6.5) g (n = 11–14). Spleen and heart sizes were normal (data not shown). Hematological variables, serum iron, transferrin saturation, and liver nonheme iron in DMT1+/int mice did not differ from those of DMT1+/+ mice (Figure 2.10A–F). Therefore, heterozygosity of intestinal DMT1 produced no observed phenotypic change in healthy mice fed a standard diet.

We examined the expression of genes involved in iron absorption and metabolism to test whether the normal-iron status of the heterozygote was achieved by upregulation of the iron-absorptive machinery. Intestinal mRNA levels of DMT1 (upstream of loxP), Cybrd1, and Fpn in DMT1+/int mice did not differ from those of DMT1+/+ mice (Figure 2.10G–I). Although we observed no regulation at the mRNA level, it is possible that protein levels of DMT1, Cybrd1, or ferroportin are elevated in the intestine of heterozygous DMT1+/int mice, e.g. as a result of increased protein stability. Nevertheless, we found no change in the hepatic expression of Hamp1 (Figure 2.10J) coding hepcidin, the primary regulator of ferroportin protein levels (Ganz, 2013).

Figure 2.10. Phenotypic characterization of male mice heterozygous for intestinal DMT1. Male control (DMT1fl/+|Cre−; A–F, n = 6–11; G–J, n = 3) and intestine-specific DMT1 heterozygous (DMT1+/int; A–F, n = 13–14; G–J, n = 4) mice at approximately 150 d of age. (A) Hematocrit. (B) Hemoglobin (Hgb) concentration. Power of the study to find a specified effect (see methods) was 0.99. (C) Mean corpuscular volume (MCV). (D) Serum iron concentration. (E) Serum transferrin saturation. (F) Liver nonheme iron content. (G–J) Expression of iron-related intestinal and hepatic genes. qPCR data are expressed as geometric mean (and SD) relative fold expression and are unitless. Student’s t tests and the FDR procedure revealed no differences between DMT1+/+ and DMT1+/int (all comparisons, 0.058 ≤ P ≤ 0.77, except liver nonheme iron, Pi = 0.011, which was found not significant by the FDR procedure).

2.4.7 Role of intestinal DMT1 in the neonate

I tested the effect of loss of intestinal DMT1 in the neonate on iron homeostasis by measuring hematocrit in males and females at 7, 14, and 21 days of age and serum iron in females at 14 and 21 days of age. I found that regardless of sex, DMT1int/int mice exhibited a reduction in hematocrit at 21 days (Figure 2.11A). I have preliminary data that DMT1int/int neonates were iron deficient as early as 14 days of age and maintained the deficiency through 21 days of age (Figure 2.11B).

Figure 2.11. Hematocrit and Serum Fe in control (DMT1+/+) and DMT1int/int neonate mice. (A) Hematocrit in male (square symbols) and female (circle symbols) control (DMT1+/+, blue fill) and intestine-specific DMT1 knockout (DMT1int/int, green fill) mice at age 7, 14, and 21 days. Data are mean, SD for n = 4-18 per group: Three-way ANOVA revealed an interaction (P = 0.003); and a 2-way interaction of age × genotype (P < 0.001). (B) Preliminary serum Fe in female DMT1+/+ (n = 2-4, blue circles) and DMT1int/int (n = 1-2, green circles) mice at age 14 and 21 days. Each point is an individual mouse and the line indicates the averages.

To examine the condition of DMT1int/int hearts, we measured left-ventricular size and function by echocardiography. I found that DMT1int/int mice exhibited dilated cardiomyopathy (Figure 2.12B) compared with DMT1+/+ (Figure 2.12A). DMT1int/int hearts exhibited moderately decreased left-ventricular function evident by the increase in stroke volume, decrease in ejection fraction, decrease in fractional shortening, and increase in cardiac output (Figure 2.12C).

Figure 2.12. Echocardiography in female DMT1+/+ and DMT1int/int mice. (A) Long-axis view and M-mode images in a DMT1+/+ mouse demonstrating normal left-ventricular size and function. (B) Long-axis view and M-mode images in a DMT1int/int mouse showing dilated cardiomyopathy with moderately decreased left-ventricular function. (C) Cardiac function in DMT1+/+ and DMT1int/int mice. Each symbol type represents an individual wildtype (blue) or DMT1int/int (green) mouse.

2.5 Discussion

2.5.1 Intestinal DMT1 is required for mammalian iron absorption and homeostasis

We have performed a comprehensive characterization of a mouse model lacking intestinal DMT1. Our key findings were that (i) intestine-specific ablation of DMT1 produces a severe iron-deficiency anemia in the mouse; (ii) absorption of 59Fe from an oral dose was virtually abolished in the DMT1int/int mouse whereas absorption of 64Cu and 54Mn did not differ from that in DMT1+/+ mice; and (iii) tissue levels of copper, manganese, and zinc in adult DMT1int/int mice were normal whereas serum iron, and tissue levels of nonheme iron and total iron were profoundly depleted.

That the anemia phenotype of the DMT1int/int mouse could be prevented or rescued by intraperitoneal iron injection, thus bypassing the intestine, indicated that the lesion was specific to the intestine. In a similar manner, intraperitoneal copper administration rescued the lethality observed in a Ctr1int/int mouse (Nose et al., 2006). We found no evidence to support a defect in the ability of the liver to regulate iron metabolism since (i) the morphology of the DMT1int/int mouse liver was normal, and (ii) hepcidin expression in the liver was strongly suppressed, as expected when serum iron and hepatic iron stores are extremely low (Ganz, 2013).

More precisely, the primary defect in the DMT1int/int mouse was an inability to take up iron at the intestinal brush-border membrane. This conclusion is consistent with the immunolocalization of DMT1 on the apical membrane of enterocytes in mouse, rat, and human intestine (Canonne-Hergaux et al., 1999; Zoller et al., 2001; Okazaki et al., 2012). No other nonheme-iron transporter has been identified at the mammalian intestinal brush border. Whereas Zrt-/Irt-like protein-14 (ZIP14) exhibits iron-uptake activity in vitro (Liuzzi et al., 2006; Pinilla-Tenas et al., 2011) and is strongly expressed in the intestine (Wang et al., 2012), its expression is enriched on the basolateral membrane (Guthrie et al., 2015). Moreover, ZIP14- mediated iron uptake in vitro is optimal at pH 7.5 (Pinilla-Tenas et al., 2011) and therefore a functional role for ZIP14 in intestinal apical iron uptake from an acidic microenvironment (Said et al., 1986) is doubtful, at least in the mature mammal.

Our study demonstrates that intestinal DMT1 is essential for iron homeostasis in the adult mouse and its ablation results in early mortality. We observed cardiomegaly in the DMT1int/int mouse by 4 months of age, and preliminary evidence (J. Rubinstein, A. Shawki, J. N. Lorenz, B. Mackenzie, unpublished data) indicates that intestinal ablation of DMT1 produces left-ventricular dilation by 180 d of age (cf. mean survival of 186 d). We therefore suspect that a dilated cardiomyopathy explains the early mortality in the DMT1int/int mouse.

Whereas the DMT1int/int mouse is born iron replete—it has normal liver nonheme iron stores at birth (Gunshin et al., 2005a)—we have nevertheless considered what may account for its survival to several months of age and for the residual 59Fe absorption from an intragastric dose. It is possible that DMT1 is not required in the suckling mammal which may obtain sufficient dietary iron by one of the following mechanisms: (i) apical uptake of lactoferrin, (ii) apical uptake of free iron via an unknown transporter active in the newborn, or (iii) absorption of iron via a paracellular pathway in the immature intestine. Nevertheless, we have preliminary evidence of reduced serum iron in the DMT1int/int mouse by 14 d of age and a mild anemia by 21 d (Prakash et al., 2015) (Figure 2.11). We think it is more likely that the residual iron absorption results from incomplete Cre-mediated excision of the floxed DMT1 gene in

51 the enterocytes of the DMT1int/int mouse, based on the following observations: (i) the expression of full length DMT1 mRNA in an enriched enterocyte preparation from untreated DMT1int/int mice was approximately 5% that of DMT1+/+ mice (Figure 2.1A), and (ii) intestinal mRNA expression of DMT1 is virtually confined to enterocytes and is not detectable in goblet cells or lamina propria (Gunshin et al., 1997). Our data therefore support the notion that DMT1 is the only mechanism by which nonheme iron is taken up at the intestinal brush border in the adult mammal.

Haploinsufficiency of intestinal DMT1 had been suggested previously (Gunshin et al., 2005a). We tested this possibility more directly in mice heterozygous for intestinal DMT1 and found no evidence of intestinal DMT1 haploinsufficiency. We found no change in hepcidin expression in liver nor of iron-related genes in the intestine; however, I did not test the possibility that transport protein levels were increased independently of hepcidin. Rare mutations in human DMT1 result in a severe hypochromic–microcytic anemia (Shawki et al., 2012). All probands identified so far have been homozygous or compound heterozygous for DMT1 mutations. That parents and siblings of the probands patients appear healthy further supports the view that one wildtype SLC11A2 allele is sufficient for mammalian iron absorption.

2.5.2 Intestinal DMT1 is not required for copper, manganese, or zinc absorption

In contrast to the deficiency in tissue iron levels in DMT1int/int mice, we found no meaningful change in tissue levels of copper, manganese, or zinc. We concluded that intestinal DMT1 is not required for zinc homeostasis, an anticipated outcome since (i) Zn2+ is a weak DMT1 substrate in vitro, relative to Fe2+ (Illing et al., 2012), and (ii) ZIP4, a metal transporter expressed at the intestinal brush border, appears to be sufficient for zinc absorption (Cousins, 2010; Jeong & Eide, 2013). Homozygous knockout of ZIP4 in the mouse is embryonic lethal and heterozygotes display hypersensitivity to zinc deficiency (Dufner-Beattie et al., 2007). Hereditary defects in human ZIP4 result in zinc deficiency and acrodermatitis enteropathica (Wang et al., 2002a).

Given that a Ctr1-independent copper-transport activity was previously characterized but not identified in Ctr1−/− mouse embryonic fibroblasts, and that DMT1 was previously suggested to import Cu+ or Cu2+ in intestinal cells or membrane vesicles, we evaluated a potential role for DMT1 in intestinal copper acquisition. The absorption of 64Cu from an intragastric dose did not differ between DMT1int/int and DMT1+/+ mice, confirming that copper does not rely on DMT1 for its absorption. We previously found that Ctr1, a high-affinity copper transporter (Lee et al., 2002a), is localized to both the apical membrane and endosomal compartments of mammalian enterocytes (Nose et al., 2010). Intestinal ablation of Ctr1 in the

52 mouse produced a severe copper deficiency and early lethality (Nose et al., 2006), establishing Ctr1 as the primary mechanism serving copper uptake at the intestinal brush border, and these studies together indicate no role for DMT1 in this process.

Moreover, intestinal ablation of DMT1 had no effect on 54Mn absorption or tissue manganese levels except for a modest decrease in renal manganese content. Others have observed decreased manganese transport in two intestinal preparations from the Belgrade rat model, by using closed duodenal loops in situ or intestinal brush-border membrane vesicles (Chua & Morgan, 1997; Knöpfel et al., 2005b). That the manganese concentrations used in the studies just cited were supraphysiological and higher than the intestinal luminal concentration expected following oral gavage in the present study may explain in part this discrepancy. Furthermore, DMT1 exhibits higher affinity for Fe2+ than it does for Mn2+ in vitro (Illing et al., 2012). Our data indicate that DMT1 plays no physiologically relevant role in manganese absorption; however, the possibility remains that DMT1 could contribute to manganese absorption in manganese overexposure or dietary iron deficiency. Alternative intestinal uptake systems for manganese have not been identified but candidates include ZIP14 which is strongly expressed in the intestine and exhibits manganese transport activity in vitro (Pinilla-Tenas et al., 2011; Wang et al., 2012). Consistent with the strong preference of DMT1 for iron over any other physiological substrate tested in vitro (Illing et al., 2012), our present data support the notion that intestinal DMT1 is specific to iron absorption, at least under normal physiological conditions.

2.5.3 Model of severe iron deficiency and iron-deficiency anemia

We attributed the severe hypochromic–microcytic anemia, splenomegaly, and cardiomegaly of the DMT1int/int mouse to iron deficiency. The effects of intestinal ablation of DMT1 on hematological and blood-iron variables were more profound than those typically observed in mice fed an iron-restricted diet (including the 129S6 strain) (Gunshin et al., 2005c). Meanwhile, more severe models such as phlebotomy or phenylhydrazine-induced hemolysis may reflect the erythropoietic stress more so than they do iron deficiency per se. The DMT1int/int mouse may therefore serve as a superior model of severe iron deficiency and iron-deficiency anemia. Moreover, the iron deficiency is so severe that we may also anticipate marked defects in biochemical pathways relying on iron-containing enzymes (Beutler, 2010), among which are enzymes involved in energetic pathways (Ganz, 2013). In addition, our data indicate that the DMT1int/int mouse ought to serve as a robust model in which to test whether (i) forms of dietary iron other than Fe(II)

53 and Fe(III) salts and chelates (e.g. transferrin, heme), and (ii) supplemental iron formulations utilize DMT1 for their intestinal absorption.

We observed compensatory regulation in the expression of some iron-related genes in untreated DMT1int/int mice. Among these was the expected downregulation of hepatic hepcidin expression. Both decreased serum iron and decreased liver nonheme iron result in suppression of this iron-regulatory hormone (Ganz, 2013). Regulation of systemic iron by hepcidin is mediated primarily by its effect on ferroportin—binding of hepcidin by ferroportin induces its internalization and degradation (Nemeth et al., 2004b). Although we did not measure protein, we expect that the marked suppression of liver hepcidin expression permits very high ferroportin protein levels in enterocytes of the DMT1int/int mouse. Expression of DMT1 (upstream of the loxP site) was strongly upregulated in enterocytes of DMT1int/int mice. A DMT1 isoform initiated in exon 1A and containing in its mRNA 3′-untranslated region an iron-responsive element (i.e. 1A/IRE(+) isoform) predominates in the intestine, and its expression is exquisitely upregulated in vitro by low-iron status (Gunshin et al., 1997; Canonne-Hergaux et al., 1999; Tabuchi et al., 2002; Hubert & Hentze, 2002).

DMT1 transports ferrous ion (Fe2+) but not ferric ion (Fe3+) (Illing et al., 2012) and, since most dietary iron is in the form of ferric ion, dietary iron must be reduced by luminal ascorbic acid (Mackenzie & Garrick, 2005) or by surface ferrireductases prior to its DMT1-mediated uptake into the enterocyte. We found that Cybrd1 (DcytB), a protein that exhibits ferrireductase activity in vitro (McKie et al., 2001; Wyman et al., 2008), was upregulated in the DMT1int/int mouse, consistent with other mammalian models of iron deficiency that reveal Cybrd1 as strongly iron-responsive (McKie et al., 2000; Collins et al., 2005; Collins, 2006). Some redundancy of intestinal ferrireductases had been suggested by the observation that ablation of Cybrd1 in the mouse had little or no observable effect even in animals fed a low-iron diet (Gunshin et al., 2005c). Another ferrireductase, Steap3, appears to be required for the reduction of endosomal iron taken up via the transferrin-receptor mediated pathway in erythroid precursors (Ohgami et al., 2005). A second family member, Steap2, is expressed in intestine and has been implicated as a putative intestinal metalloreductase (Ohgami et al., 2006; Knutson, 2007); however, our findings that the iron deficiency of the DMT1int/int mouse did not induce Steap2 expression in enterocytes, nor did parenteral iron injection suppress Steap2 expression, provide no evidence in support of a role for Steap2 in intestinal ferrireduction.

An effect of parenteral iron injection on the expression of Ctr1 was not of direct interest in this study since intestinal ablation of DMT1 had no effect on Ctr1 expression in untreated animals.

Nevertheless, our data suggest that excess systemic iron may upregulate intestinal Ctr1, possibly involving iron–copper competition in a system that is capable of driving Ctr1 expression.

We found no evidence of transcriptional regulation of intestinal ferroportin by iron deficiency in the intestinal DMT1-null model. Short-term (2-weeks on low-iron feed) or long-term (8 weeks low iron) iron deficiency has been observed to stimulate mRNA expression of Cybrd1, DMT1, and ferroportin in mouse intestinal cells primarily via a hypoxia-inducible factor-2 alpha (HIF2α)-mediated pathway (Ohgami et al., 2006; Ramakrishnan, 2002). The induction of ferroportin protein observed after an 8-week low-iron treatment, however, persisted even in the absence of intestinal HIF2α (Ramakrishnan, 2002). We expect that the marked suppression of liver hepcidin expression in the intestinal DMT1-null mouse permits very high levels of ferroportin protein. Meanwhile, we speculate that the virtual absence of iron from enterocytes of the DMT1int/int mouse may drive iron-regulatory protein-mediated repression of a ferroportin mRNA containing a 5′-iron-responsive element (FPN1A) despite the expected upregulation by the HIF2α pathway [reviewed, (Okazaki et al., 2012)].

2.5.4 How is the DMT1int/int mouse able to survive for up to 5 months?

The DMT1int/int mouse is born iron replete, with normal hepatic nonheme iron stores (Gunshin et al., 2005a). Placental maternofetal iron transfer does not appear to rely on DMT1, and we know that iron is tightly conserved in mammals. There is no doubt that the DMT1int/int mouse is unable to absorb enough iron to satisfy the erythropoietic need and, if some alternative system exists for the absorption of iron, it is not sufficient to compensate for the loss of intestinal DMT1 because the DMT1int/int mouse is anemic and cannot achieve iron homeostasis (at least after weaning).

We observed cardiomegaly in the DMT1int/int mouse by 4 months of age. We have preliminary evidence (J. Rubinstein, A. Shawki, J. N. Lorenz, B. Mackenzie, unpublished data) indicating that intestinal ablation of DMT1 in the mouse produces left-ventricular dilation—an appropriate, physiological compensatory response to the anemia—by 180 d of age (Figure 2.12B) compared with DMT1+/+ mice (Figure 2.12A). The cardiomegaly progressed to cardiomyopathy evident from an increase in cardiac output and a decrease in ejection fraction and fractional shortening around 6 months of age. We suspect that a dilated cardiomyopathy explains the early mortality in the DMT1int/int mouse (50% survival at 184 d of age), resulting in a failing heart.

Figure 2.13. Antarctic ice fish. Source: Uwe Kils, Wikimedia.org. The large hearts of these fish can easily be seen positioned between their fins.

It is remarkable that the DMT1int/int mouse survives several months. By 4 mo of age, it is essentially devoid of functional red blood cells, and presumably utilizing only dissolved oxygen, coupled with the appropriate physiological response of increase stroke volume. This is reminiscent of the Antarctic ice fish, the only vertebrate to lack erythrocytes and hemoglobin (Everson & Ralph, 1968). Dissolved oxygen in the cold waters of the Antarctic is at higher levels than that which is normally found in warmer waters. This allows for adequate passage of oxygen to the tissues of the ice fish; however this does not account entirely to the survival of these fish. The ice fish has evolved a large heart (Figure 2.13), reminiscent of the anoxic physiological response in the DMT1int/int mouse.

2.6 Conclusions

Intestine-specific ablation of DMT1 in the mouse produces a severe iron deficiency and iron-deficiency anemia, confirming a critical role for intestinal DMT1 in systemic iron homeostasis. DMT1 is essential for the uptake of iron at the intestinal brush-border membrane. Intestinal DMT1 is specific for iron absorption under physiological conditions and is not required for the absorption of copper, manganese, or zinc.

3 Roles of intestinal H+-transporters in driving iron absorption

Chapter 3 is in its entirety a draft, written by Ali Shawki, of the following manuscript in preparation

Shawki A, Engevik MA, Kim RS, Knight PB, Baik RA, Anthony SR, Worrell RT, Shull GE, Mackenzie B

(2015) Intestinal brush-border Na+/H+ exchanger-3 drives H+-coupled iron absorption in the mouse.

3.1 Abstract

Divalent metal-ion transporter-1 (DMT1), the principal or only mechanism by which nonheme iron is taken up at the intestinal brush border, is energized by the H+ electrochemical potential gradient. The provenance of the H+ gradient in vivo is unknown so we have explored a role for brush-border Na+/H+ exchangers by examining iron homeostasis and intestinal iron handling in mice lacking Na+/H+ exchanger- 2 (NHE2) or Na+/H+ exchanger-3 (NHE3). We observed modestly depleted liver iron stores in NHE2-null (NHE2–/–) mice stressed on a low-iron diet but no change in hematological or blood-iron variables, or the expression of genes associated with iron metabolism, compared with wildtype mice. Ablation of NHE3 more severely depleted liver iron stores regardless of diet. We observed decreases in blood-iron variables but no overt anemia in NHE3–/– mice on a low-iron diet. Intestinal expression of DMT1, Cybrd1, and Fpn was upregulated in NHE3–/– mice, and expression of liver Hamp1 (hepcidin) suppressed, compared with wildtype. Absorption of 59Fe from an oral dose was substantially impaired in NHE3–/– mice compared with wildtype. Our data point to an important role for NHE3 in generating the H+ gradient that drives DMT1- mediated iron uptake at the intestinal brush border.

3.2 Introduction

Iron deficiency—the most prevalent micronutrient deficiency worldwide—can result in iron-deficiency anemia, and neurological and developmental disorders in children (Beard, 2008; Beutler, 2010; Killip et al., 2007). Conversely, iron toxicity associated with iron-overload disorders such as hereditary hemochromatosis and thalassemia major can cause liver cirrhosis, cardiomyopathy, and endocrine disorders (Fleming & Ponka, 2012; Hershko, 2010). Since there exists no regulated mechanism for the excretion of iron, whole-body iron homeostasis is achieved by regulating iron absorption (Ganz, 2013).

Divalent metal-ion transporter-1 (DMT1) is the principal or only mechanism by which nonheme iron is taken up at the intestinal brush border (see Chapter 2). This widely expressed iron transporter also plays a critical role in erythroid iron assimilation (Canonne-Hergaux et al., 2002; Garrick et al., 1999) via the transferrin cycle, and functions similarly in many other cell types. We have shown that DMT1-mediated iron transport in vitro is functionally coupled with protons (Mackenzie et al., 2006; Gunshin et al., 1997); however, the provenance of the H+ electrochemical potential gradient driving DMT1-mediated iron uptake at the intestinal brush border in vivo is unknown.

The mucosal cell surface in the duodenum and proximal jejunum remains moderately acidic despite the progressive alkalization of the luminal contents. This “acidic microclimate” may reach pH 6.0 or below in human proximal jejunum in vivo and in small intestine of other mammals (Said et al., 1986; McEwan et al., 1990; Shimada & Hoshi, 1988; Lucas, 1983; McEwan et al., 1988). We propose that the intestinal brush- border acidic microclimate—thought to rely on the activity of Na+/H+ exchangers (Shimada & Hoshi, 1988)—play a crucial role in the absorption of iron.

Gastrointestinal Na+/H+ exchangers play essential roles in pH balance, and Na+ and fluid absorption (Zachos et al., 2005; Dudeja et al., 1996). Two isoforms are expressed at the brush border in the adult mammalian intestine. Na+/H+ exchanger-2 (NHE2) is localized to the apical plasma membrane and is found in enterocytes both at the absorptive surface and in the crypts (Dudeja et al., 1996; Malakooti et al., 1999; Bookstein et al., 1997; Hoogerwerf et al., 1996). Na+/H+ exchanger-3 (NHE3) is expressed predominantly in surface enterocytes in which it is localized both to the apical plasma membrane and to apical recycling endosomes (Zachos et al., 2005; Hoogerwerf et al., 1996; Repishti et al., 2001; Bookstein et al., 1994; Dudeja et al., 1996).

We have not included in our study Na+/H+ exchanger-8 (NHE8) and Na+/H+ exchanger-1 (NHE1). NHE8 is expressed in the brush-border membrane in the suckling mammal (Xu et al., 2005) prior to the ontogenic induction of NHE2 and NHE3 around weaning (Collins et al., 1997); however, NHE8 may play

58 only a minor role in acidifying the surface microclimate in the adult. Whereas NHE1 is ubiquitously expressed (Orlowski et al., 1992), it is exclusively localized to the basolateral membrane (Bookstein et al., 1994; Tse et al., 1991). Other NHE isoforms expressed in the intestine are confined to intracellular organelles (Donowitz et al., 2013; Attaphitaya et al., 1999; Orlowski et al., 1992; Numata et al., 1998) and were not therefore considered in this study.

We tested the hypothesis that the intestinal brush-border Na+/H+ exchangers NHE2 and NHE3 promote iron absorption by examining iron homeostasis and intestinal iron handling in mouse models lacking NHE2 and NHE3. Mice in which the SLC9A2 gene coding NHE2 is globally inactivated (i.e. NHE2–/–) exhibit no outwardly appearing disease phenotype (Schultheis et al., 1998a) regardless of morphological changes largely limited to gastric mucosa (Schultheis et al., 1998a), colonic crypts (Guan et al., 2006), and folliculo-stellate cell canaliculi (Miller et al., 2011). NHE2-null mice exhibit compromised recovery of intestinal barrier function (Moeser et al., 2008); however, these same mice did not exhibit intestinal morphological changes and ablation of NHE2 in NHE3-null mice does not exacerbate the phenotype of the NHE3-null mouse (Ledoussal et al., 2001). Knockout of the SLC9A3 gene coding NHE3 in mice produces defects in acid–base balance, and Na+ and fluid absorption (associated with a mild secretory diarrhea) (Schultheis et al., 1998b), and reveals NHE3 to be the predominant Na+/H+ exchanger at the intestinal brush border (Gawenis et al., 2002; Zachos et al., 2005).

3.3 Materials and methods

Reagents and media—Reagents were obtained from Sigma–Aldrich Corp. (St. Louis, MO) or Research Products International Corp. (Prospect, IL) unless otherwise indicated.

Expression of human DMT1 in Xenopus oocytes—We performed laparotomy and ovariectomy on adult female Xenopus laevis frogs (Nasco, Fort Atkinson, WI) under 3-aminoethylbenzoate methanesulfonate anesthesia (0.1% w/v in 1:1 water/ice, by immersion) following a protocol approved by the University of Cincinnati Institutional Animal Care and Use Committee. Ovarian tissue was isolated and treated with collagenase A (Roche Diagnostics Corp., Indianapolis, IN), and oocytes were isolated and stored at 17 °C in modified Barths’ medium as described (Mackenzie, 1999). We expressed in Xenopus oocytes the 1A/IRE(+) isoform of human DMT1, the product of the human SLC11A2 gene as described (Illing et al., 2012). Briefly, defolliculate stage V–VI oocytes were injected with approximately 50 ng of DMT1 RNA and incubated 6 days before being used in functional assays.

Radiotracer assays in oocytes expressing human DMT1—Oocytes were incubated at room temperature

(23 °C) in transport media containing 100 mM NaCl, 1 mM KCl, 0.6 mM CaCl2, 1 mM MgCl2, and 1 mM L- ascorbic acid, and buffered using 0–5 mM 2-(N-morpholino)ethanesulfonic acid (MES) plus 0–5 mM N′,N′- diethylpiperazine (DEPP) (both buffers from GFS Chemicals, Columbus, OH) to obtain pH 5.4–7.5 as

55 indicated. We used Fe (as FeCl3 added into ascorbate-containing medium) at final specific activity 0.24 GBq.mg−1 obtained from Perkin–Elmer Life Science Products (Boston, MA). Radiotracer metal-ion uptake was measured over 10 min, i.e. within the linear portion of the time course of 55Fe2+ uptake (Mackenzie et al., 2007). We terminated radiotracer uptake by rapidly washing the oocytes three times in radiotracer- free ice-cold pH-5.5 transport medium. Oocytes were then solubilized by using 5% (w/v) sodium dodecylsulfate and radiotracer content assayed by liquid-scintillation counting using Scintisafe–30% liquid-scintillation cocktail (Fisher Scientific, Pittsburgh, PA). Transport rates (v) were fit by a four- parameter logistic function (Equation 3.1) in which Vmin and Vmax are the minimum and maximum rates,

H + . is the H concentration at which transport was half-maximal, and a is the Hill slope.

max − min = min + H 1 + H . Equation 3.1.

Mouse models—We studied iron metabolism in adult FVB/N mice lacking NHE2, NHE3, or Hfe following a protocol approved by the University of Cincinnati Institutional Animal Care and Use Committee. Generation of the NHE2, NHE3, and Hfe knockout mouse models has been described previously (Zhou et al., 1998; Schultheis et al., 1998b; Schultheis et al., 1998a). Mice were genotyped at 21–28 d of age by obtaining DNA from tail clips treated with DirectPCR Lysis Reagent (Viagen Biotech, Inc., Los Angeles, CA) according to the manufacturer’s instructions. We analyzed PCR products using primers as described (Kiela et al., 2009; Bradford et al., 2009; Zhou et al., 1998). Mice were fed standard diet (‘Normal’) containing 350 ppm Fe (7922 NIH-07, Harlan Laboratories, Indianapolis, IN) except where indicated as having been fed a low-iron diet (“Low Fe”) containing 2–6 ppm Fe (TD.80396, Harlan–Teklad, Madison, WI). Mice were

euthanized by CO2 asphyxiation except in the studies described in Figures 3.3, 3.5 in which we collected blood (via cardiac puncture) and tissues from mice subjected to isoflurane anesthesia (by inhalation, to effect), followed by euthanization using a physical method.

Blood and tissue analyses—Standard automated complete blood count (CBC) was performed by Antech, Inc. (Chicago, IL). Serum iron (SI) and unsaturated iron-binding capacity (UIBC) were assayed by using the SL and UIBC assay kits along with calibration standards (Sekisui Diagnostics, Charlottetown, PE, Canada) according to manufacturer’s protocols, and transferrin saturation (Tfsat, %) was computed according to Equation 3.2. SI Tf = × 100 sat SI + UIBC Equation 3.2.

Liver nonheme iron content was determined by using a standard acid-digestion, chromogen- based colorimetric assay as described (Torrance & Bothwell, 1980). We performed histology of the liver to visualize stored iron. Liver tissue samples were rinsed in phosphate-buffered saline, fixed by using 4% (w/v) paraformaldehyde (Thermo Fisher Scientific, Waltham, MA), and embedded in paraffin. Sections of thickness 4 μm were stained with Perls’ Prussian blue stain. Images were acquired using the Olympus BH2 microscope and Olympus Magnafire digital image-capture system with the aid of AxioVision version 4.8.1 software (Zeiss).

qPCR analyses—Freshly isolated small intestine was rinsed and flushed with ice-cold saline. We collected an enterocyte-enriched preparation by lightly scraping the luminal surface of the proximal 1–5 cm of small intestine (i.e. duodenum) with the edge of a glass microscope slide. Enterocytes and liver tissue were collected into TRIzol reagent (Life Technologies, Carlsbad, CA), homogenized, and frozen at −80 °C prior to their use in quantitative real-time PCR (qPCR) analyses. We performed reverse transcription by using 50 µg.mL−1 oligo(dT)-20 primer and reverse transcriptase (Qiagen, Valencia, CA) according to the manufacturer’s instructions. Sample cDNA concentrations were determined by measuring absorbance at 260 nm by using the NanoDrop 2000 Spectrophotometer (Thermo Fisher Scientific, Waltham, MA). We amplified 200 ng sample cDNA with Fast SYBR Green Real-Time PCR master mix (Applied Biosystems, Life Technologies, Grand Island, NY) in a final volume of 20 µL by using the ABI Step One Machine. Gene expression was determined by using the delta delta CT method (Livak & Schmittgen, 2001), with glyceraldehyde 3-phosphate dehydrogenase (GAPDH) mRNA in each tissue sampled as reference, by using primers described in Table A.1 (Appendix) to determine gene expression. qPCR data were normalized by

61 the mean −∆∆CT in wildtype NHE2 (NHE2+/+, Figure 3.3I–L) and NHE3 (NHE3+/+, Figure 3.5I–L) mice and expressed as 2−∆∆CT, i.e. relative fold expression.

Absorption of 59Fe in NHE2-null and NHE3-null mice—We measured radiotracer iron absorption in female control, NHE2-null, and NHE3-null mice at approximately 5 months of age. Conscious mice were

59 administered Fe (0.1 µCi per gram body weight as FeSO4 in 1 mM NaCl, 1 mM L-ascorbic acid, 1 mM L- glutamine) via oral−intragastric gavage following an overnight fast.

Blood was collected into heparinized hematocrit tubes by tail-nick incision at 30 min, 1, 2 and 4 h. Blood samples were centrifuged for measurement of hematocrit and then expelled into 20-mL scintillation vials containing 1 mL SolvableTM solution (Perkin Elmer, Waltham, MA). Mice were euthanized at 4 h after which we collected approximately 100 mg liver into 1 mL Solvable. Freshly isolated small intestine was flushed with ice-cold solution of composition 130 mM NaCl, 5 mM KCl, 1 mM CaCl2, 1 mM MgCl2, 1 mM nitrilotriacetic acid (NTA), 5 mM MES, and buffered to pH 7.4 by using DEPP. Duodenal enterocytes were collected (as described above) into 1.5-mL microcentrifuge tubes containing 0.99 mL Solvable plus 10 µL protease inhibitor cocktail I (EMD Biochemicals, San Diego, CA). 0.5 mL of this sample was transferred to a 20-mL scintillation vial with 1 mL Solvable and the remainder was used to quantify protein concentration by using the Pierce BCA assay (Thermo Fisher Scientific). Tissue samples in Solvable were further processed according to the manufacturer’s protocol. We added Scintisafe−30% cocktail (Thermo Fisher Scientific) and measured sample radioactivity by using liquid-scintillation counting.

Statistical analyses—Statistical analyses were performed using SigmaPlot version 13 (Systat Software) with critical significance level α = 0.05 and β = 0.20, and applied the conventional statistical approach of hypothesis testing. Data are presented as arithmetic mean and standard deviation (SD) for n independent observations, with the following exceptions: (i) qPCR data (relative fold expression) are expressed as geometric mean and SD, and (ii) enterocyte and liver 59Fe content (Figure 3.6B,C) are presented as box plots (median, interquartile range, and whiskers representing 10th and 90th percentiles). Inferences were made by using appropriate statistical tests specified in the figure legends with all tests controlled for unequal sample size by general linear model (GLM). When between-group comparisons were made by using Student’s t tests for multiple variables (Figures 3.3 and 3.5), we used a false-discovery rate (FDR)

procedure (Curran-Everett, 2000) in which an individual P value (Pi) of less than 0.05 is ruled a false ∗ discovery if it is greater than the individual critical significance level ( ) computed for each comparison

62 tested. Given the lack of any observed effect of ablating NHE2 on hematological and blood-iron variables, and no observed effect of NHE3 on [Hgb], I determined power of the study to detect a specific effect (δ) on [Hgb] of δ = µ – a, where a is a 10% reduction from the control mean.

3.4 Results

3.4.1 DMT1-mediated iron transport in vitro is activated at low pH

We measured DMT1 activity in vitro as a function of extracellular pH. Transport of 2 µM 55Fe2+ in RNA- injected Xenopus oocytes expressing human DMT1 was stimulated at low pH (Figure 3.1). We estimated H . at 0.7 µM (equivalent to pH ≈ 6.1). As we and others have observed previously (Garrick & Dolan, 2002; Mackenzie et al., 2006), residual DMT1-mediated transport activity persisted even at pH 7.5 (Figure 3.1).

Figure 3.1. DMT1-mediated iron transport in RNA-injected Xenopus oocytes is activated at low extracellular pH. Uptake of 2 µM 55Fe2+ as a function of extracellular pH in control oocytes (n = 10–16 oocytes at each pH, gray circles) and oocytes expressing human DMT1 (n = 13–16, black circles). Data are

mean, SD. Data for oocytes expressing DMT1 were fit by Eq. 3.1 providing the following estimates: Vmin =

−1 −1 H 2 0.18 ± 0.03 pmol.min , Vmax = 0.74 ± 0.04 pmol.min , .= 0.72 ± 0.10 µM (i.e. pH ≈ 6.1), a = 2.7 ± 0.9, r = 0.98, P < 0.001, n = 8); Data for control oocytes could not be fit by Eq. 3.1 (r2 = 0.70, P = 0.15, n = 8).

3.4.2 Iron metabolism and homeostasis in NHE2-null (NHE2–/–) mice

We measured liver nonheme iron content as an indicator of iron stores in male and female mice as a function of NHE2 genotype and as a function of dietary iron. Liver iron stores in NHE2 knockout (NHE2–/–) mice were lower than those in wildtype (NHE2+/+) and heterozygous (NHE2+/–) mice, regardless of sex and whether or not the animals were stressed on a low-Fe diet for 6 weeks (Figure 3.2). Liver iron stores were lower in heterozygotes than in wildtype mice only when the animals were fed a low-Fe diet (Figure 3.2).

Figure 3.2. Liver nonheme iron content in mice as a function of NHE2 genotype. Male and female wildtype (NHE2+/+), NHE2 heterozygous (NHE2+/–), and NHE2-null mice (NHE2–/–), mean (SD) age 101 (27) d were maintained 6 weeks post-weaning on a normal diet (‘Normal’) or low-iron diet (“low-Fe”). Data are mean, SD for n mice per group, where n (left to right) = 18, 16, 26, 19, 6, 17, 17, 21, 23, 15, 12, 8. Three-way ANOVA revealed main effects of genotype, sex, and diet (all P < 0.001); 2-way interactions of sex × diet (P = 0.001) and genotype × diet (P < 0.001) but not genotype × sex (P = 0.38); and no 3-way interaction (P = 0.15). Post-hoc comparisons within diets indicated that NHE2–/– differed from NHE2+/+ and NHE2+/– regardless of diet (P < 0.001), and heterozygotes differed from wildtype mice when fed a low-Fe diet (P < 0.001) but not normal diet (P = 0.30).

We examined the effect of loss of NHE2 on hematological and blood-iron variables in mice. We limited our analysis to male NHE2+/+ and NHE2–/– mice fed a low-Fe diet since the effect of NHE2 knockout on liver iron stores was independent of sex and diet (Figure 3.2). Inspection of peripheral blood smears (not shown) did not reveal any difference in red-cell morphology between male NHE2+/+ and NHE2–/– mice fed a low-Fe diet. Automated CBC revealed no change in any hematological variable, serum Fe, or Tf saturation in NHE2–/– mice compared with NHE2+/+ mice (Figure 3.3A–H).

We examined the effect of ablation of NHE2 on the expression of genes involved in iron homeostasis and absorption. Loss of NHE2 had no effect on the intestinal expression of DMT1, the apical surface ferrireductase cytochrome-B reductase-1 (Cybrd1), or the basolateral iron exporter ferroportin (Fpn) (Figure 3.3I–K). Although we observed no regulation at the mRNA level, it is possible that protein levels of DMT1, Cybrd1, or Fpn could be elevated in the intestine of NHE2–/– mice; however, we found no change in the hepatic expression of the iron-regulatory hormone hepcidin (Hamp1) (Figure 3.3L), the primary regulator of Fpn protein levels. In summary, our data indicate that ablation of NHE2 in the mouse results in reduced iron stores; however, NHE2−/− mice maintain normal serum-Fe levels without the need for compensatory regulation of iron-related genes.

Figure 3.3. Iron metabolism and homeostasis in male NHE2-null mice fed a low-Fe diet. Male NHE2 wildtype (NHE2+/+, n = 4–5) and NHE2-null mice (NHE2−/−, n = 6–9), mean (SD) age 120 (21) d were maintained 6 weeks post-weaning on a low-Fe diet. (A) Hematocrit. (B) Red blood cell count. (C) Mean corpuscular volume. (D) Hemoglobin (Hgb) concentration. Power of the study to detect a specified effect on [Hgb] was > 0.99. (E) Mean corpuscular hemoglobin (Hgb). (F) Mean corpuscular hemoglobin (Hgb) concentration. (G) Serum iron (Fe). (H) Transferrin (Tf) saturation. Relative fold expression (geometric mean, SD) of intestinal (I–K) and hepatic (L) iron-related genes. Student’s t tests revealed no difference between NHE2+/+ and NHE2–/– in any of the variables (0.11 ≤ P ≤ 0.99).

3.4.3 Iron metabolism and homeostasis in NHE3-null (NHE3–/–) mice

Liver nonheme iron stores were depleted in male and female NHE3-null (NHE3–/–) mice compared with wildtype (NHE3+/+) or heterozygous (NHE3+/–) mice regardless of sex or of dietary iron (Figure 3.4). Iron stores were lower in female but not male heterozygotes compared with wildtype mice. Although our

65 statistical analyses did not directly compare NHE3-null with NHE2-null mice, ablation of NHE3 had a more striking effect on liver iron stores (70–88%, Figure 3.4) than did ablation of NHE2 (31–86%, Figure 3.2).

Figure 3.4. Liver nonheme iron content in mice as a function of NHE3 genotype. Male and female NHE3 wildtype (NHE3+/+), heterozygous (NHE3+/–), and null mice (NHE3–/–), mean (SD) age 113 (33) d were maintained 6 weeks post-weaning on a normal diet (‘Normal’) or low-Fe diet (“Low Fe”). Data are mean, SD for n per group, where n (left to right) = 16, 11, 8, 22, 16, 11, 16, 11, 14, 10, 12, 13. Three-way ANOVA revealed main effects of genotype, sex, and diet (all P < 0.001); 2-way interactions of genotype × sex (P < 0.001) but not genotype × diet (P = 0.53) or sex × diet (P = 0.060); and no 3-way interaction (P = 0.85). Post-hoc comparisons within sex indicated that NHE3+/– differed from NHE3+/+ among female (P < 0.001) but not male mice (P = 0.036).

We examined the effect of loss of NHE3 on hematological and blood-iron variables in mice. We limited our analysis to male NHE3+/+ and NHE3–/– mice fed a low-Fe diet since the effect of ablation of NHE3 on liver iron stores was independent of sex and diet (Figure 3.4). Inspection of peripheral blood smears from NHE3–/– mice fed a low-Fe diet for 6 weeks did not reveal any difference in red-cell morphology compared with wildtype mice (not shown). Despite having found no meaningful change in hematological variables of NHE3–/– mice compared with wildtype mice (Figure 3.5A–F) we observed modest reductions in serum Fe and Tf saturation in NHE3–/– mice (Figure 3.5G, H). Given that Na+ and fluid absorption are defective in the NHE3–/– mouse (Gawenis et al., 2002; Schultheis et al., 1998b), it is conceivable that the expected dehydration in the NHE3–/– mouse may have masked any effect of iron deficiency on hematocrit and hemoglobin concentration (we observed a trivial decrease in mean corpuscular hemoglobin concentration), and may underlie the slightly elevated red cell count.

Figure 3.5. Iron metabolism and homeostasis in male NHE3-null mice fed a low-Fe diet. Male NHE3 wildtype (NHE3+/+, n = 4–13) and NHE3-null mice (NHE3−/−, n = 5–8), mean (SD) age 115 (29) d were maintained 6 weeks post-weaning on a low-Fe diet. (A) Hematocrit. (B) Red blood cell count. (C) Mean corpuscular volume. (D) Hemoglobin (Hgb) concentration. Power of the study to detect a specified effect on [Hgb] was 0.96. (E) Mean corpuscular hemoglobin (Hgb). (F) Mean corpuscular hemoglobin (Hgb) concentration. (G) Serum iron (Fe). (H) Transferrin (Tf) saturation. Relative fold expression (geometric mean, SD) of intestinal (I–K) and hepatic (L) iron-related genes. Student’s t test revealed no difference between NHE3+/+ and NHE3–/– for hematocrit, MCV, [Hgb], or mean corpuscular Hgb (P ≥ 0.088 except bP = 0.039 ruled a false discovery by FDR). NHE3–/– differed from NHE3+/+ for the following variables: red blood cell count (aP = 0.003), mean corpuscular [Hgb] (cP = 0.006), serum Fe (dP = 0.001), and Tf saturation (eP ≤ 0.001), and all genes tested (P < 0.001).

Ablation of NHE3 resulted in upregulation of the intestinal expression of DMT1, Cybrd1, and Fpn (Figure 3.5I–K), and suppressed the hepatic expression of hepcidin to <1% that of wildtype mice (Figure 3.5L). In summary, ablation of NHE3 in the mouse results in depleted iron stores, reduced serum Fe variables, and appropriate compensatory upregulation of the iron-absorptive machinery.

3.4.4 Ablation of NHE3 but not NHE2 severely impairs intestinal iron absorption

In order to directly test roles for NHE2 and NHE3 in driving intestinal iron absorption, we measured radiotracer iron absorption from an oral–intragastric dose in wildtype (NHE+/+), NHE2-null (NHE2–/–), and NHE3-null (NHE3–/–) mice. We measured hematocrit over the 4-h time course to ensure against large changes in blood volume or signs of the onset of hypovolemic shock. We observed no change in hematocrit over time and no time-dependent differences between NHE+/+, NHE2–/– and NHE3–/– mice (data not shown).

Appearance of 59Fe in the blood of NHE2–/– mice following the oral dose did not differ from that in NHE+/+ mice; however, 59Fe appearance in the blood of NHE3–/– mice was substantially blunted, less than 30% that of NHE+/+ mice (from area under the curve) (Figure 3.6A). We obtained for NHE+/+ mice at 4 h robust 59Fe signals in enterocytes and liver (the latter tissue being expected to rapidly clear 59Fe from the portal and peripheral circulation) (Figure 3.6B,C). Whereas 59Fe content of enterocytes and liver of NHE2–/– mice did not differ from NHE+/+, enterocyte and liver 59Fe content in NHE3-null mice was only a small fraction of that observed for NHE+/+ mice. The lack of 59Fe within enterocytes of NHE3–/– mice demonstrates that the site of lesion in iron absorption in these mice is at the intestinal brush border (we assume we had bypassed the need for luminal ferrireduction by providing Fe2+ in the presence of ascorbic acid). Our data show that iron uptake at the intestinal brush border largely depends upon the activity of NHE3.

Figure 3.6. Intestinal iron absorption in female wildtype, NHE2-null, and NHE3-null mice. Wildtype (NHE+/+; n = 10), NHE2-null (NHE2–/–; n = 7) and NHE3-null (NHE3–/–; n = 7) mice, mean (SD) age 144 (140) d, were dosed with 0.1 µCi 59Fe per g body weight via oral–intragastric gavage. (A) Blood 59Fe content as a function of time after intragastric dose. Data were analyzed by Two-way ANOVA with repeated measures over time but, for clarity, are displayed as mean, SEM. We observed a main effect of genotype (P = 0.030) but not of time (P = 0.10) and no interaction (P = 0.55). Holm–Šidák pairwise comparisons versus NHE+/+ revealed that NHE3–/– (P = 0.019) but not NHE2–/– (P = 0.48) differed from NHE+/+. (B) Enterocyte 59Fe content at 4 h. One-way ANOVA on ranks (P = 0.004) and Dunn’s pairwise comparisons versus wildtype: aP = 0.13, bP = 0.002 cf. NHE+/+. (C) Liver 59Fe content at 4 h. One-way ANOVA on ranks (P = 0.024) and pairwise analysis versus NHE+/+: cP > 0.81, dP = 0.013 cf. NHE+/+.

3.4.5 Rescue of the iron-overload phenotype of the hemochromatosis mouse

Since ablation of NHE3 impaired iron absorption and reduced serum Fe and iron stores, we tested whether inactivation of NHE3 could counter the iron overload observed in the Hfe mouse model of hemochromatosis. The Hfe-null mouse (Hfe–/–) exhibits hepatic iron overload that progresses with age (Fleming et al., 2001; Zhou et al., 1998) and which results from a disruption in the hepcidin–ferroportin axis regulating iron absorption (Ahmad et al., 2002; Gao et al., 2010). We examined hepatic iron loading in male and female mice as a function of Hfe genotype and NHE3 genotype. Nonheme iron stores were elevated in Hfe–/– compared with Hfe wildtype mice (Hfe+/+) (Figure 3.7A) and substantial iron deposits were visible (by Perls’ stain) in the hemochromatosis model, i.e. Hfe–/–|NHE3+/+ (Figure 3.7B). Only a very few iron deposits were visible in liver of double-knockout mice (Hfe–/–|NHE3–/–) (Figure 3.7B), and nonheme iron stores in the double knockout were similar to those of wildtype mice (Figure 3.7A).

Therefore, our preliminary data suggest that inactivation of NHE3 may prevent or lessen hepatic iron loading in the Hfe model of hereditary hemochromatosis.

Figure 3.7. Ablation of NHE3 prevents iron overload in the hemochromatosis mouse. Hepatic iron loading in mice as a function of NHE3 genotype and Hfe (hemochromatosis) genotype (excluding heterozygotes). (A) Liver nonheme iron stores (n = 2–7; mean (SD) age 131 (61) d). Two-way ANOVA revealed main effects of NHE3 genotype (P < 0.001) and of Hfe genotype (P < 0.001), no interaction (P = 0.25). (B) Histology of the liver prepared by using Perls’ Prussian blue stain and visualized at 600× magnification (scale bars, 100 μm).

3.5 Discussion

Our study demonstrates that intestinal absorption of nonheme iron relies in large part upon the activity of NHE3, the primary determinant of the brush-border acidic microclimate in the mature mammalian intestine, and reveals NHE3 as the source of the H+ electrochemical potential gradient driving DMT1- mediated apical iron uptake.

3.5.1 Provenance of the acidic microclimate

We have shown previously that DMT1-mediated iron transport in vitro is functionally coupled to the H+ flux, and that the affinity of DMT1 for Fe2+ is increased at low pH (Mackenzie et al., 2006). In the present

+ 55 2+ H study, we found that H activated human DMT1-mediated Fe transport in vitro (Figure 3.1) with . ≈ 0.7 µM, i.e. pK ≈ 6.1, which coincides with the surface microclimate pH in the human proximal small intestine under physiological conditions (McEwan et al., 1990). Our data illustrate that even modest changes in the microclimate pH (over the ‘linear’ region, pH 6.0–6.3) will be sufficient to modulate substantially DMT1-mediated apical iron uptake. Several factors can rapidly alkalize the surface microclimate of mammalian intestine acutely, including nitric oxide and bacterial toxins (e.g. Escherichia coli STa enterotoxin and cholera toxin) (Gill et al., 2002; McEwan et al., 1988). Clostridium difficile toxin B decreases NHE3 mRNA and protein levels in human intestinal organoids (Engevik et al., 2015), and triggers the internalization of NHE3 protein in mammalian epithelial cell lines (Hayashi et al., 2004). Surface microclimate pH of proximal small intestine was 6.1 in young rats but rose to pH 6.5 in senescent rats (Ikuma et al., 1996). Whether a similar change occurs in human subjects is not known; nevertheless, this observation raises the possibility that a diminished driving force for iron absorption could be one of several factors contributing to iron deficiency common in the elderly (Fairweather-Tait et al., 2014).

Several other mammalian intestinal transporters, like DMT1, function as H+/solute symporters [reviewed, (Thwaites & Anderson, 2007)]. Among these are H+/amino acid transporter-1 (PAT1) (Anderson et al., 2004a), H+/folate transporter-1 (PCFT) (Qiu et al., 2006), and H+/peptide transporter-1 (PEPT1) (Mackenzie et al., 1996a; Fei et al., 1994). Although their dependence on NHE3 activity has not been demonstrated in vivo, PAT1 and PEPT1 transport activity in the Caco-2 intestinal cell line is inhibited by NHE3 inhibitors (Thwaites et al., 2002; Anderson & Thwaites, 2005).

Whereas gastric acid is thought to aid in iron absorption by promoting the solubility of Fe(III) complexes and the reduction of ferric ion by luminal ascorbic acid, we have found no clear requirement for gastric H+/K+ ATPase (gHKA), the primary protein responsible for the gastric H+ secretion (Engevik MA, Shawki A, Anthony SR, Baik RA, Kim RS, Worrell RT, Shull GE, Mackenzie B, manuscript in preparation).

3.5.2 Prevention or treatment of iron overload

DMT1 is a validated therapeutic target in iron-overload disorders (Darshan et al., 2010). We have found that iron absorption was disrupted in NHE3-null mice, so we postulated that ablation of NHE3 could lessen or prevent iron loading in the Hfe-deficient mouse model of hemochromatosis. Our preliminary data

71 demonstrated an absence of hepatic iron overload in double-knockout mice (Hfe–/–|NHE3–/–), a finding that warrants detailed investigation of whether specific inhibition of NHE3 in the proximal small intestine may block iron absorption and could be used therapeutically to prevent iron overload in hereditary hemochromatosis or thalassemia major.

4 Exploring the molecular physiology and structure-function of DMT1

4.1 Abstract

One of our core objectives is to understand the structure−funcon of DMT1, and idenfy crical residues and structural elements responsible for metal−ion binding, H+ coupling, and substrate translocation. Until the recent crystallization of a bacterial homologue of DMT1 (ScaDMT), we and other investigators have relied upon predictive modeling, chemical modification, and studying the molecular impact of human mutations in DMT1 associated with disease phenotypes to target potentially important residues. I have tested these predictions by mutating specific residues by site-directed mutagenesis or excisions, coupled with functional assays in RNA-injected Xenopus oocytes expressing wildtype (wt) or mutant DMT1. DMT1 is a glycoprotein, so I have tested whether glycosylation is required for the functional activity of DMT1. I found that mutating the two predicted N-linked glycosylation sites at the putative extracellular loop between TM4 and TM5 did not disrupt iron transport in DMT1. The results of treatment with a battery of chemical modifiers has identified potentially important roles for tyrosine, methionine, and proline residues. Although they are extremely rare, we know of five unrelated cases of human DMT1 mutations associated with a disease phenotype (i.e. microcytic anemia). I hypothesized that these mutations result in impaired iron-transport activity. Whereas some mutants disrupt iron transport activity, therefore explaining the patients’ phenotype, others do not. One such mutation worth highlighting is G212V. Whereas wildtype DMT1 exhibited a H+/Fe2+ coupling ratio of 17:1 (i.e. significant H+ slippage), far exceeding that expected for strict coupling ratio, G212V abolished this slippage. Slippage in cotransporters is normally thought of in terms of the energetic penalty; however, our observation that H+ slippage, but not Fe2+ transport, is disrupted in a mutant associated with disease raises the possibility that H+ slippage serve a physiological role. Further analysis of human mutants is revealing novel aspects of the molecular physiology of DMT1.

Sections 4.2, 4.4.3, and 4.5.2 contain both my original work and excerpts from the following co-

authored publication Shawki A, Knight, PB, Maliken, BD, Niespodzany EJ, Mackenzie B (2012) H+-

coupled divalent metal-ion transporter-1: Functional properties, physiological roles and therapeutics.

Curr Top Membr. 70, 169-214.

4.2 Introduction

DMT1’s vital roles in iron homeostasis are illustrated by impaired intestinal iron absorption and severe hypochromic–microcytic anemia (characteristic of iron deficiency) exhibited by (i) the Belgrade (b) rat and mk mouse models, inbred rodent strains that bear an identical (G185R) mutation in DMT1 (Fleming et al., 1997; Fleming et al., 1998), (ii) the intestine specific DMT1 knockout mouse (DMT1int/int) (Shawki et al., 2015), and (iii) human patients with mutations in DMT1 (Beaumont et al., 2006; Blanco et al., 2009; Mims et al., 2005; Iolascon et al., 2006; Lam-Yuk-Tseung et al., 2006; Bardou-Jacquet et al., 2011).

Human DMT1 shares 61% identity at the amino-acid sequence level with NRAMP1, the only other member of the SLC11 gene family in mammals (Mackenzie & Hediger, 2004). DMT1 and NRAMP1 belong to an NRAMP superfamily in animals, plants and fungi, and extending to prokaryotes (Cellier et al., 2001; Chaloupka et al., 2005). Until the recent crystallization of a bacterial homolog of mammalian DMT1, we and other investigators had relied on the following approaches to study the structure–function relationships of SLC11 proteins: (i) predicting membrane topology models and testing model predictions by site-directed mutagenesis, and (ii) functional analysis of human mutations in DMT1.

Predictive membrane topology models provide us with a starting point to understanding the structure of membrane proteins such as DMT1. Kyte–Doolittle hydrophobicity analysis predicted 12 transmembrane (TM) regions for rat (Gunshin et al., 1997), mouse (Czachorowski et al., 2009; Lam-Yuk- Tseung et al., 2003) and human DMT1 (Figure 4.1) (Shawki et al., 2012). I can test the validity of these models experimentally. For example, DMT1 is an integral membrane protein that is observed in vitro to undergo glycosylation (Mackenzie et al., 2007; Gruenheid et al., 1999; Tabuchi et al., 2002). Glycosylation of proteins has been shown to affect protein trafficking, proper folding of the protein, and biological activity and half-life of proteins (Blom et al., 2004). Two consensus sites of N-linked glycosylation (at residues Asn-336, Asn-349 according to numbering for the human 1B isoforms) are predicted to reside in the exofacial fourth extracellular loop between TM7–8 (Mackenzie et al., 2007; Czachorowski et al., 2009) (Figure 4.1), and there is speculation in the literature that glycosylation is required for DMT1 function

(Gruenheid et al., 1999; Tabuchi et al., 2002). I hypothesized that glycosylation of DMT1 is required for DMT1 function. To test the role of glycosylation in DMT1 function, I mutated Asn-336 and Asn-349 to glutamine (conserving the chemical nature of asparagine) and examined the functional impact of these mutations by using the Xenopus oocyte expression system.

Figure 4.1. Predicted topology model of human DMT1, 1B/IRE(+) isoform (based on the hydrophobicity plot generated by using the Kyte–Doolittle algorithm at http://web.expasy.org/protscale/ and default values for predicting transmembrane regions) showing sites of the human mutations (green) reported so far. Also labeled are amino acid residues corresponding to those mutated in rat and mouse DMT1 (including the G185R mutation found in the b and mk rodent models, blue). Reproduced from (Shawki et al., 2012).

Protein modification provides an alternative method in the absence of structural information to discover critical amino-acid residues (Means & Feeney, 1971; Lundblad, 2003). Although somewhat crude, this approach provides us with a starting point to target residues critical for DMT1 transport activity. Site- directed mutagenesis and expression of wildtype (wt) and mutant DMT1 in Xenopus oocytes following chemical modification and identification of critical residues provides a technique to examine functional properties of DMT1.

Studying the molecular impact of human DMT1 mutations associated with disease phenotypes provides a rapid and effective way of identifying potentially critical residues in DMT1. Mutations in human DMT1 have been reported now in five unrelated cases of hypochromic–microcytic anemia characterized

75 by lowered values for hematocrit, blood hemoglobin concentration, mean corpuscular volume, and mean corpuscular hemoglobin content, and in at least 4 out of the 5 patients hepatic iron overload (Table 4.2). Three probands are compound heterozygotes, parents and siblings of which are healthy, indicating that both mutations are contributing to the disease. The fifth proband exhibiting microcytic anemia is homozygous with no observed iron overload. I aimed to identify the molecular impact of mutations in human DMT1 in Xenopus oocytes in order to explain the disease phenotype and to better understand the structure-function of DMT1. I have tested the hypothesis that human DMT1 mutants are defective in iron- transport activity.

4.3 Materials and Methods

Reagents and media—Reagents were obtained from Sigma-Aldrich Corp. (St. Louis, MO) or Research Products International Corp. (Prospect, IL) unless otherwise indicated.

Expression of wildtype and mutant human DMT1 in Xenopus oocytes—We performed laparotomy and ovariectomy on adult female Xenopus laevis frogs (Nasco, Fort Atkinson, WI) under 2-aminoethylbenzoate methanesulfonate anesthesia (0.1% in 1:1 water/ice, by immersion) following a protocol approved by the University of Cincinnati Institutional Animal Care and Use Committee. Ovarian tissue was isolated and treated with collagenase A (Roche Diagnostics Corp., Indianapolis, IN), and oocytes were isolated and stored at 17 °C in modified Barths’ medium as described (Mackenzie, 1999). We expressed in Xenopus oocytes the 1A/IRE(+) isoform of DMT1, the product of the human SLC11A2 gene. We generated mutants by using the QuikChange 2 site-directed mutagenesis kit (Agilent Technologies, Santa Clara, CA) by PCR amplification of wildtype DMT1 (wtDMT1) using primers designed to introduce single amino-acid substitutions where noted except for Exon 5 and Exon 12 deletion (denoted as Ex5Δ and Ex12Δ hereafter). We generated the Ex5Δ and Ex12Δ mutants by splicing by overlap extension (SOE) by PCR amplification as described (Warrens et al., 1997). Competent DH5α cells (Invitrogen) were heat-shock transformed with the mutant plasmids and selected on LB/agar plates containing 100–200µg.µL–1 ampicillin. Mutations were verified by DNA sequencing of the final constructs (Cincinnati Childrens Research Center Sequencing Core, Cincinnati, Ohio).

We linearized the pOX(+) plasmid containing 1A/IRE(+) wt or mutant DMT1 cDNA under the SP6 promoter [as described (Mackenzie et al., 2007)] using SnaBI (New England Biolabs Inc, Ipswich, MA) and synthesized RNA in vitro using the mMESSAGE mMACHINE/SP6 RNA polymerase kit (Applied

Biosystems/Ambion, Austin, TX) according to the manufacturers’ protocols. Defolliculate stage V-VI oocytes were injected with approximately 50 ng of DMT1 RNA and incubated 6 to 7 days before being used in functional assays.

Oocytes were superfused or incubated at room temperature (22-24 °C) in transport assay medium (‘MeTx’, see Appendix) at pH 5.5 unless otherwise indicated.

55 −1 Radiotracer assays—We used Fe (added as FeCl3) at final specific activity 0.2–0.5 GBq.mg obtained from Perkin–Elmer Life Science Products (Boston, MA). Radiotracer metal-ion uptake was measured over 10 min, i.e. within the linear portion of the time course of 55Fe2+ uptake (Mackenzie et al., 2007). Where indicated, oocytes were pre-incubated for the time shown in solutions containing chemical modifiers. We terminated radiotracer uptake by rapidly washing the oocytes three times in radiotracer-free ice-cold pH- 5.5 transport medium. Oocytes were then solubilized by using 5% (w/v) sodium dodecylsulfate and radiotracer content assayed by liquid-scintillation counting using Scintisafe–30% liquid-scintillation cocktail (Fisher Scientific, Pittsburgh, PA).

Voltage clamp—A two-microelectrode voltage clamp (Dagan CA-1B) was used to measure currents associated with wt or mutant DMT1 in oocytes. Microelectrodes (resistance 0.5–5 MΩ) were filled with 3 M KCl. Voltage-clamp experiments comprised three protocols at room temperature (22–24°C) except

where indicated: (i) Continuous current recordings were made at holding potentials (Vh) of –70 mV, low-

pass filtered at 1 Hz, and digitized at 10 Hz. (ii) Oocytes were clamped at Vh = –50 mV, and step-changes

in membrane potential (Vm) were applied from +50 to –150 mV (in 20 mV increments) each for a duration of 200 ms, before and after the addition of Fe2+. Current was low-pass filtered at 500 Hz and digitized at 5

kHz. Steady-state data were obtained by averaging the points over the final 16.7 ms at each Vm step. (iii) Presteady-state currents were obtained at 30°C using protocol ii modified such that step-changes were applied from +90 to –130 mV. Steady-state data from protocols i or ii were fit by a modified 3-parameter

Hill relationship (Equation 4.1) for which I is the evoked current, Imax the derived current maximum, S the

+ S concentration of substrate S (H or metal ion), . the substrate concentration at which current was half-

maximal, and nH the Hill coefficient for S.

∙H = max S H . + Equation 4.1.

Following step-changes in Vm using protocol iii, I obtained presteady-state currents in oocytes expressing wt or mutant DMT1. These were isolated from capacitive currents (which decayed with half times of 0.5–0.8 ms) and steady-state currents by the fitted method (Hazama et al., 1997; Mackenzie, 1999). Briefly, the first ten points (i.e. 2 ms) after reaching maximal decay rate were fit using an exponential decay to describe the capacitive transients; these and the final steady-state currents were subtracted to obtain the ‘compensated’ currents. The compensated currents were integrated with time to obtain charge movement (Q) and fit using the Boltzmann relationship (Equation 4.2) for which maximal charge Qmax = Qdep – Qhyp (where Qdep and Qhyp represent the charge at depolarizing and hyperpolarizing

limits), V0.5 is the Vm at the midpoint of charge transfer, z is the apparent valence of the movable charge, and F, R, and T have their usual thermodynamic meanings.

−hyp 1 = − max 1+exp m .

Equation 4.2.

Transporter-mediated presteady-state currents can be used to estimate transporter density

(Zampighi et al., 1995). I estimated the number of functional wt or mutant DMT1 transporters (NT) per oocyte using Equation 4.3, in which e is the elementary charge (1.6×10–19 C).

max − T Equation 4.3.

As described previously, presteady-state currents decay with time. I calculated the presteady-

state decay time constants (τ) for wt and mutant DMT1 and fit the data by Equation 4.4, in which Vτmax is

the membrane potential at which τ is maximal, a is τmax – τmin, and b defines the steepness of the slope.

−max = + ∙ exp −0.5 min

Equation 4.4.

Determination of H+/Fe2+ coupling ratio—The specificity of radiotracer metal uptakes and the power of voltage-clamp experiments can be combined to determine the H+/Fe2+ coupling ratio in individual cells, by measuring 55Fe2+ uptake under voltage clamp.

Individual oocytes were placed in a chamber of functional volume ≈ 100 mL and voltage-clamped

at a single holding potential (Vh) throughout the experiment. The oocyte was superfused with pH 5.5

78 medium at a flow rate of 2–3 mL.min−1. Baseline current was recorded in pH 5.5 medium, before switching to 5 mM 55Fe2+ for 10 min. The oocyte was then superfused with radiotracer-free pH 5.5 medium until the current returned to baseline (≈3 min), and radioactivity counted as before.

The Fe2+-evoked current (nA) was obtained as the difference in current between baseline and after addition of 55Fe2+, and was integrated with time (trapezoidal rule) to obtain the total Fe2+-dependent charge (QFe, in nC) (1 A = 1 C.s−1). QFe was converted to a molar equivalent using the Faraday (96,500 C.mol−1) (assuming monovalency). QFe was then plotted against 55Fe2+ accumulation and fit by a linear function, in which the slope is the coupling ratio (nc).

DMT1 localization—I visually examined the localization of wildtype and mutant C-term EGFP tagged DMT1 in oocytes by using the Zeiss LSM 7 DUO confocal laser-scanning microscope (excitation at 488 nm) fitted with the LD C-Apochromat 40×/1.1 W Korr objective to measure emission in the band 500–531 nm at a pinhole setting of 36 µm.

Statistical and regression analyses—I performed statistical analyses using SigmaPlot version 13 (Systat Software) with critical significance level α = 0.05 and applied the conventional statistical approach of hypothesis testing. I have presented data as mean and standard deviation (SD) for n independent observations and used parametric tests in their analysis. Data were fit by equations 4.1, 4.2, or 4.4 using the least-squares method of nonlinear regression analysis, or by linear regression analysis, the results of which are expressed as the estimates of fit parameters ± standard error (SE); adjusted r2 is the adjusted regression coefficient and P describes the significance of the fit. Between-group comparisons were made using one-way or two-way analysis of variance (ANOVA) controlled for unequal sample size by general linear model (GLM) followed, when appropriate, by pairwise multiple comparisons by using the Holm- Šidák test.

4.4 Results

4.4.1 N-linked glycosylation sites

As a first step in testing the requirement of glycosylation for DMT1 function, I analyzed the human DMT1 1A/IRE(+) construct for putative N-linked glycosylation (N-gly) sites using the NetNGlyc 1.0 software (http://www.cbs.dtu.dk/services/NetNGlyc/, threshold 0.5). I found three predicted N-gly sites, one of which is found near the N-terminus which is expected to be intracellular (Forbes & Gros, 2003) and,

79 therefore unlikely to be glycosylated. I mutated the two putative extracellular N-gly sites in DMT1, Asn- 336 and Asn-349 to the chemically conservative amino acid glutamine (which cannot be glycosylated) and measured radiotracer iron uptake in Xenopus oocytes expressing wt and mutant DMT1. I found that neither N336Q nor N349Q abolished DMT1-mediated 55Fe2+ transport (Figure 4.2) suggesting that glycosylation is not required for DMT1 functional activity. See discussion section 4.5.1 for future experiments or next steps.

Figure 4.2. Uptake of 2µM 55Fe2+ pH 5.5 in Xenopus oocytes expressing wildtype (wt) and DMT1 in which putative sites of N-linked glycosylation were mutated (data are mean, SD for n oocytes per group). Analysis of variance (P < 0.001) followed by Holm–Šidák pairwise comparisons revealed that N336Q and N349Q were different both from wtDMT1 and control (P < 0.001).

4.4.2 Chemical modification of critical residues in DMT1

I examined the effects of a battery of chemical modifying reagents (Table 4.1) on DMT1 functional activity in RNA-injected Xenopus oocytes. Following this initial screen, I have examined the location of the amino acid residues using the predicted membrane topology of DMT1 and have selected residues that are (i) likely to be accessible to the solvent phase, or (ii) near sites known to be critical for the function of the transporter. For example, the flexible linker in TM6 is predicted to undergo rapid but subtle H+-induced conformational changes. Substitution of His-272 with Ala—which uncoupled the H+ and Fe2+ fluxes through rat DMT1 expressed in Xenopus oocytes (Mackenzie et al., 2006)—resulted in an unfolding of the N-terminal α-helix in TM6 and decreased intermolecular interactions (Xiao et al., 2011; Xiao et al., 2010). These observations confirm a critical role for His-272 in DMT1, and are consistent with the notion that this residue forms part of the mechanism by which Fe2+ transport is coupled to H+ translocation. I tested these predictions by site-directed mutagenesis and functional analysis in oocytes.

Table 4.1. Protein modification by chemical modifying reagents [Modified from (Means & Feeney, 1971)].

Reactive Amino acid Structure Modifying agent(s) with reactivity side chain residue(s)

1-Acetylimidazole (NAI) (±±) Arg N-Ethylmaleimide (NEM) (±±) Amino NH 2 Lys Carbodiimides (±) Trinitrobenzenesulfonic acid (TNBS) (+++)

1-Acetylimidazole (NAI) (+++)a N-Bromosuccinimide (NBrS) (+++) Ethyldiazoacetate (EDA) (++) N-Ethylmaleimide (NEM) (+++) p-(Chloromercuri)benzoate (pCMB) (+++) Sulfhydryl SH Cys p-(Chloromercuri)benzoate sulfonic acid (pCMBS, membrane- impermeant) (+++) Rose Bengal (RB) [photoxidation] (+++) Trinitrobenzenesulfonic acid (TNBS) (++) Trinitromethane (TNM) (+++)

Disulfide SS Cys-Cys Dithiothreitol (dTT) (+++)

Asp Ethyldiazoacetate (EDA) (+++) Carboxyl COOH Glu EDC plus GME (+++) (water-soluble carbodiimide plus nucleophile)

+ NH2 Guanidino NH C Arg 1,2-Cyclohexanedione (CHD) (+++)

NH2

1-Acetylimidazole (NAI) (+++)b N-Bromosuccinimide (NBrS) (++) Phenolic Tyr OH Rose Bengal (RB) [photoxidation] (±±) — specific at high pH Trinitromethane (TNM) (+++)

a N 1-Acetylimidazole (+++) N-Bromosuccinimide (+) Imidazole His Rose Bengal (RB) [photoxidation] (+++) — reacts at pH > pKa for His NH (ie should react at pH 7.5 but not at pH 5.5) Diethylpyrocarbonate

N-Bromosuccinimide (+++) Indole Trp Rose Bengal (RB) [photoxidation] (+++) — persists at low pH Tetranitromethane (TNM) (+) N H

Cyanogen bromide (CyBr) (+++) Trinitromethane (TNM) (+) Thioether SCH Met 3 These groups are extremely hydrophobic and usually inaccessible. See Means & Feeney for more details.

aSpontaneously reversible under the reaction conditions or on dilution, regenerating original group. bEasily reversible, regenerating original

group 81

Methionine-438 but not M240 is required for DMT1-mediated iron transport:

Methionine (Met) residues are capable of metal coordination; however, Met residues are generally inaccessible and buried within the core of the protein. Cyanogen bromide (CyBr) specifically modifies thioether groups such as that found in Met (Table 4.1). I found that DMT1 mediated 2 µM 55Fe2+ uptake at pH 5.5 was sensitive to CyBr (Figure 4.3A), pointing to a role for Met residues in DMT1. Analysis of the amino acid sequence and the predicted membrane topology of DMT1 reveal only two Met residues likely to be of interest. The first (Met-438) is located near an extracellular loop and therefore may be exposed, and the second (Met-240) is located near His-272. Whereas M240A behaved like wtDMT1 (Figure 4.3B), M438A-DMT1 disrupted 2 µM 55Fe2+ transport by 69 ± 6% (SE) (Figure 4.3C). On further examination, I

2+ found that M438A increased by 10-fold the K0.5 for Fe relative to wtDMT1 (Figure 4.3D), and altered substrate recognition such that Mn2+ was excluded (whereas wtDMT1 transports Mn2+) (Figure 4.3C). These observations suggest that M438 participates in metal-ion binding.

Figure 4.3. Effect of cyanogen bromide on DMT1-mediated iron transport points to methionine residues. Radiotracer iron uptake in control (Ctrl) oocytes and oocytes expressing wildtype (wt) and mutant DMT1 (data are mean, SD for A–C). (A) 1-h pre-incubation in 5 mM cyanogen bromide (CyBr) or vehicle (0.1% acetonitrile, pH 7.0) (n = 18-24). Two-way ANOVA revealed an interaction (P < 0.001). (B) Met–240 mutated to Ala (n = 10-16). One-way ANOVA (P < 0.001) followed by Holm–Šidák pairwise comparisons revealed that M240A-DMT1 was not different from wtDMT1 (P = 0.20) but different from Ctrl (P < 0.001). (C) Uptake at 2 µM 55Fe2+ or 54Mn2+ in Ctrl oocytes and oocytes expressing wt or M438A-DMT1 (n = 21– 30). Two-way ANOVA (P < 0.001) followed by Holm–Šidák pairwise comparisons revealed that all variables differed (P ≤ 0.002) except 54Mn2+ uptake was no different in M438A-DMT1 compared with Ctrl (P = 0.35). (D) Fe2+-evoked currents (–70mV) in Ctrl oocytes and oocytes expressing wt or M438A-DMT1. Data were Fe fit to Eq. 4.1 providing the following estimates: Imax = –156 ± 10 nA, . = 1.8 ± 0.4 µM, a = 1.8 ± 0.4 for Fe wtDMT1 (P = 0.001); Imax = –78 ± 9 nA, . = 15 ± 4 µM, a = 1.2 ± 0.2 for M438A-DMT1 (P < 0.001); Data

2 Fe for control could not be fit to Eq. 4.1 (r = –0.013, P = infinite). M438A-DMT1 exhibited an increase in . (P < 0.001 cf. wtDMT1).

Tyrosine-270 is required for DMT1-mediated iron transport:

Tyrosine (Tyr) residues can form very strong cation-π interactions with metal ions and may be expected to participate in metal coordination (Dougherty, 2007; Dougherty, 1996). I found that DMT1-mediated iron transport at pH 5.5 was sensitive to Rose Bengal (RB) photooxidation (Figure 4.4A). Specificity was demonstrated by the lack of such an effect in the dark (Figure 4.4A). Whereas RB can interact with sulfhydryl, imidazole, indole, and phenolic residues (Table 4.1), a significant effect was observed only when I preincubated oocytes with RB at pH 9.0, revealing a specific effect on the aromatic residues Tyr and tryptophan (or possibly phenylalanine). That tetranitromethane (TNM) also inhibited transport activity (Figure 4.4B) further supports the notion that Tyr residues play an important role in DMT1. There exist 22 Tyr residues in DMT1. Rather than mutate all of them, I selected Tyr residues that are at or near critical sites for the function of DMT1. My first target, Tyr-270 (1B isoform) lies in the vicinity of His-272, a residue known to be part of the H+-coupling mechanism driving Fe2+ transport (Mackenzie et al., 2006). I tested the hypothesis that Tyr-270 plays an important role in the iron-transport mechanism of DMT1. I mutated Tyr-270 to a non-aromatic un-reactive Ala and found that loss of Tyr at position 270 abolished 55Fe2+ transport activity to <1% that of wtDMT1 (Figure 4.4C).

Figure 4.4. Effect of Rose Bengal (RB)-catalyzed photooxidation on DMT1-mediated iron transport suggests a role for tyrosine residues. Radiotracer iron uptake in control (Ctrl) oocytes and oocytes expressing wildtype (wt) or mutant DMT1. (A) 1-h preincubation with 1 µM RB in the light or dark and as a function of pH (n = 9-13). Two-way ANOVA revealed an interaction (P < 0.001); within DMT1, there was no effect of RB in the dark (P ≥ 0.35), effect of RB in the light, at all pH (P ≤ 0.015) with the strongest effect at pH 9.0 (P < 0.001, cf. 5.5, 7.5). (B) 1-h preincubation in 1 or 10 mM tetranitromethane (TNM) or vehicle (1% ethanol, pH 8.0) (n = 12-16). Two-way ANOVA revealed an interaction (P < 0.001); within DMT1, there was an effect of TNM at all concentrations (P < 0.001). (C) Tyr–270 mutated to Ala (n = 10-13). One-way ANOVA (P < 0.001) followed by Holm–Šidák pairwise comparisons revealed that Y270A was different from wtDMT1 (P < 0.001) but not from Ctrl (P = 0.52).

Proline-266 is a required residue for DMT1-mediated iron transport:

Proline (Pro) residues introduce kinks into protein structures. NMR of synthetic peptides corresponding to TM1 and TM6 exhibit discontinuity of their helical structure (Wang et al., 2011; Xiao et al., 2010; Xiao et al., 2011), i.e. α-helix–flexible linker–α-helix (Figure 4.5A). I identified within the flexible linker region of TM6 a Pro residue (Pro-266) in close proximity to His-272, a residue known to be part of the H+-coupling mechanism driving Fe2+ transport (Mackenzie et al., 2006) and postulated that Pro-266 could be critical to

84 the structure of the linker region. Mutating Pro-266 to either glycine (“soft kink”) or alanine (“no kink”) disrupted by 90-95% the transport of 2 µM 55Fe2+ (pH 5.5) (Figure 4.5B).

Figure 4.5. Proline-266 is predicted to form a kink in transmembrane domain-6 (TM6). (A) Predicted location of P266 in TM6. (B) Radiotracer iron uptake in control (Ctrl) oocytes and oocytes expressing wildtype (wt) or mutant DMT1 (data are mean, SD for n oocytes per group). One-way ANOVA (P < 0.001) followed by Holm–Šidák pairwise comparisons revealed that P266A and P266G were different from wtDMT1 (P < 0.001) and not different from Ctrl (P ≥ 0.096).

4.4.3 Functional analysis of human mutations in DMT1

Mutations in human DMT1 have been reported now in five unrelated cases of hypochromic–microcytic anemia (Table 4.2). I tested the hypothesis that these mutations disrupt iron transport—thereby explaining the anemia phenotype—by using the voltage clamp and radiotracer (55Fe2+) assays in Xenopus oocytes expressing wt or mutant DMT1. Whereas in two of these cases the probands’ phenotypes could be explained by the loss of iron-transport activity in the mutated DMT1, the remainder of the mutants exhibit normal transport activity, thus requiring further analysis. The first identified was a Czech proband homozygous for a G→C substuon in the last nucleode of exon 12, with the following two consequences: (i) expression of a full-length DMT1 protein containing a conservative E399D substitution, and (ii) increased frequency of skipping exon 12 from the normal 10% to 90% in the proband (Priwitzerova et al., 2004; Mims et al., 2005). I found that deletion of exon 12 (Ex12Δ) results in ablation of iron-transport activity of DMT1 (Figure 4.6). Ex12Δ is predicted to result in a protein lacking putative TM8 and the fourth-cytosolic loop between TM8 and TM9 (Figure 4.1). I found that E399D-DMT1 exhibited similar iron transport activity compared with wtDMT1 (Figure 4.6). Therefore, the probands phenotype is explained by the predominating expression of nonfunctional Ex12Δ-DMT1.

Table 4.2. Mutations in human DMT1 and initial clinical features of the probands. Reference values are provided in the final row. Reproduced from (Shawki et al., 2012).

Mean Mutation Mean Urinary [Hb] corpuscular Liver iron Description of the impact of the mutation Proband corpuscular Hematocrit (%) hepcidin (ng / (g/dL) hemoglobin overload and appropriate references volume (fL) mg creatinine) Gene Peptide Description (pg/red cell)

Exon 12 deletion disrupts the expression, Homozygous; stability, maturation and activity of the conservative; protein; however, E399D functions normally E399D / increased Czech G1285C 7.4 54 15 25 4 +++ (Gunshin et al., 2005b; Lam-Yuk-Tseung et Ex12Δ skipping of exon al., 2005; Mims et al., 2005; Priwitzerova et 12 from normal al., 2004; Priwitzerova et al., 2005; Shawki et 10% to 90% al., 2006).

Compound Deletion of exon 5 leads to loss of iron- heterozygous; 3- transport activity; R416C exhibits defective C1246T; R416C / bp del in intron 4 processing, cell-surface expression and Italian 4.0 71 13 25 92–100 ++ CTT del Ex5Δ leads to 30–35% activity (slower kinetics) (Iolascon et al., skipping of exon 2006; Lam-Yuk-Tseung et al., 2006; Shawki et 5 al., 2006).

Compound G723T, Transport is activity disrupted in V114Δ but G212V, heterozygous; in- French [Paris] G428T429G430 8.3 64 15 26 19–43 + G212V is still capable of transporting iron V114Δ frame deletion of del (Beaumont et al., 2006; Shawki et al., 2008). Val-114

G75R is postulated to be fully defective Ecuadorian G311A G75R Homozygous 5.1 54 22 27 N/A Normal (Figure 4.10) (Blanco et al., 2009).

N491S results in a truncated protein but French G723T, G212V, Compound G212V is still capable of transporting iron 8.6 58 18 N/A N/A + [Rennes] A1560G N491S heterozygous (Bardou-Jacquet et al., 2011; Shawki et al., 2008).

Reference values for hematological variables 9.0– Reference values 70–126 23–40 28–67 10–200a (age day 1 through > 18 years in males and 22.5 females)b

a (Papanikolaou et al., 2005) b Children’s Hospitals and Clinics of Minnesota accessed at http://www.childrensmn.org/manuals/lab/hematology/018981.asp

Figure 4.6. Radiotracer iron uptake in control (Ctrl) oocytes and oocytes expressing wildtype (wt) and DMT1 human mutants (data are mean, SD, for n oocytes per group). One-way ANOVA (P < 0.001) followed by Holm–Šidák pairwise comparisons revealed that all mutants were different form wtDMT1 (P < 0.001) except for E399D-DMT1 (P = 0.64), and no mutant was different from Ctrl (P > 0.99) except for E399D- DMT1, R416C-DMT1, and G212V-DMT1 (P < 0.001), .

Hydropathy analysis (Figure 4.1) predicted that the C-terminus (C-term) of DMT1 be intracellular; however, no empirical evidence exists to support this prediction. I predict that the loss of TM8 and cytosolic loop between TM8 and TM9 in this proband will result in the flipping of the C-term from an intracellular to extracellular orientation. Ex12Δ-DMT1 provides me with an opportunity to test the hypothesis that the C-term of wtDMT1 is intracellular. I will perform immunohistochemistry of nonpermeabilized and permeabilized Xenopus oocytes expressing wt or Ex12Δ-DMT1 tagged with GFP at the C-term. I expect that I will be able to visualize wtDMT1 fluorescence in permeabilized, but not in-tact, oocytes. I expect that I will be able to visualize Ex12Δ-DMT1 in both nonpermeabilized and permeabilized oocytes. The second identified was an Italian proband compound heterozygous for a 3-bp deletion in intron-4 that led to deletion of exon-5 (Ex5Δ) and a CT nucleotide substitution resulting in the Arg-to- Cys amino acid substitution at residue 416, with the following consequences: (i) expression of a full-length DMT1 protein containing a non-conservative R416C substitution, and (ii) increased frequency of skipping exon 5 to 30-50% (Iolascon et al., 2006; Lam-Yuk-Tseung et al., 2006). I found that Ex5Δ results in complete ablation of iron transport activity of DMT1 (Figure 4.6). Ex5Δ is predicted to result in a protein lacking putative TM2 and TM3 and the cytosolic loop between TM2 and TM3 (Figure 4.1). I found that R416C- DMT1 exhibited reduced iron-transport activity to 64% ± 5% (SE) that of wtDMT1 (Figure 4.6).

To further examine the impact of the R416C mutation, I measured presteady-state currents in oocytes expressing wtDMT1 by applying step changes in membrane voltage in the absence of metal ion (Figure 4.7A). Presteady-state currents are a common feature of ion-coupled solute transporters for which mathematical modeling and simulations have permitted investigators to attribute presteady-state currents to two steps in the transport cycle, namely (i) reorientation of the empty carrier within the membrane from inward (cytoplasmic) facing to outward (extracellular) facing, and (ii) the binding and dissociation of the driving ion (Mackenzie et al., 1996b; Mackenzie et al., 1996a; Hazama et al., 1997; Loo et al., 1993; Mackenzie et al., 2008; Sacher et al., 2002). I integrated the area under the curve to obtain

charge (Q) and fit Q as a function of membrane potential (Vm) by using the Boltzmann equation (Equation

4.2). R416C-DMT1 exhibited much smaller charge movements (i.e. lower Qmax) than wtDMT1; however,

the midpoint (V0.5) of the charge distribution was unchanged (Figure 4.7B). Valence (z) estimated from the Boltzmann fit was reduced from –2 in wtDMT1 to –1 in R416C-DMT1, i.e. there is less movable net charge

in the unloaded R416C mutant. I calculated the number of functional units (NT) of the transporter using Equation 4.3. I found that the number of functional units of R416C was 1.3×1011, which was roughly similar to that of wtDMT1, 1.6×1011. Therefore, the lower activity of R416C is not due to reduced protein levels at the plasma membrane of the oocyte.

I calculated the time constants of the presteady-state (transient) currents by fitting to a double- exponential decay function, and examined the relationship of τ2 (the slower of the two time constants obtained from the double-exponential fit) as a function of Vm (Figure 4.7C). I found that R416C-DMT1 exhibited higher τ2 and higher τmax than did wtDMT1 (Figure 4.7C), i.e. slower presteady-state kinetics, which are attributable to transitions between steps 6, 1, and 2 in the model (Figure 4.7E). I measured the

+ + H H -leak current as a function of H concentration and found that the . in R416C-DMT1 is no different from wtDMT1 (Figure 4.7D). Since I found no change in the H+-binding affinity, I concluded that the R416C mutation slows the reorientation of the empty carrier, i.e. slows the rate k61 (Figure 4.7E). In essence, R416C-DMT1 is a “clumsy” transporter—slower and less mobile—a finding that suggests that either (i) Arg-416 lies within or interacts with the mobile portion of the protein, or (ii) the Cys residue introduces some constraints on the mobility of the transporter.

Figure 4.7. Transport kinetics in wildtype (wt) and R416C-DMT1 expressing oocytes in the absence of metal ion. (A) Compensated presteady-state current records for wildtype DMT1 (wtDMT1) from 3 ms after

step-changes in Vm. Illustration from (Mackenzie et al., 2007). (B) Currents from one oocyte expressing wtDMT1 and one oocyte expressing R416C-DMT1 were integrated with time to obtain charge (Q), plotted

against Vm. Data were fit by Eq. 4.2 providing the following estimates: Qmax = 50.2 ± 0.8 nC, V0.5 = 47.9 ±

2 2 0.7 mV, z = –2.0 ± 0.1, r = 0.953 for wtDMT1; Qmax = 22.3 ± 1.0 nC, V0.5 = 39.9 ± 2.9 mV, z = –1.0 ± 0.1, r =

0.920 for R416C-DMT1. (C) τ2 as a function of voltage. Presteady-state currents were fit by a double-

exponential decay; τ2 is the slower of two time constants. The τ2/Vm function was fit by a Gaussian relation

2 (Eq. 4.4) providing the following estimates: τmax = 23 ± 4 ms, Vτmax = 22 ± 3 mV, r = 0.94, P = 0.006 for

2 + wtDMT1; τmax = 47 ± 44 ms, Vτmax = 48 ± 16 mV, r = 0.57, P = 0.41 for R416C-DMT1. (D) H concentration

+ dependence of the H leak. Data were fit by Eq. 4.1 providing the following estimates: Imax = –46 ± 2 nA, H H . = 1.7 ± 0.2 µM, a = 1.1 ± 0.1 for wtDMT1; Imax = –35 ± 3 nA, . = 1.1 ± 0.3 µM, a = 0.8 ± 0.1 for R416C- H DMT1. R416C-DMT1 did not exhibit a change in . (P < 0.001 cf. wtDMT1). (E) Presteady-state currents resulted from two steps (shaded) in the transport cycle, namely reorientation of the empty carrier (τ2)

+ and/or H binding/ dissociation (τ1). R416C disrupted the reorientation step such that rate k61 may become rate-limiting.

Two other probands are compound heterozygotes with an intriguing missense mutation leading to a Gly212Val substitution in putative TM5 (Iolascon et al., 2006; Bardou-Jacquet et al., 2011; Beaumont et al., 2006) (Figure 4.1, Table 4.2). The second mutation in each of the probands results in ablation of radiotracer iron transport as follows: (i) deletion of Val-114 (V114Δ), which introduces a premature stop codon, results in a nonfunctional protein (Figure 4.6), and (ii) Asn491Ser results in loss of transport activity (Figure 4.6) despite the modestly strong expression of N491S-DMT1 at or near the oocyte plasma membrane (Figure 4.8B) compared with wtDMT1 (Figure 4.8A). The parents and siblings of each of the compound heterozygous probands are healthy indicating that mutations at both alleles contribute to the anemia phenotype. The mutant proteins are deficient in their transport activity or targeting to the endosomal or plasma membranes (summarized in Table 4.2); however, the finding that G212V-DMT1 is able to transport radiotracer iron similarly to wtDMT1 (Figure 4.6) (coupled with expected upregulation of DMT1) provides no explanation for the anemia observed in this patient, so I have further analyzed G212V mutated DMT1.

Figure 4.8. Live-cell imaging of human DMT1 tagged at the C-terminus with enhanced green fluorescent protein (EGFP) in wildtype DMT1 (wtDMT1) (A) or N491S-DMT1 (B) RNA-injected Xenopus oocytes and control (C) oocytes. Images approximately bisect the oocyte equator (optical slice ≈ 8 µm].

I measured iron-transport currents in oocytes expressing wt and G212V mutant DMT1 and found Fe that G212V-DMT1 exhibited a reduction in Vmax with no change in . (Figure 4.9A). Since iron-evoked

90 currents but not 55Fe transport were reduced in oocytes expressing the G212V mutant, I explored the H+/Fe2+ coupling ratio for wtDMT1 and G212V-DMT1 by measuring radiotracer transport in individual oocytes under voltage clamp (Figure 4.9B). From my calculation of total charge accumulation as a function of radiotracer accumulation, and taking into account the 2+ charge of iron, I found that wtDMT1 exhibited

2+ a proton-to-Fe coupling ratio (nc) of 17:1 (Figure 4.9C), far exceeding that which is expected for a tightly coupled system. I found that nc was just 1:1 for G212V-DMT1 (Figure 4.9C). Therefore, wtDMT1 exhibits substantial H+ ‘slippage’ that is abolished by the G212V mutation.

Figure 4.9. Functional properties of wildtype (wt) and G212V mutant DMT1. (A) Concentration dependence of the currents evoked by 0.1-20 µM Fe2+ in oocytes expressing wtDMT1 or the G212V Fe mutant. Data were fit by Equation 4.1 providing the following variables: Imax = –73 ± 4 nA, nH = 1 ± 0.1, .

2 Fe 2 = 1.1 ± 0.2 µM, r = 0.99 for wtDMT1; Imax = –24 ± 4 nA, nH = 0.8 ± 0.2, . = 1.4 ± 0.9 µM, r = 0.98 for Fe G212V-DMT1. The G212V mutation resulted in lesser max than for wtDMT1 (P < 0.001) but no change in Fe . (P = 0.41). (B) Current (–70mV) was measured continuously in a single oocyte expressing wtDMT1 (i) or G212V-DMT1 (ii). A stable baseline current was obtained in pHo 5.5 solution in the absence of iron

55 2+ before the addition of 5 µM Fe for 10 min and washing out with pHo 5.5 solution in the absence of iron. Current was integrated with time to obtain charge QFe and converted to a molar equivalent using the Faraday. (D) H+/Fe2+ coupling ratio was determined from B. The QFe/55Fe ratio was determined as the

+ 2+ slope of the linear relationship. The H /Fe coupling ratio (nC) was then calculated by taking into account

2+ the 2+ charge on Fe . For wtDMT1, nC ≈ 17:1, far exceeding that expected for a tightly coupled system and consistent instead with marked H+ “slippage”.

A fifth proband, homozygous for a missense mutation leading to a G75R substitution in putative TM1 (Figure 4.1), was severely anemic but displayed no hepatic iron loading (Blanco et al., 2009). Not discounting the possibility that iron loading may progress later in life, the authors in the study just cited postulated that the G75R mutant should be completely deficient in DMT1 functional activity and would more fully ablate iron transport. To begin to test this hypothesis, I expressed wildtype DMT1 (wtDMT1) and the G75R mutant in RNA-injected Xenopus oocytes and measured 55Fe2+ transport and currents. In oocytes expressing DMT1–enhanced green fluorescent protein (EGFP), I observed strong EGFP fluorescence over the entire cell perimeter, whereas I detected only very faint fluorescence in oocytes expressing G75R–EGFP (Figure 4.10A). I found no detectable iron-transport activity in the mutant (Figure 4.10B). In oocytes expressing wtDMT1 superfused first with pH-7.5 medium, then pH 5.5, I observed a small inward current (Figure 4.10C) that others have previously demonstrated to be a H+ ‘leak’ pathway (Gunshin et al., 1997; Mackenzie et al., 2006), and a large inward current upon the addition of 50 µM Fe2+. In contrast, G75R-DMT1 mediated a very large inward H+-leak current that was inhibited partly by the addition of Fe2+. Therefore, although the G75R mutant was not processed to the oocyte plasma membrane to the same extent as was wtDMT1, G75R-DMT1 exhibited a partial reaction (H+ transport) and was completely deficient in iron-transport activity.

Figure 4.10. Impact of the G75R mutation on expression and activity of human DMT1 in RNA-injected Xenopus oocytes. (A) Imaging of control oocytes, and wildtype DMT1 (wtDMT1)–enhanced green fluorescent protein (EGFP) or G75R–EGFP expression in RNA-injected Xenopus oocytes using fluorescence microscopy. Representative images are presented in which the optical slice (0.6 µm depth) approximately bisects the oocyte. Faint fluorescence at the cell perimeter of oocytes expressing G75R is marked by the white arrowheads. (B) Transport of 2 µM 55Fe2+ was measured in control (n = 10) oocytes and oocytes expressing wtDMT1 (n = 11) or G75R-DMT1 (n = 11) at pH 5.5 (data are mean, SD). One-way ANOVA (P < 0.001) followed by Holm–Šidák pairwise comparisons revealed that G75R did not differ from control (adjusted P = 0.94). (C) Continuous currents recordings in a control oocyte and oocytes expressing wtDMT1 or G75R-DMT1 and voltage clamped at −70 mV. Oocytes were continuously superfused for the periods shown by the bars with pH 7.5 medium (empty bars) or pH 5.5 medium (hatched bars) and 50 µM Fe2+ (black bars) in the presence of 1 mM L-ascorbic acid. Reproduced from (Shawki et al., 2012).

4.5 Discussion

My data extends our current understanding of the structure–function of DMT1. Having identified several residues critical to the function of DMT1, my goal now is to further explore the functional roles of the identified amino acid residues by referring back to the predicted membrane topology of DMT1, synthetic

93 peptides of transmembrane domains of DMT1, and—a very recent development—the crystal structure of a bacterial homologue (ScaDMT) (Ehrnstorfer et al., 2014). For example, I have found that DMT1 does not tolerate amino-acid substitutions at Pro-266, which I speculate provides a kink within the flexible linker region in TM6. Meanwhile, the ScaDMT structure suggests that this linker region, together with that in TM1, forms the metal-ion binding cavity, and that TM1/TM6 flexibility is required for the catalytic activity of ScaDMT.

I have curated information for all of the mutants that I have made, including some that are not discussed in this thesis, in Table 4.3 in the hope that this will serve as a useful resource in future experiments and in explaining the phenotypes arising from novel mutations in human DMT1.

4.5.1 Mapping critical structural elements in DMT1

Membrane topology models predict two consensus N-gly sites in DMT1 that reside in the fourth extracellular loop between TM7–8 (Mackenzie et al., 2007). I have begun to test the requirement of N- linked glycosylation for DMT1 functional activity and found that mutation of the predicted glycosylation sites does not abolish DMT1-mediated iron transport. The modest reduction in transport activity may possibly be explained by lower expression levels of mutant DMT1. Loss of glycosylation is expected to result in a smaller protein complex visualized by molecular weight shifts of the nascent protein (Mackenzie et al., 2007). I will check for molecular weight shifts of mutant DMT1 by western blotting to confirm the residues that undergo glycosylation. This approach will also permit me to assess protein expression levels of mutant and wtDMT1 in the membrane fraction.

The priority now is to thread the human DMT1 sequence onto known crystal structures to generate new structural models of human DMT1 (see section 5.2.2). This approach will (i) help me better interpret the results of mutagenesis experiments, and (ii) provide new targets for further analysis.

Table 4.3. Summary of mutations in DMT1 and the available functional data.

Surrounding Amino Mutation(s) Amino Acid Disease Phenotype (µM) Description of Effect Reference(s) Acid/Motif . Segment 1.42-1.58 wt hDMT1 N/A N/A N/A Wildtype, Multiple Isoforms A and B with ± Iron Regulatory Element (Mackenzie et al., 2007) (95% CI)

D86PGN AYIDPGNIES N/A N/A Highly Conserved Sequence in DMT1 homologues (Chaloupka et al., 2005) motif

Proband 4: Microcytic Chapter 4 G75 G  R WAFTGPGFL N/A Complete lack of iron transport activity. H+ leak. anemia, no iron overload (Blanco et al., 2009)

(Nevo & Nelson, 2006; D86 AYIDPGNIES Part of DPGN motif N/A Dead Protein Shawki et al., 2012) G88 G  A AYIDPGNIES Part of DPGN motif N/A Dead Protein, almost complete inhibition of uptake (Cohen et al., 2003) D93 D  A DPGNIESD N/A N/A Decreased Mn2+ uptake activity (Cohen et al., 2003) Q95D abolishes uptake activity, but double mutation with D93A, leads to the Q95 A  D ESDLQSGAV N/A N/A (Cohen et al., 2003) restoration of the Q95D mutation Chapter 4 V114 Deletion LATLVGLLL Proband 3 with G212V N/A Dead Protein (Shawki et al., 2008) mk mouse and Belgrade Leads to improper trafficking, decreased expression on the membrane, altered (Su et al., 1998; Xu et al., G185 G  R PLWGGVLIT rat, severe hypochromic, N/A glycosylation leading to increased degradation, impaired DMT1 function, Ca2+ 2004) microcytic anemia conductance (gain of function) D192 D  A ITIADTFVF N/A N/A Highly Conserved, essential for transport, located in TM 4 (Nevo & Nelson, 2006) Reduction of H+ slippage without significant changes in the metal ion transport F196 F  I DTFVFLFLD N/A N/A (Nevo & Nelson, 2004) activity or expression level Proband 3 with V114 1.7 G212 G  V EAFFGFLIT No effect on 55Fe2+ transport, disrupt H+ slippage, new H+:Fe2+ ratio is 1:1, not 17:1 (Shawki et al., 2008) Mutation (unchanged) P266A 11.1 ± 3.7 µM Chapter 4 P266 P  A, G AVIMPHNMYLH N/A Abolition of transport function, possibly altering vestibule architecture P266G 34.4 ± (Alexander et al., 2010) 3.9 H272 H  A PHNMYLHSALVK N/A 9 Loss of H+ coupling mechanism, facultative Fe2+ transport (Mackenzie et al., 2006) E330 E  A YFFIESCIA N/A N/A Highly Conserved, essential for transport, located in TM 7 (Nevo & Nelson, 2006) Proband 1 homozygous End of Exon 12 Loss of intron slice site, not recognized by the spliceosome, results in Exon 12 Chapter 4 E399 E  D for GC nucleotide N/A (GTYSGQFVME) deletion, E399D expressed mutant does not show altered function (Shawki et al., 2006) substitution Proband 2 heterozygous Disrupts reorientation of carrier leading to lower Fe2+ transport and a reduced Chapter 4 R416 R  C VVLTRSIAI 1.1 ± 0.3 with Exon 5 Deletion Vmax due to reduced turnover from 20 s-1 to 7 s-1 (Shawki et al., 2006) Chapter 4 M438 M  A HLTGMNDFL N/A 15.3 ± 4.2 70% Reduction in transport…more to come (Alexander et al., 2010) *BLUE indicates a critical residue in proximity to the selected residue in GREEN, DPGN motif is indicated by YELLOW 95

4.5.2 Functional analysis of human mutations in DMT1

For many years, investigators have known of human pedigrees containing multiple members displaying a severe microcytic-anemia phenotype, characterized by a deficiency in intestinal iron absorption that was only partially corrected with parenteral iron administration (Buchanan & Sheehan, 1981; Hartman & Barker, 1996). Whether or not such pedigrees harbor mutations in DMT1 is not known, but their phenotypes are strikingly similar to the DMT1 mutants just described. In any event, mutations in human DMT1 are extremely rare—presumably a reflection of both its critical, nonredundant functional roles and its long phylogenetic ancestry (Chaloupka et al., 2005), such that mutations in DMT1 are not generally tolerated.

Whereas some of the human mutants result in defective iron transport activity, my hypothesis was not adequate to explain the impact of the other mutations. For example, R416C-DMT1 exhibited slower reorientation and less mobility of the empty carrier from inward facing to outward facing. Judging from the ScaDMT crystal, Arg-416 of human DMT1 is not expected to reside near the predicted metal-ion binding pocket. Instead, the clumsy nature of R416C-DMT1 might be explained in part by (i) a loss of charge resulting in loss of specific interactions between Arg-416 and neighboring amino-acid residues, or (ii) the formation of a disulfide bond (at the Cys substitution) that constrains the movement of the transporter.

G212V is an exceptional mutation. Whereas wtDMT1 exhibited significant H+ slippage which is normally thought as an energetic penalty to the cell, G212V abolished H+ slippage. The observation that H+ slippage is disrupted in a mutant associated with a disease phenotype suggests that H+ slippage may serve a physiological role. I speculate that the H+ leak serves to maintain the solubility of iron once it is transported into the cell by DMT1 and before iron is incorporated into ferritin or other chaperon proteins. Although incorporation of iron into ferritin is slow, ferritin is not likely to serve as the primary chaperon; however, investigators have identified candidate cytosolic iron-chaperon proteins, poly(rC)-binding protein-1 and 2 (PCBP1 and PCBP2) (Frey et al., 2014).

The effect of G212V on slippage suggests that this residue could interact with the H+-coupling machinery, known to involve His-272 (Mackenzie et al., 2006); however, the ScaDMT crystal provides no clear support for the proximity of Gly-212 and His-272 to the putative metal-ion binding pocket formed between TM1 and TM6 (Ehrnstorfer et al., 2014). Crystallization of G212V mutated ScaDMT would provide a better tool in understanding the structural effect of the G212V mutant if one exists.

In addition to analyzing the structural effects of G212V-DMT1, I will perform the following functional experiments in oocytes to better understand the H+/Fe2+ coupling mechanism. We know that DMT1- mediated Fe2+ transport in vitro is strongly voltage dependent and pH dependent (Gunshin et al., 1997; Mackenzie et al., 2006; Mackenzie et al., 2007) (Figure 3.1), so I will measure the H+/Fe2+ coupling ratio (i) as a function of membrane potential, and (ii) as a function of extracellular pH.

Although the G75R mutant was not processed to the oocyte plasma membrane to the same extent as was wtDMT1, I found that G75R-DMT1 exhibited a partial reaction (H+ transport) and was completely deficient in iron-transport activity. I speculate that mutation of Gly-75, located within TM1, disrupts metal- ion binding but permits H+ translocation.

Human DMT1 mutations are associated with microcytic anemia. Not only were the probands anemic, all but one (G75R) showed signs of hepatic iron overload—a finding which, at first, did not appear consistent with DMT1’s perceived role in iron absorption. The first possible explanation offered was that intestinal heme-iron absorption could be upregulated, independently of DMT1 (Priwitzerova et al., 2004). In the study just cited, the authors suggested that iron overload was not apparent in the b and mk rodent models because rodents have poorly developed heme-absorptive pathways and heme iron is not typically a part of rodent chow. However, the b rat—in which a defect in intestinal metal-ion uptake has been established directly (Knöpfel et al., 2005b)—also displays hepatic iron loading (Thompson et al., 2006).

It is plausible that the mutations affect the function of various DMT1 isoforms (see section 1.4.2) differently, or that the severity of the mutation is more pronounced in the environment of the erythroid precursor cell endosome than it is at the intestinal brush border. A more straightforward explanation may be that the human mutations—like the rodent G185R mutation—possess some miniscule activity which, coupled with upregulation of the intestinal DMT1 isoform 1A/IRE(+) (see section 1.4.2) by an erythropoietic factor, permits some iron absorption to proceed in these probands. Because the absorbed iron is poorly utilized by erythroid precursors, much is diverted to the liver stores. Supporting this notion, urinary levels of the iron-regulatory hormone hepcidin, whose actions lead to a downregulation of iron absorption (Ganz & Nemeth, 2006), are low or markedly suppressed in these probands (Table 4.1).

5 General Discussion

Sections 5.1.4–5.1.6, and 5.2.2 contain both my original work and excerpts from the following co-

authored publication Shawki A, Knight, PB, Maliken, BD, Niespodzany EJ, Mackenzie B (2012) H+-

coupled divalent metal-ion transporter-1: Functional properties, physiological roles and therapeutics.

Curr Top Membr. 70, 169-214.

5.1 Physiological roles of DMT1

I have demonstrated that intestine-specific knockout of DMT1 produces a profound iron-deficiency anemia, confirming a critical role for DMT1 in iron homeostasis. I found that DMT1 is essential for the intestinal apical uptake of nonheme iron but is not required for the absorption of copper, manganese, or zinc.

I found that NHE3 knockout mice exhibit a moderate iron deficiency and compensatory regulation of iron-metabolism genes. Ablation of NHE3 impairs intestinal iron absorption, revealing a role of NHE3 in generating the H+ gradient that drives apical iron uptake.

I have demonstrated that human mutations in DMT1, in general, result in impaired DMT1 function explaining the probands anemia phenotype and analysis of other mutants has revealed novel aspects of the molecular physiology of DMT1.

Yet, there remain many unanswered questions as to the many roles of DMT1 in iron metabolism and in heavy-metal intoxication. I discuss these questions here and propose approaches that could be used to address them.

5.1.1 Is DMT1 required for intestinal iron absorption in the neonate?

I have shown that intestinal DMT1 is required for iron absorption in the adult mammal and that iron absorption relies in part on the activity of NHE3 to generate the H+ electrochemical potential gradient to drive DMT1-mediated H+-coupled iron transport at the intestinal brush border. The immature mammalian intestine, however, lacks NHE3 until around weaning (Collins et al., 1997), so I had postulated that DMT1 should not be essential in the suckling mammal. My preliminary data demonstrate, however, that DMT1 is even required in the suckling mouse. This observation raises the possibility that DMT1 participates in the absorption of iron from other sources (e.g. lactoferrin). Iron absorption in the immature intestine is

98 poorly understood, despite the prevalence of iron deficiency and iron-related developmental and neurological deficiencies in human neonates and infants [reviewed, (Collard, 2009)]. The DMT1int/int mouse model provides an excellent tool to further characterize the role of DMT1 in neonatal intestinal iron absorption.

5.1.2 Can we target NHE3 in the prevention of iron overload?

That NHE3 is required for intestinal iron absorption and that DMT1 is the principal nonheme iron transporter in the intestine raises the possibility that NHE3 may be targeted to block iron absorption indirectly rather than directly by targeting DMT1. Experimental drugs are available to further test the hypothesis that blocking NHE3 will rescue the iron overload phenotype in the HFE mouse model of hemochromatosis and other iron-overload models.

A non-absorbed NHE3 blocker may be expected to have fewer unwanted effects. Use of an NHE3 blocker that is specific to NHE3 over NHE2 and/or that is destroyed as it courses along the intestine may be effective in inhibiting iron absorption without producing diarrhea since Na+/H+ exchanger activity in the colon, along with the activity of the epithelial Na+ channel (ENaC) and colonic H+/K+ ATPase (cHKA) are expected to mediate Na+ and fluid absorption. Schultheis et al have demonstrated that small intestine luminal contents of NHE3-null mice are more alkaline than wildtype mice; however luminal contents of the colon are more acidic (Schultheis et al., 1998b), indicating that NHE2 is mainly localized to the colon and may be capable of compensating for loss of NHE3, along with cHKA and colonic ENaC, especially during treatment with an NHE3 blocker. Since ablation of NHE3 in the mouse resulted in iron deficiency without an overt anemia suggests that blocking NHE3 in hemochromatosis patients may reduce iron absorption without placing them at risk of becoming anemic.

5.1.3 What is the role of DMT1 in heme absorption?

Whereas dietary heme iron offers greater bioavailability than nonheme iron, little is known about the mechanisms of heme-iron absorption. Heme (ferrous protoporphyrin IX) is thought to be taken up intact by the enterocyte via receptor-mediated endocytosis. Investigators have identified a putative intestinal low affinity heme transporter (HCP-1) that is capable of transporting heme in HeLa cells (Shayeghi et al., 2005); however, another group discovered that HCP1 is actually a proton–coupled folate–preferring transporter (PCFT1) in the small intestine (Qiu et al., 2006), although some groups claim that this transporter may serve a dual role (Garrick & Garrick, 2009). Other investigators have discovered a human

ortholog (HRG-1) of the C. elegans heme responsive gene, CeHRG-1, to be regulated by cellular heme levels (Rajagopal et al., 2008) but no role in the intestine has been demonstrated.

The liberation of iron from heme by heme oxygenase and the acidification of the heme-containing endo–/lysosome by the V-type H+-ATPase produces a favorable environment for DMT1-mediated iron transport. I hypothesize that DMT1 is responsible for transporting iron out of the endo–/lysosome in a manner similar to that involved in transferrin-mediated iron acquisition in erythroid precursors. If my hypothesis holds up, and DMT1 is involved, then knocking out intestinal DMT1 would result in the trapping of 59Fe within enterocytes of DMT1int/int mice administered [59Fe]heme or [59Fe]hemoglobin.

5.1.4 How does DMT1 contribute to Cd absorption?

DMT1 efficiently transports cadmium in vitro (Illing et al., 2012). Iron deficiency is a risk factor for cadmium intoxication (Bressler et al., 2004; Berglund et al., 1994; Goyer, 1997; Flanagan et al., 1978; Kippler et al., 2009; Åkesson et al., 2002), suggesting that cadmium and iron share a common absorptive pathway. Indeed, investigators have ascribed to DMT1 the Cd2+ transport observed in Caco-2 cells (Bannon et al., 2003). Iron restriction increased both intestinal DMT1 mRNA and cadmium accumulation in rats and mice (Park et al., 2002; Suzuki et al., 2008), implicating DMT1 in cadmium absorption. Clouding that conclusion, in one of the studies just cited (Suzuki et al., 2008), mk mice accumulated cadmium to the same levels as did wildtype mice. The DMT1int/int model results in efficient knockout of intestinal DMT1 whereas substantial DMT1 activity persists in the mk mouse model. I will directly measure rates of 109Cd absorption in the DMT1int/int mouse to test the role of DMT1 in cadmium absorption and intoxication.

5.1.5 What is the role of DMT1 in the kidney in health and disease?

DMT1 is abundantly expressed in the kidney (Wang et al., 2012; Hubert & Hentze, 2002; Gunshin et al., 1997). The relative importance of DMT1 in the renal reabsorption of iron (a topic that, historically, has not received much attention) is yet to be fully explored (Thévenod & Wolff, 2015). Meanwhile, the literature supports an involvement of DMT1 in renal pathophysiology. Nephrotic syndromes may be associated with iron-induced damage resulting from increased glomerular permeability, increased Tf uptake in the proximal tubule, and increased DMT1-mediated iron transport into the cytosol [reviewed, (Smith & Thévenod, 2009)]. DMT1 is thought also to contribute to cadmium-induced nephrotoxicity [reviewed, (Vesey, 2010; Thévenod & Wolff, 2015)] by transporting cadmium from the late endo–/lysosome to the cytosol following its liberation from metallothionein. Cadmium–metallothionein-1 internalization by

100

WKPT-0293 Cl.2 proximal tubular cells, via the same megalin-dependent receptor-mediated endocytic pathway used by Tf, leads to cytotoxicity (Wolff et al., 2006) that can be prevented by RNAi knockdown of late endosomal/lysosomal DMT1 (Abouhamed et al., 2007). Villin is expressed in the intestine and the proximal tubule of the kidney (Gröne et al., 1986). I have confirmed by using qPCR efficient ablation of intestinal DMT1 (DMT1int/int) generated by using the villin-Cre transgenic line. I can use this same approach to examine the extent to which DMT1 is knocked down in the cortex and medulla of the kidney of the DMT1int/int mouse.

5.1.6 What are the roles of DMT1 in the brain in health and disease?

Iron transport and metabolism in the brain [reviewed, (Núñez et al., 2012; Rouault & Cooperman, 2006; Burdo & Connor, 2002; Moos & Morgan, 2004a; Snyder & Connor, 2009)] warrant special attention because of the association of neurodegenerative disease and iron accumulation. The chief pathway by which iron enters the brain is via the brain capillary endothelial cells of the blood–brain barrier (BBB); the choroid plexus of the blood–cerebrospinal fluid barrier (BCB) presents a second pathway. At both sites, iron is taken up primarily via Tf-dependent endocytosis (Bradbury, 1997), and the identification of DMT1 in BBB and BCB (Burdo et al., 2001; Siddappa et al., 2003; Rouault et al., 2009; Gunshin et al., 1997; Moos & Morgan, 2004b; Burdo et al., 2004) implicates DMT1 in that pathway. Colocalization of DMT1 with early endosomal markers in neuronal and astrocyte cell lines (Lis et al., 2004; Lis et al., 2005) suggests that DMT1 also participates in Tf-dependent uptake in these cells; although, in one study, DMT1 colocalized with late endosomal markers in neurons (Pelizzoni et al., 2012). DMT1 is expressed in neurons throughout the brain, including cortex, cerebellum, and thalamus, anterior olfactory nucleus, substantia nigra, striatum, and hippocampal pyramidal and granule cells (Gunshin et al., 1997; Williams et al., 2000; Burdo et al., 2001; Knutson et al., 2004; Moos et al., 2000).

DMT1 protein is also observed on the plasma membrane of neurons (Roth et al., 2000; Lis et al., 2005) and astrocytes (Wang et al., 2002b), thereby prompting us to consider a role for DMT1 in uptake of nontransferrin-bound iron (NTBI). Although maximally stimulated at lower pH, DMT1 is still active at pH 7.4 (Garrick et al., 2006; Mackenzie et al., 2006). Brain interstitial fluid and cerebrospinal fluid contain substantial levels of NTBI (Moos & Morgan, 1998), much of it as Fe2+ because of the high levels of ascorbate in the brain; thus, DMT1 expressed at the plasma membrane could take up NTBI without the need for surface ferrireductases. Such a role has recently been ascribed to voltage-operated Ca2+ channels largely from the observation that iron uptake was inhibited by Ca2+ (Pelizzoni et al., 2011); however,

101

because Ca2+ also inhibits DMT1 (Shawki & Mackenzie, 2010), this transporter should not be discounted. Hypoxia increased the expression of DMT1 on neuronal plasma membranes (Lis et al., 2005), and we should also consider a role for DMT1 in the transport of manganese at that location.

Neuronal iron accumulation is at least associated with prevalent neurodegenerative diseases including Parkinson disease, Alzheimer disease, amyotrophic lateral sclerosis, and Huntington disease. DMT1 is abundantly expressed in the substantia nigra pars compacta (SNpc) (Gunshin et al., 1997; Moos et al., 2000), a site of iron enrichment in healthy subjects. In Parkinson disease, substantial iron accumulation is observed in the SNpc, and this may be associated with the oxidative stress and neuronal damage characteristic of the disease (Snyder & Connor, 2009). Parkin, loss of which is characteristic of a familial form of Parkinson disease, directs the ubiquitination of DMT1 1B/IRE(−) and its proteasomal degradation (Garrick et al., 2012). There is no strong evidence with which to associate DMT1 mutations or single-nucleotide polymorphisms with the initiation of neurodegenerative diseases (Jamieson et al., 2005; Blasco et al., 2011); however, upregulation of DMT1 is implicated in their neuropathogenesis (Salazar et al., 2008; Song et al., 2007; Zheng et al., 2009; Aguirre et al., 2012; Blasco et al., 2011; Núñez et al., 2012; Huang et al., 2006) and, conversely, loss of DMT1 function in the b rat and mk mouse provides some protection against iron-mediated toxicity (Moos & Morgan, 2004b; Salazar et al., 2008).

Manganism—a syndrome resembling Parkinson disease—results from chronic exposure to manganese (e.g., via occupational exposure in miners and welders) (Roth, 2006; Rivera-Mancía et al., 2011). In addition to the typical routes by which iron may enter the brain, atmospheric manganese may gain access to the central nervous system (CNS) via olfactory or trigeminal presynaptic nerve endings and retrograde axonal transport to the brain (Roth, 2006; Tjälve & Henriksson, 1999). Absorption of a nasally instilled manganese dose in olfactory epithelium required DMT1 and was increased by anemia (Thompson et al., 2007). Within the brain, neuronal manganese uptake also appears to rely on DMT1, and this pathway may contribute to manganese-induced neurotoxicity (Roth & Garrick, 2003; Roth, 2006; Rivera- Mancía et al., 2011; Au et al., 2008).

5.2 Molecular physiology of DMT1

5.2.1 Site-directed mutagenesis and functional analysis

Structure-function analysis using data from predicted critical structures and human mutations in DMT1 associated with a disease phenotype provide information on critical features for the proper function of

102

the transporter in the absence of a crystal structure. This information will be used to build an annotated model of DMT1 as an example for other secondary active membrane transporters.

5.2.2 Structure-function analysis of DMT1

Prior to the crystallization of a DMT1 homolog (Ehrnstorfer et al., 2014), investigators have used a threading approach to predict structural features of DMT1 by using the primary sequences of DMT1 or SLC11 homologues to query known crystal structures of homologous proteins (Courville et al., 2008; Czachorowski et al., 2009). Despite very little similarity at the amino acid level, SLC11 transporters could be superimposed on the structures of the bacterial Na+/Cl−-dependent leucine transporter (LeuT) of the SLC6 family (Yamashita et al., 2005) and the Na+-coupled galactose transporter (vSGLT) of the SLC5 family (Faham et al., 2008). Modeling SLC11 homologs on the LeuT structure suggested an internal symmetry in which TM1–5 and TM6–10 fold similarly but are oriented oppositely (Courville et al., 2008). The LeuT fold also predicts the discontinuous helical structures of TM1 and TM6 observed for the short synthetic model peptides (Wang et al., 2011; Xiao et al., 2010; Xiao et al., 2011) and places the D86PG and (inverted) H267PM motifs in the extended peptide regions of TM1 and TM6 respectively (Courville et al., 2008).

Recently, the bacterial homologue of DMT1 Staphylococcus capitis, i.e. ScaDMT, has been crystallized (Ehrnstorfer et al., 2014). Although the ScaDMT architecture is similar to that of LeuT, ScaDMT exhibits 37% identity at the amino acid level and 59% homology with human DMT1 possibly making ScaDMT the better choice in structural modeling of human DMT1. ScaDMT exhibits similar substrate selectivity to human DMT1 (Illing et al., 2012) judging from a fluorescence-based assay in proteoliposomes (Ehrnstorfer et al., 2014). The metal-ion binding site of ScaDMT consists of the interaction of four amino acid residues, two of which are found in TM1 and two residues in TM6. Amino acid residues Asp-49 and Asn-52 (corresponding to Asp-86 and Asn-89 in human DMT1 1B isoform) found in TM1 are part of the absolutely conserved (from bacteria to human) D86PGN sequence. The remaining two amino acid residues A223 and M226 (corresponding to A262 and M265 in human DMT1) are found in TM6. TM6 has previously been shown to form part of the H+/Fe2+ coupling machinery (Mackenzie et al., 2006).

Models of the crystal structure of ScaDMT suggest that ScaDMT binds alkaline earth metals barium and strontium near but not at the transition metal-ion binding site. We have shown that barium and strontium inhibit iron transport activity of human DMT1 in vitro (Shawki & Mackenzie, 2010). That alkaline earth metal binding site is discrete from the transition metal binding site of ScaDMT is consistent

103

with our finding that calcium is a non-competitive inhibitor of DMT1 mediated iron transport in vitro (Shawki & Mackenzie, 2010).

In summary, experimental results, model predictions, and crystal structures provide evidence that TM1, TM4 and TM6 in particular contain structural elements that are critical for the catalytic activity of DMT1-mediated H+-coupled metal-ion transport. That mutations or the introduction of HA epitope tags elsewhere in DMT1 disrupt DMT1 catalytic activity or the correct processing of the protein would suggest that additional critical elements await their discovery.

Crystallization of the ScaDMT homolog and the availability of other structures, Vibrio parahaemolyticus sodium/galactose cotransporter (vSGLT) (Faham et al., 2008; Watanabe et al., 2010) and the Microbacterium liquefaciens cation/nucleobase cotransporter (Mhp1) (Weyand et al., 2008; Shimamura et al., 2010) present the opportunity for us to ‘thread’ human DMT1 onto the available structures to generate structural models for human DMT1. We can test these models by using state-of- the-art computational techniques that provide information to predict critical residues. We can test these critical residues by using site-directed mutagenesis and functional assays in cell systems.

Much work is yet to be done to uncover the structure of human DMT1; however, that we know DMT1 is (i) an iron-preferring H+-coupled secondary-active transporter, (ii) essential in intestinal iron absorption and erythroid iron utilization, and (iii) implicated in metal toxicity serves as impetus to increase the focus on DMT1 for the improvement of iron nutrition and or development of drugs that target DMT1 directly or indirectly by targeting a critical component to the function of DMT1.

104

References

Abboud S & Haile DJ (2000). A novel mammalian iron-regulated protein involved in intracellular iron metabolism. J Biol Chem 275, 19906-19912.

Abouhamed M, Gburek J, Liu W, Torchalski B, Wilhelm A, Wolff NA, Christensen EI, Thévenod F, & Smith CP (2006). Divalent metal transporter 1 in the kidney proximal tubule is expressed in late endosomes/lysosomal membranes: Implications for renal handling of protein-metal complexes. Am J Physiol Renal Physiol 290, F1525-F1533.

Abouhamed M, Wolff NA, Lee W-K, Smith CP, & Thévenod F (2007). Knockdown of endosomal/lysosomal divalent metal transporter 1 by RNA interference prevents cadmium-metallothionein-1 cytotoxicity in renal proximal tubule cells. Am J Physiol Renal Physiol 293, F705-F712.

Aguirre P, Urrutia P, Tapia V, Villa M, Paris I, Segura-Aguilar J, & Núñez MT (2012). The dopamine metabolite aminochrome inhibits mitochondrial complex I and modifies the expression of iron transporters DMT1 and FPN1. Biometals 25, 795-803.

Ahmad KA, Ahmann JR, Migas MC, Waheed A, Britton RS, Bacon BR, Sly WS, & Fleming RE (2002). Decreased liver hepcidin expression in the hfe knockout mouse. Blood Cells Mol Dis 29, 361-366.

Åkesson A, Berglund M, Schütz A, Bjellerup P, Bremme K, & Vahter M (2002). Cadmium exposure in pregnancy and lactation in relation to iron status. Am J Public Health 92, 284-287.

Alexander JH, Shawki A, Meller J, & Mackenzie B. Mapping critical structural elements in divalent metal- ion transporter-1 (DMT1). FASEB Journal 24, 609.7. 2010. Ref Type: Abstract

Anderson CMH, Grenade DS, Boll M, Foltz M, Wake KA, Kennedy DJ, Munck LK, Miyauchi S, Taylor PM, Campbell FC, Munck BG, Daniel H, Ganapathy V, & Thwaites DT (2004a). H+/amino acid transporter 1 (PAT1) is the imino acid carrier: An intestinal nutrient/drug transporter in human and rat. Gastroenterology 127, 1410-1422.

Anderson CMH & Thwaites DT (2005). Indirect regulation of the intestinal H+-coupled amino acid transporter hPAT1 (SLC36A1). J Cell Physiol 204, 604-613.

Anderson GJ, Ramm GA, Subramaniam VN, & Powell LW (2004b). HFE gene and hemochromatosis. J Gastroenterol Hepatol 19, 712.

Andrews NC (2005). Understanding heme transport. N Eng J Med 353, 2508-2509.

Andrews NC & Schmidt PJ (2007). Iron homeostasis. Ann Rev Physiol 69, 69-85.

105

Andriopoulos B, Corradini E, Xia Y, Faasse SA, Chen S, Grgurevic L, Knutson MD, Pietrangelo A, Vukicevic S, Lin HY, & Babitt JL (2009). BMP6 is a key endogenous regulator of hepcidin expression and iron metabolism. Nat Genet 41, 482-487.

Arredondo M, Mendiburo MJ, Flores S, Singleton ST, & Garrick MD (2014). Mouse divalent metal transporter 1 is a copper transporter in HEK293 cells. Biometals 27, 115-123.

Arredondo M, Muñoz P, Mura CV, & Núñez MT (2003). DMT1, a physiologically relevant apical Cu1+ transporter of intestinal cells. Am J Physiol Cell Physiol 284, C1525-C1530.

Attaphitaya S, Park K, & Melvin JE (1999). Molecular cloning and functional expression of a rat Na+/H+ exchanger (NHE5) highly expressed in brain. J Biol Chem 274, 4383-4388.

Au C, Benedetto A, & Aschner M (2008). Manganese transport in eukaryotes: The role of DMT1. Neurotoxicology 29, 569-576.

Bannon DI, Abounader R, Lees PS, & Bressler JP (2003). Effect of DMT1 knockdown on iron, cadmium, and lead uptake in Caco-2 cells. Am J Physiol Cell Physiol 284, C44-C50.

Bardou-Jacquet E, Island ML, Jouanolle AM, Détivaud L, Fatih N, Ropert M, Brissot E, Mosser A, Maisonneuve H, Brissot P, & Loréal O (2011). A novel N491S mutation in the human SLC11A2 gene impairs protein trafficking and in association with the G212V mutation leads to microcytic anemia and liver iron overload. Blood Cells Mol Dis 47, 243-248.

Barrientos A, Barros MH, Valnot I, Rotig A, Rustin P, & Tzagoloff A (2002). Cytochrome oxidase in health and disease. Gene 286, 53-63.

Beard JL (2008). Why iron deficiency is important in infant development. J Nutr 138, 2534-2536.

Beaumont C, Delaunay J, Hetet G, Grandchamp B, de Montalembert M, & Tchernia G (2006). Two new human DMT1 gene mutations in a patient with microcytic anemia, low ferritinemia, and liver iron overload. Blood 107, 4168-4170.

Berglund M, Åkesson A, Nermell B, & Vahter M (1994). Intestinal absorption of dietary cadmium in women depends on body iron stores and fiber intake. Environ Health Perspect 102, 1058-1066.

Bernard C, Hoff HE, Guillemin R, & Guillemin L (1974). Second Lecture: The Three Forms of Life. In Lectures On The Phenomena Of Life Common To Animals And Plants pp. 46-91. Charles C. Thomas Publisher, Springfield, IL.

Bessis M & Weed RI (1973). Chapter 2: The Erythrocytic Series. In Living Blood Cells And Their Ultrastructure pp. 85-284. Springer-Verlag New York.

106

Beutler E (2010). Disorders of iron metabolism. In Williams Hematology, eds. Kaushansky K, Beutler E, Seligsohn U, Lichtman MA, Kipps TJ, & Prchal JT, pp. 565-606. McGraw-Hill Medical, New York, NY.

Blanco E, Kannengiesser C, Grandchamp B, Tasso M, & Beaumont C (2009). Not all DMT1 mutations lead to iron overload. Blood Cells Mol Dis 43, 199-201.

Blasco H, Vourc'h P, Nadjar Y, Ribourtout B, Gordon PH, Guettard YO, Camu W, Praline J, Meininger V, Andres CR, Corcia P, & French ALS Study Group (2011). Association between divalent metal transport 1 encoding gene (SLC11A2) and disease duration in amyotrophic lateral sclerosis. J Neurol Sci 303, 124-127.

Blom N, Sicheritz-Pontén T, Gupta R, Gammeltoft S, & Brunak S (2004). Prediction of post-translational glycosylation and phosphorylation of proteins from the amino acid sequence. Proteomics 4, 1633-1649.

Bookstein C, DePaoli AM, Xie Y, Niu P, Musch MW, Rao MC, & Chang EB (1994). Na+/H+ exchangers, NHE- 1 and NHE-3, of rat intestine: Expression and localization. J Clin Invest 93, 106-113.

Bookstein C, Xie Y, Rabenau K, Musch MW, McSwine RL, Rao MC, & Chang EB (1997). Tissue distribution of Na+/H+ exchanger isoforms NHE2 and NHE4 in rat intestine and kidney. Am J Physiol Cell Physiol 273, C1496-C1505.

Bradbury MW (1997). Transport of iron in the blood-brain-cerebrospinal fluid system. J Neurochem 69, 443-454.

Bradford EM, Sartor MA, Gawenis LR, Clarke LL, & Shull GE (2009). Reduced NHE3-mediated Na+ absorption increases survival and decreases the incidence of intestinal obstructions in cystic fibrosis mice. Am J Physiol Gastrointest Liver Physiol 296, G886-G898.

Brasse-Lagnel C, Karim Z, Letteron P, Bekri S, Bado A, & Beaumont C (2011). Intestinal DMT1 cotransporter is down-regulated by hepcidin via proteasome internalization and degradation. Gastroenterology 140, 1261-1271.

Bressler JP, Olivi L, Cheong JH, Kim Y, & Bannon D (2004). Divalent metal transporter 1 in lead and cadmium transport. Ann N Y Acad Sci 1012, 142-152.

Brown JX, Buckett PD, & Wessling-Resnick M (2004). Identification of small molecule inhibitors that distinguish between non-transferrin bound iron uptake and transferrin-mediated iron transport. Chem Biol 11, 407-416.

Buchanan GR & Sheehan RG (1981). Malabsorption and defective utilization of iron in three siblings. J Pediatr 98, 723-728.

Buckett PD & Wessling-Resnick M (2009). Small molecule inhibitors of divalent metal transporter-1. Am J Physiol Gastrointest Liver Physiol 296, G798-G804.

107

Burdo JR & Connor JR (2002). Iron transport in the central nervous system. In Molecular and Cellular Iron Transport, ed. Templeton DM, pp. 487-508. Marcel-Dekker, New York.

Burdo JR, Menzies SL, Simpson IA, Garrick LM, Garrick MD, Dolan KG, Haile DJ, Beard JL, & Connor JR (2001). Distribution of divalent metal transporter 1 and metal transport protein 1 in the normal and Belgrade rat. J Neurosci Res 66, 1198-1207.

Burdo JR, Simpson IA, Menzies S, Beard J, & Connor JR (2004). Regulation of the profile of iron- management proteins in brain microvasculature. J Cereb Blood Flow Metab 24, 67-74.

Byrnes V, Barrett S, Ryan E, Kelleher T, O'Keane C, Coughlan B, & Crowe J (2002). Increased duodenal DMT- 1 expression and unchanged HFE mRNA levels in HFE-associated hereditary hemochromatosis and iron deficiency. Blood Cells Mol Dis 29, 251-260.

Cadieux JA, Zhang Z, Mattice M, Brownlie-Cutts A, Fu J, Ratkay LG, Kwan R, Thompson J, Sanghara J, Zhong J, & Goldberg YP (2012). Synthesis and biological evaluation of substituted pyrazoles as blockers of divalent metal transporter 1 (DMT1). Bioorg Med Chem Lett 22, 90-95.

Canonne-Hergaux F, Calafet J, Richer E, Cellier M, Grinstein S, Borregaard N, & Gros P (2002). Expression and subcellular localization of NRAMP1 in human neutrophil granules. Blood 100, 268-275.

Canonne-Hergaux F, Gruenheid S, Ponka P, & Gros P (1999). Cellular and subcellular localization of the Nramp2 iron transporter in the intestinal brush border and regulation by dietary iron. Blood 93, 4406- 4417.

Cellier MFM, Bergevin I, Boyer E, & Richer E (2001). Polyphyletic origins of bacterial Nramp transporters. Trends Genet 17, 365-370.

Cesta MF (2006). Normal structure, function, and histology of the spleen. Toxicol Pathol 34, 455-465.

Chaloupka R, Courville P, Veyrier F, Knudsen B, Tompkins TA, & Cellier MFM (2005). Identification of functional amino acids in the Nramp family by a combination of evolutionary analysis and biophysical studies of metal and proton cotransport in vivo. Biochemistry 44, 726-733.

Cherukuri S, Potla R, Sarkar J, Nurko S, Harris ZL, & Fox PL (2005). Unexpected role of ceruloplasmin in intestinal iron absorption. Cell Metab 2, 309-319.

Chua ACG & Morgan EH (1997). Manganese metabolism is impaired in the Belgrade laboratory rat. J Comp Physiol [B] 167, 361-369.

Cohen A, Nevo Y, & Nelson N (2003). The first external loop of the metal ion transporter DCT1 is involved in metal ion binding and specificity. Proc Natl Acad Sci USA 100, 10694-10699.

108

Collard KJ (2009). Iron homeostasis in the neonate. Pediatrics 123, 1208-1216.

Collins JF (2006). Gene chip analyses reveal differential genetic responses to iron deficiency in rat duodenum and jejunum. Biol Res 39, 25-37.

Collins JF, Franck CA, Kowdley KV, & Ghishan FK (2005). Identification of differentially expressed genes in response to dietary iron deprivation in rat duodenum. Am J Physiol Gastrointest Liver Physiol 288, G964- G971.

Collins JF, Xu H, Kiela PR, Zeng J, & Ghishan FK (1997). Functional and molecular characterization of NHE3 expression during ontogeny in rat jejunal epithelium. Am J Physiol Cell Physiol 273, C1937-C1946.

Conrad ME, Benjamin BI, Williams HL, & Foy AL (1967). Human absorption of hemoglobin-iron. Gastroenterology 53, 5-10.

Conrad ME, Umbreit JN, Moore EG, Hainsworth LN, Porubcin M, Simovich MJ, Nakada MT, Dolan K, & Garrick MD (2000). Separate pathways for cellular uptake of ferric and ferrous iron. Am J Physiol Gastrointest Liver Physiol 279, G767-G774.

Courville P, Urbankova E, Rensing C, Chaloupka R, Quick M, & Cellier MFM (2008). Solute carrier 11 cation symport requires distinct residues in transmembrane helices 1 and 6. J Biol Chem 283, 9651-9658.

Cousins RJ (2010). Gastrointestinal factors influencing zinc absorption and homeostasis. Int J Vitam Nutr Res 80, 243-248.

Curran-Everett D (2000). Multiple comparisons: Philosophies and illustrations. Am J Physiol Regul Integr Comp Physiol 279, R1-R8.

Czachorowski M, Lam-Yuk-Tseung S, Cellier M, & Gros P (2009). Transmembrane topology of the mammalian Slc11a2 iron transporter. Biochemistry 48, 8422-8434.

Darshan D, Frazer DM, & Anderson GJ (2010). Molecular basis of iron-loading disorders. Expert Rev Mol Med 12, e36.

Davidsson L (2003). Approaches to improve iron bioavailability from complementary foods. J Nutr 133, 1560S-1562S.

Donowitz M, Ming Tse C, & Fuster D (2013). SLC9/NHE gene family, a plasma membrane and organellar family of Na+/H+ exchangers. Mol Aspects Med 34, 236-251.

Dostalikova-Cimburova M, Kratka K, Balusikova K, Chmelikova J, Hejda V, Hnanicek J, Neubauerova J, Vranova J, Kovar J, & Horak J (2011). Duodenal expression of iron transport molecules in patients with hereditary hemochromatosis or iron deficiency. J Cell Mol Med 16, 1816-1826.

109

Dougherty DA (1996). Cation-π interactions in chemistry and biology: a new view of benzene, Phe, Tyr, and Trp. Science 271, 163-168.

Dougherty DA (2007). Cation-pi interactions involving aromatic amino acids. J Nutr 137, 1504S-1508S.

Du X, She E, Gelbart T, Truksa J, Lee P, Xia Y, Khovananth K, Mudd S, Mann N, Moresco EM, Beutler E, & Beutler B (2008). The serine protease TMPRSS6 is required to sense iron deficiency. Science 320, 1088- 1092.

Dudeja PK, Rao DD, Syed I, Joshi V, Dahdal RY, Gardner C, Risk MC, Schmidt L, Bavishi D, Kim KE, Harig JM, Goldstein JL, Layden TJ, & Ramaswamy K (1996). Intestinal distribution of human Na+/H+ exchanger isoforms NHE-1, NHE-2, and NHE-3 mRNA. Am J Physiol Gastrointest Liver Physiol 271, G483-G493.

Dufner-Beattie J, Weaver BP, Geiser J, Bilgen M, Larson M, Xu W, & Andrews GK (2007). The mouse acrodermatitis enteropathica gene Slc39a4 (Zip4) is essential for early development and heterozygosity causes hypersensitivity to zinc deficiency. Hum Mol Genet 16, 1391-1399.

Edwards JA & Hoke JE (1972). Defect of intestinal mucosal iron uptake in mice with hereditary microcytic anemia. Proc Soc Exp Biol Med 141, 81-84.

Ehrnstorfer IA, Geertsma ER, Pardon E, Steyaert J, & Dutzler R (2014). Crystal structure of a SLC11 (NRAMP) transporter reveals the basis for transition-metal ion transport. Nat Struct Mol Biol 21, 990-996. el Marjou F, Janssen K-P, Chang BH-J, Li M, Hindie V, Chan L, Louvard D, Chambon P, Metzger D, & Robine S (2004). Tissue-specific and inducible Cre-mediated recombination in the gut epithelium. Genesis 39, 186- 193.

Engevik MA, Engevik KA, Yacyshyn MB, Wang J, Hassett DJ, Darien B, Yacyshyn BR, & Worrell RT (2015). Human Clostridium difficile infection: inhibition of NHE3 and microbiota profile. Am J Physiol Gastrointest Liver Physiol 308, G497-G509.

Espinoza A, Le BS, Olivares M, Pizarro F, Ruz M, & Arredondo M (2012). Iron, copper, and zinc transport: Inhibition of divalent metal transporter 1 (DMT1) and human copper transporter 1 (hCTR1) by shRNA. Biol Trace Elem Res 146, 281-286.

Everson I & Ralph R (1968). Blood anayses of some antarctic fish. Bull Brit Antarct Surv.

Faham S, Watanabe A, Besserer GM, Cascio D, Specht A, Hirayama BA, Wright EM, & Abramson J (2008). The crystal structure of a sodium galactose transporter reveals mechanistic insights into Na+/sugar symport. Science 321, 810-814.

Fairweather-Tait SJ, Wawer AA, Gillings R, Jennings A, & Myint PK (2014). Iron status in the elderly. Mech Ageing Dev 136-137, 22-28.

110

Fei YJ, Kanai Y, Nussberger S, Ganapathy V, Leibach FH, Romero MF, Singh SK, Boron WF, & Hediger MA (1994). Expression cloning of a mammalian proton-coupled oligopeptide transporter. Nature 368, 563- 566.

Flanagan PR, McLellan JS, Haist J, Cherian G, Chamberlain MJ, & Valberg LS (1978). Increased dietary cadmium absorption in mice and human subjects with iron deficiency. Gastroenterology 74, 841-846.

Fleming MD, Romano MA, Su MA, Garrick LM, Garrick MD, & Andrews NC (1998). Nramp2 is mutated in the anemic Belgrade (b) rat: Evidence of a role for Nramp2 in endosomal iron transport. Proc Natl Acad Sci USA 95, 1148-1153.

Fleming MD, Trenor CC, Su MA, Foernzler D, Beier DR, Dietrich WF, & Andrews NC (1997). Microcytic anaemia mice have a mutation in Nramp2, a candidate iron transporter gene. Nat Genet 16, 383-386.

Fleming RE, Holden CC, Tomatsu S, Waheed A, Brunt EM, Britton RS, Bacon BR, Roopenian DC, & Sly WS (2001). Mouse strain differences determine severity of iron accumulation in Hfe knockout model of hereditary hemochromatosis. Proc Natl Acad Sci USA 98, 2707-2711.

Fleming RE, Migas MC, Zhou X, Jiang J, Britton RS, Brunt EM, Tomatsu S, Waheed A, Bacon BR, & Sly WS (1999). Mechanism of increased iron absorption in murine model of hereditary hemochromatosis: Increased duodenal expression of the iron transporter DMT1. Proc Natl Acad Sci USA 96, 3143-3148.

Fleming RE & Ponka P (2012). Iron overload in human disease. N Eng J Med 366, 348-359.

Folgueras AR, de Lara FM, Pendas AM, Garabaya C, Rodriguez F, Astudillo A, Bernal T, Cabanillas R, Lopez- Otin C, & Velasco G (2008). Membrane-bound serine protease matriptase-2 (Tmprss6) is an essential regulator of iron homeostasis. Blood 112, 2539-2545.

Foot NJ, Dalton HE, Shearwin-Whyatt LM, Dorstyn L, Tan SS, Yang B, & Kumar S (2008). Regulation of the divalent metal ion transporter DMT1 and iron homeostasis by a ubiquitin-dependent mechanism involving Ndfips and WWP2. Blood 112, 4268-4275.

Foot NJ, Leong YA, Dorstyn LE, Dalton HE, Ho K, Zhao L, Garrick MD, Yang B, Hiwase D, & Kumar S (2011). Ndfip1-deficient mice have impaired DMT1 regulation and iron homeostasis. Blood 117, 638-646.

Forbes JR & Gros P (2003). Iron, manganese, and cobalt transport by Nramp1 (Slc11a1) and Nramp2 (Slc11a2) expressed at the plasma membrane. Blood 102, 1884-1892.

Franich NR, Fitzsimons HL, Fong DM, Klugmann M, During MJ, & Young D (2008). AAV vector-mediated RNAi of mutant huntingtin expression is neuroprotective in a novel genetic rat model of Huntington's disease. Mol Ther 16, 947-956.

111

Frey AG, Nandal A, Park JH, Smith PM, Yabe T, Ryu MS, Ghosh MC, Lee J, Rouault TA, Park MH, & Philpott CC (2014). Iron chaperones PCBP1 and PCBP2 mediate the metallation of the dinuclear iron enzyme deoxyhypusine hydroxylase. Proc Natl Acad Sci USA 111, 8031-8036.

Ganz T (2013). Systemic iron homeostasis. Physiol Rev 93, 1721-1741.

Ganz T & Nemeth E (2006). Iron Imports. IV. Hepcidin and regulation of body iron metabolism. Am J Physiol Gastrointest Liver Physiol 290, G199-G203.

Gao J, Chen J, De Domenico I, Koeller DM, Harding CO, Fleming RE, Koeberl DD, & Enns CA (2010). Hepatocyte-targeted HFE and TFR2 control hepcidin expression in mice. Blood 115, 3374-3381.

Garrick LM, Dolan KG, Romano MA, & Garrick MD (1999). Non-transferrin-bound iron uptake in Belgrade and normal rat erythroid cells. J Cell Physiol 178, 349-358.

Garrick MD & Dolan KG (2002). An expression system for a transporter of iron and other metals. In Oxidants and Antioxidants: Ultrastructure and Molecular Biology Protocols, ed. Armstrong D, pp. 147-154. Humana Press, Totowa, NJ.

Garrick MD & Garrick LM (2009). Cellular iron transport. Biochim Biophys Acta 1790, 309-325.

Garrick MD, Kuo HC, Vargas F, Singleton S, Zhao L, Smith JJ, Paradkar P, Roth JA, & Garrick LM (2006). Comparison of mammalian cell lines expressing distinct isoforms of divalent metal transporter 1 in a tetracycline-regulated fashion. Biochem J 398, 539-546.

Garrick MD, Zhao L, Roth JA, Jiang H, Feng J, Foot NJ, Dalton H, Kumar S, & Garrick LM (2012). Isoform specific regulation of divalent metal (ion) transporter (DMT1) by proteasomal degradation. Biometals 25, 787-793.

Garry DJ, Ordway GA, Lorenz JN, Radford NB, Chin ER, Grange RW, Bassel-Duby R, & Williams RS (1998). Mice without myoglobin. Nature 395, 905-908.

Gawenis LR, Stien X, Shull GE, Schultheis PJ, Woo AL, Walker NM, & Clarke LL (2002). Intestinal NaCl transport in NHE2 and NHE3 knockout mice. Am J Physiol Gastrointest Liver Physiol 282, G776-G784.

Gill RK, Saksena S, Syed IA, Tyagi S, Alrefai WA, Malakooti J, Ramaswamy K, & Dudeja PK (2002). Regulation of NHE3 by nitric oxide in Caco-2 cells. Am J Physiol Gastrointest Liver Physiol 283, G747-G756.

Golubov J, Flanagan P, & Adams P (1991). Inhibition of iron absorption by omeprazole in rat model. Dig Dis Sci 36, 405-408.

Goswami T & Andrews NC (2006). Hereditary hemochromatosis protein, HFE, interaction with transferrin receptor 2 suggests a molecular mechanism for mammalian iron sensing. J Biol Chem 281, 28494-28498.

112

Goyer RA (1997). Toxic and essential metal interactions. Ann Rev Nutr 17, 37-50.

Gröne HJ, Weber K, Helmchen U, & Osborn M (1986). Villin--a marker of brush border differentiation and cellular origin in human renal cell carcinoma. Am J Pathol 124, 294-302.

Gruenheid S, Canonne-Hergaux F, Gauthier S, Hackam DJ, Grinstein S, & Gros P (1999). The iron transport protein NRAMP2 is an integral membrane glycoprotein that colocalizes with transferrin in recycling endosomes. J Exp Med 189, 831-841.

Guan Y, Dong J, Tackett L, Meyer JW, Shull GE, & Montrose MH (2006). NHE2 is the main apical NHE in mouse colonic crypts but an alternative Na+-dependent acid extrusion mechanism is upregulated in NHE2- null mice. Am J Physiol Gastrointest Liver Physiol 291, G689-G699.

Gunshin H, Allerson CR, Polycarpou-Schwarz M, Rofts A, Rogers JT, Kishi F, Hentze MW, Rouault TA, Andrews NC, & Hediger MA (2001). Iron-dependent regulation of the divalent metal ion transporter. FEBS Lett 509, 309-316.

Gunshin H, Fujiwara Y, Custodio ÁO, Direnzo C, Robine S, & Andrews NC (2005a). Slc11a2 is required for intestinal iron absorption and erythropoiesis but dispensable in placenta and liver. J Clin Invest 115, 1258- 1266.

Gunshin H, Jin J, Fujiwara Y, Andrews NC, Mims M, & Prchal J (2005b). Analysis of the E399D mutation in SLC11A2. Blood 106, 2221-2222.

Gunshin H, Mackenzie B, Berger UV, Gunshin Y, Romero MF, Boron WF, Nussberger S, Gollan JL, & Hediger MA (1997). Cloning and characterization of a mammalian proton-coupled metal-ion transporter. Nature 388, 482-488.

Gunshin H, Starr CN, Direnzo C, Fleming MD, Jin J, Greer EL, Sellers VM, Galica SM, & Andrews NC (2005c). Cybrd1 (duodenal cytochrome b) is not necessary for dietary iron absorption in mice. Blood 106, 2879- 2883.

Guthrie GJ, Aydemir TB, Troche C, Martin AB, Chang SM, & Cousins RJ (2015). Influence of ZIP14 (slc39A14) on intestinal zinc processing and barrier function. Am J Physiol Gastrointest Liver Physiol 308, G171-G178.

Harding C, Heuser J, & Stahl P (1983). Receptor-mediated endocytosis of transferrin and recycling of the transferrin receptor in rat reticulocytes. J Cell Biol 97, 329-339.

Harrison PM & Arosio P (1996). The ferritins: molecular properties, iron storage function and cellular regulation. Biochim Biophys Acta 1275, 161-203.

Harrison SA & Bacon BR (2003). Hereditary hemochromatosis: Update for 2003. J Hepatol 38, S14-S23.

113

Hartman KR & Barker JA (1996). Microcytic anemia with iron malabsorption: An inherited disorder of iron metabolism. Am J Hematol 51, 269-275.

Hayashi H, Szászi K, Coady-Osberg N, Furuya W, Bretscher AP, Orlowski J, & Grinstein S (2004). Inhibition and redistribution of NHE3, the apical Na+/H+ exchanger, by Clostridium difficile toxin B. J Gen Physiol 123, 491-504.

Hazama A, Loo DDF, & Wright EM (1997). Presteady-state currents of the rabbit Na+/glucose cotransporter (SGLT1). J Membrane Biol 155, 175-186.

Herrmann T, Muckenthaler M, Van Der HF, Brennan K, Gehrke SG, Hubert N, Sergi C, Grone HJ, Kaiser I, Gosch I, Volkmann M, Riedel HD, Hentze MW, Stewart AF, & Stremmel W (2004). Iron overload in adult Hfe-deficient mice independent of changes in the steady-state expression of the duodenal iron transporters DMT1 and Ireg1/ferroportin. J Mol Med 82, 39-48.

Hershko C (2010). Pathogenesis and management of iron toxicity in thalassemia. Ann N Y Acad Sci 1202, 1-9.

Hoogerwerf WA, Tsao SC, Devuyst O, Levine SA, Yun CHC, Yip JW, Cohen ME, Wilson PD, Lazenby AJ, Tse C-M, & Donowitz M (1996). NHE2 and NHE3 are human and rabbit intestinal brush-border proteins. Am J Physiol 270, G29-G41.

Horonchik L & Wessling-Resnick M (2008). The small-molecule iron transport inhibitor ferristatin/NSC306711 promotes degradation of the transferrin receptor. Chem Biol 15, 647-653.

Huang E, Ong W-Y, Go M-L, & Connor JR (2006). Upregulation of iron regulatory proteins and divalent metal transporter-1 isoforms in the rat hippocampus after kainate induced neuronal injury. Exp Brain Res 170, 376-386.

Hubert N & Hentze MW (2002). Previously uncharacterized isoforms of divalent metal transporter (DMT)- 1: Implications for regulation and cellular function. Proc Natl Acad Sci USA 99, 12345-12350.

Ikuma M, Hanai H, Kaneko E, Hayashi H, & Hoshi T (1996). Effects of aging on the microclimate pH of the rat jejunum. Biochim Biophys Acta 1280, 19-26.

Illing AC, Shawki A, Cunningham CL, & Mackenzie B (2012). Substrate profile and metal-ion selectivity of human divalent metal-ion transporter-1. J Biol Chem 287, 30485-30496.

Iolascon A, d'Apolito M, Servedio V, Cimmino F, Piga A, & Camaschella C (2006). Microcytic anemia and hepatic iron overload in a child with compound heterozygous mutations in DMT1. Blood 107, 349-354.

Jamieson SE, White JK, Howson JM, Pask R, Smith AN, Brayne C, Evans JG, Xuereb J, Cairns NJ, Rubinsztein DC, & Blackwell JM (2005). Candidate gene association study of solute carrier family 11a members 1 (SLC11A1) and 2 (SLC11A2) genes in Alzheimer's disease. Neurosci Lett 374, 124-128.

114

Jeong J & Eide DJ (2013). The SLC39 family of zinc transporters. Mol Aspects Med 34, 612-619.

Jiang L, Garrick MD, Garrick LM, Zhao L, & Collins JF (2013). Divalent metal transporter 1 (Dmt1) mediates copper transport in the duodenum of iron-deficient rats and when overexpressed in iron-deprived HEK- 293 cells. J Nutr 143, 1927-1933.

Johnson-Wimbley TD & Graham DY (2011). Diagnosis and management of iron deficiency anemia in the 21st century. Therap Adv Gastroenterol 4, 177-184.

Kelleher T, Ryan E, Barrett S, Sweeney M, Byrnes V, O'Keane C, & Crowe J (2004). Increased DMT1 but not IREG1 or HFE mRNA following iron depletion therapy in hereditary haemochromatosis. Gut 53, 1174-1179.

Kiela PR, Laubitz D, Larmonier CB, Midura-Kiela MT, Lipko MA, Janikashvili N, Bai A, Thurston R, & Ghishan FK (2009). Changes in mucosal homeostasis predispose NHE3 knockout mice to increased susceptibility to DSS-induced epithelial injury. Gastroenterology 137, 965-75, 975.

Killip S, Bennett JM, & Chambers MD (2007). Iron deficiency anemia. Am Fam Physician 75, 671-678.

Kippler M, Goessler W, Nermell B, Ekström EC, Lönnerdal B, El Arifeen S, & Vahter M (2009). Factors influencing intestinal cadmium uptake in pregnant Bangladeshi women--a prospective cohort study. Environ Res 109, 914-921.

Knöpfel M, Smith C, & Solioz M (2005a). ATP-driven copper transport across the intestinal brush border membrane. Biochem Biophys Res Comm 330, 645-652.

Knöpfel M, Zhao L, & Garrick MD (2005b). Transport of divalent transition-metal ions is lost in small- intestinal tissue of b/b Belgrade rats. Biochemistry 44, 3454-3465.

Knutson M, Menzies S, Connor J, & Wessling-Resnick M (2004). Developmental, regional, and cellular expression of SFT/UbcH5A and DMT1 mRNA in brain. J Neurosci Res 76, 633-641.

Knutson MD (2007). Steap proteins: Implications for iron and copper metabolism. Nutr Rev 65, 335-340.

Kozyraki R, Fyfe J, Verroust PJ, Jacobsen C, Dautry-Varsat A, Gburek J, Willnow TE, Christensen EI, & Moestrup SK (2001). Megalin-dependent cubilin-mediated endocytosis is a major pathway for the apical uptake of transferrin in polarized epithelia. Proc Natl Acad Sci USA 98, 12491-12496.

Kwik-Uribe CL, Reaney S, Zhu Z, & Smith D (2003). Alterations in cellular IRP-dependent iron regulation by in vitro manganese exposure in undifferentiated PC12 cells. Brain Res 973, 1-15.

Lam-Yuk-Tseung S, Camaschella C, Iolascon A, & Gros P (2006). A novel R416C mutation in human DMT1 (SLC11A2) displays pleiotropic effects on function and causes microcytic anemia and hepatic iron overload. Blood Cells Mol Dis 36, 347-354.

115

Lam-Yuk-Tseung S, Govoni G, & Gros P (2003). Iron transport by NRAMP2/DMT1: pH regulation of transport by two histidines in transmembrane domain 6. Blood 101, 3699-3707.

Lam-Yuk-Tseung S, Mathieu M, & Gros P (2005). Functional characterization of the E399D DMT1/NRAMP2/SLC11A2 protein produced by an exon 12 mutation in a patient with microcytic anemia and iron overload. Blood Cells Mol Dis 35, 212-216.

Latunde-Dada GO, Van der WJ, Vulpe CD, Anderson GJ, Simpson RJ, & McKie AT (2002). Molecular and functional roles of duodenal cytochrome B (dcytb) in iron metabolism. Blood Cells Mol Dis 29, 356-360.

Ledoussal C, Woo AL, Miller ML, & Shull GE (2001). Loss of the NHE2 Na+/H+ exchanger has no apparent effect on diarrheal state of NHE3-deficient mice. Am J Physiol Gastrointest Liver Physiol 281, G1385- G1396.

Lee J, Peña MM, Nose Y, & Thiele DJ (2002a). Biochemical characterization of the human copper transporter Ctr1. J Biol Chem 277, 4380-4387.

Lee J, Petris MJ, & Thiele DJ (2002b). Characterization of mouse embryonic cells deficient in the Ctr1 high affinity copper transporter: Identification of a Ctr1-independent copper transport system. J Biol Chem 277, 40253-40259.

Lee P, Peng H, Gelbart T, Wang L, & Beutler E (2005). Regulation of hepcidin transcription by interleukin- 1 and interleukin-6. Proc Natl Acad Sci USA 102, 1906-1910.

Lee PL, Gelbart T, West C, Halloran C, & Beutler E (1998). The human Nramp2 gene: characterization of the gene structure, alternative splicing, promoter region and polymorphisms. Blood Cells Mol Dis 24, 199- 215.

Lill R, Hoffmann B, Molik S, Pierik AJ, Rietzschel N, Stehling O, Uzarska MA, Webert H, Wilbrecht C, & Mühlenhoff U (2012). The role of mitochondria in cellular iron-sulfur protein biogenesis and iron metabolism. Biochim Biophys Acta 1823, 1491-1508.

Lin C, Zhang Z, Wang T, Chen C, & James Kang Y (2015). Copper uptake by DMT1: A compensatory mechanism for CTR1 deficiency in human umbilical vein endothelial cells. Metallomics 7, 1285-1289.

Linder MC, Moriya M, Whon A, Kassa A, & Gilley C (2006). Vesicular transport of Fe and interaction with other metal ions in polarized Caco2 cell monolayers. Biol Res 39, 143-156.

Lis A, Barone TA, Paradkar PN, Plunkett RJ, & Roth JA (2004). Expression and localization of different forms of DMT1 in normal and tumor astroglial cells. Brain Res Mol Brain Res 122, 62-70.

Lis A, Paradkar PN, Singleton S, Kuo HC, Garrick MD, & Roth JA (2005). Hypoxia induces changes in expression of isoforms of the divalent metal transporter (DMT1) in rat pheochromocytoma (PC12) cells. Biochem Pharmacol 69, 1647-1655.

116

Liuzzi JP, Aydemir F, Nam H, Knutson MD, & Cousins RJ (2006). Zip14 (Slc39a14) mediates non-transferrin- bound iron uptake into cells. Proc Natl Acad Sci USA 103, 13612-13617.

Livak KJ & Schmittgen TD (2001). Analysis of relative gene expression data using real-time quantitative PCR and the 2–ΔΔCT method. Methods 25, 402-408.

Loo DDF, Hazama A, Supplisson S, Turk E, & Wright EM (1993). Relaxation kinetics of the Na+/glucose cotransporter. Proc Natl Acad Sci USA 90, 5767-5771.

Lucas M (1983). Determination of acid surface pH in vivo in rat proximal jejunum. Gut 24, 734-739.

Lundblad RL (2003). Chemical Reagents for Protein Modification, 3 ed. CRC Press, Boca Raton.

Mackenzie B (1999). Selected techniques in membrane transport. In Biomembrane Transport, ed. Van Winkle LJ, pp. 327-342. Academic Press, San Diego, CA.

Mackenzie B & Garrick MD (2005). Iron Imports. II. Iron uptake at the apical membrane in the intestine. Am J Physiol Gastrointest Liver Physiol 289, G981-G986.

Mackenzie B & Hediger MA (2004). SLC11 family of H+-coupled metal-ion transporters NRAMP1 and DMT1. Pflügers Arch Eur J Physiol 447, 571-579.

Mackenzie B, Illing AC, & Hediger MA (2008). Transport model of the human Na+-coupled L-ascorbic acid (vitamin C) transporter SVCT1. Am J Physiol Cell Physiol 294, C451-C459.

Mackenzie B, Loo DDF, Fei YJ, Liu W, Ganapathy V, Leibach FH, & Wright EM (1996a). Mechanisms of the human intestinal H+-coupled oligopeptide transporter hPEPT1. J Biol Chem 271, 5430-5437.

Mackenzie B, Loo DDF, Panayotova-Heiermann M, & Wright EM (1996b). Biophysical characteristics of the pig kidney Na+/glucose cotransporter SGLT2 reveal a common mechanism for SGLT1 and SGLT2. J Biol Chem 271, 32678-32683.

Mackenzie B, Shawki A, Ghio AJ, Stonehuerner JD, Zhao L, Ghadersohi S, Garrick LM, & Garrick MD (2010). Calcium-channel blockers do not affect iron transport mediated by divalent metal-ion transporter-1. Blood 115, 4148-4149.

Mackenzie B, Takanaga H, Hubert N, Rolfs A, & Hediger MA (2007). Functional properties of multiple isoforms of human divalent metal-ion transporter 1 (DMT1). Biochem J 403, 59-69.

Mackenzie B, Ujwal ML, Chang M-H, Romero MF, & Hediger MA (2006). Divalent metal-ion transporter DMT1 mediates both H+-coupled Fe2+ transport and uncoupled fluxes. Pflügers Arch Eur J Physiol 451, 544- 558.

117

Malakooti J, Dahdal RY, Schmidt L, Layden TJ, Dudeja PK, & Ramaswamy K (1999). Molecular cloning, tissue distribution, and functional expression of the human Na+/H+ exchanger NHE2. Am J Physiol 277, G383- G390.

Mastrogiannaki M, Matak P, Keith B, Simon MC, Vaulont S, & Peyssonnaux C (2009). HIF-2α, but not HIF- 1α, promotes iron absorption in mice. J Clin Invest 119, 1159-1166.

McEwan GTA, Daniel H, Fett C, Burgess MN, & Lucas ML (1988). The effect of Escherichia coli STa enterotoxin and other secretagogues on mucosal surface pH of rat small intestine in vivo. Proc R Soc Lond B Biol Sci 234, 219-237.

McEwan GTA, Lucas ML, & Mathan VI (1990). A combined TDDA-PVC pH and reference electrode for use in the upper small intestine. J Med Eng Tech 14, 16-20.

McKie AT, Barrow D, Latunde-Dada GO, Rolfs A, Sager G, Mudaly E, Mudaly M, Richardson C, Barlow D, Bomford A, Peters TJ, Raja KB, Shirali S, Hediger MA, Farzaneh F, & Simpson RJ (2001). An iron-regulated ferric reductase associated with the absorption of dietary iron. Science 291, 1755-1759.

McKie AT, Marciani P, Rolfs A, Brennan K, Wehr K, Barrow D, Miret S, Bomford A, Peters TJ, Farzaneh F, Hediger MA, Hentze MW, & Simpson RJ (2000). A novel duodenal iron-regulated transporter, IREG1, implicated in the basolateral transfer of iron to the circulation. Mol Cell 5, 299-309.

Means GE & Feeney RE (1971). Chemical Modification of Proteins Holden-Day, San Francisco.

Mena NP, Esparza A, Tapia V, Valdes P, & Núñez MT (2008). Hepcidin inhibits apical iron uptake in intestinal cells. Am J Physiol Gastrointest Liver Physiol 294, G192-G198.

Meynard D, Kautz L, Darnaud V, Canonne-Hergaux F, Coppin H, & Roth MP (2009). Lack of the bone morphogenetic protein BMP6 induces massive iron overload. Nat Genet 41, 478-481.

Miller ML, Andringa A, Schultheis PJ, & Shull GE (2011). Loss of the NHE2 Na+/H+ exchanger in mice results in dilation of folliculo-stellate cell canaliculi. J Biomed Biotechnol 2011, 510827.

Mims MP, Guan Y, Pospisilova D, Priwitzerova M, Indrak K, Ponka P, Divoky V, & Prchal JT (2005). Identification of a human mutation of DMT1 in a patient with microcytic anemia and iron overload. Blood 105, 1337-1342.

Moeser AJ, Nighot PK, Ryan KA, Simpson JE, Clarke LL, & Blikslager AT (2008). Mice lacking the Na+/H+ exchanger 2 have impaired recovery of intestinal barrier function. Am J Physiol Gastrointest Liver Physiol 295, G791-G797.

Moos T & Morgan EH (2004a). The metabolism of neuronal iron and its pathogenic role in neurological disease: review. Annals of the New York Academy of Science 1012:14-26., 14-26.

118

Moos T & Morgan EH (2004b). The significance of the mutated divalent metal transporter (DMT1) on iron transport into the Belgrade rat brain. J Neurochem 88, 233-245.

Moos T, Trinder D, & Morgan EH (2000). Cellular distribution of ferric iron, ferritin, transferrin and divalent metal transporter 1 (DMT1) in substantia nigra and basal ganglia of normal and beta2-microglobulin deficient mouse brain. Cell Mol Biol 46, 549-561.

Moos T & Morgan EH (1998). Evidence for low molecular weight, non-transferrin-bound iron in rat brain and cerebrospinal fluid. J Neurosci Res 54, 486-494.

Munro HN & Linder MC (1978). Ferritin: structure, biosynthesis, and role in iron metabolism. Physiol Rev 58, 317-396.

Nemeth E, Rivera S, Gabayan V, Keller C, Taudorf S, Pedersen BK, & Ganz T (2004a). IL-6 mediates hypoferremia of inflammation by inducing the synthesis of the iron regulatory hormone hepcidin. J Clin Invest 113, 1271-1276.

Nemeth E, Tuttle MS, Powelson J, Vaughn MB, Donovan A, Ward DM, Ganz T, & Kaplan J (2004b). Hepcidin regulates cellular iron efflux by binding to ferroportin and inducing its internalization. Science 306, 2090- 2093.

Nevo Y & Nelson N (2004). The mutation F227I increases the coupling of metal-ion transport in DCT1. J Biol Chem 279, 53056-53061.

Nevo Y & Nelson N (2006). The NRAMP family of metal-ion transporters. Biochim Biophys Acta 1763, 609- 620.

Nose Y, Kim BE, & Thiele DJ (2006). Ctr1 drives intestinal copper absorption and is essential for growth, iron metabolism, and neonatal cardiac function. Cell Metab 4, 235-244.

Nose Y, Wood LK, Kim B-E, Prohaska JR, Fry RS, Spears JW, & Thiele DJ (2010). Ctr1 is an apical copper transporter in mammalian epithelial cells in vivo that is controlled at the level of protein stability. J Biol Chem 285, 32385-32392.

Numata M, Petrecca K, Lake N, & Orlowski J (1998). Identification of a mitochondrial Na+/H+ exchanger. J Biol Chem 273, 6951-6959.

Núñez MT, Urrutia P, Mena N, Aguirre P, Tapia V, & Salazar J (2012). Iron toxicity in neurodegeneration. Biometals 25, 761-776.

Ohgami RS, Campagna DR, Greer EL, Antiochos B, McDonald A, Chen J, Sharp JJ, Fujiwara Y, Barker JE, & Fleming MD (2005). Identification of a ferrireductase required for efficient transferrin-dependent iron uptake in erythroid cells. Nat Genet 37, 1264-1269.

119

Ohgami RS, Campagna DR, McDonald A, & Fleming MD (2006). The Steap proteins are metalloreductases. Blood 108, 1388-1394.

Okazaki Y, Ma Y, Yeh M, Yin H, Li Z, Yeh KY, & Glass J (2012). DMT1 (IRE) expression in intestinal and erythroid cells is regulated by peripheral benzodiazepine receptor-associated protein 7. Am J Physiol Gastrointest Liver Physiol 302, G1180-G1190.

Ordway GA & Garry DJ (2004). Myoglobin: an essential hemoprotein in striated muscle. J Exp Biol 207, 3441-3446.

Orlowski J, Kandasamy RA, & Shull GE (1992). Molecular cloning of putative members of the Na/H exchanger gene family: cDNA cloning, deduced amino acid sequence, and mRNA tissue expression of the rat Na/H exchanger NHE-1 and two structurally related proteins. J Biol Chem 267, 9331-9339.

Pantopoulos K (2004). Iron metabolism and the IRE/IRP regulatory system: an update. Annals of the New York Academy of Science 1012, 1-13.

Papanikolaou G, Tzilianos M, Christakis JI, Bogdanos D, Tsimirika K, Macfarlane J, Goldberg YP, Sakellaropoulos N, Ganz T, & Nemeth E (2005). Hepcidin in iron overload disorders. Blood 105, 4103-4105.

Paradkar PN & Roth JA (2006). Nitric oxide transcriptionally down-regulates specific isoforms of divalent metal transporter (DMT1) via NF-kappaB. Journal of Neurochemistry 96, 1768-1777.

Park CH, Valore EV, Waring AJ, & Ganz T (2001). Hepcidin, a urinary antimicrobial peptide synthesized in the liver. J Biol Chem 276, 7806-7810.

Park JD, Cherrington NJ, & Klaassen CD (2002). Intestinal absorption of cadmium is associated with divalent metal transporter 1 in rats. Toxicol Sci 68, 288-294.

Pelizzoni I, Macco R, Morini MF, Zacchetti D, Grohovaz F, & Codazzi F (2011). Iron handling in hippocampal neurons: Activity-dependent iron entry and mitochondria-mediated neurotoxicity. Aging Cell 10, 172-183.

Pelizzoni I, Zacchetti D, Smith CP, Grohovaz F, & Codazzi F (2012). Expression of divalent metal transporter 1 in primary hippocampal neurons: Reconsidering its role in non-transferrin-bound iron influx. J Neurochem 120, 269-278.

Picard V, Govoni G, Jabado N, & Gros P (2000). Nramp 2 (DCT1/DMT1) expressed at the plasma membrane transports iron and other divalent cations into a calcein-accessible cytoplasmic pool. J Biol Chem 275, 35738-35745.

Pietrangelo A, Dierssen U, Valli L, Garuti C, Rump A, Corradini E, Ernst M, Klein C, & Trautwein C (2007). STAT3 is required for IL-6-gp130-dependent activation of hepcidin in vivo. Gastroenterology 132, 294-300.

120

Pinilla-Tenas JJ, Sparkman BK, Shawki A, Illing AC, Mitchell CJ, Zhao N, Liuzzi JP, Cousins RJ, Knutson MD, & Mackenzie B (2011). Zip14 is a complex broad-scope metal-ion transporter whose functional properties support roles in the cellular uptake of zinc and nontransferrin-bound iron. Am J Physiol Cell Physiol 301, C862-C871.

Prakash S, Shawki A, Niespodzany E, & Mackenzie B. Intestinal divalent metal-ion transporter-1 is required for iron homeostasis in the neonatal mouse. FASEB Journal 29[1 Suppl], 1011.5. 2015. Ref Type: Abstract

Preza GC, Ruchala P, Pinon R, Ramos E, Qiao B, Peralta MA, Sharma S, Waring A, Ganz T, & Nemeth E (2011). Minihepcidins are rationally designed small peptides that mimic hepcidin activity in mice and may be useful for the treatment of iron overload. J Clin Invest 121, 4880-4888.

Priwitzerova M, Pospisilova D, Prchal JT, Indrak K, Hlobilkova A, Mihal V, Ponka P, & Divoky V (2004). Severe hypochromic microcytic anemia caused by a congenital defect of the iron transport pathway in erythroid cells. Blood 103, 3991-3992.

Priwitzerova M, Pospisilova D, Sheftel AD, Divoky V, & Ponka P (2005). Functional consequences of the human DMT1 mutation on protein expression and iron uptake. Blood 106, 3985-3987.

Qiao B, Sugianto P, Fung E, del-Castillo-Rueda A, Moran-Jimenez M-J, Ganz T, & Nemeth E (2012). Hepcidin-induced endocytosis of ferroportin is dependent on ferroportin ubiquitination. Cell Metab 15, 918-924.

Qiu A, Jansen M, Sakaris A, Min SH, Chattopadhyay S, Tsai E, Sandoval C, Zhao R, Akabas MH, & Goldman ID (2006). Identification of an intestinal folate transporter and the molecular basis for hereditary folate malabsorption. Cell 127, 917-928.

Raffin SB, Woo CH, Roost KT, Price DC, & Schmid R (1974). Intestinal absorption of hemoglobin iron-heme cleavage by mucosal heme oxygenase. J Clin Invest 54, 1344-1352.

Rajagopal A, Rao AU, Amigo J, Tian M, Upadhyay SK, Hall C, Uhm S, Mathew MK, Fleming MD, Paw BH, Krause M, & Hamza I (2008). Haem homeostasis is regulated by the conserved and concerted functions of HRG-1 proteins. Nature 453, 1127-1131.

Ramakrishnan U (2002). Prevalence of micronutrient malnutrition worldwide. Nutr Rev 60, S46-S52.

Ramey G, Deschemin JC, Durel B, Canonne-Hergaux F, Nicolas G, & Vaulont S (2010). Hepcidin targets ferroportin for degradation in hepatocytes. Haematologica.

Rapisarda C, Puppi J, Hughes RD, Dhawan A, Farnaud S, Evans RW, & Sharp PA (2010). Transferrin receptor 2 is crucial for iron sensing in human hepatocytes. Am J Physiol Gastrointest Liver Physiol.

121

Repishti M, Hogan DL, Pratha V, Davydova L, Donowitz M, Tse CM, & Isenberg JI (2001). Human duodenal mucosal brush border Na+/H+ exchangers NHE2 and NHE3 alter net bicarbonate movement. Am J Physiol Gastrointest Liver Physiol 281, G159-G163.

Rivera-Mancía S, Ríos C, & Montes S (2011). Manganese accumulation in the CNS and associated pathologies. Biometals 24, 811-825.

Rockey DC (1999). Occult gastrointestinal bleeding. N Engl J Med 341, 38-46.

Rolfs A, Bonkovsky HL, Kohlroser JG, McNeal K, Sharma A, Berger UV, & Hediger MA (2002). Intestinal expression of genes involved in iron absorption in humans. Am J Physiol Gastrointest Liver Physiol 282, G598-G607.

Ross SL, Tran L, Winters A, Lee K-J, Plewa C, Foltz I, King C, Miranda LP, Allen J, Beckman H, Cooke KS, Moody G, Sasu BJ, Nemeth E, Ganz T, Molineux G, & Arvedson TL (2012). Molecular mechanism of hepcidin-mediated ferroportin internalization requires ferroportin lysines, not tyrosines or JAK-STAT. Cell Metab 15, 905-917.

Roth JA (2006). Homeostatic and toxic mechanisms regulating manganese uptake, retention, and elimination. Biol Res 39, 45-57.

Roth JA & Garrick MD (2003). Iron interactions and other biological reactions mediating the physiological and toxic actions of manganese. Biochem Pharmacol 66, 1-13.

Roth JA, Horbinski C, Feng L, Dolan KG, Higgins D, & Garrick MD (2000). Differential localization of divalent metal transporter 1 with and without iron response element in rat PC12 and sympathetic neuronal cells. J Neurosci 20, 7595-7601.

Rouault TA, Zhang DL, & Jeong SY (2009). Brain iron homeostasis, the choroid plexus, and localization of iron transport proteins. Metab Brain Dis 24, 673-684.

Rouault TA & Cooperman S (2006). Brain iron metabolism. Semin Pediatr Neurol 13, 142-148.

Ruud JT (1954). Vertebrates without erythrocytes and blood pigment. Nature 173, 848-850.

Sacher A, Nelson N, Ogi JT, Wright EM, Loo DDF, & Eskandari S (2002). Presteady-state and steady-state kinetics and turnover rate of the mouse γ-aminobutyric acid transporter (mGAT3). J Membrane Biol 190, 57-73.

Said HM, Blair JA, Lucas ML, & Hilburn ME (1986). Intestinal surface acid microclimate in vitro and in vivo in the rat. J Lab Clin Med 107, 420-424.

122

Salazar J, Mena N, Hunot S, Prigent A, Alvarez-Fischer D, Arredondo M, Duyckaerts C, Sazdovitch V, Zhao L, Garrick LM, Núñez MT, Garrick MD, Raisman-Vozari R, & Hirsch EC (2008). Divalent metal transporter 1 (DMT1) contributes to neurodegeneration in animal models of Parkinson's disease. Proc Natl Acad Sci USA 105, 18578-18583.

Saroussi S & Nelson N (2009). Vacuolar H+-ATPase—an enzyme for all seasons. Pflügers Arch Eur J Physiol.

Schultheis PJ, Clarke LL, Meneton P, Harline M, Boivin GP, Stemmermann G, Duffy JJ, Doetschman T, Miller ML, & Shull GE (1998a). Targeted disruption of the murine Na+/H+ exchanger isoform 2 gene causes reduced viability of gastric parietal cells and loss of net acid secretion. J Clin Invest 101, 1243-1253.

Schultheis PJ, Clarke LL, Meneton P, Miller ML, Soleimani M, Gawenis LR, Riddle TM, Duffy JJ, Doetschman T, Wang T, Giebisch G, Aronson PS, Lorenz JN, & Shull GE (1998b). Renal and intestinal absorptive defects in mice lacking the NHE3 Na+/H+ exchanger. Nat Genet 19, 282-285.

Shah YM, Matsubara T, Ito S, & Gonzalez FJ (2009). Intestinal hypoxia-inducible transcription factors are essential for iron absorption following iron deficiency. Cell Metab 9, 152-164.

Shaw GC, Cope JJ, Li L, Corson K, Hersey C, Ackermann GE, Gwynn B, Lambert AJ, Wingert RA, Traver D, Trede NS, Barut BA, Zhou Y, Minet E, Donovan A, Brownlie A, Balzan R, Weiss MJ, Peters LL, Kaplan J, Zon LI, & Paw BH (2006). Mitoferrin is essential for erythroid iron assimilation. Nature 440, 96-100.

Shawki A, Anthony SR, Nose Y, Engevik MA, Niespodzany EJ, Barrientos T, Öhrvik H, Worrell RT, Thiele DJ, & Mackenzie B (2015). Intestinal DMT1 is critical for iron absorption in the mouse but is not required for the absorption of copper or manganese. Am J Physiol Gastrointest Liver Physiol 309, G635-G647.

Shawki A, Illing AC, & Mackenzie B. Molecular impact of human divalent metal-ion transporter DMT1 mutations associated with disease phenotypes. FASEB Journal 20, A1278. 2006. Ref Type: Abstract

Shawki A, Illing AC, & Mackenzie B. Molecular impact of divalent metal-ion transporter (DMT1) mutations (V114del and G212V) found in a compound heterozygote with microcytic anemia and hepatic iron overload. FASEB Journal 22, 1192.3. 2008. Ref Type: Abstract

Shawki A, Knight PB, Maliken BD, Niespodzany EJ, & Mackenzie B (2012). H+-coupled divalent metal-ion transporter-1: Functional properties, physiological roles and therapeutics. Curr Top Membr 70, 169-214.

Shawki A & Mackenzie B (2010). Interaction of calcium with the human divalent metal-ion transporter-1. Biochem Biophys Res Comm 393, 471-475.

Shayeghi M, Latunde-Dada GO, Oakhill JS, Laftah AH, Takeuchi K, Halliday N, Khan Y, Warley A, McCann FE, Hider RC, Frazer DM, Anderson GJ, Vulpe CD, Simpson RJ, & McKie AT (2005). Identification of an intestinal heme transporter. Cell 122, 789-801.

123

Sheftel AD, Richardson DR, Prchal J, & Ponka P (2009). Mitochondrial iron metabolism and sideroblastic anemia. Acta Haematol 122, 120-133.

Shemin D & Rittenberg D (1946). The life span of the human red blood cell. J Biol Chem 166, 627-636.

Shimada T & Hoshi T (1988). Na+-dependent elevation of the acidic cell surface pH (microclimate pH) of rat jejunal villus cells induced by cyclic nucleotides and phorbol ester: possible mediators of the regulation of the Na+/H+ antiporter. Biochim Biophys Acta 937, 328-334.

Shimamura T, Weyand S, Beckstein O, Rutherford NG, Hadden JM, Sharples D, Sansom MSP, Iwata S, Henderson PJF, & Cameron AD (2010). Molecular basis of alternating access membrane transport by the sodium-hydantoin transporter Mhp1. Science 328, 470-473.

Siddappa AJ, Rao RB, Wobken JD, Casperson K, Leibold EA, Connor JR, & Georgieff MK (2003). Iron deficiency alters iron regulatory protein and iron transport protein expression in the perinatal rat brain. Pediatr Res 53, 800-807.

Smith CP & Thévenod F (2009). Iron transport and the kidney. Biochim Biophys Acta 1790, 724-730.

Snyder AM & Connor JR (2009). Iron, the substantia nigra and related neurological disorders. Biochim Biophys Acta 1790, 606-614.

Song N, Jiang H, Wang J, & Xie JX (2007). Divalent metal transporter 1 up-regulation is involved in the 6- hydroxydopamine-induced ferrous iron influx. J Neurosci Res 35, 3118-3126.

Su MA, Trenor CC, Fleming JC, Fleming MD, & Andrews NC (1998). The G185R mutation disrupts function of the iron transporter Nramp2. Blood 92, 2157-2163.

Suzuki T, Momoi K, Hosoyamada M, Kimura M, & Shibasaki T (2008). Normal cadmium uptake in microcytic anemia mk/mk mice suggests that DMT1 is not the only cadmium transporter in vivo. Toxicol Appl Pharmacol 227, 462-467.

Tabuchi M, Tanaka N, Nishida-Kitayama J, Ohno H, & Kishi F (2002). Alternative splicing regulates the subcellular localization of divalent metal transporter 1 isoforms. Mol Biol Cell 13, 4371-4387.

Thévenod F & Wolff NA. Iron transport in the kidney: Implications for physiology and cadmium nephrotoxicity. Metallomics . 2015. Ref Type: In Press

Thompson BAV, Sharp PA, Elliott R, & Fairweather-Tait SJ (2010). Inhibitory effect of calcium on non-heme iron absorption may be related to translocation of DMT-1 at the apical membrane of enterocytes. J Agric Food Chem 58, 8414-8417.

124

Thompson K, Molina RM, Brain JD, & Wessling-Resnick M (2006). Belgrade rats display liver iron loading. J Nutr 136, 3010-3014.

Thompson K, Molina RM, Donaghey T, Schwob JE, Brain JD, & Wessling-Resnick M (2007). Olfactory uptake of manganese requires DMT1 and is enhanced by anemia. FASEB J 21, 223-230.

Thwaites DT & Anderson CMH (2007). H+-coupled nutrient, micronutrient and drug transporters in the mammalian small intestine. Exp Physiol 92, 603-619.

Thwaites DT, Kennedy DJ, Raldua D, Anderson CMH, Mendoza ME, Bladen CL, & Simmons NL (2002). H+/dipeptide absorption across the human intestinal epithelium is controlled indirectly via a functional Na+/H+ exchanger. Gastroenterology 122, 1322-1333.

Tjälve H & Henriksson J (1999). Uptake of metals in the brain via olfactory pathways. Neurotoxicology 20, 181-195.

Torrance JD & Bothwell TH (1980). Tissue iron stores. In Iron, ed. Cook JD, pp. 90-115. Churchill Livingstone, New York, NY.

Tse CM, Ma AI, Yang VW, Watson AJ, Levine S, Montrose MH, Potter J, Sardet C, Pouyssegur J, & Donowitz M (1991). Molecular cloning and expression of a cDNA encoding the rabbit ileal villus cell basolateral membrane Na+/H+ exchanger. EMBO J 10, 1957-1967.

Vesey DA (2010). Transport pathways for cadmium in the intestine and kidney proximal tubule: Focus on the interaction with essential metals. Toxicol Lett 198, 13-19.

Vulpe CD, Kuo YM, Murphy TL, Cowley L, Askwith C, Libina N, Gitschier J, & Anderson GJ (1999). Hephaestin, a ceruloplasmin homologue implicated in intestinal iron transport, is defective in the sla mouse. Nat Genet 21, 195-199.

Wallander ML, Leibold EA, & Eisenstein RS (2006). Molecular control of vertebrate iron homeostasis by iron regulatory proteins. Biochim Biophys Acta 1763, 668-689.

Wang CY & Knutson MD (2013). Hepatocyte divalent metal-ion transporter-1 is dispensable for hepatic iron accumulation and non-transferrin-bound iron uptake in mice. Hepatology 58, 788-798.

Wang C-Y, Jenkitkasemwong S, Duarte S, Sparkman BK, Shawki A, Mackenzie B, & Knutson MD (2012). ZIP8 is an iron and zinc transporter whose cell-surface expression is up-regulated by cellular iron loading. J Biol Chem 287, 34032-34043.

Wang D, Song Y, Li J, Wang C, & Li F (2011). Structure and metal ion binding of the first transmembrane domain of DMT1. Biochim Biophys Acta 1808, 1639-1644.

125

Wang K, Zhou B, Kuo YM, Zemansky J, & Gitschier J (2002a). A novel member of a zinc transporter family is defective in acrodermatitis enteropathica. Am J Hum Genet 71, 66-73.

Wang XS, Ong WY, & Connor JR (2002b). A light and electron microscopic study of divalent metal transporter-1 distribution in the rat hippocampus, after kainate-induced neuronal injury. Exp Neurol 177, 193-201.

Warrens AN, Jones MD, & Lechler RI (1997). Splicing by overlap extension by PCR using asymmetric amplification: An improved technique for the generation of hybrid proteins of immunological interest. Gene 186, 29-35.

Watanabe A, Choe S, Chaptal V, Rosenberg JM, Wright EM, Grabe M, & Abramson J (2010). The mechanism of sodium and substrate release from the binding pocket of vSGLT. Nature 468, 988-991.

Wessling-Resnick M (2006). Iron Imports. III. Transfer of iron from mucosa into the circulation. Am J Physiol Gastrointest Liver Physiol 290, G1-G6.

Wetli H, Buckett PD, & Wessling-Resnick M (2006). Chemical genetic screening identifies the selanazal drug ebselen as a potent inhibitor of DMT1-mediated iron uptake. Chem Biol 13, 965-972.

Weyand S, Shimamura T, Yajima S, Suzuki S, Mirza O, Krusong K, Carpenter EP, Rutherford NG, Hadden JM, O'Reilly J, Ma P, Saidijam M, Patching SG, Hope RJ, Norbertczak HT, Roach PCJ, Iwata S, Henderson PJF, & Cameron AD (2008). Structure and molecular mechanism of a nucleobase-cation-symport-1 family transporter. Science 322, 709-713.

WHO (2011). Haemoglobin concentrations for the diagnosis of anaemia and assessment of severity. WHO/NMH/NHD/MNM/11.1, Vitamin and Mineral Nutrition Information System, World Health Organization, Geneva. http://www.who.int/vmnis/indicators/haemoglobin, accessed on 21 October 15 A.D.

Williams K, Wilson MA, & Bressler J (2000). Regulation and developmental expression of the divalent metal-ion transporter in the rat brain. Cell Mol Biol 46, 563-571.

Wolff NA, Abouhamed M, Verroust PJ, & Thévenod F (2006). Megalin-dependent internalization of cadmium-metallothionein and cytotoxicity in cultured renal proximal tubule cells. J Pharmacol Exp Ther 318, 782-791.

Wolff NA, Garrick LM, Zhao L, Garrick MD, & Thévenod F (2014). Mitochondria represent another locale for the divalent metal transporter 1 (DMT1). Channels 8, 458-466.

Wrighting DM & Andrews NC (2006). Interleukin-6 induces hepcidin expression through STAT3. Blood 108, 3204-3209.

126

Wyman S, Simpson RJ, McKie AT, & Sharp PA (2008). Dcytb (Cybrd1) functions as both a ferric and a cupric reductase in vitro. FEBS Lett 582, 1901-1906.

Xiao S, Li J, Wang Y, Wang C, Xue R, Wang S, & Li F (2010). Identification of an "α-helix-extended segment- α-helix" conformation of the sixth transmembrane domain in DMT1. Biochim Biophys Acta 1758, 1556- 1564.

Xiao S, Wang C, Li J, & Li F (2011). Folding and assembly of TMD 6-related segments of DMT 1 in trifluoroethanol aqueous solution. J Pept Sci 17, 505-511.

Xu H, Jin J, DeFelice LJ, Andrews NC, & Clapham DE (2004). A spontaneous, recurrent mutation in divalent metal transporter-1 exposes a calcium entry pathway. PLoS Biol 2, E50.

Xu H, Chen R, & Ghishan FK (2005). Subcloning, localization, and expression of the rat intestinal sodium- hydrogen exchanger isoform 8. Am J Physiol Gastrointest Liver Physiol 289, G36-G41.

Yamashita A, Singh SK, Kawate T, Jin Y, & Gouaux E (2005). Crystal structure of a bacterial homologue of Na+/Cl–-dependent neurotransmitter transporters. Nature 437, 215-223.

Yanatori I, Tabuchi M, Kawai Y, Yasui Y, Akagi R, & Kishi F (2010). Heme and non-heme iron transporters in non-polarized and polarized cells. BMC Cell Biol 11, 39.

Zachos NC, Tse M, & Donowitz M (2005). Molecular physiology of intestinal Na+/H+ exchange. Annu Rev Physiol 67, 411-443.

Zampighi GA, Kreman M, Boorer KJ, Loo DDF, Bezanilla F, Chandy G, Hall JE, & Wright EM (1995). A method for determining the unitary functional capacity of cloned channels and transporters expressed in Xenopus laevis oocytes. J Membrane Biol 148, 65-78.

Zhang A-S, Canonne-Hergaux F, Gruenheid S, Gros P, & Ponka P (2008). Use of Nramp2-transfected Chinese hamster ovary cells and reticulocytes from mk/mk mice to study iron transport mechanisms. Exp Hematol 36, 1227-1235.

Zhang Z, Kodumuru V, Sviridov S, Liu S, Chafeev M, Chowdhury S, Chakka N, Sun J, Gauthier SJ, Mattice M, Ratkay LG, Kwan R, Thompson J, Cutts AB, Fu J, Kamboj R, Goldberg YP, & Cadieux JA (2012). Discovery of benzylisothioureas as potent divalent metal transporter 1 (DMT1) inhibitors. Bioorg Med Chem Lett 22, 5108-5113.

Zhao N, Gao J, Enns CA, & Knutson MD (2010). ZRT/IRT-like protein 14 (ZIP14) promotes the cellular assimilation of iron from transferrin. J Biol Chem 285, 32141-32150.

Zheng W, Xin N, Chi ZH, Zhao BL, Zhang J, Li JY, & Wang ZY (2009). Divalent metal transporter 1 is involved in amyloid precursor protein processing and Aβ generation. FASEB J 23, 4207-4217.

127

Zhou XY, Tomatsu S, Fleming RE, Parkkila S, Waheed A, Jiang J, Fei Y, Brunt EM, Ruddy DA, Prass CE, Schatzman RC, O'Neill R, Britton RS, Bacon BR, & Sly WS (1998). HFE gene knockout produces mouse model of hereditary hemochromatosis. Proc Natl Acad Sci USA 95, 2492-2497.

Zoller H, Koch RO, Theurl I, Obrist P, Pietrangelo A, Montosi G, Haile DJ, Vogel W, & Weiss G (2001). Expression of the duodenal iron transporters divalent-metal transporter 1 and ferroportin 1 in iron deficiency and iron overload. Gastroenterology 120, 1412-1419.

128

Appendix

Table A.1. Primers used in real-time quantitative PCR assays of isolated mouse tissues.

Gene NCBI Forward primer Reverse primer Ref. Reference Sequence

Cybrd1 AF_354666 5′–GCA GCG GGC TCG AGT TTA–3′ 5′–TTC CAG GTC CAT GGC AGT CT–3′ (Meynard et al., 2009)

Ctr1 NM_175090 5'- TAT GAA CCA CAC GGA CGA CAA -3' 5'- GCC ATT TCT CCA GGT GTA TTG A -3' —

DMT1 NM_008732

Upstream 5′–CGC TCG GTA AGC ATC TCG AA–3′ 5′–TGT TGC CAC CGC TGG TAT CT–3′ (Meynard of first loxP et al., 2009) site

Within 5′–TCC TCA TCA CCA TCG CAG ACA CTT–3′ 5′–TCC AAA CGT GAG GGC CAT GAT AGT–3′ (Wang & floxed Knutson, region 2013)

Fpn NM_016917 5′–CAT TGC TGC TAG AAT CGG TCT T–3′ 5′–GCA ACT GTG TCA CCG TCA AAT–3′ (Meynard et al., 2009)

Hamp1 AF_503444 5′–AAG CAG GGC AGA CAT TGC GAT–3′ 5′–CAG GAT GTG GCT CTA GGC TAT GT–3′ (Meynard et al., 2009)

GAPDH NM_008084 5′–CAT GGC CTT CCG TGT TCC TA–3′ 5′–CCT GCT TCA CCA CCT TCT TGA–3′ (Franich et al., 2008)

Steap2 BC150881.1 5′–ACG GAA AAC TGA AGG ACA GAA GA–3′ 5′–GAG ATG GAT CGG GGG TTG TG–3′ —

∙H = max S H . + Equation A.1.

129

Media

Intestinal apical wash medium (for radiotracer gavage studies in mice)

130 mM NaCl

5 mM KCl

1 mM CaCl2

1 mM MgCl2

1 mM nitrilotriacetic acid (NTA) (for 59Fe, 54Mn) a

5 mM 2-(N-morpholino)ethanesulfonic acid (MES)

pH 7.4 buffered by adding N′,N′-diethylpiperazine (DEPP) a 1 mM L-ascorbate and 1 mM L-histidine were used in place of NTA for the 64Cu study.

Modified Barths’ medium (MBM) b

88 mM NaCl

1 mM KCl

2.4 mM NaHCO3

0.82 mM MgSO4

0.66 mM NaNO3

0.75 mM CaCl2

10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES)

10 mg.mL−1 gentamicin

pH 7.5 buffered by adding Tris base b 2+ Ca -free MBM was used for collagenase treatment to defolliculate oocytes, in which case CaCl2 was

replaced by equimolar MgCl2.

130

Metal-ion transport assay solution (MeTx)

100 mM NaCl

2 mM KCl

0.6 mM CaCl2

1 mM MgSO4

0.66 mM NaNO3

0.1 or 1 mM c L-ascorbic acid

0–5 mM 2-(N-morpholino)ethanesulfonic acid (MES)

0–5 mM N′,N′-diethylpiperazine (DEPP)

pH 5.4−7.5 d buffered by mixing stock solutions containing either MES or DEPP c 0.1 mM L-ascorbic acid was used in voltage clamp experiments, 1 mM in radiotracer experiments with

55 55 Fe (to ensure complete reduction of the FeCl3 stock). d Unless indicated, media for transport assays were at pH 5.5.

131

Publications arising from work during the tenure of my PhD

1. Carvalho S, Barreira da Silva R, Shawki A, Castro H, Lamy M, Eide D, Costa V, Mackenzie B and Tomás AM. LiZIP3 is a cellular zinc transporter that mediates the tightly regulated import of zinc in Leishmania infantum parasites. Mol Microbiol 96: 581-595, 2015.

Abstract: Cellular zinc homeostasis ensures that the intracellular concentration of the element is kept within limits that enable its participation in critical physiological processes without exerting toxic effects. We report here the identification and characterization of LiZIP3, a novel mediator i zinc homeostasis in the human parasite Leishmania infantum. LiZIP3 is a member of the ZIP family of divalent metal-ion transporters. Expression of LiZIP3 in Saccharomyces cerevisiae rescued the growth of strains that were deficient in zinc acquisition. Expression of LiZIP3 in Xenopus laevis oocytes stimulated the uptake of a

2+ broad range of metal ions, among which, Zn appeared to be the preferred substrate (K0.5 ≈ 0.1 µM). Evidence that in LiZIP3 functions in the cellular import of zinc in L. infantum derives from the observations that (1) LiZIP3 protein was localized to the cell membrane, and (2) LiZIP3 overexpression leads to augmented zinc internalization. Expression and cell-surface location of LiZIP3 were absent in parasites that were exposed to high zinc bioavailability. An early event modulating LiZIP3 under zinc-replete conditions is the arrest of LiZIP3 synthesis which entails down-regulation of mRNA stability in a process involving a short-lived protein(s). Our data reveal a role for LiZIP3 in the acquisition of zinc by L. infantum in a highly regulated manner, contributing to zinc homeostasis in the parasite.

2. Engevik MA, Shawki A, Anthony SR, Baik RR, Kim RS, Worrell RT, Shull GE and Mackenzie B. Iron metabolism in a mouse model lacking gastric H+/K+ ATPase. Am J Physiol Gastrointest Liver Physiol, in preparation.

Abstract: Associations have been drawn between iron deficiency and conditions of decreased gastric acid, lending support to the premise that gastric acid is required for iron absorption; however, little has been done to test this directly. We have tested the hypothesis that gastric acid is required for iron absorption by examining iron metabolism in young adult mice lacking the gastric H+/K+ ATPase α subunit (Atp4a−/−), a model of achlorhydria. We found that male Atp4a−/− mice maintained nonheme iron stores equivalent to those of wildtype mice whether fed a normal or low-iron diet (6 weeks). Female Atp4a−/− mice fed a low- iron diet, but not those fed a normal diet, exhibited depleted iron stores compared with wildtype. No anemia was apparent in female Atp4a−/− mice, and serum iron and transferrin saturation did not differ between wildtype and Atp4a–/– mice on either diet. qPCR analysis revealed no changes in the expression of iron-related genes in female Atp4a−/− mice fed a normal diet compared with wildtype. Feeding female

132

mice a low-iron diet produced changes in gene expression—downregulation of hepcidin in the liver, and upregulation of the ferrireductase Cybrd1 and divalent metal-ion transporter-1 (DMT1) in the small intestine—that were more pronounced in Atp4a−/− mice than in wildtype. Therefore, although female Atp4a−/− mice challenged on a low-iron diet exhibited decreased iron stores, iron-appropriate compensatory changes in the expression of iron-related genes were sufficient to maintain serum iron and hematological variables at wildtype levels. Our data reveal no strict requirement for gastric acid in iron absorption.

3. Illing AC, Shawki A, Cunningham CL and Mackenzie B. Substrate profile and metal-ion selectivity of human divalent metal-ion transporter-1. J Biol Chem 287: 30485-30496, 2012.

Abstract: Divalent metal-ion transporter-1 (DMT1) is a H+-coupled metal-ion transporter that plays essential roles in iron homeostasis. DMT1 exhibits reactivity (based on evoked currents) with a broad range of metal ions; however, direct measurement of transport is lacking for many of its potential substrates. We performed a comprehensive, substrate-profile analysis for human DMT1 expressed in RNA-injected Xenopus oocytes by using radiotracer assays and the continuous measurement of transport by fluorescence (CMTF) with the metal-sensitive PhenGreen SK (PGSK) fluorophore. We provide validation for the use of PGSK fluorescence quenching as a reporter of cellular metal-ion uptake. We determined metal-ion selectivity under fixed conditions using the voltage clamp. Radiotracer and CMTF assays revealed that DMT1 mediates the transport of several metal ions that were ranked in selectivity by using

2+ 2+ 2+ 2+ 2+ 2+ 2+ the ratio Imax/K0.5 (determined from evoked currents at –70 mV): Cd > Fe > Co , Mn >> Zn , Ni , VO . DMT1 expression did not stimulate the transport of Cr2+, Cr3+, Cu1+, Cu2+, Fe3+, Ga3+, Hg2+, or VO1+. 55Fe2+ transport was competitively inhibited by Co2+ and Mn2+. Zn2+ only weakly inhibited 55Fe2+ transport. Our data reveal that DMT1 selects Fe2+ over its other physiological substrates and provide a basis for predicting the contribution of DMT1 to intestinal, nasal, and pulmonary absorption of metal ions and their cellular uptake in other tissues. Whereas DMT1 is a likely route of entry for the toxic heavy metal Cd, and may serve the metabolism of Co, Mn, and V, we predict that DMT1 should contribute little to the absorption or uptake of Zn. The conclusion in previous reports that Cu is a substrate of DMT1 is not supported.

4. Mitchell CJ, Shawki A, Ganz T, Nemeth E and Mackenzie B. Functional properties of human ferroportin, a cellular iron exporter reactive also with cobalt and zinc. Am J Physiol Cell Physiol 306: C450-C459, 2014.

Abstract: Iron homeostasis is achieved by regulating the intestinal absorption of the metal and its recycling by macrophages. Iron export from enterocytes or macrophages to blood plasma is thought to be mediated

133

by ferroportin under the control of hepcidin. Although ferroportin was identified over a decade ago, little is understood about how it works. We expressed in Xenopus oocytes a human ferroportin-EGFP fusion protein and observed using confocal microscopy its exclusive plasma-membrane localization. As a first step in its characterization, we established an assay to detect functional expression of ferroportin by microinjecting oocytes with 55Fe and measuring efflux. Ferroportin expression increased the first-order rate constants describing 55Fe efflux up to 300-fold over control. Ferroportin-mediated 55Fe efflux was

-1 saturable, temperature-dependent (activation energy, Ea ≈ 17 kcal.mol ), maximal at extracellular pH ≈ 7.5, and inactivated at extracellular pH < 6.0. We estimated that ferroportin reacts with iron at its intracellular aspect with apparent affinity constant < 10-7 M. Ferroportin expression also stimulated efflux of 65Zn and 57Co but not of 64Cu, 109Cd, or 54Mn. Hepcidin treatment of oocytes inhibited efflux of 55Fe, 65Zn, and 57Co. Whereas hepcidin administration in mice resulted in a marked hypoferremia within 4 h, we observed no effect on serum zinc levels in those same animals. We conclude that ferroportin is an iron- preferring cellular metal-efflux transporter with a narrow substrate profile that includes cobalt and zinc. Whereas hepcidin strongly regulated serum iron levels in the mouse, we found no evidence that ferroportin plays an important role in zinc homeostasis.

5. Pinilla-Tenas JJ, Sparkman BK, Shawki A, Illing AC, Mitchell CJ, Zhao N, Liuzzi JP, Cousins RJ, Knutson MD and Mackenzie B. Zip14 is a complex broad-scope metal-ion transporter whose functional properties support roles in the cellular uptake of zinc and nontransferrin-bound iron. Am J Physiol Cell Physiol 301: C862-C871, 2011.

Abstract: Recent studies have shown that overexpression of the transmembrane protein Zrt- and Irt-like protein 14 (Zip14) stimulates the cellular uptake of zinc and nontransferrin-bound iron (NTBI). Here, we directly tested the hypothesis that Zip14 transports free zinc, iron, and other metal ions by using the Xenopus laevis oocyte heterologous expression system, and use of this approach also allowed us to characterize the functional properties of Zip14. Expression of mouse Zip14 in RNA-injected oocytes stimulated the uptake of 55Fe in the presence of L-ascorbate but not nitrilotriacetic acid, indicating that Zip14 is an iron transporter specific for ferrous ion (Fe2+) over ferric ion (Fe3+). Zip14-mediated 55Fe2+ uptake was saturable (K0.5 ≈ 2 µM), temperature-dependent (apparent activation energy, Ea ≈ 15

2+ 2+ 2+ 2+ 55 2+ kcal/mol), pH-sensitive, Ca -dependent, and inhibited by Co , Mn , and Zn . HCO3- stimulated Fe transport. These properties are in close agreement with those of NTBI uptake in the perfused rat liver and in isolated hepatocytes reported in the literature. Zip14 also mediated the uptake of 109Cd2+, 54 Mn2+, and

65 2+ 64 65 2+ Zn but not Cu (I or II). Zn uptake also was saturable (K0.5 ≈ 2 µM) but, notably, the metal-ion inhibition profile and Ca2+ dependence of Zn2+ transport differed from those of Fe2+ transport, and we

134

propose a model to account for these observations. Our data reveal that Zip14 is a complex, broad-scope metal-ion transporter. Whereas zinc appears to be a preferred substrate under normal conditions, we found that Zip14 is capable of mediating cellular uptake of NTBI characteristic of iron-overload conditions.

6. Shawki A, Knight PB, Maliken BD, Niespodzany EJ and Mackenzie B. H+-coupled divalent metal-ion transporter-1: Functional properties, physiological roles and therapeutics. Curr Top Membr 70: 169- 214, 2012.

Abstract: Divalent metal-ion transporter-1 (DMT1) is a widely expressed, iron-preferring membrane transport protein. Animal models establish that DMT1 plays indispensable roles in intestinal nonheme- iron absorption and iron acquisition by erythroid precursor cells. Rare mutations in human DMT1 result in severe microcytic-hypochromic anemia. When we express DMT1 in RNA-injected Xenopus oocytes, we observe rheogenic Fe2+ transport that is driven by the proton electrochemical potential gradient. In that same preparation, DMT1 also transports cadmium and manganese but not copper. Whether manganese metabolism relies upon DMT1 remains unclear but DMT1 contributes to the effects of overexposure to cadmium and manganese in some tissues. There exist at least four DMT1 isoforms that arise from variant transcription of the SLC11A2 gene. Whereas these isoforms display identical functional properties, N- and C-terminal variations contain cues that direct the cell-specific targeting of DMT1 isoforms to discrete subcellular compartments (plasma membrane, endosomes, and lysosomes). An iron-responsive element (IRE) in the mRNA 3'-untranslated region permits the regulation of some isoforms by iron status, and additional mechanisms by which DMT1 is regulated are emerging. Natural-resistance-associated macrophage protein-1 (NRAMP1)—the only other member of the mammalian SLC11 gene family— contributes to antimicrobial function by extruding from the phagolysosome divalent metal ions (e.g. Mn2+) that may be essential cofactors for bacteria-derived enzymes or required for bacterial growth. The principal or only intestinal nonheme-iron transporter, DMT1 is a validated therapeutic target in hereditary hemochromatosis (HH) and other iron-overload disorders.

7. Shawki A, Engevik MA, Kim RS, Anthony SR, Knight PB, Baik RR, Worrell RT, Shull GE and Mackenzie B. Intestinal brush-border Na+/H+ exchanger NHE3 drives DMT1-mediated H+-coupled iron absorption in the mouse. Blood, in preparation.

Abstract: Divalent metal-ion transporter-1 (DMT1), the principal or only mechanism by which nonheme iron is taken up at the intestinal brush border, is energized by the H+ electrochemical potential gradient. The provenance of the H+ gradient in vivo is unknown so we have explored a role for brush-border Na+/H+ exchangers by examining iron homeostasis and intestinal iron handling in mice lacking Na+/H+ exchanger-

135

2 (NHE2) or Na+/H+ exchanger-3 (NHE3). We observed modestly depleted liver iron stores in NHE2-null (NHE2–/–) mice stressed on a low-iron diet but no change in hematological or blood-iron variables, or the expression of genes associated with iron metabolism, compared with wildtype mice. Ablation of NHE3 more severely depleted liver iron stores regardless of diet. We observed decreases in blood-iron variables but no overt anemia in NHE3–/– mice on a low-iron diet. Intestinal expression of DMT1, Cybrd1, and Fpn was upregulated in NHE3–/– mice, and expression of liver Hamp1 (hepcidin) suppressed, compared with wildtype. Absorption of 59Fe from an oral dose was substantially impaired in NHE3–/– mice compared with wildtype. Our data point to an important role for NHE3 in generating the H+ gradient that drives DMT1- mediated iron uptake at the intestinal brush border.

8. Shawki A, Anthony SR, Nose Y, Engevik MA, Niespodzany EJ, Barrientos T, Öhrvik H, Worrell RT, Thiele DJ and Mackenzie B. Intestinal DMT1 is critical for iron absorption in the mouse but is not required for the absorption of copper or manganese. Am J Physiol Gastrointest Liver Physiol 309: G635-G647, 2015.

Abstract: Divalent metal-ion transporter-1 (DMT1) is a widely expressed iron-preferring membrane- transport protein that serves a critical role in erythroid iron utilization. We have investigated its role in intestinal metal absorption by studying a mouse model lacking intestinal DMT1 (i.e. DMT1int/int). DMT1int/int mice exhibited a profound hypochromic–microcytic anemia, splenomegaly and cardiomegaly. That the anemia was due to iron deficiency was demonstrated by the following observations in DMT1int/int mice: (1) blood iron and tissue nonheme-iron stores were depleted; (2) mRNA expression of liver hepcidin (Hamp1) was depressed; and (3) intraperitoneal iron injection corrected the anemia, and reversed the changes in blood iron, nonheme-iron stores, and hepcidin expression levels. We observed decreased total iron content in multiple tissues from DMT1int/int mice compared with DMT1+/+ mice but no meaningful change in copper, manganese, or zinc. DMT1int/int mice absorbed 64Cu and 54Mn from an intragastric dose to the same extent as did DMT1+/+ mice but the absorption of 59Fe was virtually abolished in DMT1int/int mice. This study reveals a critical function for DMT1 in intestinal nonheme-iron absorption for normal growth and development. Further, this work demonstrates that intestinal DMT1 is not required for the intestinal transport of copper, manganese, or zinc.

9. Shawki A, Illing AC, and Mackenzie B. Molecular impact of divalent metal-ion transporter (DMT1) mutations (V114del and G212V) found in a compound heterozygote with microcytic anemia and hepatic iron overload. Blood Cells Mol Dis, in preparation.

136

Abstract: DMT1 is a H+-coupled divalent metal-ion transporter critical for intestinal iron (Fe) absorption and erythroid Fe utilization. Beaumont et al (Blood 107, 4168 [2006]) described a patient compound heterozygous for two new DMT1 mutations (V114 deletion and G212V) associated with hypochromic- microcytic anemia and hepatic Fe overload. The proband’s sibling and parents each possess one mutated allele but were asymptomatic. We expressed mutant and wildtype (wt) DMT1 in Xenopus oocytes and used voltage-clamp and radiotracer assays to examine the molecular impact of these novel mutations. The V114 deletion abolished 55Fe2+ transport activity whereas 55Fe2+ transport mediated by G212V-DMT1 did not differ from wtDMT1. G212V had no effect on the pH dependence of 55Fe2+ transport, pre-

2+ steadystate kinetics, or the affinity for Fe (K0.5 = 1.7 ± 1.1 μM c.f. 1.3 ± 0.5 μM in wtDMT1). To our surprise, however, G212V-DMT1 mediated much smaller Fe2+-evoked currents than did wtDMT1. From simultaneous measurement of currents and 55Fe2+ fluxes, we arrived at a wildtype H+/Fe2+ ratio of 17 ± 1 which greatly exceeds that expected for strict stoichiometric transport, i.e. most of the wtDMT1-mediated current arises from H+ slippage. Such slippage is normally thought of in terms of the energetic penalty to the cell; however, indications that H+ slippage, but not Fe2+ transport, is disrupted in the G212V mutant associated with disease raise the possibility that H+ slippage serve a physiological role.

10. Wang C-Y, Jenkitkasemwong S, Duarte S, Sparkman BK, Shawki A, Mackenzie B and Knutson MD. ZIP8 is an iron and zinc transporter whose cell-surface expression is up-regulated by cellular iron loading. J Biol Chem 287: 34032-34043, 2012.

Abstract: ZIP8 (SLC39A8) belongs to the ZIP family of metal ion transporters. Among the ZIP proteins, ZIP8 is most closely related to ZIP14, which can transport iron, zinc, manganese, and cadmium. Here we investigated the iron transport ability of ZIP8, its subcellular localization, pH dependence, and regulation by iron. Transfection of HEK293T cells with ZIP8 cDNA enhanced the uptake of 59Fe and 65Zn by 200 and 40%, respectively, compared with controls. Excess iron inhibited the uptake of zinc and vice versa. In RNA-

55 2+ injected Xenopus oocytes, ZIP8-mediated Fe transport was saturable (K0.5 of ≈ 0.7 µM) and inhibited by zinc. ZIP8 also mediated the uptake of 109Cd2+, 57Co2+, 65Zn2+ > 54Mn2+, but not 64Cu (I or II). By using immunofluorescence analysis, we found that ZIP8 expressed in HEK 293T cells localized to the plasma membrane and partially in early endosomes. Iron loading increased total and cell surface levels of ZIP8 in H4IIE rat hepatoma cells. We also determined by using site-directed mutagenesis that asparagine residues 40, 88, and 96 of rat ZIP8 are glycosylated and that N-glycosylation is not required for iron or zinc transport. Analysis of 20 different human tissues revealed abundant ZIP8 expression in lung and placenta and showed that its expression profile differs markedly from ZIP14, suggesting nonredundant functions. Suppression of endogenous ZIP8 expression in BeWo cells, a placental cell line, reduced iron uptake by ≈

137

40%, suggesting that ZIP8 participates in placental iron transport. Collectively, these data identify ZIP8 as an iron transport protein that may function in iron metabolism.

Published abstracts arising during the tenure of my PhD

1. Carvalho S, Silva R, Shawki A, Mackenzie B, Castro H, Eide D, Costa V and Tomas AM. Leishmania infantum ZIP3 is a zinc transporter that is tightly regulated by zinc status. FASEB J 26: 1112.4, 2012.

2. Mackenzie B, Shawki A, Kim R, Anthony SR, Knight PB, Bradford EM and Shull GE. Intestinal brush- border Na+/H+ exchangers are required for iron homeostasis in the mouse. FASEB J 25: 238.1, 2011.

3. Mitchell CJ, Shawki A, Nemeth E, Ganz T and Mackenzie B. Functional expression in Xenopus oocytes reveals that human ferroportin is an iron exporter shared with zinc. Am J Hematol 86: E87, 2011.

4. Prakash S, Shawki A, Niespodzany E and Mackenzie B. Intestinal divalent metal-ion transporter-1 is required for iron homeostasis in the neonatal mouse. FASEB J 29: 1011.5, 2015.

5. Shawki A, Engevik MA, Kim R, Anthony SR, Knight PB, Baik R, Worrell RT, Shull GE and Mackenzie B. Ablation of intestinal brush-border Na+/H+ exchanger NHE3 impairs iron absorption in the mouse. FASEB J 28: 900.1, 2014.

6. Shawki A, Anthony SR, Niespodzany EJ, Amratia AA and Mackenzie B. Ablation of intestinal divalent metal-ion transporter-1 produces iron-deficiency anemia. FASEB J 27: 950.3, 2013.

7. Shawki A, Ruwe TA, Mitchell CJ, Prakash S, Nemeth E, Ganz T and Mackenzie B. Ferroportin- mediated cellular iron efflux requires extracellular calcium. FASEB J 29: 566.15, 2015.

8. Shawki A, Kim R, Anthony SR, Knight PB, Baik R, Bradford EM, Shull GE and Mackenzie B. Intestinal brush-border Na+/H+ exchanger NHE3 is required for iron homeostasis in the mouse. Genes Nutr 6: S46, 2012.

9. Shawki A, Kim R, Anthony SR, Knight PB, Bradford EM, Shull GE and Mackenzie B. Intestinal brush- border Na+/H+ exchangers are required for iron homeostasis in the mouse. Am J Hematol 86: E87, 2011.

10. Shawki A, Anthony SR, Nose Y, Barrientos De Renshaw T, Thiele DJ and Mackenzie B. Intestinal divalent metal-ion transporter-1 is critical for iron homeostasis but is not required for maintenance of Cu or Zn. FASEB J 26: 1112.2, 2012.

138

11. Shawki A, Anthony SR, Nose Y, Niespodzany EJ, Amratia AA, Barrientos T, Thiele DJ and Mackenzie B. Intestinal divalent metal-ion transporter-1 is critical for iron homeostasis but is not required for maintenance of Cu or Zn. Am J Hematol 88: E195, 2013.

12. Shawki A, Anthony SR, Engevik MA, Niespodzany EJ, Worrell RT and Mackenzie B. Intestinal divalent metal-ion transporter-1 is required for the absorption of iron but not copper. FASEB J 28: 996.3, 2014.

13. Shawki A, Illing AC and Mackenzie B. Molecular impact of a human divalent metal-ion transporter-1 (DMT1) mutation (G212V) found in two compound heterozygotes with microcytic anemia. FASEB J 28: 893.40, 2014.

14. Shawki A, Niespodzany EJ and Mackenzie B. No evidence that copper is a transported substrate of the iron transporter DMT1. FASEB J 26: 1112.3, 2012.

15. Wang C-Y, Jenkitkasemwong S, Sparkman BK, Shawki A, Mackenzie B and Knutson MD. Metal transport, subcellular localization, and tissue distribution of Zip8, a Zip14 homologue. FASEB J 26: 641.32, 2012.

139