The Significance of the Evolutionary Relationship of Prion Proteins and ZIP Transporters in Health and Disease

Home , GFRAL, PRNP, Retroposon, Zinc transporter protein, Zinc transporter ZIP10, Zinc transporter ZIP4

Sepehr Ehsani

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Department of Laboratory Medicine and Pathobiology University of Toronto

The Significance of the Evolutionary Relationship of Prion Proteins and ZIP Transporters in Health and Disease

Sepehr Ehsani

Doctor of Philosophy

Department of Laboratory Medicine and Pathobiology University of Toronto

2012 Abstract

The cellular prion protein (PrPC) is unique amongst mammalian proteins in that it not only has the capacity to aggregate (in the form of scrapie PrP; PrPSc) and cause neuronal degeneration, but can also act as an independent vector for the transmission of disease from one individual to another of the same or, in some instances, other species. Since the discovery of PrPC nearly thirty years ago, two salient questions have remained largely unanswered, namely, (i) what is the normal function of the cellular protein in the central nervous system, and (ii) what is/are the factor(s) involved in the misfolding of PrPC into PrPSc? To shed light on aspects of these questions, we undertook a discovery-based interactome investigation of PrPC in mouse neuroblastoma cells (Chapter 2), and among the candidate interactors, identified two members of the ZIP family of zinc transporters (ZIP6 and ZIP10) as possessing a PrP-like domain.

Detailed analyses revealed that the LIV-1 subfamily of ZIP transporters (to which ZIPs 6 and 10 belong) are in fact the evolutionary ancestors of prions (Chapter 3). We were further able to demonstrate that PrPC likely emerged from a ZIP ancestor molecule nearly half-a-billion years ago via a retrotransposition event (Chapter 4). Moreover, biochemical investigations on ZIP10, as a model LIV-1 ZIP transporter, demonstrated that the ectodomain shedding of ZIP10 observed in prion-infected mice resembles a cellular response to transition metal starvation and suggested ii that prion disease in mice might phenocopy a transition metal starvation status (Chapter 5).

These studies have opened a new angle to study prion biology in health and disease. Biochemical investigations on other LIV-1 ZIPs and attempts at the structural elucidation of the PrP-like domain of LIV-1 ZIP proteins are ongoing and have not been included in this thesis.

iii

Acknowledgments

I would like to first and foremost thank my thesis supervisor Dr. Gerold Schmitt-Ulms for providing me with a creative, challenging and intellectually-satisfying research project which not only contributed to our collective understanding of aspects of cellular biology, but also had on its horizon the hope of lessening the suffering of those grappling with a dreadful neurodegenerative disease.

I would also like to greatly thank the present and past members of my Ph.D. advisory committee, Dr. Rod Bremner, Dr. Sidney E. Croul, Dr. Philip A. Marsden and Dr. Janice Robertson, and the Department of Laboratory Medicine and Pathobiology’s Graduate Coordinator Dr. Harry P. Elsholtz, for their valuable support and guidance. Thank you also to Dr. Glen K. Andrews (University of Kansas Medical Center), Dr. Avi Chakrabartty (University of Toronto) and Dr. John R. Glover (University of Toronto) for evaluating my thesis.

During my four years in the graduate program, I was fortunate to be able to collaborate with Joel C. Watts, Hairu Huo and Yu Bai on the prion interactome project, Ashkan Salehzadeh, Hairu Huo and Mohadeseh Mehrabian on the prion/ZIP biochemistry project, Cosmin L. Pocanschi on the ZIP structural project, and Renzhu Tao and Hezhen Ren on the prion/ZIP bioinformatics project. I would like to extend my sincere thanks to them all.

Lastly, I would like to thank Robert Strome for teaching me many molecular cloning techniques and Hairu Huo for training me in a number of laboratory techniques.

Work on the projects which formed parts of this thesis was funded through support from the Ontario Graduate Scholarship program, the University of Toronto Fellowship program, the Garfield Weston Foundation, the Canadian Institutes of Health Research (MOP-74734) and PrioNet Canada.

I would like to dedicate this thesis to my parents.

Table of Contents

Acknowledgments ...... iv

Table of Contents ...... v

List of Tables ...... xi

List of Figures ...... xii

List of Appendices ...... xv

Abbreviations ...... xvi

Chapter 1 Prion and ZIP Proteins: An Introduction ...... 1

1.1 Preamble ...... 1

1.2 Ancestral ties and diversification ...... 2

1.3 Function – novel links to EMT and cancer ...... 5

1.4 Expression – crosstalk and division of labor ...... 11

1.5 Signaling – from Tinman to Snail and beyond ...... 13

1.6 Post-translational modifications ...... 15

1.7 Binding of divalent cations – a common denominator ...... 18

1.8 Conclusions ...... 21

Chapter 2 Interactome Analyses Identify Ties of PrPC and Its Mammalian Paralogs to Oligomannosidic N-Glycans and Endoplasmic Reticulum-Derived Chaperones ...... 22

2.1 Introduction ...... 23

2.2 Materials and methods ...... 25

2.2.1 Ethics Statement ...... 25

2.2.2 Antibodies ...... 25

2.2.3 Clones ...... 25

2.2.4 Cell culture, in vivo crosslinking, inhibitor treatments and cell viability assay ... 26

2.2.5 Affinity purification of bait proteins ...... 26

2.2.6 Protein reduction, alkylation and trypsinization ...... 26

2.2.7 iTRAQ labeling ...... 27

2.2.8 Two-dimensional liquid chromatography ...... 27

2.2.9 Electrospray ionization QqTOF mass spectrometry analysis ...... 28

2.2.10 Database searches ...... 28

2.2.11 Snowdrop lectin affinity purification ...... 29

2.2.12 Cell surface biotinylation ...... 29

2.2.13 Proteinase K digestion ...... 30

2.3 Results ...... 30

2.3.1 Large-scale quantitative and comparative interactome investigation of members of the mammalian prion protein family ...... 30

2.3.2 Interactions amongst members of the mammalian prion protein family ...... 37

2.3.3 Direct versus indirect PrP interactors ...... 38

2.3.4 PrP forms high-molecular weight complexes with proteins carrying oligomannosidic N-glycans ...... 39

2.3.5 A subset of cellular P4hb, Pdia3 and calreticulin reside at the cell surface of neuroblastoma cells ...... 41

2.3.6 Inhibitors of protein disulfide isomerases increase PrPSc levels in a subset of ScN2a cell clones ...... 44

2.4 Discussion ...... 46

2.5 Conclusion ...... 52

Chapter 3 Evolutionary Descent of Prion Genes from the ZIP Family of Metal Ion Transporters ...... 54

3.1 Introduction ...... 54

3.2 Materials and Methods ...... 56

3.2.1 Molecular clones ...... 56

3.2.2 Cell culture ...... 56

3.2.3 Western blotting ...... 56

3.2.4 In vivo crosslinking ...... 57

3.2.5 Affinity purification of bait proteins ...... 57

3.2.6 Protein reduction, alkylation and trypsinization ...... 57

3.2.7 iTRAQ labeling ...... 57

3.2.8 Two-dimensional liquid chromatography ...... 58

3.2.9 ESI-QqTOF mass spectrometry analysis ...... 58

3.2.10 Database searches ...... 59

3.2.11 Multiple sequence alignments ...... 60

3.2.12 Structural threading ...... 60

3.2.13 Accession numbers ...... 60

3.3 Results ...... 62

3.3.1 Quantitative interactome analyses ...... 62

3.3.2 Structural and sequence similarities of mammalian prion proteins and ZIPs ...... 67

3.3.3 Biological similarities between PrP and ZIPs ...... 70

3.3.4 Convergent evolution versus common evolutionary origin ...... 74

3.4 Discussion ...... 77

3.4.1 Protein-protein interactions amongst mammalian prion proteins and ZIPs ...... 80

3.4.2 N-terminal duplication versus PL domain insertion ...... 80

3.4.3 Transmembrane-to-GPI-anchor attachment ...... 83

3.4.4 Analyses of genomic sequences ...... 83

3.4.5 Structure and function of PL domain ...... 84

3.5 Conclusions ...... 85

vii

Chapter 4 Evidence for Retrogene Origins of the Prion Gene Family ...... 86

4.1 Introduction ...... 86

4.2 Methods ...... 88

4.2.1 Multiple sequence alignments ...... 88

4.2.2 Intron-exon genomic organization ...... 89

4.2.3 Synteny analysis ...... 92

4.2.4 Pseudogene discovery ...... 92

4.2.5 Accession numbers ...... 92

4.3 Results ...... 93

4.3.2 ZIP genes of all evolutionary lineages are characterized by complex intron- exon structures not observed in prion gene sequences ...... 96

4.3.3 No shared genes in the genomic neighborhoods of ZIP and prion genes ...... 98

4.3.4 Pseudogene analyses uncover instances of genomic insertions of spliced and reverse-transcribed ZIP transcripts in vertebrates ...... 101

4.4 Discussion ...... 105

4.4.1 Emergence of prion-like ZIP ectodomain in early metazoa ...... 107

4.4.2 Generation of prion founder gene in vertebrates ...... 107

4.4.3 Other retropositional events ...... 110

4.5 Conclusions ...... 112

Chapter 5 LIV-1 ZIP Ectodomain Shedding in Prion-Infected Mice Resembles Cellular Response to Transition Metal Starvation ...... 113

5.1 Introduction ...... 114

5.2 Materials and Methods ...... 115

5.2.1 Cell viability assay ...... 115

viii

5.2.2 Confocal immunofluorescence analysis ...... 115

5.2.3 Immunohistochemical staining ...... 116

5.2.4 Divalent metal ion measurements ...... 116

5.2.5 Glycosylation analysis ...... 117

5.2.6 Metal chelation ...... 117

5.2.7 Molecular Cloning ...... 117

5.2.8 Polyclonal antibody generation ...... 117

5.2.9 Reverse transcription polymerase chain reaction (RT-PCR) ...... 118

5.2.10 Preparation of mouse brain homogenates ...... 118

5.2.11 SDS-PAGE and immunoblotting ...... 119

5.2.12 Proteinase K digestion ...... 119

5.2.13 Sucrose gradient isolation of lipid rafts ...... 120

5.2.14 Transient overexpression and knockdown ...... 120

5.3 Results ...... 121

5.3.1 Prion disease in mice is accompanied by abnormal endoproteolysis of ZIP10 .. 121

5.3.2 ZIP10 is one of many LZTs expressed in N2a cells that localizes to the plasma membrane and undergoes complex posttranslational maturation ...... 124

5.3.3 ZIP10 is N-glycosylated at ‘NxT’ acceptor site that it shares with PrPC ...... 131

5.3.4 Appearance of truncated and N-glycosylated variants of ZIP10 depends on extracellular divalent metal ion status ...... 132

5.3.5 Appearance of truncated ZIP10 can be rescued by manganese and zinc but not copper replenishment ...... 137

5.3.6 A zinc coordination site within TMD5 regulates N-glycosylation and endoproteolysis of ZIP10 ...... 138

5.3.7 Transition metal starvation or RML infection causes PrPC to shed distinct N1 or N2 fragments but triggers shedding of identical N_ecto fragments from immature ZIP10 ...... 142

5.4 Discussion ...... 145

5.4.1 Functional significance of ZIP10 ectodomain shedding ...... 147 ix

5.4.2 Alpha-site cleavage of PrPC ...... 149

5.5 Conclusions ...... 150

Chapter 6 Future Directions of the ZIP-Prion Connection ...... 151

References ...... 157

Appendices ...... 184

List of Tables

Table 1.1: Characteristics of human PrP and ZIP paralogs.

Table 2.1: Prion protein family interactome in mouse neuroblastoma cells.

Table 3.1: Sequences from Entrez, Ensembl and UniProt databases utilized for alignments.

Table 3.2: Quantitative analysis of mouse Dpl, PrP and Sho interactomes identifies metal ion transporters of the ZIP protein family in spatial proximity to all three members of the mammalian prion protein family.

Table 3.3: Comparison of mouse PrP gene family paralogs with ZIPs 5/6/10.

Table 3.4: Summary of evidences presented in support of evolutionary descent of PrP gene family from ZIP metal ion transport ancestor gene.

Table 4.1: Protein accession numbers and abbreviations of species names.

Table 5.1: Inductively coupled plasma atomic emission spectroscopy analysis of transition metals in relevant media and solutions.

List of Figures

Figure 1.1: Comparison of PrP and ZIPs 5, 6 and 10.

Figure 1.2: Simplified phylogenetic tree and table depicting both the wide distribution of ZIP sequences in most organisms and the restricted distribution of prion genes to the Chordata lineage.

Figure 1.3: Cartoon comparing properties of ZIP and CDF (ZnT) zinc ion transporters. (A) ZIP- mediated passive zinc import. (B) Antiport model of CDF (ZnT).

Figure 1.4: ZIP-dependent signaling upstream of E-cadherin expression in three biological paradigms.

Figure 2.1: Expression analysis of FLAG-tagged mouse prion proteins.

Figure 2.2: Flow chart depicting strategy for semi-quantitative comparison of prion protein family interactomes.

Figure 2.3: The mammalian prion protein family interactome.

Figure 2.4: Evidence for PrPC in high-molecular weight protein complexes.

Figure 2.5: Evidence for PrPC in high molecular weight protein complexes captured by a lectin with specificity for oligomannosidic glycans.

Figure 2.6: Evidence for cell surface localization of a subset of PDIs and calreticulin in mouse neuroblastoma cells.

Figure 2.7: Inhibition of protein disulfide isomerases causes accumulation of PrPSc in a subset of ScN2a cell clones.

Figure 2.8: Inhibition of protein disulfide isomerases shows no effect on PrPSc formation in a subset of ScN2a cell clones.

Figure 2.9: Mammalian PrPC interactome.

xii

Figure 3.1: Evidence for specific co-enrichment of ZIP10 and ZIP6 with all three members of the mammalian prion protein family.

Figure 3.2: Structural similarity between mouse ZIP10, PrP and Dpl.

Figure 3.3: Molecular organization, mode of membrane attachment and phylogenetic relationship of ZIP and prion gene families.

Figure 3.4: Sequence evidence for common origin and divergent sequence evolution of members of ZIP and prion protein families.

Figure 3.5: Multiple sequence alignment of cysteine-flanked core sequence segment within PL domain.

Figure 3.6: Models depicting evolutionary origin and topology of members of prion protein family, hypothetical ZIP ancestor and ZIP10 transporter.

Figure 4.1: Tree diagram depicting species utilized for genomic analyses in Chapter 4.

Figure 4.2: Broad phylogenetic distribution of LIV-1 ZIP metal ion transporters contrasts narrow distribution of prion genes in Chordata lineage.

Figure 4.3: The cysteine-flanked core within the prion-like domain of ZIP proteins is confined to metazoa.

Figure 4.4: Multiple introns observed in the coding regions of ZIP genes are missing from prion genes.

Figure 4.5: Multiple introns observed in the coding regions of ZIP genes are missing from prion genes.

Figure 4.6: Lack of shared genes in proximity of PrP and ZIP genes.

Figure 4.7: Evidence for the existence of ZIP pseudogenes in the human genome.

Figure 4.8: Evidence for a C-terminally truncated ZIP6 pseudogene in the opossum genome.

Figure 4.9: Two-step model of emergence of prion gene from a ZIP ancestor.

xiii

Figure 4.10: Precedent of retroposition event leading to a subbranch of GPI-anchored proteins within family of transmembrane proteins.

Figure 5.1: Prion disease in mice causes appearance of novel protein band which can be detected with ZIP10-directed antibody.

Figure 5.3: Multiple LZTs but not ZIP5 are expressed in N2a cells.

Figure 5.4: ZIP10 is localized at the plasma membrane and undergoes complex posttranslational modifications.

Figure 5.5: ZIP10 exhibits low level of co-localization with PrPC in N2a cells.

Figure 5.6: N-glycosylation acceptor sites within ZIP10.

Figure 5.7: ZIP10 is N-glycosylated at ‘NxT’ acceptor site it shares with PrPC and is partially shed from the membrane by at least two distinct cleavages.

Figure 5.8: Depletion of zinc and manganese but not copper from the cell culture medium causes shedding of the ZIP10 ectodomain and cleavage of PrPC in N2a cells.

Figure 5.9: A conserved zinc coordination site within TMD5 influences N-glycosylation and shedding of the ZIP10 ectodomain.

xiv

List of Appendices

Abbreviations

Supplemental Table 2.1

Supplemental Figure 3.1

Supplemental Table 3.1

Supplemental Figure 4.1

Supplemental Figure 4.2

Abbreviations

AE acrodermatitis enteropathica bps base pairs

CDF cation diffusion facilitator

CFC cysteine-flanked core

CID collision-induced dissociation

CTD carboxy-terminal domain

Dm Drosophila melanogaster

Dpl Doppel

Dr Danio rerio

DTPA diethylene triamine pentaacetic acid

EMT epithelial-to-mesenchymal transition

Esg Escargot

FA formaldehyde

GPI glycosylphosphatidylinositol

HA hemagglutinin

HMM hidden Markov model

HPLC high performance liquid chromatography i.c. intracerebral

IP immunoprecipitation

xvi iTRAQ isobaric tagging for relative and absolute quantitation

LC liquid chromatography

LZT LIV-1 subfamily of ZIP zinc transporters

MEM minimal essential medium

Mm Mus musculus

MS/MS tandem mass spectrometry

MW molecular weight

N2a Neuro2a mouse neuroblastoma

ORF open reading frame

PBS phosphate buffered saline p.i. post-inoculation

PL PrP-like

PrPC cellular prion protein

PVDF polyvinylidene fluoride

RT room temperature

Sho Shadoo

SLC solute carrier

Ta Trichoplax adhaerens

TM transmembrane

TMD transmembrane domain

xvii

TPEN N,N,N’,N’-tetrakis (2-pyridylmethyl) ethylenediamine

Tr Takifugu rubripes

ZIP Zrt- Irt-like protein

xviii 1

Chapter 1 Prion and ZIP Proteins: An Introduction

Please note that all parts of this chapter, excluding minor modifications/updates, were published in the following article [1]: Sepehr Ehsani, Hairu Huo, Ashkan Salehzadeh, Cosmin L. Pocanschi, Joel C. Watts, Holger Wille, David Westaway, Ekaterina Rogaeva, Peter H. St. George-Hyslop, Gerold Schmitt-Ulms (2011) Family reunion – the ZIP/prion gene family. Prog Neurobiol 93(3):405-20.

Candidate’s role: Co-wrote manuscript

Summary: Prion and ZIP proteins are introduced, and an overview of their function, expression, signaling, post-translational modification and metal biology is provided. Furthermore, aspects of prion and ZIP ancestral ties are explained in anticipation of their full elucidation in subsequent chapters.

1.1 Preamble

Over the past four years we have proposed an evolutionary link between prion genes and ZIP metal ion transporters [1, 2] (Chapter 3), thereby ending the ‘orphan’ status of vertebrate prion genes and merging two protein families that until then had no explicit connection to each other. Due to its significance for a range of fatal human and livestock diseases, the prion protein has been under intense scrutiny for nearly thirty years, during which more than 10,000 articles have been published on the subject. In contrast, very few reports on ZIP proteins appeared before 1996 [3-7] and to this day the literature on ZIP proteins has remained comparably small, with approximately 100 articles published. Many updates on the status quo of prion science have been disseminated throughout the years and some of these reviews provide a wealth of information on the topic [8, 9]. A number of insightful review articles have also covered the gene family of ZIP transporters and their proposed connections to cancer biology [10, 11]. Particular attention has been paid to ZIP4, a ZIP paralog genetically linked to acrodermatitis enteropathica (AE) [12, 13], an inherited childhood zinc-deficiency disorder [14].

The intent of this introductory chapter is to provide an advance overview of the biology of prion proteins and ZIP transporters from the perspective of their phylogenetic relatedness that will act as an introduction to Chapters 2 to 5, which demonstrate the progression of investigations that laid the foundation for the prion-ZIP evolutionary and biochemical connections. In light of both the sheer volume of published work on the prion protein and the considerable diversity of ZIP transporters, ZIPs 5, 6 and 10 will be highlighted. This restricts the literature review to topics that, in our mind, warrant particular scrutiny for ongoing efforts aimed at shedding light on the significance of the relationship between prion and ZIP genes in health and disease. The topics covered are: the (1) diversification, (2) function, (3) spatial distribution and (4) regulation of protein expression of ZIP/prion gene family members, a review of (5) posttranslational modifications and (6) N-terminal repeat sequences and their metal binding specificities.

1.2 Ancestral ties and diversification

Solute carrier (SLC) proteins comprise the second largest group of membrane proteins in humans after G protein-coupled receptors (GPCRs) [15]. The Gene Nomenclature Committee of the Human Genome Organization (HUGO) [16] and other investigators have assigned almost 400 human genes to 47 SLC families based on extensive sequence comparisons and data that suggest these proteins serve as exchangers, coupled transporters or passive transporters [17-19]. Members of a given family have minimally 20-25% amino acid sequence identity to at least one other member within that family and typically share a similar substrate. It has been suggested that some SLC families may share a common evolutionary origin [18]. Two distinct SLC families, SLC30 and SLC39, comprise multi-spanning transmembrane proteins, which contribute to cellular zinc homeostasis by transporting zinc ions across cellular membranes out of and into the cytosol, respectively. The genes encoding members of the ancient SLC39 (solute carrier 39) protein family, historically referred to as the Zrt-, Irt-like (ZIP) family [20], can be found in all organisms investigated, from archaea and eubacteria to eukaryotes [21]. The human SLC39 gene family is comprised of at least fourteen members that have been grouped into four subfamilies [11], and genetic studies for half of mammalian ZIP genes have thus far been reported. Based on multiple alignments of the membrane-spanning domain of ZIP transporters, however, an

3 argument can be made for members of this protein family to fall into two main groups: ZIP transporters that (i) harbor a putative intramembrane metalloproteinase signature sequence, also referred to as the LIV-1 subfamily of ZIP zinc transporters (LZT) [11], and (ii) those which do not. Whereas Archaea and Plantae genomes appear to code for a disproportionate number of non- LIV-1 subfamily members of ZIP transporters, a striking expansion of the LZT branch seems to have occurred in most metazoan lineages and, in particular, during early Chordata speciation, with the result that the genomic complement of LZT genes in tetrapods and teleosts is most pronounced and largely identical [22]. The majority of LZT sequences are equipped with N- terminal domains which are predicted to protrude into the extracellular/luminal space. For a subset of LZT paralogs this ectodomain contains a PrP-like (PL) domain harboring a cysteine- flanked core (CFC), based on sequence similarities and structural predictions [2] (Figure 1.1 and Chapter 3). More specifically, prion gene sequences are most closely related in sequence to one phylogenetic branch within this subset of LZTs comprised of ZIPs 5, 6 and 10 (Table 1.1). Based on the apparent restriction of prion genes to the genomes of chordates, the emergence of the prion gene family from ZIP transporters is likely to have taken place early during Chordata radiation and prior to the divergence of teleost and tetrapod lineages. This conclusion is supported by sequence comparisons of fish and tetrapod prion gene sequences [23-26], comparative genomics and synteny analyses [24, 27-29] (Figure 1.2, and discussed in more detail in Chapter 4).

Figure 1.1: Comparison of PrP and ZIPs 5, 6 and 10. (A) Domain organization and posttranslational modifications. (B) Topology models.

4 Table 1.1: Characteristics of human PrP and ZIP paralogs

MEMBRANE MP IN METAL TISSUE GENE LOCUS PROTEIN CFC AA LOCALIZATION DISEASES REFERENCES ATTACHMENT TM5 SPECIF DISTRIBUTION PRNP/ PRNP 20p13 GPI Yes No 253 Cu, Zn widespread PM CJD, vCJD, GSS, FFI, kuru [28, 30-40] hPrP SPRN/ [28, 33-35, 39, SPRN 10q26.3 GPI No No 151 ND CNS PM unknown hSho 40] PRND/ [28, 32-36, 39, PRND 20p13 GPI Yes No 176 Cu testis PM unknown hDpl 40] PM ([Zn]- SLC39A1 1q21.3 hZIP1 Type III TM No No 324 Zn widespread prostate cancer (correlative) [41-48] dependent) PM ([Zn]- SLC39A2 14q11.2 hZIP2 Type III TM No No 309 Zn liver, prostate, uterus prostate cancer (correlative) [41-46, 48-50] dependent) PM ([Zn]- [41-43, 45, 46, SLC39A3 19p13.3 hZIP3 Type III TM No No 314 Zn blood, breast, prostate, prostate cancer (correlative) dependent) 48, 49] cecum, colon, kidney, liver, AE, pancreatic and hepato- [10, 13, 41-44, SLC39A4 8q24.3 hZIP4 Type III TM Yes Yes 622 Zn PM small intestine, stomach cellular cancers (correlative) 46, 51-57] colon, kidney, liver, PM (baso- [10, 41-43, 45, SLC39A5 12q13.3 hZIP5 Type III TM Yes Yes 540 Zn unknown pancreas, spleen, stomach lateral) 46] hZIP6/ [10, 41-44, 46, SLC39A6 18q12.2 Type III TM Yes Yes 755 Zn Widespread PM breast cancer LIV-1 58, 59] hZIP7/ breast cancer [10, 41-43, 45, SLC39A7 6p21.32 Type III TM No Yes 469 Zn, Mn Widespread ER, Golgi HKE4 (Tamoxifen resistance) 46, 59, 60] hZIP8/ Cd, Fe, Widespread (including in PM, vesicles, breast cancer [10, 41, 42, 46, SLC39A8 4q24 Type III TM Yes Yes 460 BIGM103 Mn, Zn utero) mitochondria (Fulvestrant resistance), CAD 61-64]

SLC39A9 14q24.1 hZIP9 Type III TM No No 307 Zn Widespread ND unknown [41, 42, 46, 65]

[10, 42, 46, 66, SLC39A10 2q32.3 hZIP10 Type III TM Yes Yes 831 Zn Widespread PM breast cancer 67]

SLC39A11 17q24.3 hZIP11 Type III TM No No 335 Zn Widespread ND unknown [41, 42, 46]

SLC39A12 10p12.33 hZIP12 Type III TM Yes Yes 691 Zn Widespread PM asthma [10, 42, 46]

SLC39A13 11p11.2 hZIP13 Type III TM Unclear Yes 371 Zn Widespread Golgi connective tissue disorders [10, 42, 46, 68]

Zn, Cd, asthma, colorectal cancer, [10, 42, 46, 69- SLC39A14 8p21.3 hZIP14 Type III TM Yes Yes 492 Widespread PM Fe inflammatory hypozincemia 73]

AA: amino acids; AE: acrodermatitis enteropathica; CAD: coronary artery disease; CFC: cysteine-flanked core; CJD: Creutzfeldt-Jakob disease; CNS: central nervous system; ER: endoplasmic reticulum; FFI: fatal familial insomnia; GPI: glycophosphatidylinositol; GSS: Gerstmann-Sträussler-Scheinker syndrome; MP: metalloprotease; ND: not determined; PM: plasma membrane; TM: transmembrane; vCJD: variant Creutzfeldt-Jakob disease

Figure 1.2: Simplified phylogenetic tree and table depicting both the wide distribution of ZIP sequences in most organisms and the restricted distribution of prion genes to the Chordata lineage. The paralog distribution was deduced from alignments published by the Wellcome Trust Sanger Institute (TreeFam, http://www.treefam.org) or was determined by aligning representative ZIP and prion protein sequences to genomic sequences. The number of paralogs containing a cysteine-flanked core (CFC) domain is indicated in brackets.

1.3 Function – novel links to EMT and cancer

In light of the wealth of biochemical, genetic, transgenic and structural data available for the prion protein, its elusive function is not only surprising but represents a formidable challenge to the field [8, 74, 75]. Over the past twenty years, several prnp-knockout mouse lines have been generated and subjected to intense scrutiny for phenotypes. The persistence with these studies has led to multiple interesting observations [76, 77] and revealed a highly penetrant myelin maintenance defect in the peripheral nerves of PrP-deficient mice [78]. Other functions ascribed to PrPC include a role in olfactory behavior [79], a positive influence on neuronal progenitor cell

6 proliferation [80] and participation in the self-renewal of hematopoietic stem cells [81]. Thus, at this time there is no lack of biological systems to which PrPC has been tied; the still-missing facet, however, is information about how precisely PrPC exerts its role in these diverse biological systems at the molecular level. A parallel and orthogonal quest for insights into possible functions has been based on a similarly intense search for interacting proteins. Because this body of work has been reviewed before [40, 75, 82-84], we will limit our discussion to aspects of the biology of PrP that may be derived from its connection to ZIP proteins.

The mere observation that ZIP genes appear to exist in the genomes of all organisms is indicative of an indispensable role for cell survival. A close inspection of amino acid sequences of individual ZIP transporters reveals a series of hydrophobic stretches of approximately 20-25 amino acids and the presence of multiple HX-repeat motifs. These features suggested early on that the ZIP proteins acquire a type III membrane topology and may be equipped with the ability to bind zinc or other divalent cations, respectively. Although in the case of the prion gene knockout, investigations have often been inconclusive, gene-targeting approaches remain the gold standard for the functional assignment of genes [85]. With S. cerevisiae being the most systematically investigated eukaryotic model organism, a look at its ZIP genes and their presumed functions may be informative. The S. cerevisiae genome codes for five ZIP genes, namely ZRT1, ZRT2, ZRT3, ATX2 and YKE4 [86], with YKE4 being the only yeast paralog that can be classified as belonging to the LZT subfamily. Individual yeast ZIP transporter genes can be deleted without affecting viability in standard media [5, 87, 88] and even a ZRT1/ZRT2 double knockout has been reported to be viable [5], suggesting that some functional redundancy may exist amongst yeast ZIP paralogs. Alternatively, an essential quality of individual ZIP members may only be teased out if cells are subjected to the right kind of stress. Indeed, slow growth can be observed when ZRT1- or YKE4-knockout cells are grown in media containing limiting or excessive amounts of zinc, respectively [4, 89]. Taken together, the above data strongly hinted at a role of these proteins in cellular zinc homeostasis, a functional assignment which was further consolidated by the subsequent demonstration that mutations in human ZIP4 can cause the hereditary zinc uptake deficiency AE. Only a few mammalian ZIP genes have been subjected to gene targeting investigations to date. The fourteen human SLC39 paralogs have been mapped to 12 different chromosomes, with chromosome 8 coding for SLC39A4 and SLC39A14, and chromosome 14 coding for the distantly-related SLC39A2 and SLC39A9 genes. SLC39A5, 6

7 and 10 are encoded on the long arms of chromosomes 12, 18 and 2, respectively. No ZIP paralog is found in the vicinity of the human PRNP-PRND gene complex [90] on chromosomal arm 20p or SPRN encoded by its gene locus on 10q (Table 1.1, and see Chapter 4 for a more detailed discussion). ZIPs 1, 2 and 3 have been demonstrated to be nonessential in mice when zinc is replete but the knockout of these genes dramatically compromised development when zinc was limiting [48, 91]. Homozygous ZIP4-knockout mice, generated by crossing ZIP4-heterozygous mice, display severe malformations and die during embryogenesis on day 10, a phenotype that cannot be rescued by feeding pregnant mothers a zinc-enriched diet [92]. Haploinsufficiency of ZIP4 creates pleiotropic phenotypes that may range from retarded growth to domed head shapes or hydrocephali to missing eyes. Whereas the underlying molecular mechanisms that mediate these gross morphological abnormalities are not currently understood, the severity of these phenotypes correlates inversely with dietary zinc levels during pregnancy, consistent with the proposed role of ZIP4 in intestinal zinc uptake [92]. Cumulatively, the above studies corroborated the conclusion that the primary cellular function of ZIP transporters might be the influx of zinc into the cytosol. What drives this import activity? Because zinc exists primarily in a protein-bound state within the cell, with cytosolic levels of free zinc estimated in the nanomolar range, a gradient of unbound zinc exists across the plasma membrane which might be sufficient to drive passive zinc uptake by facilitated diffusion. Indeed, at least for human ZIP2, available data suggest that its zinc import activity is neither energy-dependent, nor requires existing K+ or Na+ gradients across the plasma membrane [21]. Interestingly, however, the - authors observed an increase in zinc uptake in response to treatment with HCO3 , arguing that 2+ - this process may be facilitated by a Zn -HCO3 symport mechanism (as also reported for ZIP8 [93]). These data were reminiscent of an earlier report which documented a profound bicarbonate-dependent increase in zinc uptake in fibroblasts harvested from normal but not AE- afflicted individuals [94] but conflict with a recent report on a functionally reconstituted proteobacterial ZIP ortholog that argues against a symport mechanism [95].

A high-resolution structure for members of the SLC39 family is needed to elucidate mechanistic details of their transport activity. In the absence of such data, a look at SLC30 transporters, also known as cation diffusion facilitators (CDFs) or ZnTs, zinc transporters which do not appear to be related in sequence but share with SLC39 family members the ability to transport zinc across cellular membranes, may offer some insights into possible implementations

8 of such a transport function. Because CDFs transport zinc against the physiological zinc gradient out of the cytosol, their activity cannot (despite their name) be based on facilitated diffusion (Figure 1.3). The arguably best-understood member of this family is a protein referred to as YiiP, one of two CDFs coded by the E. coli genome [96]. YiiP exists as a homodimer of two 33kDa subunits, each anchored in the membrane by six TM domains and carrying a carboxy- terminal domain (CTD) that protrudes into the cytoplasm. YiiP has been shown to employ a mechanism that relies on the antiport of protons to drive the export of Zn2+ from native membrane vesicles [97]. High resolution X-ray scattering and NMR data for YiiP [98] and its molecular cousin CzrB in Thermus thermophilus [99, 100] revealed that CTDs acquire a metallochaperone fold and contribute to homodimerization by coordinating the binding of multiple zinc ions at the protein-protein interface. Site-directed fluorescence energy transfer (FRET) measurements suggest that zinc binding triggers hinge movements of the CTDs relative to the six-helix TM bundle and leads to subtle rearrangements of the transmembrane helices that may alter zinc coordination in the active site and thus promote transport [101]. It is likely that the ectodomain in the LZT subfamily of SLC39 members may mimic aspects of the CTD role in SLC30 members. However, it is to be expected that the influx activity of SLC39 family members may not recapitulate all facets of events proposed for SLC30 exporters because free versus protein-bound divalent cations, available in the extracellular and cytosolic environments, respectively, may pose different challenges for metal capture. Aside from conspicuous differences in the number of helices of transmembrane domains, protruding domains of the two protein families may have developed distinct folds as an adaptation to distinct metal availabilities and redox environments. Thus, the metallochaperone-like fold of the CTD may be an absolute requirement for wresting a divalent cation from a cytosolic metal-binding protein but may be of lesser relevance for a zinc importer which accesses the extracellular pool of free divalent cations.

Figure 1.3: Cartoon comparing properties of ZIP and CDF (ZnT) zinc ion transporters. (A) ZIP-mediated passive zinc import. (B) Antiport model of CDF (ZnT).

Although it is currently unknown whether ZIP proteins, like SLC30 transporters, operate as functional dimers (see Chapters 5 and 6), evidence is accumulating that proteins harboring a PrP-like domain are capable of engaging in direct interactions with each other. Genetically, a direct interaction between PrPC and Doppel (Dpl) or Dpl-like N-terminally truncated mutants of PrP (ΔPrP) has been postulated based on the ability of PrPC to counteract the toxicity associated with cerebellar expression of either Dpl or ΔPrP [40]. However, evidence of biochemical interactions between Dpl and PrPC had been lacking until recently when two independent studies reported the co-immunoprecipitation of Dpl and PrPC [102, 103] (see Chapter 2), suggesting that the two proteins either interact directly with each other or exist in close spatial proximity. Although PrPC may not exist at the cell membrane as a constitutive dimer, dimeric forms of PrPC, the formation of which appears to be mediated by the well-conserved hydrophobic tract region, have been observed in cell culture and shown to be required for PrP to exert its protective role [104, 105]. Consistent with this observation, PrP deletion mutants lacking the hydrophobic domain (and therefore possibly impaired in dimerization) are highly neurotoxic in vivo [106, 107]. Whereas hydrophobic tract-mediated dimerization may also explain the ability of PrPC to interact with Sho [102, 108], additional elements within the PrP-like domain may have an

10 intrinsic ability to contribute to dimerization. This is also suggested by the aforementioned data documenting the ability of PrPC to bind to Dpl, ZIP6 and ZIP10 (Chapter 2), proteins that do not contain a PrP/Sho-like hydrophobic tract.

With zinc being increasingly recognized as a second messenger [109], research into cellular zinc homeostasis and zinc-related diseases has experienced a renaissance in recent years. Beyond their contribution to the overall health of a cell through maintaining stable zinc levels, proteins involved in zinc homeostasis are now thought to influence all major cell-fate decisions [41, 110, 111]. At the molecular level, this influence is primarily mediated by zinc providing structural stability (e.g., in zinc-finger-motif-containing transcriptional regulators) or acting as a catalytic cofactor (e.g., in superoxide dismutase, carbonic anhydrase and metalloenzymes). Recently, free-zinc waves have been reported to regulate the activity of receptor protein tyrosine phosphatase-β [112], suggesting a clear pathway for the second-messenger function of zinc. From a prion research perspective, gene-targeting experiments that may shed light on the specific functions of ZIPs 5, 6 or 10 in rodents would seem to be most relevant but are currently lacking. At present, the most illuminating data in this regard may originate from knockout studies of ZIP orthologs conducted in fruitflies. The D. melanogaster genome has been reported to code for a total of nine ZIP orthologs [11, 113]. While the exact orthologous relationships of fruitfly and human ZIPs remain to be determined, a relatively detailed functional characterization has been conducted for a gene referred to as fear-of-intimacy (foi) that exhibits the strongest sequence similarity to the branch of mammalian ZIPs 5, 6 and 10 [114, 115]. The naming of this gene originally referred to an impairment seen in FOI-deficient fruitfly embryos which prevented a small group of specialized cells from coalescing following morphogenetic rearrangements and cell migration processes underlying gonad and trachea formations [51, 114, 116, 117]. These developmental processes themselves are reminiscent of epithelial-to-mesenchymal transitions (EMT) occurring during gastrulation. The connection to EMT is intriguing in this context as independent developmental investigations in zebrafish implicated both ZIP6 [118] and the prion protein [119] in morphogenetic cell movements during zebrafish gastrulation. Finally, it appears as if lessons learned from these model organisms also apply, at least partially, to their mammalian cousins. Aggressive forms of cancers with a propensity to turn malignant have long been proposed to hijack an already-in-place molecular EMT program for their transition to invasiveness [120]. ZIP6 and ZIP10 have not only been found to be upregulated in certain

11 cancers [10, 121, 122], but they have also been linked repeatedly to a role during the EMT of a subset of human carcinomas in recent years [67, 123-126].

1.4 Expression – crosstalk and division of labor

The five yeast ZIP transporters localize to the plasma membrane (ZRTs 1 and 2) [5], endoplasmic reticulum (ER) (YKE4) [89], Golgi apparatus (ATX2) [87] and vacuolar membrane (ZRT3) [88], suggesting that cells may gain a selective advantage by dispatching specialized ZIP transporters to distinct locales. However, the assignment of more than one ZIP transporter to the yeast plasma membrane requires a different explanation: it has been observed that ZRT1 and ZRT2 exhibit dramatically different zinc uptake affinities. Whereas ZRT1 is active in zinc- limited cells as a high-affinity zinc uptake transporter [4], ZRT2 is active in zinc-replete cells [5]. Current insights into the tissue expression of human ZIP (hZIP) genes reflect a series of independent observations rather than systematic analyses. Available data suggest widespread expression of most hZIP paralogs except for hZIPs 2, 3, 4 and 5. Of the three ZIP paralogs most closely related to prion genes, ZIP5 is primarily expressed in visceral organs (kidney, liver, colon, pancreas), ZIP6 levels are highest in hormonal tissues (breast, prostate, pituitary gland) but can also be detected in the brain [11, 127], and ZIP10 is predominantly expressed in the brain and spinal cord [10]. Reminiscent of the situation in yeast, the majority of hZIP transporters appear to be targeted to the plasma membrane, and only a few select ZIPs, for example hZIP7 and hZIP9, seem to localize predominantly to membranes surrounding the ER [128], Golgi [129], or the trans-Golgi network [65]. Whereas all eukaryotic cells appear to express multiple ZIP paralogs, relatively little is known about the biology that determines the subset of ZIP paralogs expressed in a given cell or tissue. Observations of reciprocal expression levels and complementary targeting of apical and basolateral cell surfaces have been reported for ZIP4 and ZIP5, suggesting that an intricate balance and crosstalk may exist amongst individual ZIP transporters in response to extracellular stimuli [130, 131].

In light of their emerging role in the cellular homeostasis of zinc and other divalent cations, it is not surprising that the expression of individual ZIPs can be shown to respond to zinc levels. For example, in yeast the expression of ZRT transporters correlates inversely with zinc levels in the medium [132]. More specifically, the transcription of ZRT1 which exhibits higher affinity to zinc than its paralog ZRT2 [5] is increased when zinc levels in the medium are low,

12 but is subject to rapid ubiquitin-mediated degradation once zinc is replete [133, 134]. However, responsiveness to zinc levels is not a generic characteristic of ZIP transporters, as no such correlation has been observed for the yeast paralog ATX2 (which appears to primarily function in cellular manganese homeostasis [87]). A systematic analysis of zinc transporter mRNA levels in a range of tissues with varying levels of zinc in the aquatic environment has been reported for zebrafish [22]. The study uncovered complex relationships between aquatic zinc and mRNA expression levels for individual ZIPs, with some ZIP genes being unresponsive to varying zinc levels (e.g., ZIP6) and others responding profoundly but in an inconsistent manner in a diverse spectrum of tissues investigated (e.g., ZIPs 1, 3, 4, 7 and 10). Similar patterns of zinc responsiveness appear to exist for ZIP transporters expressed in mammalian cells [135]. For example, ZIP10 levels in rat renal brush border cell membrane have been shown to correlate with extracellular zinc levels and zinc uptake has been reported to be time-, temperature- and ZIP10- concentration-dependent in this cell model [66]. For ZIP5, it has been proposed that cells might increase responsiveness to zinc levels by employing a translational stalling mechanism that attaches ZIP5 mRNA to polysomes when zinc is limiting [136].

Although the highest levels of PrPC expression are observed in the mammalian central nervous system, substantial levels are also observed in cardiac muscle and lung, and lower levels can be found in other tissues including the intestine, limb muscles and spleen, and on lymphocytes. As such, the expression profile of PrP may more closely resemble the expression of ZIP6 and ZIP10 than the expression profile of ZIP5; however, cell-specific characterizations of expression levels are needed to dissect whether levels of PrP and individual ZIP proteins are correlated. Like many ZIPs, the expression of PrP appears to respond to the presence of metals in complex ways [137, 138]. For example, copper in the medium was shown to decrease PrPC expression in immortalized murine GN11 neurons [139], but translated into increases in PrPC expression in human HeLa and mouse N2a cells [140], mouse fibroblast cells [141] and primary rat hippocampal and cortical neurons [142]. Based on the data reviewed here, it is likely that PrP may encounter distinct combinations of multiple ZIP transporters in the wide range of cell types in which it is known to be expressed. It will be of interest to observe whether individual ZIPs and PrP can reciprocally influence each other’s expression levels. This scenario would not be surprising given the complementary nature of PrP and Sho expression observed in select areas of the mouse brain [143].

1.5 Signaling – from Tinman to Snail and beyond

What can insights into signaling pathways operating upstream or downstream of ZIPs 5, 6 or 10 tell us about PrP? Very little is known about the regulation of expression of ZIP10. Given the zinc-responsiveness of expression observed for many zinc transporters, it is not surprising that metal response elements (MREs) exist in regulatory introns of ZIP10 [144, 145]. Although no putative roles for MREs have been reported for ZIP5, evidence for microRNA-mediated 3’-UTR regulatory elements has recently been documented [146]. The reported observations for ZIP10 are similar to the prion gene for which it has been shown that copper-replete conditions can promote its expression in certain cell types, including human fibroblast cells, through the activity of metal-responsive transcription factor-1 (MTF-1) [147]. Additional promoter elements may also come into play; in the case of PrP, for example, a well-known pathway that involves the ataxia telangiectasia mutated (ATM) protein kinase and components of the MAP kinase pathway culminates in the transcription factor SP1 occupying its corresponding binding sites on the PrP promoter and causing an upregulation of PrP expression in response to an increase in intracellular copper [140]. The 5’-upstream region of the human PRNP gene contains putative binding sites for many additional transcription factors including AP-1 (FOS) and AP-2 (TFAP2A), p53 (TP53), MyoD (MYOD1) and NKX2-5 [148, 149]. Comparably little is known about promoter elements within ZIP genes. While compelling data have established a role for STAT3 acting upstream of ZIP6 in zebrafish [118], it is currently not known whether STAT3 acts immediately upstream of ZIP6, for example as an activator of its transcription, or whether signaling mediators are positioned between STAT3 and ZIP6 (Figure 1.4). In the fruitfly, FOI is transcriptionally activated by Tinman [116, 150], the D. melanogaster ortholog to the mammalian NKX2-5 transcription factor [151] for which, as mentioned, consensus binding elements have also been described in the prion promoter.

Figure 1.4: ZIP-dependent signaling upstream of E-cadherin expression in three biological paradigms.

The similarity of gastrulation phenotypes linked to PrP deficiency and ZIP6 (LIV-1) inactivation in zebrafish may be a particularly rewarding angle from which to explore overlaps in signaling pathways downstream of ZIP6 and/or PrP [118, 119]. It is noteworthy that both publications tied PrP or ZIP6 functions to the regulation of cell adherence and the E-cadherin protein. How does ZIP6 expression influence E-cadherin levels? A first indication of what might be occurring is the observation of a ZIP6-dependent nuclear translocation of the transcriptional repressor Snail [118], a master regulator of EMT [152]. As a zinc finger protein, Snail requires zinc as a structural cofactor for binding to elements present within the E-cadherin promoter. Consequently, the possibility arises that a local increase in zinc levels, mediated by ZIP6- dependent zinc import, charges cytoplasmic Snail with zinc and thus triggers its translocation into the nucleus. Consistent with observations in zebrafish, the fruitfly ZIP ortholog FOI acts upstream of the Snail family member Escargot (Esg) to exert the aforementioned tracheal branch fusion process [153]. Surprisingly, whereas both the tracheal and gonadal FOI-mediated phenotypes in the fruitfly depend on E-cadherin expression, the gonad coalescence phenotype appears to be independent of Esg. In light of the data reviewed herein, the link between ZIP6 and cancer metastasis may be explained by a body of literature that ties the molecular biology revolving around STAT3, EMT and Snail [154] to cancer malignancy [120, 155-158]. Cells with ectopic Snail expression adopt a fibroblastoid phenotype and acquire tumorigenic and invasive properties [152]. In line with this model, recent reports document that inhibition of ZIP6 in both the cervical cancer-derived HeLa cell model and pancreatic cancer cells causes the deactivation

15 of Snail and interferes with their metastatic features [126, 159]. However, it is already emerging that the underlying phenomena are considerably more complex than we currently understand and likely differ amongst experimental paradigms. Thus, additional stimuli, such as the exposure of cells to bacterial lipopolysaccharides, have been shown to feed into the signaling pathways that control ZIP6 expression [160]. Furthermore, E-cadherin levels correlate inversely with ZIP6 levels in some paradigms but in others appear to follow ZIP6 levels directly as, for example, in breast tumor cells (T47D or MCF-7) [124, 161] or the previously-mentioned FOI-dependent gonad formation phenotype in the fruitfly. Interestingly, ZIP6 also appears to have opposite effects when compared to PrP on E-cadherin levels in the zebrafish gastrulation paradigm. Whereas the knockdown of PrP appears to cause a destabilization of the mature E-cadherin pool, transcriptional inactivation of ZIP6 is linked to E-cadherin stabilization. The observed influence of PrP on cadherin may well be independent of ZIP zinc transport activities and instead be based on PrP’s ability to influence cellular copper and zinc homeostasis [162]. Alternatively, PrP may exert its influence on EMT through binding to NCAM [163], an interaction known to facilitate the recruitment of NCAM into lipid rafts and to promote activation of Fyn kinase [164]. Fyn has been shown to phosphorylate focal adhesion kinase and thus promote the assembly of integrin- mediated focal adhesions, cell spreading and EMT [165]. Ablation of NCAM inhibits EMT and overexpression of NCAM correlates with tumor invasion [166]. Finally, an intriguing possibility is that PrP may directly influence the signaling events outlined above based on its capacity to bind to ZIP6, consistent with our data documenting co-affinity purification of PrP and ZIP6 following in vivo crosslinking with formaldehyde [102] (Chapter 2). The latter explanation would suggest that the influence of PrP on ZIP6 is that of a negative regulator.

1.6 Post-translational modifications

The existence of a highly-conserved disulfide bridge spanning alpha helices B and C within the globular domain of the mammalian prion protein is well-documented [167]. Not surprisingly, Sho, devoid of the corresponding domain and of cysteine residues in general, lacks a disulfide bridge. Dpl, on the other hand, has been shown to harbor a second disulfide bridge serving to further stabilize its globular domain [168, 169]. The existence of disulfide bridges within ZIP proteins has not been rigorously documented to date. However, except for human ZIP11, whose

16 amino acid sequence is devoid of cysteine residues, all human ZIP paralogs contain multiple cysteines. Most of these are scattered throughout the respective primary structures with little conservation across ZIP paralogs. However, two pairs of cysteines conserved in a subset of paralogs within the LIV-1 branch of ZIP transporters are the exception. One of the pairs (flanking the CFC), with regard to its position within the predicted globular domain in these proteins, matches the location of the disulfide bridge in PrP [2] (Chapter 3). The remaining two conserved cysteines are found within the CPALLY and CHELPHELGD motifs in the predicted extracellular globular domain and the putative transmembrane domain V, respectively, and have been hypothesized to interact in regulating the movement of zinc across the membrane [10].

Members of the mammalian prion gene family are post-translationally modified on up to two N-glycan acceptor sites [143, 169]. However, only one of the two N-glycosylation motifs found in each of PrP and Dpl is conserved between these proteins. This site is positioned two residues C-terminal to the first cysteine residue (which participates in the formation of the aforementioned conserved disulfide bridge). Whereas a subset of human ZIP paralogs contains ‘Nx(T/S)’ consensus acceptor site motifs for N-glycosylation primarily on their N-terminal ectodomains, the occupancy of these sites has not been experimentally determined. However, indirect evidence for N-glycosylation of these transporters has been obtained for ZIP4 [170], ZIP6 [127], ZIP8 [63, 171] and the closely-related transporter ZIP14 [172, 173] by documenting faster migration of denatured proteins via SDS-PAGE following the global removal of N-linked sugars by treatment of proteins with N-glycosidase F (PNGase F) or after culturing cells in the presence of tunicamycin. A similar gel shift could not be observed for ZIP1 and ZIP3 following N-glycosidase F treatment [174], adding weight to the impression that N-glycosylation may be restricted to transporters with extended N-terminal domains. It will be of interest to determine whether ZIPs 5, 6 and 10, which contain a conserved ‘Nx(T/S)’ motif in the same relative position as the conserved N-glycosylation site shared between PrP and Dpl, are in fact N- glycosylated at this acceptor site (Figure 1.1, and see Chapter 5).

It has repeatedly been shown that PrP can be endoproteolytically processed through a so- called alpha-site cleavage of the peptide bond connecting amino acids 110 and 111 (corresponding to human PrP), causing the release of an N-terminal fragment (N1) into the medium [175-181]. ADAM10 and ADAM17, the latter also known as TACE, have been proposed to catalyze this cleavage in response to phorbol ester stimulation [182-184] in a late

17 compartment of the secretory pathway [185]. Interestingly, the same stimuli and metalloendoproteinases have also been implicated [186] in a second PrPC cleavage activity (between amino acids 228 and 229), leaving behind a very short three-amino-acid stub and causing the shedding of a soluble ectodomain form of PrPC from the cell surface [187, 188]. The entire ectodomain has been observed to be released when PrPC is subjected to a distinct phospholipase activity which can cleave the glycosylphosphatidylinositol (GPI) anchor directly [189]. Finally, a scissile bond within PrP, termed the beta-cleavage site [190], has been identified in proximity to the end of the octapeptide repeat region [175, 178]. At least in a subset of experimental paradigms, cleavage at this site, leading to the secretion of a shorter N-terminal fragment (N2) into the medium, appears to be directly mediated by free reactive oxygen radicals [191, 192]. Cleavage at or near this site may further be mediated by endolysosomal cathepsin- like proteases [193]. Little is known about proteolytic cleavages within ZIP transporters. However, it has recently been shown that the N-terminal domain of ZIP4 is shed through an endoproteolytic event when cells are grown for prolonged durations in zinc-depleted cell culture media, and in vivo in the visceral yolk sac and the intestine as a consequence of zinc deficiency [194]. Levels of cleaved ZIP4 were reduced in the presence of endocytosis inhibitors, possibly indicating that the N-terminally truncated ZIP4 protein can be recycled to the plasma membrane. An additional regulatory sequence has been mapped to a histidine-rich motif within the intracellular loop connecting transmembrane domains III and IV of human ZIP4. Ubiquitination of this domain has been proposed to regulate human ZIP4 endocytosis and its priming for proteasomal degradation [195].

No systematic endoproteolytic analyses have been undertaken for other ZIP proteins. Unpublished results from our laboratory suggest, however, that ZIPs 5, 6 and 10 are subject to distinct and complex internal cleavages (see Chapter 5 for a more discussion on ZIP10 endoproteolytic cleavages). As mentioned above, members of the LIV-1 subfamily of ZIP transporters contain a ‘HEXXH’ amino acid sequence motif (‘CHELPHELGD’) within their predicted transmembrane domain V (Figure 1.1) [196]. The presence of this motif, best known for its occurrence in the active center of metalloproteinases (MPs), has provoked the tantalizing hypothesis that these proteins may not only bind and import zinc and other divalent cations into the cytosol, but possibly serve a second proteolytic function [11, 127]. When present in MPs, this motif is typically embedded in extended active site signature sequence patterns shared amongst

18 closely related proteins [197, 198]. Thus, in addition to the two ion-coordinating histidine residues and the catalytic acid embedded in the pentapeptide core motif ‘HEXXH’, a glutamine, asparagine or histidine residue contributed by different structural elements work together with the help of a tetrahedrally-coordinated zinc (or rarely an alternative metal) to polarize the metal- bound water molecule for its nucleophilic attack on a nearby peptide bond. Detailed scrutiny of the sequence environment of the consensus ‘HEXXH’ motif within the LIV-1 branch of ZIP proteins revealed additional similarities to metzincin and deformylase groups of zinc MPs [127, 196] but did not meet discriminative criteria for assignment to any known MP subfamily. This observation is perhaps not surprising given the topological restraints of intramembrane proteolysis, also exemplified by X-ray crystallographic data of a Site-2 protease (S2P) [199]. This family of intramembrane MPs appears to coordinate zinc based on an alpha-helical membrane-embedded ‘HEXXH’ pentapeptide and a highly-conserved aspartic acid that is distant in sequence but spatially close in the tertiary structure. Thus, whether members of the LIV-1 branch of the ZIP protein family (which, except for the shared ‘HEXXH’ motif, bear no apparent resemblance to S2P proteases) contain a novel MP active site awaits further investigation.

Whereas PrP has been shown to be post-translationally modified by the addition of a GPI anchor, ZIP transporters are multi-spanning transmembrane proteins. The orientation and distance of the CFC domain to the respective membrane attachment sites is similar in PrP and ZIP proteins. Furthermore, the amino acid sequences of the first transmembrane domain in ZIPs and the GPI anchor attachment signal in PrP are similar. These observations led to the inference that the signal peptide for the attachment of a GPI anchor present in PrP may have evolutionarily originated from the first transmembrane domain of a ZIP ancestral molecule [2] (Chapter 3). This conclusion was supported by previous reports documenting that small changes within or in proximity to the first transmembrane domain can be sufficient to cause a shift to a GPI-anchored mode of membrane attachment [200-202].

1.7 Binding of divalent cations – a common denominator

Although, beginning with metal selectivity studies in bacterial ZIP orthologues [203-205], the predominant preference of ZIP proteins for zinc transport has been well-established, the metal binding characteristics of these proteins are complex. Thus, whereas good agreement amongst

19 available studies exists to suggest that ZIPs 5, 6 and 10 are primarily responsible for the cytosolic import of zinc, a few select ZIP transporters have long been known to contribute to the transport of different metals [3]. In particular, ZIP8 and ZIP14 have been shown to transport, in addition to zinc [172, 206], cadmium, manganese and iron (Table 1.1) [62, 173, 207, 208]. These observations suggest that the core molecular architecture of ZIP transporters is intrinsically adaptable toward the transport of a range of divalent metals. It further suggests that around the time in evolution when the most pronounced radiation of the LZT branch of ZIP genes occurred, i.e., prior to the divergence of teleosts and tetrapods [2, 86], the metal transport selectivity of ancient LZT proteins may have been less well-developed. The structural components which contribute to metal-binding and specificity of individual ZIP transporters are currently poorly understood. Various authors have noted the presence of histidine-rich repeat motifs found within N-terminal extracellular domains and in loop sequences connecting putative TM domains 2-3 and 3-4 of many ZIP transporters [196, 209]. Most often, these histidine-rich motifs are of the type [HX]n (or derivatives of this motif), whereby ‘X’ represents any amino acid and ‘n’ may range from 2 to ~30. Consistent with a possible role in metal binding, similar motifs can also be found in many members of the SLC30 family of zinc exporters [196]. However, the absence of some or all of these histidine-rich stretches in individual ZIP transporter paralogs suggests that these sequences are not essential for transport but instead may serve other functions, for example, in the sensing or initial capture of metal ions and/or interaction with other proteins. In the IRT1 ZIP transporter from thale cress (Arabidopsis thaliana), a specific histidine residue and two nearby acidic residues were identified by alanine scanning mutagenesis to play a role in the metal specificity of this transporter [203]. Two highly conserved histidine residues within putative transmembrane domains IV and V have been proposed to line the metal transport pathway and serve as parts of an intramembraneous metal binding site [209]. Consistent with the notion that the histidine residue in TM domain V may contribute to metal transport specificity, this residue is replaced with glutamic acid in ZIPs 8 and 14, two mammalian paralogs shown to exhibit relatively broad metal ion transport specificities.

The metal binding characteristics of mammalian prion proteins have been under intense scrutiny ever since the protein sequence of PrPC was shown to harbor four or five copies of a histidine-containing octapeptide motif (‘PHGGGWGQ’) in its N-terminal domain (reviewed in [137, 138, 210]). Although considerable discord exists in the literature with regard to which

20 metals bind to which sites within PrPC, the predominant view has been that mammalian octapeptide repeat sequences preferentially bind copper and have lesser affinities to zinc, manganese, iron and other divalent metals. However, it has repeatedly been pointed out that binding of PrP to zinc, albeit shown to occur with lower affinity in vitro, may become significant in an in vivo environment where zinc levels may greatly exceed local copper concentrations (reviewed in [162]). Consistent with this notion, PrPC appears to respond equally well to zinc or copper but not to other divalent cations in both a well-known endocytosis paradigm [211, 212] and an assay which can monitor the ability of unilamellar-embedded N-terminal fragments of PrP to engage in direct interactions [213]. A more detailed understanding of the metal-binding characteristics of the octapeptide repeat domain and an adjacent metal binding site involving human PrP histidine residues 96 and 111 has emerged in recent years, in particular from structural characterizations of bacterially-expressed prion proteins or synthetic peptides [214, 215]. A refined model which takes into account physiological zinc and copper levels posits that both zinc and copper may bind to the octarepeat domain when copper levels are low. As copper levels increase, copper may displace residual PrP-bound zinc, which in turn may continue to influence the copper coordination mode of the prion protein [216]. A more detailed discussion on PrPC’s metal binding capacity follows in Chapter 5. Like PrPC, Dpl can also bind copper ions, albeit in the conspicuous absence of histidine-containing octarepeat motifs [217, 218]. While Dpl appears to form complexes with zinc transport proteins [102], it remains to be established whether zinc plays a role in these interactions or whether Dpl can capture zinc ions independently. Investigations into a possible involvement of Dpl in cellular zinc homeostasis may shed light on a male sterility phenotype linked to Dpl function [219, 220], a research direction supported by a literature that ties cellular zinc biology to sperm cell proliferation and motility [221].

Much less is known about the metal binding specificities of prion protein orthologs found in teleosts. It has been noted before that little sequence conservation exists in N-terminal repeat sequences between tetrapod and teleost prion proteins [24]. Significantly, the tandem repeat region in zebrafish PrP-related2 (annotated as PrP-related3 in some databases) has been shown to harbor histidine clusters of the type [HX]n seen in ZIP proteins that can efficiently bind to zinc [222, 223]. Thus, in both ZIP transporters and prion genes there is evidence for a parallel

21 divergent evolution of N-terminal sequences that may have created distinct specializations in metal binding preferences.

1.8 Conclusions

This section aimed to provide an anticipatory introduction to the next four chapters, which will demonstrate the presence of ZIP6 and ZIP10 in the mouse PrPC interactome (Chapter 2), the discovery of the evolutionary origin of prions from LIV-1 ZIP transpters (Chapter 3), the details of the mechanism of evolution (Chapter 4) and investigations into the biochemistry of a model LIV-1 ZIP protein (ZIP10) vis-à-vis its shared biology with PrPC in a mouse neuroblastoma cell line and prion-infected mouse brains (Chapter 5). Finally, some thoughts on future directions in this avenue of research are presented in Chapter 6.

Chapter 2 Interactome Analyses Identify Ties of PrPC and Its Mammalian Paralogs to Oligomannosidic N-Glycans and Endoplasmic Reticulum-Derived Chaperones

Please note that all parts of this chapter, except parts of the Introduction and Figure 2.9, were published in the following article [102]: Joel C. Watts*, Hairu Huo*, Yu Bai*, Sepehr Ehsani*, Amy Hye Won Jeon, Tujin Shi, Nathalie Daude, Agnes Lau, Rebecca Young, Lei Xu, George A. Carlson, David Williams, David Westaway, Gerold Schmitt-Ulms (2009) Interactome Analyses Identify Ties of PrPC and Its Mammalian Paralogs to Oligomannosidic N-Glycans and Endoplasmic Reticulum-Derived Chaperones. PLoS Pathog 5(10):e1000608. [* equal contribution]

Candidate’s role: Contributed to bioinformatic interactome analyses, biochemical validation, manuscript assembly and editing

Summary: A quantitative investigation of the PrPC interactome was conducted in a cell culture model permissive to prion replication. To facilitate recognition of relevant interactors, the study was extended to Doppel (Prnd) and Shadoo (Sprn), two mammalian PrPC paralogs. Interestingly, this work not only established a similar physiological environment for the three prion protein family members in neuroblastoma cells, but also suggested direct interactions amongst them. Furthermore, multiple interactions between PrPC and the neural cell adhesion molecule, the laminin receptor precursor, Na/K ATPases and protein disulfide isomerases (PDI) were confirmed, thereby reconciling previously separate findings. Subsequent validation experiments established that interactions of PrPC with PDIs may extend beyond the endoplasmic reticulum and may play a hitherto unrecognized role in the accumulation of PrPSc. A simple hypothesis is presented which accounts for the majority of interactions observed in uninfected cells and suggests that PrPC organizes its molecular environment on account of its ability to bind to adhesion molecules harboring immunoglobulin-like domains, which in turn recognize oligomannose-bearing membrane proteins.

2.1 Introduction

As alluded to in the previous chapter, prion diseases are fatal neurodegenerative diseases which can affect a relatively broad range of host organisms including humans, sheep, cattle and deer. The normal cellular prion protein, denoted PrPC, and coded for by the prion gene (prnp), is found in most cell types within the body. In disease, this protein undergoes a structural transition to its disease-causing scrapie form (PrPSc) with profoundly altered physicochemical properties [9]. The accumulation of PrPSc is toxic to cells and may eventually lead to widespread cell death that is characteristically accompanied by a spongiform degeneration of the brains of afflicted individuals. Despite a wealth of data on the evolutionary conservation, cellular localization, structure, molecular environment and metal-binding properties of PrPC, its precise cellular functions are still debated [1]. PrPC is a relatively small protein which assembles into an unstructured N-terminal domain and a globular C-terminal half characterized by the presence of an internal disulfide bridge, up to two N-linked glycans and a glycosylphosphatidylinositol (GPI) anchor for insertion into the cellular plasma membrane. Multiple lines of investigation have led to the conclusion that mature PrPC is embedded in specialized membrane domains, so-called raft- like domains, rich in cholesterol and sphingolipids [224]. It has been suggested that these raft- like domains host the self-perpetuated accumulation of PrPSc which subsequently triggers a poorly-understood cascade of events that ultimately leads to cell death. Whereas significant progress has been made in the past few years in defining the minimal requirements for PrPC conversion in vitro [225], the molecular environment which hosts the earliest steps in prion disease manifestation in neuronal cells remains enigmatic. This shortcoming is not due to a lack of proteins proposed to interact directly with PrPC. In fact, more than three dozen proteins have been suggested to reside in spatial proximity to PrPC using multiple experimental paradigms [40]. Additional candidate interactors have been proposed to bind preferentially to PrPSc [226- 228]. In surveying this body of literature, however, it is apparent that very few of the candidate interactors have been independently verified by multiple investigators and, overall, little agreement exists as to their relative importance for prion protein biology. Notable exceptions may represent the 37-kDa/67-kDa laminin receptor precursor (herein referred to as LRP; also known as ribosomal protein SA (Rpsa), and not to be confused with Lrp1, the low density lipoprotein receptor-related protein 1) [229], one of the first proteins identified to bind to the prion protein in a yeast two-hybrid (Y2H) screen [230]; the neural cell adhesion molecule

(Ncam1; herein referred to as NCAM), which was initially identified in a cellular crosslinking study based on a co-immunoprecipitation methodology [231] ; and heparin sulfate proteoglycans (HSPGs) [232, 233]. Since its original discovery, LRP has been proposed to act as a cell surface receptor for PrPC that may play a role in prion propagation [234]. In the case of NCAM, binding to PrP has been shown to play a role in neuritic outgrowth possibly mediated through an interaction with Fyn tyrosine kinase [235, 236].

In studying the molecular environment of membrane proteins, a limitation exists in that the proteins must be solubilized by the addition of detergents in a manner that does not disrupt the protein-protein interactions under investigation. A solution to this obstacle constitutes the covalent stabilization of interactions by chemical crosslinking prior to the disruption of cellular integrity. In this regard, we previously reported on a large-scale investigation of the molecular PrPC neighborhood in mice following limited in vivo crosslinking by time-controlled transcardiac perfusion (tcTPC)-based delivery of formaldehyde (FA) to the brain [237]. A conspicuous feature of this PrPC interactome dataset was the relative abundance of membrane proteins which harbor immunoglobulin (Ig)-like folding motifs in their extracellular domains confirming previous data describing a binding domain of PrPC within these folds [231].

Given the large diversity of cell types present in the brain, we questioned whether an equivalent concentration of Ig-like domain-harboring proteins would have been found if the investigation of the PrPC interactome had been restricted to a single cell type. A particular concern was that the relative abundance of Ig-like domain-harboring proteins in the previous dataset may have masked the ability to identify biologically important interactions of PrPC that are more transient in nature or involve less abundant proteins. We therefore chose to reinvestigate the PrPC interactome in mouse neuroblastoma cells (N2a), which are by far the best-characterized cell model for the study of the biology and conversion of PrPC to PrPSc [238, 239]. To facilitate the discrimination of unspecific binders from specific interactors in this study, we incorporated quantitative mass spectrometry based on isotopic labeling and extended our investigation to the mammalian PrPC paralogs Doppel (encoded by the Prnd gene; protein product herein referred to as Dpl) and Shadoo (encoded by the Sprn gene; protein product herein referred to as Sho). To avoid the use of non-identical affinity chromatography steps, we equipped the three bait proteins with the same N-terminal FLAG epitope [240], with the awareness that a similar PrPC expression construct had in prior investigations neither interfered with the

posttranslational processing nor the conversion of PrPC in cell or animal models [239]. We now present data which suggest that members of the mammalian prion protein family may populate highly similar molecular environments when expressed in the neuroblastoma cell model system. We further document that a subset of endoplasmic reticulum (ER) chaperones which interact strongly with PrPC escape from the ER to reside in spatial proximity to PrPC at the plasma membrane. Pharmacological inhibition of these chaperones could increase PrPSc levels in prion- infected N2a sublines, suggesting a protective role for the chaperones in this paradigm. Finally, our data consolidate multiple previously controversial interactions and suggest a scenario whereby PrPC may organize its molecular environment by its ability to recognize a specialized subset of cell adhesion molecules which recruit membrane proteins carrying high-mannose glycans into spatial proximity with PrPC.

2.2 Materials and methods

2.2.1 Ethics Statement

All procedures which required the handling of mice were conducted in accordance with an animal use protocol (No. 20006633) approved by the University of Toronto Animal Care Committee.

2.2.2 Antibodies

The mouse monoclonal antibody directed against PrP (clone 7A12) was a generous gift from Dr. Man-Sun Sy (Case Western Reserve University School of Medicine, Cleveland, OH, USA). The anti-FLAG M2 antibody (F3165) was obtained from Sigma-Aldrich (Oakville, ON, Canada). Anti-ERp57 (Pdia3) antiserum was raised against glutathione S-transferase-fused mouse ERp57. Anti-P4hb antiserum (SPA-891) was purchased from StressGen Biotechnologies (Victoria, BC, Canada).

2.2.3 Clones

FLAG affinity tags (N-DYKDDDDK-C) were inserted before residue 29 of mouse PrP, residue 27 of mouse Dpl, and residue 26 of mouse Sho (all in the pcDNA3 mammalian expression

vector) using standard PCR-based mutagenesis techniques. The identity of all constructs was verified by DNA sequencing.

2.2.4 Cell culture, in vivo crosslinking, inhibitor treatments and cell viability assay

Cell culture and formaldehyde crosslinking of N2a cells followed a protocol described before [231]. For bulk selection of stably transfected cells, cultures were expanded in the presence of 1 mg/mL G418 and maintained at a concentration of 0.2 mg/mL G418. RML-infected N2a cells (ScN2a) were maintained as above. For PDI inhibition experiments, DTNB or bacitracin (Sigma- Aldrich) was added to the cell culture medium at indicated concentrations followed by a two-day incubation at 37°C. The effect of inhibitor treatments on cell viability was assessed by the Trypan blue dye exclusion assay. Briefly, 0.4% Trypan blue was gently mixed with trypsinized cells and incubated for 5 minutes at room temperature. Subsequently, the percentage of viable cells per volume equivalent was determined with the use of a hemocytometer.

2.2.5 Affinity purification of bait proteins

Approximately 109 in vivo FA-crosslinked cells each of control and FLAG-prion expressing N2a

cell lines were lysed in homogenization buffer (50 mM NH4Cl, 80 mM Tris, pH 8.0) supplemented with 1× Complete Protease Inhibitor Cocktail (Roche, Palo Alto, CA, USA). To ensure near quantitative extraction of membrane proteins, an equal volume of extraction buffer (20 mM NaCl, 1% sodium deoxycholate, 1% NP-40, 20 mM Tris, pH 8.0) was added, followed by a 30-min incubation and 5-min sonication in a water bath sonicator. Insoluble cellular debris was removed by high-speed centrifugation (100,000 × g, 1 h). Subsequently, the crosslinked bait protein complexes were immunoaffinity-captured on anti-FLAG-agarose (Sigma-Aldrich). During this step samples were gently agitated on a turning wheel for 12 h, then washed extensively with 0.5 M NaCl, 0.05% SDS, 1% NP-40, 20 mM HEPES, pH 7.3, and detergents

removed by a pre-elution wash with 10 mM NH4HCO3, pH 8.0. Proteins were eluted by acidification with 0.2% trifluoroacetic acid, 20% acetonitrile, pH 2.0.

2.2.6 Protein reduction, alkylation and trypsinization

Protein-containing fractions were denatured in the presence of 6 M urea, 20 mM NH4HCO3, pH 8.0, followed by reduction with 1 mM tris-(2-carboxyethyl)-phosphine for 30 min at 60°C and

alkylation with 2.5 mM 4-vinylpyridine for 1 h at room temperature in the dark. Samples were diluted four-fold to ensure that the concentration of urea did not exceed 1.5 M. Tryptic digestion was initiated by the addition of 1% (wt/wt) of side chain-modified, TPCK-treated porcine trypsin and allowed to proceed at 37°C for 6 h.

2.2.7 iTRAQ labeling

Following trypsinization, equal quantities of tryptic peptide mixtures were spiked with 1 pmol of synthetic (Glu1)-Fibrinopeptide B (GluFib) (Sigma-Aldrich) to serve in the downstream analysis as an internal control for the efficiency of individual labeling reactions. Equal labeling with all four reagents was confirmed by equal intensities of 114:115:116:117 signature peaks upon forced fragmentation of the GluFib [M+2H]2+ parent ion at m/z 785.85. Any strong deviation from this ratio would have indicated problems with the labeling reaction or recovery of individual samples prior to the sample mixing step. Individual iTRAQ labeling reagents (Applied Biosystems, Foster City, CA, USA) were reconstituted in ethanol, added to peptide mixtures derived from the tryptic digestion of IP eluates (Control: iTRAQ 114, Dpl: iTRAQ 115, PrP: iTRAQ 116 , Sho: iTRAQ 117) and incubated at room temperature in the dark for 3 h.

2.2.8 Two-dimensional liquid chromatography

Strong cation exchange (SCX) chromatography was used to achieve peptide fractionation of the complex digest mixture. Samples digested with trypsin were adjusted to 25% acetonitrile and

acidified (pH 3.0) by 20-fold dilution in 25% acetonitrile, 20 mM KH2PO4, pH 3.0. High performance liquid chromatography (HPLC) was carried out using the Ultimate System (Dionex, Sunnyvale, CA, USA) equipped with a microflow calibration cartridge, a Valco injection port and a 180 nL volume UV cell. Separation was achieved on a self-packed 0.5 mm × 110 mm Luna SCX column (Phenomenex, Torrance, CA, USA) at a flow rate of 18 µl/min with a steep

salt gradient from 0–400 mM NH4Cl in 25% acetonitrile, 20 mM KH2PO4, pH 3.0. Fractions eluted from the SCX column were desalted with C18 Empore (3M, Minneapolis, MN, USA) stop and go extraction (STAGE) tips and subsequently subjected to nano-flow RP-HPLC using the Ultimate LC system (Dionex, Sunnyvale, CA, USA) equipped with a nanoflow calibration cartridge at a flow rate of 250 nL/min. Peptides were separated on a 75-µm ID self-packed column containing Proteo C12 reverse-phase matrix (Phenomenex) using a 100-min gradient from 2%–34% acetonitrile in water, with 0.1% (wt/vol) formic acid as the ion-pairing agent.

2.2.9 Electrospray ionization QqTOF mass spectrometry analysis

The column effluent was coupled directly via a fused silica capillary transfer line to a QSTAR XL hybrid quadrupole/time-of-flight tandem mass spectrometer (Applied Biosystems; MDS Sciex, Concord, ON, Canada) equipped with a MicroIonSpray source. The progress of each LC/MS run was monitored by recording the total ion current (TIC) as a function of time for ions in the m/z range 300 to 1800. At 5-s intervals through the gradient, a mass spectrum was acquired for 1 s, followed by one collision-induced dissociation (CID) acquisition of 4 s each on ions selected by preset parameters of the information-dependent acquisition (IDA) method, using nitrogen as the collision gas. Singly-charged ions were excluded from CID selection. The collision energy was adjusted automatically for each CID spectrum using an empirically- optimized formula which considers the charge state and m/z value of the precursor ion.

2.2.10 Database searches

Peak lists for database searching were created using Mascot Distiller (Version 1; MatrixScience, London, UK). Searches were performed using designated MS/MS data interpretation algorithms within Protein Prospector (Version 4.21.3; University of California, San Francisco, CA, USA) [241] and Mascot (Version 2.2; MatrixScience). Searches considered up to one missed cleavage and charge states ranging from +2 to +4. The analysis of iTRAQ data was assisted by the software program ProteinPilot (Version 2.0; Applied Biosystems; MDS Sciex). For a protein to be listed in the data tables, it had to be identified by all three search algorithms. All proteins listed in the table were identified with high confidence by the application of the following filters: (i) CID spectra with individual confidence scores of less than 80% were not included (the determination of these confidence scores is documented in the Protein Pilot 2.0 Software Help); (ii) assignments to non-iTRAQ-labeled peptides were not considered; (iii) all identifications of proteins had to be based on at least two CID spectra. Raw iTRAQ ratios were corrected for impurity levels of individual reagent lots determined by the manufacturer. In instances where only two peptides supported the identification of a protein, we required the underlying CID spectra to generate a Mascot score indicating a <5% probability that the match could be considered a random event and further confirmed matches by manual interpretation of spectra. As searches were carried out without species restriction, the correct assignment of matches to mouse entries served as an additional internal control. Identifications were confirmed in repeat experiments performed at a two-fold lower scale. It should be noted that the vast majority of

proteins were identified with Mascot scores exceeding thresholds conventionally applied for confident identifications. The mass tolerance range between expected and observed masses used for database searches was ±150 ppm for MS peaks, and ±0.15 Da for MS/MS fragment ions. These relatively large thresholds were used to capture more of the low intense peaks that frequently display broader distribution and are therefore assigned with lower mass accuracy. Threshold levels were optimized based on LC/MS/MS datasets of tryptic digests of standard proteins. All samples were searched against the National Center for Biotechnology Information nonredundant database (nrNCBI, release: June 2008) and a ‘decoy’ database in which all entries of the above NCBI database had been inverted. The interactome diagram was constructed using Cytoscape 2.6.0 [242].

2.2.11 Snowdrop lectin affinity purification

Mice were subjected to time-controlled transcardiac perfusion crosslinking and brain extracts generated as described before [237]. Brain extracts which had been cleared of all insoluble content by high speed centrifugation (100,000 × g, 1 hr) were added to snowdrop (Galanthus nivalis) lectin agarose (Sigma-Aldrich), which had been pre-equilibrated in Lysis buffer (50 mM Tris, pH 8.3, 150 mM NaCl, 0.5% NP-40, 0.5% sodium deoxycholate, 1× Complete Protease Inhibitor), and incubated on an end-over-end turning wheel for 12 h at 4°C. Following extensive washes with lysis buffer, proteins were eluted by incubation of lectin agarose beads with 1× Laemmli sample buffer containing 100 mM DTT. To generate the crosslink reversal series, snowdrop lectin agarose eluate fractions were subjected to 90°C heat treatment for the indicated durations.

2.2.12 Cell surface biotinylation

N2a and ScN2a cells, grown to 90% confluency, were washed three times with 10 ml of ice-cold PBS and subsequently subjected to biotinylation with 0.5 mg/ml HOOK-Sulfo-NHS-SS-Biotin (Pierce Biotechnology, Rockford, IL, USA) in PBS for a duration of 30 min at 4°C. Residual biotin was quenched by the addition of 100 mM glycine in PBS and further incubation for 15 min. Following cell lysis in detergent containing buffer (0.5% NP40, 0.5% DOC, 150 mM NaCl, 1× Complete Protease Inhibitor Cocktail) the capture of biotinylated proteins on streptavidin agarose (Sigma-Aldrich) occurred during a 12 h incubation. Following extensive washing with lysis buffer, captured proteins were eluted by reducing the disulfide bond within the HOOK-

Sulfo-NHS-SS Biotin conjugation reagent in the presence of 100 mM DTT, 62.5 mM Tris, pH 6.8, 2% SDS.

2.2.13 Proteinase K digestion

Confluent cells were lysed in 10 mM Tris-HCl, pH 8.0; 100 mM NaCl; 0.5% NP-40; 0.5% sodium deoxycholate (DOC). Equal amounts of protein determined using the bicinchoninic acid (BCA) reagent were digested with 20 µg/ml PK (GIBCO, Carlsbad, CA, USA) at a ratio of 1:25 (w/w protease to protein) for 1 h at 37°C. Reactions were stopped by the addition of 2 mM phenylmethylsulfonyl fluoride (PMSF). Following ultracentrifugation, pellets were resuspended in SDS-loading buffer and subjected to immunoblot analysis.

2.3 Results

2.3.1 Large-scale quantitative and comparative interactome investigation of members of the mammalian prion protein family

In preparation for this study, mouse neuroblastoma cells (N2a) stably expressing N-terminally FLAG-tagged full-length versions of Dpl, PrP and Sho were generated (Figure 2.1). Small-scale expression tests followed by diagnostic N-glycosidase F or phosphatidylinositol-specific phospholipase C digestions confirmed that all three bait proteins mimicked their untagged parent molecules in (i) the presence of N-glycans, and (ii) membrane attachment by means of a GPI anchor [143] (and data not shown). To generate the biological source material for a large-scale comparative interactome investigation, the three cell lines and an “empty-vector” stably- transfected N2a cell line serving as a negative control were expanded to 109 cells each using cell culture conditions which promote adherent growth. To covalently stabilize protein-protein interactions prior to the disruption of cellular integrity, cells were subjected to a 15-min treatment with FA [231]. Subsequently, cells were lysed by the addition of detergents and the four extracts (each containing approximately 500 mg of cellular protein) purified side-by-side on anti-FLAG affinity agarose matrices. The presence of covalent linkages between proteins permitted the use of highly stringent washing conditions to minimize the presence of unspecific binders. Following elution from the affinity matrix (with a yield of approximately 100 µg of protein material per sample), protein complexes were denatured and trypsinized. Finally, peptide

mixtures were tagged with isobaric tags for relative and absolute quantitation (iTRAQ) [243], and the samples were combined and subjected to a comprehensive analysis by tandem mass spectrometry (MS/MS) (Figure 2.2). A query of mouse protein databases led to the identification of more than 100 proteins. All identifications can be considered confident by multiple measures: (i) identical identifications were made by two matching algorithms; (ii) scores assigned by the algorithms exceeded significance thresholds (P < 0.05) for all identifications and were based on a minimum of two peptides; and (iii) a search of a decoy database generated by the inversion of sequences for all mouse protein entries resulted in no identifications which shared any of the above features [244]. A reduction of the total list of identified proteins to the subset of proteins whose identification correlated with the presence of at least one of the three bait proteins was based on peak intensities of iTRAQ reporter ions found in the low mass range of individual collision induced dissociation (CID) spectra. The relative intensity of these ions is indicative of the relative contribution of each of the iTRAQ-labeled samples to the generation of a given CID spectrum. Thus, by calculating the ratio of reporter ion intensities for each of the three bait- specific reporter ions (iTRAQ 115: Dpl; iTRAQ 116: PrP; iTRAQ 117: Sho) and the negative control ion (iTRAQ 114 reporter), peptides purifying with at least one of the three baits were recognized by iTRAQ reporter ion ratios greater than 1. The above analysis revealed that more than 50 proteins co-purified with at least one of the three bait proteins (Table 2.1). Interestingly, the majority of these proteins co-purified with all three bait proteins, suggesting that the molecular environment of the three members of the mammalian prion protein family assessed in the context of mouse neuroblastoma cells is highly similar.

Figure 2.1: Expression analysis of FLAG-tagged mouse prion proteins. A. Schematic representation of murine prion proteins with FLAG tags inserted near the N-terminus. B. Expression of transiently-transfected FLAG-prion proteins in N2a cells as assessed by Western blotting with the anti-FLAG M2 antibody. The presence of a non-specific band in N2a lysates recognized by the M2 antibody is denoted by an asterisk.

Figure 2.2: Flow chart depicting strategy for semi-quantitative comparison of prion protein family interactomes. In vivo formaldehyde crosslinked protein complexes containing N- terminally FLAG-tagged bait proteins are stringently purified on anti-FLAG agarose parallel to a negative control sample derived from an empty vector expression clone. Following alkylation, reduction and trypsinization, digests are side-by-side iTRAQ labeled and subsequently combined. Two-dimensional liquid chromatography of peptides is coupled to online ESI- MS/MS, which is followed by computationally-aided protein identification and quantitative analysis.

34 Table 2.1: Prion protein family interactome in mouse neuroblastoma cells.

IPI accession % 114 115 116 117 Number Symbol Identified proteinsa Pept.b Uniquec Cov.d Controle Dpl PrP Sho

IPI:IPI00131622.1 Prnd prion protein dublet (doppel, Dpl) 8 13 39.1 2.0 92.5 3.0 2.5 IPI:IPI00120793.1 Prnp prion protein (PrP) 5 5 26.8 4.0 20.5 65.1 10.5 IPI:IPI00226455.1 Sprn shadow of prion protein (shadoo, Sho) 3 6 69.4 2.9 11.4 20.5 65.2 IPI:IPI00751369.1 Ldha lactate dehydrogenase A 15 19 67.3 6.0 26.3 44.0 23.7 IPI:IPI00230108.6 Pdia3 protein disulfide isomerase associated 3 14 15 61.2 6.0 46.7 18.8 28.4 IPI:IPI00229517.5 Lgals1 galectin-1 (lectin, galactose binding, soluble 1) 5 7 57.8 5.3 32.8 37.4 24.4 IPI:IPI00122971.1 Ncam1 neural cell adhesion molecule 1 (NCAM) 23 34 54.3 3.4 58.6 16.6 21.3 IPI:IPI00670985.3 Gm9234 predicted gene 9234 (EG668548) 6 6 53.7 13.6 29.8 36.7 20.0 IPI:IPI00128973.1 Gap43 growth associated protein 43 (neuromodulin) 2 2 46.7 7.0 30.5 24.5 38.0 IPI:IPI00850840.1 Rpsa ribosomal protein SA (laminin receptor precursor, LRP) 7 8 43.4 11.8 33.8 26.9 27.5 IPI:IPI00319992.1 Hspa5 heat shock protein 5 15 20 42.7 7.4 32.9 22.5 37.3 IPI:IPI00515173.1 H2-K1 histocompatibility 2, K1, K region 6 6 40.6 3.6 60.9 17.0 18.5 IPI:IPI00762203.2 Ftl1 ferritin light chain 1 3 4 39.9 10.4 36.9 29.4 23.3 IPI:IPI00454042.2 Fam3c family with sequence similarity 3, member C 4 4 35.7 4.7 64.4 12.5 18.5 IPI:IPI00133522.1 P4hb protein disulfide-isomerase (prolyl 4-hydroxylase beta) 8 9 35.4 5.7 22.5 28.4 43.4 IPI:IPI00135686.2 Ppib peptidylprolyl isomerase B 4 4 34.3 10.3 30.3 24.2 35.1 IPI:IPI00132950.1 Rps21 ribosomal protein S21 2 2 31.3 8.6 18.7 37.9 34.9 IPI:IPI00110805.1 H2-D1 histocompatibility 2, D region locus 1 5 5 29.3 3.4 56.7 19.9 20.0 IPI:IPI00119618.1 Canx calnexin precursor 8 8 28.8 7.1 40.7 18.8 33.4 IPI:IPI00554929.2 Hsp90ab1 heat shock protein 90 alpha (cytosolic), class B member 1 10 11 28.6 11.5 15.6 33.4 39.6 IPI:IPI00123639.1 Calr calreticulin 3 3 27.9 7.2 32.5 24.7 35.6 IPI:IPI00116498.1 Ywhaz 14-3-3 protein zeta 4 5 27.8 10.0 21.6 35.4 33.0 IPI:IPI00462199.1 Bsg basigin 5 5 27.3 4.1 32.0 23.2 40.7 IPI:IPI00230682.6 Ywhab 14-3-3 protein beta 2 2 27.0 10.5 26.3 23.4 39.8 IPI:IPI00857709.1 Tmed2 transmembrane emp24 domain trafficking protein 2 2 2 26.9 6.0 33.6 26.1 34.4 IPI:IPI00854971.1 Pdia6 protein disulfide isomerase associated 6 2 2 25.6 7.8 29.5 27.9 34.9 IPI:IPI00466570.4 Tmed10 transmembrane emp24-like trafficking protein 10 4 5 25.1 3.5 35.7 16.8 44.0 IPI:IPI00230707.6 Ywhag 14-3-3 protein gamma 2 2 23.9 9.8 12.0 48.7 29.6 IPI:IPI00132799.4 C1qbp complement component 1, q subcomponent binding protein 4 4 23.3 9.8 17.4 22.2 50.7 IPI:IPI00124700.1 Tfrc transferrin receptor protein 1 5 5 22.4 6.6 49.4 19.1 24.9 IPI:IPI00129526.1 Hsp90b1 heat shock protein 90, beta (Grp94), member 1 (endoplasmin) 3 3 20.9 7.7 23.4 28.7 40.2 IPI:IPI00123342.4 Hyou1 hypoxia up-regulated 1 2 2 20.6 10.4 25.3 31.4 32.9 IPI:IPI00762435.2 Fyn fyn proto-oncogene 3 3 20.5 13.3 24.2 28.6 33.9 IPI:IPI00129519.3 Basp1 brain abundant, membrane attached signal protein 1 7 10 19.9 8.2 18.8 37.4 35.5

IPI:IPI00187289.2 Tmem206 transmembrane protein 206 (C1orf75) 2 2 18.9 3.5 18.2 11.1 67.2 IPI:IPI00416577.1 Gdi2 guanosine diphosphate (GDP) dissociation inhibitor 2 2 2 18.8 6.3 32.8 28.0 33.0 IPI:IPI00132474.3 Itgb1 integrin beta 1 5 5 18.4 7.1 33.3 28.1 31.6 IPI:IPI00265291.6 Enpp1 ectonucleotide pyrophosphatase/phosphodiesterase 1 4 4 18.3 9.5 37.6 24.2 28.7 IPI:IPI00114641.2 Slc3a2 solute carrier family 3, member 2 (CD98) 2 4 18.2 10.2 43.6 21.4 24.8 IPI:IPI00420835.3 Itga6 integrin alpha 6 3 3 15.6 8.1 40.2 20.6 31.1 IPI:IPI00762897.1 Ugcgl1 UDP-glucose ceramide glucosyltransferase-like 1 3 3 15.5 7.4 30.0 32.4 30.1 IPI:IPI00115546.4 Gnao1 guanine nucleotide binding protein, alpha O 2 2 15.3 10.1 32.6 27.3 29.9 IPI:IPI00120245.1 Itgav integrin alpha V 2 2 14.5 9.0 37.4 23.5 30.1 IPI:IPI00311682.5 Atp1a1 sodium/potassium transporting ATPase alpha 1 polypeptide 3 3 14.2 6.5 43.3 19.0 31.2 IPI:IPI00124221.1 Atp1b3 sodium/potassium transporting ATPase beta 3 polypeptide 3 4 13.7 5.1 33.9 28.3 32.7 IPI:IPI00652902.1 Gnai2 guanine nucleotide binding protein, alpha inhibiting 2 2 2 13.5 5.1 35.8 24.3 34.8 IPI:IPI00271951.5 Pdia4 protein disulfide isomerase associated 4 4 4 13.3 10.1 25.1 27.0 37.8 IPI:IPI00123996.1 Nrp1 neuropilin 1 2 3 12.8 7.0 49.3 19.6 24.1 IPI:IPI00752565.1 Igsf8 immunoglobulin superfamily, member 8 3 3 12.2 4.8 49.0 21.2 25.0 IPI:IPI00118569.1 Gna13 guanine nucleotide binding protein, alpha 13 3 3 10.1 14.6 35.1 32.8 17.6 IPI:IPI00845689.1 Tmed9 transmembrane emp24 transport domain containing 9 (gp25L2) 5 7 9.7 7.2 44.2 26.1 22.5 IPI:IPI00469000.4 Slc39a6 metal ion transporter ZIP6 3 3 7.3 3.2 29.9 32.8 34.2 IPI:IPI00458003.1 Enpp3 ectonucleotide pyrophosphatase/phosphodiesterase 3 2 2 5.4 5.9 43.4 25.1 25.5 IPI:IPI00273801.3 Slc39a10 metal ion transporter ZIP10 2 2 4.3 2.4 36.3 16.7 44.6 IPI:IPI00831568.1 L1cam L1 cell adhesion molecule 3 3 3.5 10.9 30.9 26.7 31.5 aCandidate interactors are listed in order, with the position of a given protein in the table reflecting the percentage of primary structure corresponding to the combined unique CID spectra. In instances where a subset of CID spectra were matched to more than one isoform or member of a protein family, only the highest scoring entry was selected unless an independent identification was supported by at least two unique CID spectra. Proteins were sorted into specific versus unspecific binder categories based on their iTRAQ distribution, i.e. proteins were considered unspecific interactors if their derived CID spectra revealed iTRAQ114 signature mass peak signal intensities which exceeded 15% of combined and normalized (100%) intensities for iTRAQ114-117 mass peaks. Only specific interactors are shown in this table.The complete dataset can be viewed in Supplemental Table 2.1. bOnly CID spectra underlying different peptides were considered, i.e. if the same peptide was identified with different charge states or modifications it counted as one hit. cTotal number of unique CID spectra. Please note that the same peptide was only counted more than once if it was identified with different charge states or modifications. dPercent sequence coverage based on the presence of peptides for which no higher ranked assignment to other proteins could be made. eFor the calculation of iTRAQ values the intensity of individual peptide associated iTRAQ signature peaks was normalized to combine to 100% per peptide and subsequently averaged. Standard deviations were determined and are listed in Supplemental Table 2.1.

We next explored the known cellular localization of candidate interactors based on bioinformatic methods and literature mining. This investigation revealed that multiple proteins in the dataset were likely to encounter the three bait proteins during their early passage through the secretory pathway, since they constitute (i) classical ER chaperones (heat shock protein 5, Hspa5; calnexin, Canx; calreticulin, Calr; endoplasmin, Hsp90b1), (ii) isomerases which facilitate disulfide or proline cis-trans rearrangements (protein disulfide isomerase associated 3, Pdia3; protein disulfide isomerase associated 4, Pdia4; prolyl 4-hydroxylase beta polypeptide, P4hb; peptidylprolyl isomerase B, Ppib), or (iii) proteins involved in the trafficking between ER and Golgi compartments (transmembrane emp24 transport domain containing 9, Tmed9; and transmembrane emp24-like trafficking protein 10, Tmed10). Other proteins in the dataset were likely to reside in spatial proximity to the mature bait proteins at the plasma membrane, because they are themselves known to be either (iv) embedded in the plasma membrane through transmembrane (TM) domains (NCAM; transferrin receptor, Tfrc; integrins; neuropilin 1, Nrp1; L1 cell adhesion molecule, L1cam; basigin, Bsg), or constitute (v) secreted proteins (galectin-1, Lgals1; family with sequence similarity 3, member C, Fam3c; ectonucleotide pyrophosphatase/phosphodiesterase 1 and 3, Enpp1/3) or (vi) cytosolic proteins (14-3-3 proteins zeta/beta/gamma, Ywhaz/Ywhab/Ywhag; guanine nucleotide binding proteins, Gnao1/Gnai2/Gna13; and Fyn proto-oncogene) that attach to the plasma membrane from outside or within the cell, respectively (Figure 2.3). In contrast to the above proteins that were designated as specific interactors by their iTRAQ reporter ion ratios, proteins which were equally represented in control and specific interactome cells and therefore constituted unspecific binders could be mapped to a wide range of cellular compartments including the nucleus and mitochondria. This result was expected based on the known cellular biology of PrPC and therefore served as strong biological validation of the iTRAQ-based discrimination of specific interactors from unspecific contaminants.

Figure 2.3: The mammalian prion protein family interactome. Network depicting proteins identified in spatial proximity to Dpl, PrPC and Sho as nodes and their direct or indirect interactions as spokes. Components identified in previous studies by other investigators are shown with red lines surrounding and red edges connecting the nodes. The intensity of gray shading of nodes correlates directly with the fractional sequence coverage with which individual components were identified. Increased spoke thickness indicates stronger co-enrichment with the connected bait protein. Node shapes indicate predominant cellular localization of proteins, i.e. ER and Golgi compartments (rectangular), plasma membrane (oval), cytosol (diamond) or secreted (octagonal). Nodes depicting proteins whose presence in the samples correlated strongest with the presence of a given bait protein are arranged in spatial proximity to the respective bait protein nodes in the bottom half of the network.

2.3.2 Interactions amongst members of the mammalian prion protein family

It was anticipated that iTRAQ reporter ion signals in CID spectra derived from tryptic peptides of the three bait proteins would reflect the logic of the iTRAQ conjugation step. For example, CID spectra which supported the identification of Dpl had to be mainly contributed by the sample which was labeled with the iTRAQ 115 reagent. A close inspection of the low mass range of CID spectra confirmed that this expectation was met for all three bait proteins and

thereby served as an internal control for the reliability with which relative iTRAQ reporter ion intensities could be employed to trace the contribution of samples to the identification of proteins (Table 2.1). Interestingly, however, a closer inspection of the above spectra also revealed that approximately one-third of the peptides which supported the identification of PrP were contributed by the FLAG-Dpl and FLAG-Sho affinity purified samples. This observation was supported by five CID spectra consistent with the interpretation that the purification of FLAG- Dpl and FLAG-Sho caused a significant co-purification of endogenous wild-type (wt) PrPC (P < 0.05). A similar co-purification of endogenous wt Sho in N2a cells was observed with FLAG- Dpl and FLAG-PrP. A co-purification of wt Dpl with FLAG-PrP and FLAG-Sho baits was neither observed nor expected because no endogenous expression of Dpl can be reliably detected in N2a cells (or indeed in the adult mouse central nervous system) [245]. Taken together, these data suggest that PrP, Dpl and Sho are capable of residing in close spatial proximity at the cell membrane and may engage in reciprocal contacts with each other.

2.3.3 Direct versus indirect PrP interactors

It is unlikely that members of the mammalian prion protein family engage in direct interactions with all candidate interactors we identified. For example, some retrieved proteins may represent complexes of the type PrPC - crosslink - interactor 1 - crosslink - interactor 2. Such indirect interactors are often less strongly represented in such datasets, due to both the non-stoichiometric crosslinking that can be achieved and the stringent washing conditions applied. A good correlate for the relative abundance of a protein in proteomics investigations of this nature is the sequence coverage with which it was identified. Based on this rationale, the three bait proteins may bind directly to lactate dehydrogenase (Ldha), protein disulfide isomerases Pdia3 and P4hb, galectin- 1, NCAM, Gm9234 (also known as EG668548), neuromodulin (Gap43), LRP and its ribosomal protein partner Rps21, heat shock proteins Hspa5 and Hsp90ab1, histocompatibility antigens H2- K1 and H2-D1, Fam3c, Ppib, calnexin, calreticulin, 14-3-3 zeta/delta and basigin, as these proteins were identified with the highest sequence coverage (25-75%).

Due to their relatively small size, members of the mammalian prion protein family would be expected to contain a small number of distinct surface domains that can engage in direct protein-protein interactions. Consequently, at least a subset of binders may bind to the same face within the bait proteins and may even resemble each other in their primary structure or structural

folds. When candidate interactors were subjected to scrutiny from this angle, similarities between NCAM and basigin, two of the proteins represented with the highest sequence coverage in the dataset, became apparent. Not only do both proteins harbor Ig-like domains, but a subset of these domains bear a striking resemblance in their sequence, an observation made previously [246, 247]. Interestingly, these very same Ig-like domains are capable of binding to membrane proteins carrying oligomannosidic N-glycans [248-251]. The latter modification is only found on a relatively small number of mature N-glycosylated proteins expressed mainly in the brain, because oligomannose structures typically represent transient intermediates subject to additional posttranslational modification during their passage through the secretory pathway [248, 251, 252]. Amongst the few known proteins carrying oligomannosidic N-glycan structures are the cell adhesion molecule L1, the β-subunit of the Na/K ATPase, the transferrin receptor, integrins [253] and members of the ectonucleotide pyrophosphatase/phosphodiesterase protein family [252]. Remarkably, not only were all the above proteins present in our dataset, but their iTRAQ ratios also documented their specific co-purification with members of the mammalian prion protein family. In light of the relatively low sequence coverage observed for proteins carrying oligomannosidic N-glycans, their presence in the dataset likely reflects indirect binding to the bait proteins mediated by the mentioned molecules that harbor Ig-like domains with an affinity for oligomannosidic N-glycans.

2.3.4 PrP forms high-molecular weight complexes with proteins carrying oligomannosidic N-glycans To investigate whether PrPC also resides in spatial proximity to membrane proteins equipped with oligomannosidic N-glycans in the mouse brain, we subjected anaesthetized wild-type mice to brief crosslinking with formaldehyde to stabilize protein-protein interactions in vivo prior to the disruption of cellular integrity [237]. Perfused mouse brains were then rapidly dissected and homogenized in the presence of a quenching reagent. To verify the success of the crosslinking reaction and characterize the size distribution of crosslinked protein complexes containing PrPC an aliquot of extracts was passed through a pre-calibrated size-exclusion chromatography column. This pilot analysis documented that PrPC was found in protein complexes co-eluting with calibration markers which span a range of 160 kDa to 669 kDa. Uncrosslinked PrPC could be recovered from the crosslinked complexes by a crosslink reversal procedure based on the administration of heat and reducing agents (Figure 2.4). Subsequently, the crosslinked extracts

were passed over snowdrop lectin (Galanthus nivalis agglutinin, GNA) agarose, a lectin which specifically recognizes oligomannosidic N-glycans [252]. Eluate fractions, when subjected to Western blotting analysis, documented the presence of a subset of PrPC in high-molecular weight protein complexes that were specifically enriched by snowdrop lectin agarose (Figure 2.5). The binding to the affinity matrix was determined to be specific by (i) the PrPC-specific band pattern in the eluate fraction demonstrating the selective capture of PrPC in high-molecular weight crosslinked complexes, rather than in its uncrosslinked form; and (ii) the observation that pre- incubation of the extract with soluble snowdrop lectin prevented affinity capture. A crosslink reversal time series revealed that the high-molecular weight complexes preferentially contain fully-glycosylated PrPC consistent with the notion that this interaction involves mature PrPC present at the cell surface.

Figure 2.4: Evidence for PrPC in high-molecular weight protein complexes. Brain extracts from mice subjected to transcardiac perfusion crosslinking were passed over a size exclusion chromatography column. Western blot analysis of eluate fractions with 7A12 antibody documented the presence of PrPC in high-molecular weight complexes which co-eluted together with calibration proteins spanning a molecular weight range of 669 kDa (lane 1) to 440 kDa

(lane 3) to 160 kDa (lane 5). Crosslink reversal by 20-min heat treatment of samples caused a shift of PrPC reactive bands to levels of fully glycosylated PrPC consistent with the release of PrPC from crosslinked complexes (lanes 2, 4 and 6). Th, thyroglobulin; Ft, ferritin; Ig, immunoglobulin; Lg, lactoglobulin.

Figure 2.5: Evidence for PrPC in high molecular weight protein complexes captured by a lectin with specificity for oligomannosidic glycans. Brain extracts from mice subjected to transcardiac perfusion crosslinking were passed over snowdrop lectin agarose. Western blot analysis of eluate fractions with 7A12 antibody documented presence of PrPC in high molecular weight complexes (lane 5). Only weak bands were detected when the affinity capture step was preceded by incubation of extracts with soluble snowdrop lectin (sGNA, lane 2). Crosslink reversal by 5-min (lanes 3 and 6) or 20-min (lanes 4 and 7) heat treatment of samples caused a shift of PrPC reactive bands to levels of fully glycosylated PrPC consistent with the release of PrPC from crosslinked complexes.

2.3.5 A subset of cellular P4hb, Pdia3 and calreticulin reside at the cell surface of neuroblastoma cells A conspicuous feature of the Dpl/PrP/Sho-interactome dataset constituted the identification of multiple PDIs. In particular, Pdia3 (also known as ERp57) was represented with strong sequence

coverage (61%) and its identification was further corroborated by similarly strong identifications for calnexin and calreticulin, known interactors of this PDI. Whereas binding of PDIs to both Dpl and PrP was expected due to the presence of internal disulfide bridges in these proteins, co- segregation with Sho (Pdia3 117:114 iTRAQ reporter ion ratio of 2:1), which does not carry a disulfide bridge, was an unexpected finding. PDIs are known to harbor both isomerase and chaperone functions. In light of previous calls for an involvement of chaperone-like activities in the conversion of PrPC to PrPSc, we asked whether a subset of PDIs may play a role in this process. It is currently thought that the conversion of PrPC occurs only after it has reached the cell surface [254, 255], i.e. well-removed from its presence within the ER where it would be expected to encounter PDIs. Thus, if PDIs were involved in PrPC conversion, these proteins would also have to be present outside the ER, preferably at the plasma membrane. Intriguingly, a small number of reports suggest that, at least in some cells, this might indeed be the case (please refer to Discussion section). To investigate whether Pdia3, P4hb and calreticulin are also present at the cell surface in mouse neuroblastoma cells, a cell surface biotinylation experiment was undertaken. Cells were incubated for a brief duration with a cell-impermeable biotin-labeling reagent, which conjugates to primary amines through an N-hydroxysuccinimide ester group, or were subjected to mock treatment. To determine the degree to which broken cells were present at the time of cell surface biotinylation, the downstream analysis was extended to histone 2B. If a large amount of broken cells had been available, histone 2B would have been present at appreciable levels in the cell culture medium resulting in an unequal signal, absent in the mock- treated sample and strong in the biotinylated sample. The observation that histone 2B signals in streptavidin agarose eluates were equal documented that low levels of non-biotinylated histone 2B had been captured unspecifically and served as evidence that levels of extract employed in the side-by-side streptavidin agarose capture steps had indeed been equal. Western blot analysis of extract and streptavidin eluate fractions probed with a PrPC-specific antibody served as a positive control (Figure 2.6). Taken together this experiment revealed that a subset of Pdia3, P4hb and calreticulin was enriched in streptavidin eluate fractions derived from in vivo biotinylated cells compared to the corresponding samples derived from mock-treated control cells. The possibility that this distribution may have been the result of an unspecific binding to the affinity matrix was excluded based on a comparison with histone 2B signals observed evenly across all streptavidin eluate fractions, a distribution expected for a protein which binds to the affinity matrix unspecifically. A calculation based on a semi-quantitative titration analysis of

Western blot signal intensities revealed that the relative amounts of PDIs and calreticulin at the plasma membrane are small compared to their respective quantities in the ER, but constitute at least 1:1000 of the total amounts in these cells. Actual amounts may exceed these estimates as the calculation did not account for the non-stoichiometric labeling achieved during the biotin conjugation step and losses of biotin-derivatized PDIs incurred during the streptavidin affinity purification step.

Figure 2.6: Evidence for cell surface localization of a subset of PDIs and calreticulin in mouse neuroblastoma cells. N2a and ScN2a cells were subjected to cell surface biotinylation or mock treatment. Subsequently, cellular extracts were side-by-side affinity purified on a streptavidin agarose matrix. Extracts and streptavidin agarose eluate fractions were analyzed by Western blotting and membranes probed with antibodies directed against histone 2B, PrP, P4hb, Pdia3 and calreticulin. The relative strength of P4hb-, Pdia3- and calreticulin-specific signals present in eluate fractions from biotinylated (lanes 6 and 8) versus non-biotinylated samples (lanes 5 and 7) is consistent with the conclusion that a subset of these proteins resided at the cell surface during the biotinylation step. The relative intensity of P4hb- and Pdia3-derived signals in extract (40 µg total protein loaded per lane) and eluate fractions can be estimated from the concomitant analysis of 1/4 and 1/20 of the extract fraction shown in lane 1. Thus, P4hb and Pdia3 signals in the eluate fraction (captured from 2 mg total extract) were equivalent to or exceeded the amount of these proteins present in 2 µg of extract. Please note the presence of

histone 2B in all eluate fractions (lanes 5-8), in support of the conclusion that the protein binds unspecifically to the affinity matrix.

2.3.6 Inhibitors of protein disulfide isomerases increase PrPSc levels in a subset of ScN2a cell clones We next addressed whether PDI-inherent chaperone activities influence PrP conversion in prion- infected mouse neuroblastoma cells (ScN2a). The identification of four different PDIs in the dataset suggested that the knockdown of just one PDI may be uninformative. We therefore employed inhibitors known to interfere broadly with the activity of PDIs. One possible outcome of PDI inhibition in vivo would include a depression of PrPSc levels by inhibiting PDI-mediated unfolding of an obligate precursor form of PrPC. Interestingly, both the addition of the dodecapeptide antibiotic bacitracin or 5,5’-dithiobis(2-nitrobenzoic acid) (DTNB), membrane- impermeable reagents known to inhibit PDIs through noncovalent and covalent modes of binding, respectively [256], caused an increase in the level of proteinase K-resistant PrPSc (Figure 2.7). The effect sizes correlated with inhibitor concentrations applied, and in the presence of up to 1 mM of either of the inhibitors, the viability of ScN2a cells (assessed by the trypan blue dye exclusion assay) was not impaired during the course of the experiment. A direct inhibitory effect of these reagents on the proteinase K digestion itself can be ruled out, since levels of the faster migrating PrPSc-derived signals, characteristic for the presence of PrPSc in this disease model (attributed to the action of endogenous proteases that produce a “trimming” of PrP N-terminal sequences [257, 258]), were similarly elevated even prior to the proteinase K digestion step. This observation was made with two clones of ScN2a cells obtained from independent sources; surprisingly, however, it could not be observed with an isolate of chronically prion-infected GT1 cells or two additional ScN2a clones derived from yet other sources (Figure 2.8), possibly indicating the existence of cell line- or subclone-specific PDI family member expression profiles. We note that variations in the behavior of different ScN2a cell clones are well-documented, as is the occurrence of heteroploidy. While variations in our experimental paradigm preclude a generalized observation, they nonetheless put a spotlight on ER-derived chaperones that may engage in physiological interactions with PrPC even at the cell surface, as discussed below.

Figure 2.7: Inhibition of protein disulfide isomerases causes accumulation of PrPSc in a subset of ScN2a cell clones. A. Bacitracin or DTNB, reagents known to inhibit PDIs through orthogonal modes of binding, were added at the indicated concentrations to the cell culture medium and left on the ScN2a cells for a duration of 2 days. Following cell lysis, protein levels were adjusted and a subset of samples subjected to digestion with proteinase K (lanes 411). All samples were analyzed by Western blotting with a PrP specific antibody. Please note the increase in intensity of low molecular weight bands characteristic for the presence of PrPSc even without proteinase K treatment (lanes 1–3). B. A two-sample, two-tailed Student’s t-test assuming unequal variance was utilized to determine the statistical significance of bacitracin-treated band intensities compared to normalized untreated samples. The result of this analysis indicated a significant increase of PrPSc in ScN2a clones 1 and 2 treated with 1 mM bacitracin (P-value = 0.01).

Figure 2.8: Inhibition of protein disulfide isomerases shows no effect on PrPSc formation in a subset of ScN2a cell clones. Bacitracin was added at the indicated concentrations to the cell culture medium and left on the ScN2a clone 3 cells for a duration of 2 days. Following cell lysis, protein levels were adjusted and a subset of samples subjected to digestion with proteinase K. All samples were analyzed by Western blotting with a PrP specific antibody. Please note the absence of an effect on PrPSc levels in this cell clone.

2.4 Discussion The work presented here may comprise the most comprehensive cell-based analysis of the PrPC interactome to date. Furthermore, it is, to our knowledge, the first study to extend the PrPC interactome characterization to other members of the mammalian prion protein family. Critical technical advances from previous PrPC interactome work constitute the incorporation of FLAG affinity purification tags and the use of iTRAQ-based quantitative mass spectrometry for the discrimination of specific from unspecific interactors. These conceptual choices greatly facilitated bioinformatic analyses and afforded an in-depth glimpse into the molecular environment of these proteins.

The data presented support the conclusion that all three members of the mammalian prion protein family populate highly similar molecular environments when expressed in neuroblastoma

cells. This interpretation is based on (i) the high degree of overlap amongst candidate interactors identified for each of the three bait proteins, and (ii) the reciprocal co-capture of endogenous prion proteins in affinity purifications of FLAG-tagged bait proteins. Due to the stringent washing steps applied following the FLAG tag-dependent capture step, this observation is unlikely to represent an unspecific co-enrichment of raft-like membrane domains rich in GPI- anchored proteins. Instead, it points at a close spatial proximity, if not direct interaction, of members of the prion protein family at the plasma membrane. The possibility of a direct interaction between Dpl and PrPC had been raised previously in a report demonstrating rescue of a Dpl-dependent ataxic phenotype in two PrP-deficient (Prnp0/0) mouse lines by reintroduction of PrPC [259]. Recently, a similar rescue of Dpl-mediated toxicity was documented upon overexpression of PrPC or Sho in a primary cell culture model of cerebellar granule neurons which exhibits undetectable endogenous levels of these proteins [143]. However, it has remained unresolved whether these observations depend on the spatial proximity of the respective proteins as opposed to action upon other proteins in shared or intersecting biochemical pathways.

The comparison of protein-protein interaction data reported here with previous data based on either the immunocapture of endogenous PrPC [237] or myc-tagged transgenic PrPC [260] harvested from mouse brains suggests that the move to a cell model may, together with the increase in scale, contributed to a reduction in protein ‘noise’ and thereby an enhanced analysis depth. As a result, the study revealed spatial proximity of the bait proteins to multiple proteins (NCAM, LRP, Na/K ATPase and PDIs) which had previously been proposed to bind to PrPC in the context of widely-differing experimental paradigms.

A comparison of input and eluent protein amounts in this study suggests that less than 0.1% of membrane-associated proteins present in the crosslinked cellular extracts were withdrawn by individual immunoaffinity capture steps. The strength with which a given protein was identified in proximity to our bait proteins may at least partially reflect its relative abundance in the vicinity of the bait protein. Multiple lines of evidence in our dataset support the notion that candidate proteins we identified do indeed reside naturally in spatial proximity to members of the mammalian prion protein in vivo: (1) the proteins we identified originate from cellular compartments to which PrPC is expected to localize; (2) some highly abundant cellular proteins (tubulin, actin, histones) were only seen in the list of unspecific binders to the affinity matrix; (3) no nuclear proteins were found despite the use of deoxycholate in the cell lysis

buffer, a detergent known to solubilize nuclear membranes; and (4) ten out of about fifty candidate interactors in the dataset had previously been reported to interact with the cellular prion protein. The majority of candidate interactors identified in this work (e.g. PDIs, calreticulin, LRP, 14-3-3 proteins and Na/K ATPase) are known to be widely expressed across a large number of diverse tissues and as such are plausible interactors of the bait proteins also outside of N2a cells. Other interactors exhibit a more restricted tissue expression and are primarily known for their expression in cell types belonging to the neuronal lineage (e.g. NCAM and L1), in agreement with the predominant and partially overlapping expression of PrPC and Sho in a subset of these cell types. For Dpl, a protein normally confined to the testis, a physiological proximity to some of these proteins is unlikely, except in the artificial experimental circumstances (described above) which result in ectopic expression of Dpl in the brain.

Proteins that are expected to engage in short-lived interactions are typically underrepresented in interactome datasets of this kind, and as such it was not surprising to find that enzymes known to modify N-glycans or responsible for attachment of the GPI anchor to PrPC were not identified. Moreover, it is likely that the crosslinking chemistry we employed may have precluded the detection of some interactions which lack suitably-positioned functional groups required for the generation of productive covalent bonds. A bias towards covalent interactions that do not sterically block the FLAG epitope may have further skewed the dataset. The above note of caution should also be extended to instances in which candidate interactors co-purify only with a subset of the bait proteins, as seen here for the largely uncharacterized proteins Fam3C and C1orf75 that co-purified with Dpl and Sho, respectively. More work will be needed to determine whether this apparent selective binding to a subset of bait proteins represents a technical artifact or reflects underlying differences in the biology of the bait proteins.

The conversion of the cellular prion protein has repeatedly been proposed to involve additional host-encoded factors. In particular, the participation of chaperones in this process has remained an attractive hypothesis based on theoretical considerations and species barrier effects [261]. Any chaperone to fulfill this role would be expected to bind to PrPC and reside at the cellular plasma membrane. The latter conclusion can be drawn from multiple lines of investigation in cell models of prion disease [254, 255]. This study revealed multiple chaperone molecules such as Pdia3, P4hb and calreticulin that may share these characteristics. The

subsequent observation that pharmacological inhibition of PDIs causes an increase in proteinase K-resistant PrPSc levels in a subset of ScN2a cell clones is exciting as it suggests PDIs may, under certain circumstances, be protective in the context of prion disease, thereby extending observations from an earlier report which documented an increase in Pdia3 levels in prion infected mice and an inverse correlation of PrPSc toxicity and Pdia3 expression levels in mouse neuroblastoma cells [262]. Consistent with a more direct role in PrP conversion reactions, PDIs have been shown to localize to lipid rafts in other cell systems [263], and experiments which employed an in vitro PrPSc amplification methodology suggested that free sulfhydryl groups — through an unidentified mechanism — play a critical role in this reaction [264]. The cellular disulfide exchange homeostasis is a complex aspect of cell biology due to the existence of more than ten PDI family members in mammals with distinct but only partially understood substrate specificities, cellular localizations and intricate regulations [265]. The scope and emphasis of this study was such that it could only begin to address the role these molecules may play in prion biology. However, it is attractive to hypothesize that some cells may put a low ER leakage rate of PDIs to good use by employing these molecules to “patrol” their plasma membrane and prevent undesired occurrences of protein misfolding. In disease, or when pharmacologically suppressed, the rate of misfolding may exceed the rate of PDI-mediated rescue and thus produce a net accumulation of PrPSc. Similar concepts implicating rates of synthesis and degradation have been proposed in a ‘dynamic susceptibility’ model based on observations in N2a cells [266].

Similar to PDIs, calreticulin — in addition to its well established role in the ER — may play an important function in cell-cell recognition in apoptosis and cell-surface recognition during cellular attachment to laminin [267]. This is relevant in the current context as multiple lines of evidence suggest that PrPC confers protection against apoptosis and plays a role in adhesion to the extracellular matrix [268]. While it remains to be determined whether the spatial proximity of PrPC and calreticulin plays a role in these cellular activities, we noted with interest that calreticulin binding to laminin depends on the presence of oligomannosidic N-glycans on laminin [269, 270]. The link to oligomannose-carrying N-glycans was intriguing because the same posttranslational structures had also been signaled by a comprehensive bioinformatic analysis to which our dataset was subjected. The latter pointed to a previously reported similarity in the ability of NCAM and basigin, two proteins we identified with high sequence coverage, to bind to oligomannose-bearing membrane proteins through a specialized Ig-like domain.

Surprisingly, scrutiny of the dataset for proteins reported to carry this posttranslational modification subsequently established their strong enrichment. A snowdrop lectin binding experiment then confirmed that PrPC indeed resides in spatial proximity to proteins carrying oligomannosidic N-glycans in the intact mouse brain. From these observations, a model emerged based on which members of the mammalian prion protein family may primarily ‘organize’ their molecular microenvironment by binding to NCAM and basigin which in turn recognize a small subset of proteins carrying N-glycans of the rare oligomannose type. The latter proteins may themselves be part of protein assemblies containing more than one component.

One of these assemblies surrounds the oligomannosidic carbohydrate-carrying β-subunit of the Na/K ATPase (Atp1b3), a protein complex known to also consist of an α-subunit and a tightly associated G-protein complex, both of which were observed in this dataset and an earlier PrPC interactome investigation based on in vivo crosslinked proteins in the mouse brain [237]. Recently, a role for the cellular prion protein in lactate metabolism has been proposed [271]. The authors also documented interactions of PrPC within a molecular assembly involving basigin and sodium/potassium channels. Our data corroborate key interactions proposed in this initial study and documented that this novel concept is not restricted to PrPC but also extends to Dpl and Sho.

A molecule which may participate in multiple sub-nodes is the integrin β1 subunit (Itgb1). In addition to Itgb1, our data documented spatial proximity of all three bait proteins to the laminin-specific integrin α6 (Itga6) and the integrin αv (Itgav). The binding of bait proteins to these integrins is unlikely to have occurred independently of Itgb1 binding, because stable Itgb1–Itgav (integrin αvβ1) or Itgb1–Itga6 (integrin α6β1, also known as VLA-6) heterodimeric complexes are well described in the literature [272]. VLA-6 plays an important role in the adhesion of cells to laminin and in this capacity has been shown to associate with ecto- calreticulin [273] or LRP [274]. The identification of LRP (Rpsa), better known as the 37- kDa/67-kDa laminin receptor in the prion research field and currently under investigation as a therapeutic target for prion disease [275-277], corroborated previous experiments documenting its spatial proximity to PrPC (and possibly Dpl [278, 279]) and was complemented by the identification of ribosomal protein S21 (Rps21), a protein which not only can bind to LRP but shares with the latter the ability to exert a ribosome-independent function at the plasma membrane [280, 281]. Finally, Itgb1-containing integrins may functionally interact with CD98 (Slc3a2), a type II TM protein [282, 283]. The circle to basigin can be closed by a well-

established interaction between CD98 and basigin [284, 285]. Although the details of this interaction are far from understood, it has been suggested that binding of CD98 to basigin may play a role in cell fusion control, including the specialized fusion of a sperm to the membrane of an egg that precedes fertilization [286-288]. Given the independently proposed role of Dpl in sperm fusion events [289], it will be of interest to determine whether the direct interaction of Dpl with the above protein machinery bears functional significance in this context.

During the assembly of this work, a systems biology study reported on global changes in protein expression during the course of prion disease in mice [290]. Not surprisingly, a comparison of the two datasets revealed little overlap in the proteins which we found to reside in proximity to PrPC and the proteins whose (mRNA) expression is affected in disease. This finding is consistent with an interpretation that global responses to prion disease may not center on proteins which are spatially close to PrPC and also the distinct technological bases of the two datasets. A notable exception to this observation may represent proteins from the ZIP metal ion transporter family of which two paralogs were found in the vicinity of PrPC in this study (ZIP6 and ZIP10), and another paralog ZIP14 was found to be upregulated during the course of prion disease in five independent mouse strains. Interestingly, these three ZIP proteins belong to a common branch (referred to as the LIV-1 subfamily) within mammalian ZIP paralogs whose N- terminal extracellular domain harbors features that are strikingly reminiscent to sequence features found within members of the mammalian prion protein family itself. Extensive follow- up investigations have led to the surprising conclusion that the mammalian prion gene family is phylogenetically derived from a ZIP ancestor molecule (Chapter 3).

The possible participation of PrPC family members in both lactate metabolism and laminin adhesion are unlikely to represent independent functional specializations exerted in the setting of distinct protein microenvironments. Arguing against such a scenario are (i) the known interactions between molecules we identified in close proximity to members of the mammalian prion proteins, and (ii) the recurring theme of oligomannosidic N-glycans which emerged from the bioinformatic analysis. Instead, a unifying model came into focus which suggests that the involvement of PrPC in these biological functions is mediated by remarkably interconnected molecular machineries. In proposing the model, it was not our intent to oversimplify a complex molecular biology that surely must underlie interactions in this network. It is, for example, likely that connections amongst proteins in this network are not arranged in a chain-like and static

fashion through a series of single contact sites. Rather, a more realistic representation of such a network would predict a system built upon highly cooperative and dynamic interactions.

2.5 Conclusion

Taken together, this work suggests that members of the prion protein family may be embedded in specialized membrane domains characterized by an extended molecular network which supports functions in adhesion control, lactate metabolism and cell fusion events (Figure 2.9). Our investigation leads to the conclusion that biochemical interactions among the members of the mammalian prion protein family exist. It confirms multiple candidate interactors of PrPC in a unified paradigm including an independent verification of the laminin receptor precursor (LRP) as an authentic PrP interactor. It thereby not only reconciles a fragmented literature, but should facilitate future efforts to further define the relative contribution of these interactors to the biology of PrPC. Novel interactors uncovered in this study paint a coherent picture based on which an argument for the spatial proximity of NCAM, basigin, Na/K ATPase and LRP can be made, even independent of their ability to bind to PrPC. Moreover, data presented here raise many new questions as to (i) whether biological outputs of PrPC implied by the “guilt-by- association” logic will be confirmed, (ii) the importance of oligomannosidic N-glycans for the structural integrity and assembly of the molecular environment of PrPC, (iii) the role which a subset of cellular chaperones may play in PrP biology and disease, and (iv) the exact environment which hosts the conversion of PrPC. It is hoped that this work will stimulate investigations which will not only resolve these questions, but lead to insights from which a therapeutic strategy for addressing prion diseases can be derived.

Figure 2.9: Mammalian PrPC interactome. Diagrammatic representation of PrPC interactors as identified in Figure 2.3.

Chapter 3 Evolutionary Descent of Prion Genes from the ZIP Family of Metal Ion Transporters

Please note that all parts of this chapter were published in the following article [2]: Gerold Schmitt-Ulms*, Sepehr Ehsani, Joel C. Watts, David Westaway, Holger Wille (2009) Evolutionary Descent of Prion Genes from the ZIP Family of Metal Ion Transporters. PLoS ONE 4(9):e7208. [* corresponding author]

Candidate’s role: Contributed to bioinformatic analysis, manuscript assembly and editing

Summary: Quantitative interactome data in Chapter 2 demonstrated the spatial proximity of two metal ion transporters of the ZIP family, ZIP6 and ZIP10, to mammalian prion proteins in vivo. A subsequent bioinformatic analysis presented below revealed the unexpected presence of a PrP- like amino acid sequence within the N-terminal, extracellular domain of a distinct sub-branch of the ZIP protein family that includes ZIP5, ZIP6 and ZIP10. Additional structural threading and orthologous sequence alignment analyses argued that the prion gene family is phylogenetically derived from a ZIP-like ancestral molecule. The level of sequence homology and the presence of prion protein genes in most chordate species place the split from the ZIP-like ancestor gene at the base of the chordate lineage. This relationship explains structural and functional features found within mammalian prion proteins as elements of an ancient involvement in the transmembrane transport of divalent cations.

3.1 Introduction

As described in Chapter 2, the prion protein is one of the most intensively studied mammalian proteins. Yet surprisingly, the evolutionary origin and physiological function of this protein have remained largely elusive [8, 291]. The function of a protein can sometimes be inferred from genomic investigations which may reveal proteins with similar sequences or sequence modules of known function. Extensive investigations of this kind have provided evidence for PrP-related

sequences in most species of the vertebrate lineage [25, 26, 292] and revealed two mammalian paralog sequences encoding for the proteins Doppel (Dpl) and Shadoo (Sho), which together with PrPC constitute the mammalian prion protein family [40]. Despite these advances, no conclusive biological role has emerged for PrPC from the characterization of the additional members of the mammalian prion protein family [143].

Alternatively, by comparing the structural features of particular proteins, otherwise cryptic functional similarities may be inferred. Investigations of PrPC have revealed an unusual dichotomy of its structure, which consists of an extended, largely unstructured N-terminus and a globular and relatively stable C-terminal domain formed by two short β-strands and three α- helices [167]. While proteins with mixed β-strands and α-helices are widespread in nature, aside from Sho and Dpl which share structural features with the N- and C-terminal regions of PrP, respectively, no other protein has been identified which could be used to infer the physiological function of PrPC based on structural similarities. In fact, the ‘prion fold’, present in both PrPC and Dpl, is thus far unique to these two proteins amongst all high-resolution protein structures obtained to date. Extensive evidence has, however, been accumulated which demonstrates that the prion protein can bind a subset of divalent metal cations through two types of histidine- containing motifs embedded within its N-terminal domain [30].

Most proteins do not act in isolation but partner with other proteins to exert their biological roles [293]. Thus, the function of a protein can sometimes be deduced by characterizing its binding partners. Following this ‘guilt-by-association’ logic, we set out to identify the function of the cellular prion protein through a comprehensive interactome investigation (Chapter 2). Surprisingly, this work not only revealed a subset of the ZIP family of zinc transporters as novel prion protein interactors, but also shed light on the evolutionary origins of the prion gene family and offers an explanation for the ability of prion proteins to bind divalent cations.

3.2 Materials and Methods

3.2.1 Molecular clones

FLAG affinity tags were inserted before residue 29 of mouse PrP, residue 27 of mouse Dpl, and residue 26 of mouse Sho (all in the pcDNA3 mammalian expression vector) using standard PCR- based mutagenesis techniques. The identity of all constructs was verified by DNA sequencing. The expression of FLAG-tagged bait proteins was confirmed by Western blotting using a FLAG- directed primary antibody (Sigma-Aldrich, Oakville, ON, Canada).

3.2.2 Cell culture

Mouse neuroblastoma cells (N2a, clone CCL 131, American Type Culture Collection, Rockville, MD) were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) containing 10% fetal bovine serum (FBS) and 0.2 × penicillin/streptomycin and maintained in a humidified

environment in the presence of 5% CO2. N2a cells were either transiently or stably transfected with the FLAG-prion constructs or a native pcDNA3 vector (negative control) in OptiMEM using Lipofectamine 2000 (Invitrogen Canada, Burlington, ON, Canada) according to the manufacturer’s protocol. For bulk selection of stably transfected cells, cultures were expanded in the presence of 1 mg/mL G418 and maintained at a concentration of 0.2 mg/mL G418.

3.2.3 Western blotting

Cells were lysed in 0.5% sodium deoxycholate, 1% NP-40, 150 mM NaCl, 50 mM Tris/HCl, pH 8.0 and extracts clarified by centrifugation at 20,000 × g for 10 min at 4°C. Protein concentrations were determined using the bicinchoninic acid assay (BCA) (Pierce Biotechnology, Rockford, IL, USA). Proteins were separated on 4-12% NuPAGE gels (Invitrogen) and subsequently transferred to polyvinylidene fluoride (PVDF) membranes. Membranes were blocked with 5% milk in Tris-buffered saline containing 0.05% Tween-20 (TBST). Membranes were incubated overnight at 4°C with primary antibody, washed three times with TBST, and incubated with horseradish peroxidase-conjugated secondary antibody (Bio-Rad Laboratories, Hercules, CA, USA) for 2 h at room temperature. Following three washes with TBST, membranes were developed using Western Lightning ECL (PerkinElmer, Woodbridge, ON, Canada).

3.2.4 In vivo crosslinking

Mild formaldehyde crosslinking of N2a cells followed a protocol described before[163]. Briefly, cells grown to confluency were washed with phosphate buffered saline (PBS) and subjected to 15 min crosslinking with 2% w/v formaldehyde in PBS at room temperature. The crosslinking reaction was quenched by incubating cells for 15 min with 125 mM glycine in PBS.

3.2.5 Affinity purification of bait proteins

Approximately 108 in vivo formaldehyde crosslinked cells each of control and FLAG-prion expressing N2a cell lines were lysed in homogenization buffer (50 mM NH4Cl, 80 mM Tris, pH 8.0) supplemented with 1× Complete Protease Inhibitor Cocktail (Roche, Palo Alto, CA, USA). To ensure near quantitative extraction of membrane proteins, an equal volume of extraction buffer (20 mM NaCl, 1% deoxycholate, 1% NP-40, 20 mM Tris, pH 8.0) was added, followed by a 30-min incubation and 5-min sonication in a water bath sonicator. Insoluble cellular debris was removed by high-speed centrifugation (100,000 × g, 1 h). Subsequently, the crosslinked bait protein complexes were immunoaffinity-captured on anti-FLAG-agarose (Sigma-Aldrich, Oakville, ON, Canada). During this step samples were gently agitated on a turning wheel for 12 h, then washed extensively with 0.5 M NaCl, 0.05% SDS, 1% NP-40, 20 mM HEPES, pH 7.3,

and detergents removed by a pre-elution wash with 10 mM NH4HCO3, pH 8.0. Proteins were eluted by acidification with 0.2% trifluoroacetic acid, 20% acetonitrile, pH 2.0.

3.2.6 Protein reduction, alkylation and trypsinization

Protein-containing fractions were denatured in the presence of 6 M urea, 20 mM NH4HCO3, pH 8.0, followed by reduction with 1 mM Tris-(2-carboxyethyl)-phosphine for 30 min at 60°C and alkylation with 2.5 mM 4-vinylpyridine for 1 h at room temperature in the dark. Samples were diluted four-fold to ensure that the concentration of urea did not exceed 1.5 M. Tryptic digestion was initiated by the addition of 1% (wt/wt) side chain-modified, TPCK-treated porcine trypsin and allowed to proceed at 37°C for 6 h.

3.2.7 iTRAQ labeling

Following trypsinization, equal quantities of tryptic peptide mixtures were spiked with 1 pmol of synthetic (Glu1)-Fibrinopeptide B (GluFib) (Sigma-Aldrich) to serve in the downstream analysis

as an internal control for the efficiency of individual labeling reactions. Equal labeling with all four reagents was confirmed by the equal intensities of 114:115:116:117 signature peaks upon forced fragmentation of the GluFib [M+2H]2+ parent ion at m/z 785.85. Any strong deviation from this ratio would have indicated problems with the labeling reaction or recovery of individual samples prior to the sample mixing step. Individual iTRAQ labeling reagents (Applied Biosystems, Foster City, CA, USA) were reconstituted in ethanol, added to peptide mixtures derived from the tryptic digestion of immunoprecipitation (IP) eluates (Control: iTRAQ 114, Dpl: iTRAQ 115, PrP: iTRAQ 116 , Sho: iTRAQ 117) and incubated at room temperature in the dark for 3 h.

3.2.8 Two-dimensional liquid chromatography

Strong cation exchange (SCX) chromatography was used to achieve peptide fractionation of the complex digest mixture. Samples digested with trypsin were adjusted to 25% acetonitrile and

acidified (pH 3.0) by 20-fold dilution in 25% acetonitrile, 20 mM KH2PO4, pH 3.0. HPLC was carried out using the Ultimate System (Dionex, Sunnyvale, CA, USA) equipped with a microflow calibration cartridge, a Valco injection port and a 180 nL volume UV cell. Separation was achieved on a self-packed 0.5 mm × 110 mm Luna SCX (Phenomenex, Torrance, CA, USA)

column at a flow rate of 18 µL/min with a steep salt gradient from 0–400 mM NH4Cl in 25% acetonitrile, 20 mM KH2PO4, pH 3.0. Fractions eluted from the SCX column were desalted with C18 Empore (3M, Minneapolis, MN, USA) stop and go extraction (STAGE) tips[294] and subsequently subjected to nano-flow RP-HPLC using the Ultimate LC system (Dionex) equipped with a nanoflow calibration cartridge at a flow rate of 250 nL/min. Peptides were separated on a 75-µm ID self-packed column containing Proteo C12 reverse-phase matrix (Phenomenex) using a 100-min gradient from 2%–34% acetonitrile in water, with 0.1% (wt/vol) formic acid as the ion-pairing agent.

3.2.9 ESI-QqTOF mass spectrometry analysis

The analysis of samples by tandem mass spectrometry was essentially done as described before [295]. Briefly, the column effluent was coupled directly via a fused silica capillary transfer line to a QSTAR XL hybrid quadrupole/time-of-flight tandem mass spectrometer (Applied Biosystems; MDS Sciex, Concord, ON, Canada) equipped with a MicroIonSpray source. The progress of each LC/MS run was monitored by recording the total ion current (TIC) as a function

of time for ions in the m/z range of 300 to 1800. At 5-s intervals through the gradient, a mass spectrum was acquired for 1 s, followed by one collision-induced dissociation (CID) acquisition of 4 s each on ions selected by preset parameters of the information-dependent acquisition (IDA) method, using nitrogen as the collision gas. Singly-charged ions were excluded from CID selection. The collision energy was adjusted automatically for each CID spectrum using an empirically optimized formula which considers the charge state and m/z value of the precursor ion.

3.2.10 Database searches

Peak lists for database searching were created using Mascot Distiller (Version 1; MatrixScience, London, UK). Searches were performed using designated MS/MS data interpretation algorithms within Protein Prospector (Version 4.21.3; University of California, San Francisco, CA, USA) [241] and Mascot (Version 2.2; MatrixScience). Modifications considered were oxidation of methionine, phosphorylations of serine and threonine, N-terminal (pyro)Glu and alkylation with 4-vinylpyridine. Searches further considered up to one missed cleavage and charge states ranging from +2 to +4. For a protein to be listed in the data tables it had to be identified by both search algorithms. In the few instances where only two peptides supported the identification of a protein, we required the underlying CID spectra to generate a Mascot score indicating a <5% probability that the match could be considered a random event [296] and further confirmed matches by the peptide sequence tag approach [297] and manual interpretation of spectra. As searches were carried out without species restriction, the correct assignment of matches to mouse entries served as an additional internal control. The mass tolerance range between expected and observed masses used for database searches was ±150 ppm for MS peaks, and ±0.15 Da for MS/MS fragment ions. These relatively large thresholds were used to capture more of the low intense peaks that frequently display broader distribution and thus are assigned with lower mass accuracy. Threshold levels were optimized based on LC/MS/MS datasets of tryptic digests of standard proteins. All samples were searched against the National Center for Biotechnology Information nonredundant database (nrNCBI, release: June 2008) and a ‘decoy’ database in which all entries of the above NCBI database had been inverted. The analysis of iTRAQ data was assisted by the software program ProteinPilot (Applied Biosystems; MDS Sciex). A feature of this software package was used to correct raw iTRAQ ratios for impurity levels of individual reagent lots determined by the manufacturer.

3.2.11 Multiple sequence alignments

Multiple sequence alignments were obtained using a combination of MAFFT [298] and ClustalX2’s implementation of ClustalW2[299]. Manual adjustments were made to further improve alignments. Sequence identity and conservation were determined using the AlignX feature of Vector NTI Advance 10.3.1 (Invitrogen, Carlsbad, CA, USA) [300].

This algorithm calculates identity values based on the percentage of identical residues among all ungapped positions between sequence pairs. Similarity values are calculated based on the percentage of identical plus similar amino acids among all ungapped positions between pairs. To determine the statistical significance of alignments, the COMPASS program was used [301]. Analyses were based on the ‘SCOP40 iter5’ database, gap opening and extension penalties of 10 and 1, respectively, and the BLOSUM62 substitution matrix.

3.2.12 Structural threading

The sequences of the PL domains of ZIP5, ZIP6 and ZIP10 were submitted to the FFAS03 fold and function assignment server [302] for structure prediction and structural threading. In order to eliminate threading artifacts based on uncertainties in determining the boundaries of the PL domain, we submitted a number of alternatively truncated sequences to the FFAS03 server. This approach helped to better define the boundaries of the folded region of the PL domain. The FFAS03 server provided a ranked list of structural threading models based on various Protein Data Bank (PDB) entries; invariably the highest scoring models were based on PrP and Dpl structures [303-305]. We analyzed all sensible models for the position of secondary structure elements and derived a consensus prediction for the secondary structure of the PL domain of ZIP10. The closest model to the consensus prediction was chosen for the corresponding figure. The figures of the PrP and Dpl structures and of the ZIP10 PL domain were produced using the Chimera package from the Computer Graphics Laboratory, University of California, San Francisco (supported by NIH P41 RR-01081) [306].

3.2.13 Accession numbers

Please refer to Table 3.1 below.

Table 3.1. Sequences from Entrez, Ensembl and UniProt databases utilized for alignments.

SEQUENCES Entrez Ensembl UniProt

Ag_ZIP5/6/10 XP_317935.4 Dm_ZIP5/6/10 (1) NP_001097608.1 Dm_ZIP5/6/10 (2) NP_523974.3 Dr_PrP NP_991149.1 Dr_Sho2 ENSDARP00000074841 Dr_ZIP5 XP_690258.2 Dr_ZIP6 NP_001001591.1 Dr_ZIP8 ENSDARG00000028214 Dr_ZIP10 NP_956965.1 Dr_ZIP14 XP_001340102.2 Ga_PrP-1 CAL64057.1 Gg_PrP ENSGALP00000040285 Hm_ZIP12 XP_002160698.1 Hs_PrP ENSP00000368752 Hs_ZIP4 ENSP00000276833 Hs_ZIP5 ENSP00000266980 Hs_ZIP6 ENSP00000269187 Hs_ZIP8 ENSP00000378310 Hs_ZIP10 ENSP00000352655 Hs_ZIP12 ENSP00000366588 Hs_ZIP13 ENSP00000346956 Hs_ZIP14 ENSP00000352779 Mm_PrP NP_035300.1 Mm_Sho CAF18554.1 Mm_Dpl AAF02544.1 Mm_ZIP4 ENSMUSP00000073134 Mm_ZIP5 AAH28990.1 Mm_ZIP6 NP_631882.2 Mm_ZIP8 ENSMUSP00000029810 Mm_ZIP10 ENSMUSP00000027131 Mm_ZIP12 ENSMUSP00000080911 Mm_ZIP13 ENSMUSP00000073263 Mm_ZIP14 ENSMUSP00000066108 Tn_PrP Q4SJ93 Tr_PrP AAN38988.1 Tr_Sho1 CAG34291.1 Tr_Sho2 CAG34292.1 Tr_ZIP6 ENSTRUP00000006592 Tr_ZIP8 ENSTRUP00000045288 Tr_ZIP10 ENSTRUP00000007404 Tr_ZIP12 ENSTRUP00000019248 Tr_ZIP14 ENSTRUP00000038591 Ts_PrP Q9I9C0 Xl_PrP NP_001082180.1

Ag → Anopheles gambiae (African malaria mosquito) Dm → Drosophila melanogaster (fruit fly) Dr → Danio rerio (zebrafish) Ga → Gasterosteus aculeatus (three-spined stickleback) Gg → Gallus gallus (chicken) Hm → Hydra magnipapillata (hydra)

Hs → Homo sapiens (human) Mm → Mus musculus (house mouse) Tn → Tetraodon nigroviridis (spotted green pufferfish) Tr → Takifugu rubripes (Japanese pufferfish) Ts → Trachemys scripta (red-eared slider turtle) Xl → Xenopus laevis (African clawed frog)

3.3 Results

3.3.1 Quantitative interactome analyses

Various earlier attempts by us and others had already led to the identification of a few dozen proteins that co-purify with the cellular prion protein under a range of experimental conditions [237]. Often investigations of this kind result in long lists of candidate interactors. The challenge then remains to discriminate specific from unspecific binding partners. Here we incorporated quantitative mass spectrometry based on isobaric tagging of peptides into the workflow to overcome this limitation [307]. Furthermore, investigations were extended to all three members of the mammalian prion protein family to further facilitate discrimination of potential interactors by differential interactome comparison. A murine neuroblastoma cell line (N2a), a cell model widely used for studying prion replication, served as the biological source material [238]. Cells were stably transfected with expression plasmids coding for individual members of the mammalian prion protein family that had been FLAG-tagged in the vicinity of the N-terminus of the mature protein or with a negative control vector [240]. The conceptual choice of both the FLAG-tag and N-terminal attachment site was guided by data documenting that such insertions do not interfere with either PrPC processing or conversion in transgenic animals [239], as well as studies of glycosylation and trafficking in Dpl- and Sho-transfected cells (Coomaraswamy, J., et al., in preparation). To stabilize physiologically relevant interactions, adherent cells were crosslinked by a short treatment with formaldehyde prior to the cell harvest step [163]. Following cell lysis in the presence of detergents, crosslinked protein complexes were affinity- purified based on the presence of the FLAG-affinity tag, trypsinized in solution, and subjected to labeling with isobaric tags for relative and absolute quantitation (iTRAQ) [243]. The samples for the separate prion proteins and the control were mixed and jointly analyzed by tandem mass spectrometry. It should be noted that subsequent to iTRAQ conjugation, the contribution of each sample to the identification of a peptide can be readily calculated by determining the relative intensity of signature iTRAQ mass signals in the relevant collision-induced dissociation (CID)

63 spectra. A comprehensive analysis of samples led to the identification of approximately one hundred proteins, more than thirty of which were observed with iTRAQ signature mass intensity ratios that suggested specific co-enrichment with members of the prion protein family (Figure 3.1, Chapter 2). Strikingly, two proteins, murine ZIP10 (Slc39a10) and ZIP6 (Slc39a6), which were unequivocally identified with non-overlapping peptides in Dpl, Sho and PrP interactome samples but not in the negative control, represent members of the ZIP family (Zrt-, Irt-like Protein) of transmembrane zinc ion transporter proteins (Table 3.2). In humans and mice, the slc39a gene family codes for fourteen distinct proteins (ZIPs 1-14). Amino acid sequence comparisons of the human ZIP proteins argue that ZIP10 and ZIP6, together with their phylogenetically closest paralog ZIP5, constitute a distinct sub-branch within this family [10].

Figure 3.1: Evidence for specific co-enrichment of ZIP10 and ZIP6 with all three members of the mammalian prion protein family. Comparison of FLAG-affinity chromatography eluates by quantitative tandem mass spectrometry. Side-by-side affinity purified bait protein complexes were trypsinized and subjected to iTRAQ labeling of peptides as follows: iTRAQ114 label: empty vector; iTRAQ115 label: FLAG-Dpl; iTRAQ116 label: FLAG-PrP; and iTRAQ117 label: FLAG-Sho. A. ZIP10 co-purified specifically with the three bait proteins. Collision- induced dissociation (CID) spectrum from ZIP10 derived peptide with amino acid sequence QSTEEIGR ([M+2H]2+, m/z 626.35). Inset: Low mass iTRAQ reporter ion region documenting relative contribution to the identification of this peptide by samples labeled with iTRAQ115 (FLAG-Dpl), iTRAQ116 (FLAG-PrP) and iTRAQ117 (FLAG-Sho) reagents but not negative control sample labeled with iTRAQ114 reagent. B. CID spectrum derived from ZIP6 peptide with amino acid sequence ESASSSEVTSAVYNAVSEGTR ([M+3H]3+, m/z 759.03). iTRAQ reporter ions document that ZIP6 co-purified specifically with the three bait proteins. C. Actin co-purified unspecifically in all four samples including the negative control. CID spectrum derived from actin peptide with amino acid sequence TTGIVMDSGDGVTHTVPIQEGYALPHAILR ([M+4H]4+, m/z 666.35).

Identified proteinsa IPI accession Amino Peptidesd Uniquee % Covf 114b 115 116 117 number acidsc Control Dpl PrP Sho Specific binders Dpl IPI00131622.1 178 8 13 39.1 2.0 92.5 3.0 2.5 PrP IPI00120793.1 254 5 5 26.8 4.0 20.5 65.1 10.5 Sho IPI00226455.1 147 3 6 50.4 2.9 11.4 20.5 65.2 ZIP10 IPI00273801.3 833 2 2 4.3 2.4 36.3 16.7 44.6 ZIP6 IPI00469000.4 765 3 3 7.3 3.2 29.9 32.8 34.2 Unspecific binders Actin IPI00110850.1 375 21 46 88.5 19.1 21.6 31.8 27.5 aProteins were sorted into specific versus unspecific binder categories based on their iTRAQ distribution, i.e. proteins were considered unspecific interactors if their derived CID spectra revealed iTRAQ114 to 117 signature mass peak signal intensities which exceeded 10% of combined intensities for all samples including the unspecific control. bFor the calculation of iTRAQ values the intensity of individual peptide associated iTRAQ signature peaks was normalized to combine to 100% per peptide and subsequently averaged. Standard deviations were determined and are listed in Supplemental Table 3.1. cLength of precursor molecules prior to posttranslational processing. dOnly CID spectra underlying different peptides were considered i.e. if the same peptide was identified with different charge states or modifications it counted as one hit. eTotal number of unique CID spectra. Please note that the same peptide was only counted more than once if it was identified with different charge states or modifications. fPercent sequence coverage based on the presence of peptides for which no higher ranked assignment to other proteins could be made.

3.3.2 Structural and sequence similarities of mammalian prion proteins and ZIPs

Intrigued by the presence of the aforementioned ZIP proteins in the dataset (which we had never observed in similarly generated datasets with different bait proteins), these proteins were subjected to an extensive bioinformatic analysis. Surprisingly, the results revealed the presence of a domain within this subset of mouse ZIP proteins that displayed a substantial amino acid sequence similarity to both PrP and Dpl. This initial search was conducted with the SUPERFAMILY database of structural and functional protein annotations (version 1.69) against SCOP (Structural Classification of Proteins) [308] linear hidden Markov models (HMMs; Superfamily model 0037705), a subset of which had been trained on multiple sequence alignments of prion protein ortholog sequences (Superfamily 54098, designated as ‘Prion-like’) [309]. More specifically, a 111-amino acid fragment within the N-terminal extracellular domain of murine ZIP10 (residues 285 to 395) not only showed good general alignment (16% identity, 42% similarity) with the C-terminal globular domain of mouse PrP, but also demonstrated positional agreement of both conserved cysteine residues (which form a disulfide bridge in PrPC) and the first N-glycosylation site ‘NxT’ motif found within this domain (Figure 3.2A). Highly similar motifs were also found within the corresponding regions of ZIP6 and ZIP5. Remarkably, the sequence similarity between ZIP10 and PrP in this region is comparable to the similarity observed between Dpl and PrP (18% identity, 44% similarity). No SCOP linear HMM outside the ‘prion-like’ family of HMMs aligned better to this region (from a total of 982 different models in this SCOP release). Indicative of the specificity of this annotation, aside from ZIPs and the aforementioned PrP family members, no other protein among the more than 120,000 mouse and human proteins contained in the LOCATE subcellular localization database (URL: http://locate.imb.uq.edu.au) [310] is recognized to contain a ‘Prion-like’ domain. The E-value for this assignment was 0.011 and therefore had to be qualified as ambiguous. However, SCOP Superfamily algorithms are not optimized to determine remote homologies amongst protein families but instead are largely employed for an initial detection and annotation of protein domains. To further assess the statistical significance of the sequence similarity between PrP and ZIPs in this domain and to ensure that an overemphasis was not placed on outlier sequences, we applied a method dedicated to the detection of remote homologies which uses profile-profile alignments and is embedded in an algorithm that can be initiated from the COMPASS program

[301]. A query of this algorithm with the mouse ZIP10 segment 285-385 returned the SCOP40 database entry ‘Prion protein domain’ (designator d.6.1.1) as the only hit (E-value = 4.51e-4) which passed the statistical significance threshold of 5e-3. The low E-value strongly indicates that sequence similarities are not restricted to a pair of spurious outlier ZIP and PrP sequences and classifies the respective domains within ZIPs and PrP as homologous.

Figure 3.2: Structural similarity between mouse ZIP10, PrP and Dpl. A. Structural threading of the PL domain within ZIP10 predicts striking resemblance to PrP and Dpl with regard to relative position and order of secondary structure motifs. The secondary structure for the PL domain of ZIP10 was ranked according to the frequency of the prediction in separate threadings (H 75-100% and H 50-75% for α-helices; the same scale applies to β-sheeted structures “E/E”). Also shown is a matrix of amino acid identities and similarities between depicted sequences of ZIP10, PrP and Dpl. B. Comparison of high-resolution nuclear magnetic resonance structures of PrP (PDB entry: 1ag2) and Dpl (PDB entry:1I17) with a predicted structure for the ZIP10 PL domain. Dark green, grey and black highlights depict conserved, similar and identical residues, respectively. ‘S’ and ‘G’ labels indicate sites of disulfide linkages and glycosylation, respectively.

Proteins of common evolutionary origin are frequently characterized by the presence of similar folds. Well-studied examples include proteins of the immunoglobulin and fibronectin type III superfamilies. Such proteins are often the result of divergent evolution during which they have accumulated differences in their primary structure but continue to sustain similar folds. The

comparison of protein structures therefore constitutes an orthogonal approach in studying the relationship of proteins. Whereas a multitude of high-resolution structures of prion proteins from many species have been solved, similar information is currently not available for metal ion transporters of the ZIP family.

We employed the FFAS03 fold and function assignment server [302] to carry out a search for any deposited protein structure that would allow (i) a meaningful threading of the ‘prion-like’ (PL) domain in ZIPs 5, 6 or 10 onto any structure template, and (ii) would meet the stringent threshold criteria for significant homology matches embedded in the underlying FFAS03 fold and function assignment algorithm. The published structures of the prion protein and its paralog Doppel were the only structures that fulfilled both of these criteria. More specifically, the profile-based sequence alignment revealed prominent sequence similarities between mouse PrP, Dpl and ZIP10 (Figure 3.2A), which allowed the threading of the sequence of the PL domain of ZIP10 onto the structures of both PrP and Dpl (Figure 3.2B).

The FFAS03 structural alignment score for the ZIP10 PL domain (murine ZIP10 residues 285-395 achieved statistical significance for an alignment with the structure of Xenopus laevis PrP (PDB entry: 1xu0) [305] with a value of -9.6 (and a threshold of -9.5 indicating less than 3% false positives). Similarly, the ZIP5 PL domain (murine ZIP5 residues 96-212) also generated a FFAS03 structural alignment score with the same PrP structure in the statistically significant range (with a value of -10.3). Consistently, the models with the strongest scores were based on alignments of PL domains of ZIP proteins with PrP or Dpl structures and all attempts to align the ZIP PL domain to structural templates outside the prion protein family produced insignificant fits. The PL domain of ZIP10 (and also of ZIP5 and ZIP6, not shown) is predicted to contain a structural arrangement very similar to that of PrP or Dpl with three α-helices and possibly a small β-sheet composed of two short β-strands (Figure 3.2B). The two C-terminal α-helices of the ZIP5, ZIP6, and ZIP10 PL domains (helices B and C) are predicted to be stabilized by a disulfide bridge as is the case with the structures of PrP and Dpl. The root mean square deviation (RMSD) between backbone carbon atoms of the nuclear magnetic resonance (NMR) structures for PrP and Dpl is 3.7 Å (as determined by the DaliLite server [311]). Surprisingly, the RMSD between the predicted structure of the ZIP10 PL domain and the PrP and Dpl structures returned even lower values of 2.6 Å and 2.9 Å, respectively, indicating that the primary structure of the ZIP10 PL domain is highly compatible with the basic prion protein fold. Consequently, this

observation argues that the secondary structure elements found within the C-terminal domains of PrP and Dpl originated from similar structural features in the ZIP PL domain.

3.3.3 Biological similarities between PrP and ZIPs

We next compared the known biological features of prion proteins and ZIP proteins of the phylogenetic branch comprising ZIPs 5, 6 and 10 (Table 3.3). Like PrPC, ZIP6 and ZIP10 exhibit widespread expression in biological tissues with high transcript levels in the brain [10]. ZIP6 transcripts can also be found in the testis, the predominant site of Dpl expression. The expression of ZIP5 is more restricted and no ZIP5 transcripts can be found in N2a cells by RT- PCR (Chapter 5), consistent with the absence of ZIP5 peptides in our interactome dataset. Reminiscent of the N-terminal repeat motifs in PrP, ZIPs 5, 6 and 10 are equipped with histidine- rich sequences N-terminal to their PL domains. Also reminiscent of PrP, the cysteine-flanked core (CFC) domains (Figure 3.3A) in ortholog sequences of these ZIPs are more highly conserved than the more N-terminal portions of their extracellular sequences. The diversification of the latter is likely aided by an increased chance of recombination events generally associated with tandem repeat regions and previously suggested for prion sequences from different species [24].

Figure 3.3: Molecular organization, mode of membrane attachment and phylogenetic relationship of ZIP and prion gene families. A. Schematic drawing depicting the molecular organization of ZIPs 5, 6, 10 (only consensus features shown) and PrP. B. Comparison of orientation and membrane topology of mammalian prion family proteins, and a consensus ZIP5/6/10 molecule. C. Tree diagram depicting the human ZIP family of zinc metal ion transporters. The ZIP protein family in humans (and mice) consists of fourteen paralogs which can be grouped into four subfamilies based on sequence similarities (indicated by different background shading) [10]. Green shading indicates the sub-branch of paralog ZIP sequences most similar to prion family gene sequences. The right side of the panel compares the molecular organization of N-terminal domains within human ZIPs. Please note the divergence in lengths and molecular organization of N-terminal sequences which is contrasted by the presence of a relatively well-conserved cysteine-flanked core (present in 8 out of 14 paralogs) and highly conserved transmembrane domains. D. Simplified phylogenetic tree (modeled after [312]; arbitrary branch lengths) and table depicting the wide distribution of ZIP sequences in most organisms. Green shading indicates the phylogenetic branch of chordates with widespread existence of prion gene ortholog sequences in their genome. The paralog distribution was deduced from an ortholog alignment of 633 ZIP sequences published by the Sanger Institute (TreeFam release 7.0, http://www.treefam.org) or was determined by aligning representative ZIP protein sequences to genomic sequences (proteobacteria, archaea, Cnidaria and fish). The number of ZIP subfamily IV (LIV-1) paralogs containing a cysteine-flanked core (CFC) domain (indicated in brackets) was determined by inspection of sequences for the presence of cysteines that (i) flank a PALLY-like signature motif, and (ii) adhere to CFC consensus distance constraints derived from the multiple alignment of confirmed CFC domains: the cysteine-to- cysteine distance and the cysteine-to-transmembrane attachment site distance. Please note that ZIP proteins harboring CFC sequences N-terminal to their transmembrane domains can be found in organisms with relatively primitive body plans such as hydra (H. magnipapillata) or protostomia such as the fruitfly (D. melanogaster).

Table 3.3: Comparison of mouse PrP gene family paralogs with ZIPs 5/6/10.

Name Localization Lengtha Expression Function Mode of Extra- PL domain to membrane cellular membrane Metal binding/ Zebrafish knockout Protein Gene Compartment attachment N-terminus Total domain distance Tissue N2a transport phenotype Kidney, liver, Plasma mZIP5 slc39a5 Type-3 TM Extracellular 535 ~210 ~15 spleen, colon No Zn ndb Membrane stomach, pancreas Widespread, Impaired mesoderm Plasma mZIP6 slc39a6 Type-3 TM Extracellular 765 ~334 ~22 prominent in Yes Zn formation during gastrulation, Membrane brain and testis altered E-cadherin expression [118]

Plasma Widespread, mZIP10 slc39a10 Type-3 TM Extracellular 833 ~407 ~30 Yes Zn nd Membrane prominent in brain

Impaired mesoderm Plasma Widespread, mPrP Prnp GPI-anchor Extracellular 254 230 17 Yes Cu, Zn, Fe formation during gastrulation, Membrane prominent in brain altered E-cadherin expression [119]

Plasma mSho Sprn GPI-anchor Extracellular 147 122 n/a Brain Yes ndb nd Membrane

Plasma mDpl Prnd GPI-anchor Extracellular 179 157 14 Testis, heart No Cu nd Membrane aIn amino acids. Numbers for ZIP proteins are tentative because the exact N-terminal boundary of their transmembrane domain 1 is not known. N-terminal signal peptides included. bNot determined.

The PL domains within ZIPs 5, 6 and 10 are part of their extracellular N-terminal domains and, with regard to orientation and relative distance to their downstream membrane anchorage sites, resemble PrPC (Figure 3.3A, B, C; Table 3.3). Both PrPC and the above ZIPs belong to a small group of proteins known to be able to concomitantly bind multiple divalent metal ions through histidine-containing motifs embedded in N-terminal repeat sequences [10, 313]. While it is unknown whether PrPC’s metal binding ability primarily serves a purpose for sensing, scavenging, or transport of divalent cations in vivo, roles of the prion protein in the cellular response to copper-induced oxidative stress [140] and zinc homeostasis [162, 314] have been proposed before, and members of the ZIP protein family are well-characterized in their ability to transport zinc and other divalent metals across membranes [315]. Finally, ZIPs 5, 6 and 10 reside, like PrPC, in the plasma membrane and are expected to facilitate import of extracellular zinc into the cytoplasm [10, 315]. Taken together, this context analysis uncovered multiple additional commonalities between ZIP proteins and PrPC consistent with the interpretation that these molecules are evolutionarily related.

3.3.4 Convergent evolution versus common evolutionary origin

It is well-established that similar secondary and tertiary protein structures can evolve independently, as is the case with structural similarity of prokaryotic subtilisin and eukaryotic chymotrypsin [316]. However, instances of convergent evolution at the protein sequence level are rare. Whenever reported, observations appear restricted to convergent adjustments of individual amino acids rather than independent evolutionary inventions of extensive blocks of similar primary structure. A frequently cited case represents the independent adaptation of the protein lysozyme to the acidic milieu found in the digestive tracts of unrelated species [317]. To investigate whether the similarities between PrP and ZIP10 constitute a case of convergent evolution, we analyzed ortholog sequences of both PrP and ZIP10 across a wide range of species within the chordate lineage. This was based on the conception that a divergent trend in sequence similarity would be indicative of evolution from a common phylogenetic ancestor. The opposite scenario, in which the highest sequence similarity is seen in phylogenetically distant species, would constitute evidence for convergent evolution between PrPC and ZIP10 sequences. PrP- related sequences can be found in all mammals and most species of the vertebrate lineage. In contrast, ZIP proteins date back much further and are highly conserved throughout evolution.

Related sequences can be found in all kingdoms of life, including bacteria and plants (Figure 3.3D). We therefore compared PrP and ZIP10 ortholog sequences from mammals and fish. Alignment of PrP and ZIP10 sequences from pufferfish (Takifugu rubripes) revealed multiple additional amino acid residues conserved between fish ZIP10 and PrP-1, primarily in highly conserved sequence positions (28% sequence identity between pufferfish sequences compared to 16% identity observed in the respective murine sequences) (Figure 3.4A). Consistent with the structural threading (Figure 3.2), small gaps in the alignment of pufferfish sequences corresponded to predicted loops between helices B and C, and the region between helix C and the signal peptide for GPI anchor attachment in PrP (equivalent to the first transmembrane domain in ZIP proteins). Furthermore, a highly conserved ‘PALxxQ’ motif noted by Taylor et al. [10] is present within the PL domain of ZIP10 orthologs but is not found in mouse or human PrP sequences. Interestingly, both pufferfish and Tetraodon nigroviridis PrP-1 sequences contain this motif (PAL(V/I)(D/E)Q) in perfect positional agreement with the ZIP10 sequence. PrP sequences from turtle (Trachemys scripta) (Figure 3.4A), chicken (Gallus gallus) and frog (Xenopus laevis) (Figure 3.5), organisms that lie between fish and mammals on the evolutionary scale, gave intermediate percent identity values when aligned with ZIP10 sequences (e.g. 18% sequence identity between turtle PrP and pufferfish ZIP10). The elevated sequence identity observed between ZIP10 and PrP protein sequences from fish argues against convergence as an explanation for the similarity of the ZIP and PrP PL domain sequences. Instead, this analysis suggests that PrP descended from a progenitor in the more ancient ZIP family of zinc ion transporters. Another alternative to this parsimonious explanation would be a viewpoint favoring a reverse phylogenetic relationship between these gene families, with PrP as the progenitor. However, here it would be necessary to posit that prion gene sequences disappeared independently in most kingdoms of life except in chordates.

Figure 3.4: Sequence evidence for common origin and divergent sequence evolution of members of ZIP and prion protein families. A. Multiple sequence alignment of PrP globular domain with ZIP10 PL of selected ortholog sequences. B. Multiple sequence alignment of N- terminal Sho and ZIP5 sequences from pufferfish, zebrafish and mouse. Please note the greater divergence of ZIP10 and PrP in human (Hs: H. sapiens) and murine (Mm: M. musculus) sequences relative to the respective sequence pairs in turtle (Ts: T. scripta) and pufferfish species (Tn: T. nigroviridis and Tr: T. rubripes). Highlight colors used in the alignments are as in Figure 3.2. ‘N’, ‘S’, ‘G’ and ‘C’ labels indicate N-termini, disulfide linkages, glycosylation sites and C- termini, respectively. Colors used in the schematic drawing are as in Figure 3.3A. For full-length multiple alignments of a subset of these and related sequences, please see Supplemental Figure 3.1.

Figure 3.5: Multiple sequence alignment of cysteine-flanked core sequence segment within PL domain. Our data specifically point at a prion ancestor gene in the ZIP family sub-branch containing ZIPs 5, 6 and 10. Please note that all relevant ZIP protein sequences in this branch harbor both flanking cysteines, consistent with the interpretation that these cysteines may engage in a direct disulfide bridge (analogous to the situation in prion proteins). Similarly, the NxT glycosylation motif is shared amongst ZIPs 5, 6, 10 and prion sequences ranging from pufferfish to humans but not found in more distantly related ZIP paralogs. Amino acid-specific colors are as in Supplemental Figure 3.1.

3.4 Discussion

Multiple lines of evidence place a ZIP5/ZIP6/ZIP10-like ancestor gene at the root of the PrP gene family. No single evidence we uncovered is on its own sufficient to firmly establish the phylogenetic relationship between ZIP and prion genes. However, the many orthogonal and corroborating pieces of evidence we collected and, equally important, the absence of any conflicting observations, cumulatively support this conclusion (Table 3.4, see also discussion

78 points below). Beyond explaining the origin of prion genes in the vertebrate lineage, this finding provides a rationale for structural and biological features within modern-day prion proteins as remnants of an ancient involvement in the sensing and/or import of metal ions from the extracellular milieu.

Table 3.4. Summary of evidences presented in support of evolutionary descent of PrP gene family from ZIP metal ion transport ancestor gene.

Evidence Category Description of Evidences 1. Sequence ZIP10 constitutes the only non-prion gene hit by SCOP “prion-like” HMM (out of 120,000 entries in LOCATE human-mouse protein database). COMPASS profile-profile analysis passed homology E-value threshold and confirmed that similarity of PL domain sequences is not merely restricted to spurious outliers. The GPI-attachment sequence of prion gene sequences shows sequence similarity with the TM1 domain found in ZIPs. Precedents exist for the transformation of a transmembrane sequence into a signal peptide for GPI anchor attachment. An additional pair of ZIP and prion gene sequences (zebrafish ZIP5 / pufferfish Sho2) exhibits a degree of sequence identity/similarity which falls on the significance threshold indicating homology. ZIP genes contain histidine-rich repeat motifs reminiscent of octarepeats in prion sequences.

A zebrafish PrP sequence has been documented which shares the presence of N-terminal [HX]n clusters with ZIPs. 2. Structure A common distance of cysteine-flanked core domains to membrane attachment sites is observed in both prion and ZIP protein families. Precedents of protein families exist with individual members employing transmembrane domains or GPI anchors for membrane attachment. A systematic attempt to thread ZIPs 5/6/10 to any protein structure in the PDB led to the independent assignment of the prion fold. The prion protein structures are the only fold templates onto which ZIP sequences can be threaded with scores that pass the threshold for significant homology. ZIPs 5/6/10 are expected to display dichotomy of disordered N-terminal sequences and globular PL domains, reminiscent of prion proteins. 3. Function Consistent with multiple lines of evidence suggesting that proteins harboring the prion fold can bind to each other, ZIP proteins co-purified with prion proteins in this study. While many proteins are known to bind divalent cations, PrP and ZIPs belong to a small group of proteins known to capture divalent cations at multiple binding sites embedded within disordered extracellular domains. Both the prion protein and ZIPs 5/6/10 have been shown to transport zinc ions across the plasma membrane. ZIP6 and PrP knockouts have been shown to display a rare common phenotype in zebrafish (inhibition of gastrulation / altered E-cadherin expression). 4. Localization ZIPs 5/6/10 and prion proteins share localization to the plasma membrane. ZIPs 5/6/10 display common orientation of shared sequence motifs with regard to the plasma membrane. Predominant tissues of expression of ZIPs 5/6/10 are reminiscent of PrP/Sho/Dpl gene expression profiles. 5. Phylogenetics Comparison of orthologous ZIP and prion sequences indicates divergent sequence evolution consistent with phylogenetic relationships. ZIPs 6/10 identified to bind to members of mammalian prion protein family populate a common phylogenetic branch and represent, along with ZIP5, the subset of mouse ZIPs (out of fourteen ZIP paralogs) which objectively display the strongest sequence similarity to prion gene sequences. A simple and plausible model for the emergence of the prion gene family in Chordata exists.

3.4.1 Protein-protein interactions amongst mammalian prion proteins and ZIPs

The evolutionary relationship between ZIP and prion proteins was discovered in this study on the basis of an interaction between ZIP6 or ZIP10 and PrP, Dpl, or Sho in N2a cells. The fact that these molecules are phylogenetically related and likely possess similar globular folds in their extracellular domains (except Sho which lacks the folded domain) suggests that there may be an inherent ability of proteins containing the prion protein fold to interact with each other. In support of this idea, the ability of PrPC to bind to PrPSc during prion replication implies that under certain circumstances PrP may be able to recognize itself. . It was even reported that Cu2+ and Zn2+ ions could modulate the self-association of PrP by binding to the octarepeat region and do so in a collaborative fashion despite their different binding affinities [213, 216]. Furthermore, it has been suggested that a direct interaction between PrPC and Dpl may explain the ability of PrPC to protect against Dpl-induced neurotoxicity [259]. Our observation that ZIP10 and ZIP6 co-purified not just with FLAG-PrP and FLAG-Dpl but also with FLAG-Sho, as evidenced by strong iTRAQ117 signals in the CID spectra which led to the identification of ZIP10 and ZIP6 (Figure 3.1), requires a different explanation because FLAG-Sho does not harbor a PL domain. One possibility is that the interaction between Sho and ZIP6 or ZIP10 is indirect (i.e. Sho may bind to a protein which interacts with ZIP6/ZIP10, but does not itself directly interact with the ZIP protein). A second possibility is that multiple binding domains govern the interaction between prion proteins and the ZIP family members. In this scenario, distinct domains present in the N-termini of Sho and PrPC and the C-termini of Dpl and PrPC may be involved in complex formation. Interestingly, the notion of two independent interaction sites between PrP and a hypothetical partner (dubbed “LPrP”) was deduced earlier from transgenic analyses of internally deleted forms of PrP [318].

3.4.2 N-terminal duplication versus PL domain insertion

While it is generally accepted that the Prnd gene, which encodes Dpl, originated from a duplication of the Prnp gene, a conclusive phylogenetic link between Prnp and Sprn (the gene that encodes Sho) has not yet been established. In fish, the genes encoding PrP-1 and Sho-2 are located next to each other, which is indicative of a possible phylogenetic relationship [28]. Furthermore, there is strong sequence similarity between the hydrophobic domains of the two

proteins, and Sho has been shown to possess PrPC-like protective abilities [143]. In light of the observation that the interactomes of Sho and PrPC in N2a cells are strikingly similar (Chapter 2), a phylogenetic relationship between Sprn and Prnp seems increasingly likely. One evolutionary model has suggested that Prnp is derived from Sprn by means of an ancient gene duplication event [28]. In this model, Prnp acquired the sequences that code for the α-helical C- terminal domain present in modern PrP sequences following its divergence from Sprn. Ortholog sequence alignments and structural threading data described in this report suggest that this sequence came from the ancestral protein that gave rise to modern-day ZIP5, 6 and 10.

Two alternative scenarios seem plausible for the origin of modern-day prion proteins: (i) a gene rearrangement may have caused the fusion of the PL domain of a ZIP ancestor gene with a Sho-like precursor molecule, thereby giving rise to a PrP founder gene (Figure 3.6). This model, as described above, requires that Sprn existed prior to Prnp. Alternatively, (ii) the complete Prnp gene, including promoter elements and sequences N-terminal to the PL domain, could have originated from a ZIP ancestor molecule. If the latter had occurred, similarities between ZIP proteins and PrP should not be restricted to the PL domain but include the respective N-terminal sequences. This model posits that both Sprn and Prnp are derived from a ZIP-like precursor gene. In support of this theory, the sequences N-terminal to both PrP and ZIP proteins contain similar signal peptides and numerous histidine residues that are involved in binding metal ions. In ZIPs 5, 6 and 10, the histidines are primarily found in clusters whereas in most PrP genes they are found within the context of the tandem octarepeat motifs and a second type of binding site involving histidine residues 95 and 110. A notable exception to this observed segregation represents the amino acid sequence of zebrafish PrP (also known as D. rerio PrP-rel-

2 or PrP3) which contains extended histidine-rich sequence clusters of the [HX]n variety seen in ZIPs (Supplemental Figure 3.1) [222, 223]. Although sequence conservation is low N-terminal to the cysteine-flanked core domain between both PrP and ZIP, it should be noted that this region is not well conserved even within orthologs of these protein families, thereby potentially masking evolutionary relationships. Intriguingly, alignment of a Sho-2 sequence from pufferfish showed considerable sequence conservation to ZIP5 from zebrafish (34% identity, 41% similarity) and close examination revealed that (i) a group of basic residues, (ii) the position of a largely hydrophobic segment and (iii) the abovementioned cluster of histidines are shared between these sequences (Figure 3.4B). This sequence alignment falls on the empirical threshold

which distinguishes instances of significant similarity from uncertain alignments [319]. Nonetheless, in the context of other similarities detailed above, it further supports the notion that the homology between ZIP and prion protein families extends beyond the globular domain present in PrP and Dpl to include sequences N-terminal to this domain, which are present in Sho (Figure 3.6). To firmly establish which model, if either, is correct, additional Prnp and/or Sprn ortholog sequences from early chordates will need to be uncovered and compared to ZIP sequences.

Figure 3.6: Models depicting evolutionary origin and topology of members of prion protein family, hypothetical ZIP ancestor and ZIP10 transporter. Cartoon depicting the emergence of members of the prion gene family from ZIP ancestor gene(s): (i) hypothetical Sho ancestor gene, (ii) prion gene founder and (iii) evolutionary intermediate prion gene family ancestor. Alternative hypotheses for the origin of prion genes: (1) insertion of ZIP ancestor-derived PL domain into Sho ancestor molecule; or (2) duplication of N-terminal ZIP ancestor fragment followed by expansion of hydrophobic domain and differentiation of former transmembrane domain into signal peptide for attachment of GPI anchor. Based on the above models, Sho genes either (1) evolved independently or (2) were themselves derived from a ZIP ancestor. Please note that our current analyses favor model (2).

3.4.3 Transmembrane-to-GPI-anchor attachment

Whereas PrPC is inserted into the membrane by the presence of a C-terminal glycosylphosphatidylinositol (GPI) anchor [320], ZIP5, ZIP6 and ZIP10 are type III transmembrane proteins characterized by a relatively large N-terminal extracellular domain, encompassing about half of the molecule’s primary structure, and a C-terminal domain harboring up to eight transmembrane spanning segments (Figure 3.3A, B). Precedents exist in other protein families in which some members are inserted into the membrane through transmembrane domains while others employ a GPI anchor for membrane attachment. This dichotomy of membrane attachment modes has, for example, been documented among members of the carcinoembryonic antigen (CEA) protein family [321] and cadherins, a diverse family of GPI- anchored (T-cadherin), single-spanning (classical cadherins and protocadherins) and multi- spanning (7TM-cadherins) transmembrane proteins [322]. In some instances surprisingly small changes in the protein sequence within or in proximity to the first transmembrane domain have been shown to cause a shift to GPI anchor-based modes of membrane attachment [200-202]. Thus, the GPI anchor signal sequence present in modern PrP may have derived from the first transmembrane domain of a ZIP ancestral molecule (Figure 3.6). In support of this idea, (i) a good alignment is observed between the PrP GPI anchor signal sequence and the ZIP10 TM1 domain for both fish and mammals (Figure 3.4A), and (ii) a consistent distance can be observed between the CFC domains and membrane attachment sites in the two protein families.

3.4.4 Analyses of genomic sequences

The availability of genomic sequences from the teleost lineage proved highly useful for this work. The presence of multiple copies of prion genes in this vertebrate branch may have helped to preserve original ZIP-like features in a subset of these genes. A phylogenetic linkage of teleost and mammalian prion genes has been established earlier [28] and therefore conclusions drawn from teleost sequence alignments are also relevant for establishing evolutionary relationships of mammalian prions. Interestingly, mouse PrP has recently been shown to partially rescue the phenotypic defects caused by down-regulation of zebrafish PrP-1, arguing that contemporary mammalian and fish PrP molecules have retained some degree of functional similarity throughout their divergent evolution [119].

Whereas the genomic organization of genes encoding for proteins of the prion family has been subjected to close scrutiny previously [323], relatively little is known about regulatory elements and the exon/intron structures of the slc39a gene family members. A subset of slc39a10 ortholog genes and a group of genes from the prion family have in common a first exon containing non-coding regulatory sequence elements and a downstream exon in which a very short segment of non-coding sequence is followed by the respective ATG start codons. The protein coding region of all members of the prion protein family appears to be confined to a single exon. ZIP proteins, in contrast, are encoded by multiple exons. In particular, exon/intron arrangements of ZIP gene sequences encoding for protein sequences upstream of the PL domain are most diverse and may be encoded by a single exon (e.g. mouse ZIP10) or multiple exons with non-conserved exon/intron boundaries across ortholog ZIP sequences. A more complete understanding of genomic rearrangements underlying the emergence of prion sequences from a ZIP ancestor gene should be based on a comparison of all aspects of the respective genomic segments in multiple organisms.

3.4.5 Structure and function of PL domain

The evolutionary history of the prion protein C-terminal domain has remained a mystery as no other protein for which high-resolution structures have been determined contains the prion protein fold. Our structural threading data suggest that the amino acid sequences within the PL domain region of ZIP5, 6 and 10 are highly compatible with the prion protein fold and the predicted structures are strikingly similar to those previously obtained for PrP and Dpl from various organisms. Since no high resolution structural data currently exists for members of the ZIP family, it will be of great interest to solve the structures of the extracellular domains of ZIP5, 6 and 10 and thereby confirm or refute the structural threading presented here.

The function of the extracellular PL domain within the ZIP family branch containing ZIPs 5, 6 and 10 has yet to be characterized. However, multiple more distantly related paralogs (ZIPs 4, 8, 12, 13 and 14) contain a homologous domain (Figures 3.3C and 3.5) and in the case of ZIP4, a protein that has been genetically linked to a rare recessive zinc deficiency disorder (acrodermatitis enteropathica), a role for N-terminal sequences in the sensing of metal ions has recently been suggested [194]. Interestingly, the abovementioned histidine-rich repeat PrP sequence found in zebrafish has recently been shown to exhibit preferential zinc binding

properties [223] (Supplemental Figure 3.1). Although the preferential binding of copper ions to mammalian PrP has been repeatedly demonstrated, earlier PrP molecules may have harbored different metal ion binding specificities, perhaps reflecting their phylogenetic relationship to zinc transporters. The ability to bind copper has also been described for Dpl, which possesses the prion fold but does not contain an N-terminal histidine-rich domain [169, 245]. Although a knockout of PrP in mice does not result in overt phenotypic deficits, a recent paper has shown that a knockdown of PrP-1 in zebrafish causes a rare gastrulation arrest defect during embryonic development that was linked by the authors to abnormal E-cadherin expression [119]. Interestingly, a strikingly similar defect was observed when ZIP6 (also known as LIV-1) was knocked-down during zebrafish development [118]. Thus, in zebrafish, ZIP6 and PrP-1 may be involved in similar developmental pathways or possess similar functions, possibly related to the regulation of cell adherence and E-cadherin protein levels [118, 119]. For ZIP6, a causal link of this phenotype to zinc transport has been uncovered that involves the zinc-dependent nuclear translocation of the zinc-finger protein Snail, a master regulator of epithelial-mesenchymal transition [118].

3.5 Conclusions

It remains to be determined whether the physical and evolutionary link to ZIP transporters will contribute to efforts towards the elucidation of the physiological functions of members of the prion protein family. It is hoped that a mechanistic understanding of the workings of ZIP transporters as well as high-resolution structures of their extracellular domains may provide insights into the origins of and constraints underlying the conformational changes associated with prion diseases. Additional work could also help to reveal whether membrane-inserted ZIP transporters or N-terminal fragments thereof shed into the extracellular space play a direct role in the manifestation or propagation of prion disease.

Chapter 4 Evidence for Retrogene Origins of the Prion Gene Family

Please note that all parts of this chapter, except Supplemental Figures 4.1 and 4.2, were published in the following article [324]: Sepehr Ehsani*, Renzhu Tao, Cosmin L. Pocanschi, Hezhen Ren, Paul M. Harrison, Gerold Schmitt-Ulms* (2011) Evidence for retrogene origins of the prion gene family. PLoS ONE 6(10):e26800. [* co-corresponding authors]

Candidate’s role: Contributed to bioinformatic analysis, manuscript assembly and editing

Summary: A computational investigation of the molecular mechanism of prion-ZIP evolution based on sequence, intron-exon, synteny and pseudogene analyses is presented. Our data suggest that during the emergence of metazoa, a cysteine-flanked core domain was modularly inserted, or arose de novo, in a preexisting ZIP ancestor gene to generate a prion-like ectodomain in a subbranch of ZIP genes. Approximately a half-billion years later, a genomic insertion of a spliced transcript coding for such a prion-like ZIP ectodomain may have created the prion founder gene. We document that similar genomic insertions involving ZIP transcripts, and probably relying on retropositional elements, have indeed occurred more than once throughout evolution.

4.1 Introduction

Extensive genomic investigations have provided evidence for additional PrP-related sequences in the vertebrate lineage [24-26]. In mammals, two paralogs of the prion gene, the genes encoding for the proteins Doppel/Dpl (prnd) and Shadoo/Sho (sprn) have been described [40]. Interestingly, the existence of prion genes and their paralogs appears to be restricted to vertebrates and therefore represents, on the evolutionary timescale, a relatively recent genomic development. Where did the prion founder gene originate from? The previous chapters demonstrated the evolutionary descent of the prion gene from the Zrt-, Irt-like protein (ZIP)

family of metal ion transporters (Chapter 3) [2] and documented that members of the mammalian prion protein family reside in spatial proximity to their ZIP molecular cousins in neuroblastoma cells (Chapter 2) [102]. More specifically, sequence alignments, structural threading data and multiple additional pieces of evidence placed a ZIP5/ZIP6/ZIP10-like ancestor gene at the root of the PrP gene family (Table 3.4). Amino acid sequence comparisons of the human ZIP proteins argue that ZIP6 and ZIP10, together with their phylogenetically closest paralog ZIP5, constitute a distinct subbranch in this family [10]. What we termed the prion-like (PL) domains of these ZIPs are predicted to form ectodomains that resemble PrPC with regard to orientation and relative distance to their downstream membrane anchorage sites [1]. Within these PL domains, one can readily identify a sequence segment that is characterized by stronger species-to-species sequence conservation than surrounding segments and is flanked by a pair of cysteine residues. These cysteines (which form a disulfide bridge in prion family protein structures) are universally conserved across all known prion or prion-like domains. Throughout this report we will refer to the sequence segment bounded by these cysteines as the cysteine- flanked core (CFC) domain.

ZIP genes date back much further than prion gene sequences. Indeed, related sequences can be found in all kingdoms of life, including bacteria and plants, and the ZIP gene family has undergone independent expansions within the distinct evolutionary lineages. Thus, whereas the genomes of humans and the plant species Arabidopsis thaliana code for similar numbers of distinct ZIP proteins (14 and 17 paralogs, respectively), the evolutionary subbranch of the ZIP family with members harboring a prion-like ectodomain underwent a profound expansion only during the early stages of Chordata emergence that was not mirrored in the plant lineage. This development preceded the emergence of the prion gene family and may serve as an explanation for its restricted existence in vertebrates. Today, based on sequence comparisons, four ZIP subbranches can be distinguished. The branch which contains ZIP transporters with a prion-like ectodomain can also be distinguished from other ZIP sequences on the basis of a putative intramembrane metalloprotease signature sequence and is frequently referred to as the LIV-1 subfamily of ZIP zinc transporters (LZTs).

The question arises as to precisely how the prion founder gene was created. Although a number of scenarios regarding the mode of evolution was presented in our original article [2], insights into the mechanistic aspects of the emergence of the prion founder gene based on an in-

88 depth analysis of relevant sequences were lacking. Here we undertook systematic bioinformatic analyses of select prion and ZIP genes to explore whether the mechanism of prion gene evolution can be deduced. We distinguish two genomic rearrangements: (i) the emergence of a first prion- like ectodomain harboring a cysteine-flanked core in a ZIP gene, and (ii) the formation of the prion founder gene. We document that as much as a half-billion years may have separated these two genomic rearrangement events. Surprisingly, our results point to a genomic insertion of processed and reverse-transcribed ZIP-ancestor mRNA as the most parsimonious explanation for the origin of the founder of the prion gene subfamily. We further document that similar insertions involving ZIP transcripts that probably relied on retropositional elements have occurred at other time points in vertebrate evolution.

4.2 Methods

4.2.1 Multiple sequence alignments

Sequence alignments were carried out using the AlignX feature of Vector NTI Advance 11.0 (Invitrogen, Carlsbad, CA, USA) [300]. A gap opening penalty of 10, gap extension penalty of 0.05 and gap separation penalty range of 8 were utilized in conjunction with the blosum62mt2 score matrix. Local adjustments were made in instances where visual inspection suggested an alternative alignment to the one returned by the algorithm. Sequences were selected for inclusion in this analysis with a view to (i) cover a broad spectrum of organisms ranging from pre- metazoan yeast and choanoflagellates to invertebrates to humans; (ii) depict all mammalian ZIP paralogs that contain a CFC domain; and (iii) represent a broad spectrum of prion sequences from fish to humans. Please see Figure 4.1 below for a simplified phylogenetic tree that identifies organisms selected for this and subsequent analyses.

Figure 4.1: Tree diagram depicting species utilized for genomic analyses in this study. In all instances, species included for a given analysis were to provide a broad and most informative sample and, at the same time, minimize redundancy. Because the questions which were addressed differed from analysis to analysis, the most relevant sample of gene sequences differed accordingly. MA, multiple alignment; EI, exon-intron; SA, synteny analysis.

4.2.2 Intron-exon genomic organization PrP and ZIP genes from a variety of organisms were selected based on their relevance to the ZIP- PrP evolutionary hypothesis [2]. Whenever multiple paralogs of a certain gene were available, the gene with the highest homology (based on protein sequence alignments) to other sequences in the figure was chosen. Intron-exon structures (sequences and information on the lengths of gene segments) and the start and stop codon positions were systematically extracted for each gene of interest from Ensembl (European Molecular Biology Laboratory-European Bioinformatics Institute, EMBL-EBI, and Wellcome Trust Sanger Institute, release 59) and Entrez (National Center for Biotechnology Information, NCBI, GenBank release 180.0) genomic

90 databases (Table 4.1). Transmembrane (TM) region boundaries and CFCs were identified as described previously [2]. The scales of the intron-exon figures were based on the length of the longest gene. The genes were arranged by aligning the 5’ end of their respective CFC domains.

Table 4.1. Protein accession numbers and abbreviations of species names

SEQUENCES Ensembl Entrez

Sp_ZIP NP_594942.1 Sc_YKE4 NP_012241.1 Mb_ZIP XP_001750768.1 Ta_ZIP XP_002108729.1*† Hm_ZIP12 XP_002160698.1 Aa_ZIP XP_001648141.1 Ag_ZIP XP_317935.4 Dm_ZIP (1) NP_001097608.1 Dm_ZIP (2) NP_523974.3 Bf_ZIP (1) XP_002594126.1 Bf_ZIP (2) XP_002608436.1 Ci_ZIP ENSCINP00000006884 Dr_PrP-rel3 ENSDARP00000017287 Dr_ZIP5 XP_690258.2 Dr_ZIP6 NP_001001591.1 Dr_ZIP8 ENSDARP00000056149 Dr_ZIP10 NP_956965.1 Dr_ZIP14 XP_001340102.2 Tn_PrP1 Q4SJ93 Tn_ZIP6 CAF89895.1 Tr_PrP1 AAN38988.1 Tr_PrP2 NP_001072109.1 Tr_PrP-like ENSTRUP00000016346 Tr_ZIP8 ENSTRUP00000045288 Tr_ZIP10 ENSTRUP00000007404 Tr_ZIP12 ENSTRUP00000019247 Tr_ZIP14 ENSTRUP00000038591 Ol_ZIP10 ENSORLP00000011004 Ga_PrP1 CAL64057.1 Ga_PrP2A ENSGACP00000007465 CAL64056.1 Xl_PrP NP_001082180.1

Xt_ZIP5 NP_001120272.1 Ts_PrP Q9I9C0 Gg_PrP ENSGALP00000040285 Gg_ZIP10 ENSGALP00000012604 Md_PrP ENSMODP00000014477 Md_ZIP6 ENSMODP00000000564 Mm_Dpl AAF02544.1 Mm_PrP ENSMUSP00000088833 Mm_Sho ENSMUSP00000053901 Mm_ZIP4 ENSMUSP00000073134 Mm_ZIP5 ENSMUSP00000037753 Mm_ZIP6 ENSMUSP00000064667 Mm_ZIP8 ENSMUSP00000029810 Mm_ZIP10 ENSMUSP00000027131 Mm_ZIP12 ENSMUSP00000080911 Mm_ZIP14 ENSMUSP00000066108 Hs_PrP ENSP00000368752 Hs_ZIP4 ENSP00000276833 Hs_ZIP5 ENSP00000266980 Hs_ZIP6 ENSP00000269187 Hs_ZIP8 ENSP00000378310 Hs_ZIP10 ENSP00000352655 Hs_ZIP12 ENSP00000366586 Hs_ZIP14 ENSP00000370635

Aa → Aedes aegypti (yellow fever mosquito) Ag → Anopheles gambiae (African malaria mosquito) Bf → Branchiostoma floridae (Florida lancelet) Ci → Ciona intestinalis (vase tunicate) Dm → Drosophila melanogaster (fruitfly) Dr → Danio rerio (zebrafish) Ga → Gasterosteus aculeatus (three-spined stickleback) Gg → Gallus gallus (chicken) Hm → Hydra magnipapillata (hydra) Hs → Homo sapiens (human) Mb → Monosiga brevicollis (marine choanoflagellate) Md → Monodelphis domestica (gray short-tailed opossum) Mm → Mus musculus (house mouse) Ol → Oryzias latipes (medaka) Sc → Saccharomyces cerevisiae (baker’s yeast) Sp → Schizosaccharomyces pombe (fission yeast) Ta → Trichoplax adhaerens Tn → Tetraodon nigroviridis (spotted green pufferfish) Tr → Takifugu rubripes (Japanese pufferfish) Ts → Trachemys scripta (red-eared slider turtle) Xl → Xenopus laevis (African clawed frog) Xt → Xenopus tropicalis (Western clawed frog)

*Exon 1 identified using an alignment of TRIADDRAFT_62928 and TRIADDRAFT_18721 genes.

†An additional CFC-containing ZIP sequence was identified in Trichoplax adhaerens (TRIADDRAFT_28345, CFC = CYSASQIFNIFKVSSTGADVKQFNSLSPALVQQVATSAC) based on an alignment with TRIADDRAFT_18721. However, due to incomplete exon annotation, the sequence was not used in the present analyses.

4.2.3 Synteny analysis For synteny analyses, the chromosomal locations, lengths and the directionality of the three neighboring genes upstream (5’) and downstream (3’) of prion or ZIP genes of interest were extracted from Ensembl and Entrez genomic databases. To facilitate side-by-side comparisons, genomic regions were depicted with 5’ boundaries of prion or ZIP genes aligned. In instances of uncertain identity (e.g., genes annotated with numerical identifiers), BLAST searches were conducted to establish possible homology relationships amongst genes.

4.2.4 Pseudogene discovery Protein sequences of human ZIP6 and ZIP10 were submitted to the PseudoGeneQuest online program (Institute of Medical Technology, Tampere, Finland, version 0.4) [325] to search for known human pseudogenes, pseudogene fragments and interrupted processed pseudogenes. The program used the human genome build 37.1 and known pseudogenes were retrieved from the Pseudogene.org database (version 71). The results were then individually BLAST-searched to determine if the hits indeed constituted ZIP pseudogenes or represented misannotated ZIP paralogs. To identify possible pseudogenes in other organisms, different domains (and combinations thereof) of LIV-1 ZIP sequences from different chordate species, which spanned more than one exon, were BLAST-searched against all genomes available in the NCBI database (GenBank release 180.0) and results showing contiguity in one or more exonic areas were flagged for further analysis. Repetitive elements were identified using the RepeatMasker online interface (Institute for Systems Biology, Seattle, WA, USA, version open-3.2.9) [326].

4.2.5 Accession numbers A list of accession numbers for sequences mentioned in this manuscript and the key to species name abbreviations appear in Table 4.1.

4.3 Results

4.3.1 The cysteine-flanked core within prion-like domains of metazoan ZIP proteins is set apart from surrounding sequences by a high level of positional sequence conservation and a pair of flanking introns Our previous analyses revealed the existence of a PL domain in a subset of genes belonging to the LIV-1 subfamily of ZIP zinc transporters (LZT) in diverse non-vertebrate organisms for which complete genomic data were available at the time, including D. melanogaster (fruitfly) and H. magnipapillata (jellyfish) [2]. The ongoing international genome sequencing activities have in recent times generated additional genome depositories for a range of organisms with more primitive body plans. Thus, to refine the evolutionary time point at which the first CFC domain may have emerged in a ZIP ancestor, we extended our search to the genomes of fungi, other relevant unicellular eukaryotes and early metazoa. These genomic queries made use of the PSI-BLAST algorithm and interrogated the respective genomic databases with sequence templates that forced perfect matching in highly conserved sequence positions (derived from a multiple alignment of prion-like domains we had identified in ZIP proteins earlier) but allowed variation in other positions of the sequence. This approach failed to detect ZIP gene sequences with a predicted prion-like domain in all genomes of unicellular organisms we investigated but revealed the existence of a ZIP sequence with a characteristic CFC domain in Trichoplax adhaerens (Ta) (Figure 4.2 and Table 4.1). A multiple alignment of prion-like domains of ZIP gene sequences from diverse organisms (please see Figure 4.1 for a summary of species used in this and subsequent analyses), including Ta, and prion genes revealed a dichotomy in the degree of sequence conservation within the globular PL domain itself, i.e., sequences N-terminal to the CFC domain are conspicuously enriched in charged residues but show, in contrast to sequences within the CFC, relatively little positional conservation (Figure 4.3). Interestingly, highly conserved intron-exon boundaries can be found immediately N-terminal and in close C-terminal proximity to the CFC not only in all human LZTs (LIV-1 ZIP zinc transporters), which contain this domain, but also in distant LZT sequences found in the genomes of species that range from Trichoplax to fruitfly to pufferfish. Consistent with their ancient origins, the lengths of these positionally-conserved introns are known or predicted (in instances where no transcripts are available) to vary widely from a few nucleotides to thousands of base pairs. This analysis further revealed that ZIP zinc transporter genes of unicellular organisms neither code for a CFC nor

feature introns in the respective segments of their genes. In fact, their protein sequences do not align N-terminal of the transmembrane domain and were merely included in this analysis to document these observations.

Figure 4.2: Broad phylogenetic distribution of LIV-1 ZIP metal ion transporters contrasts narrow distribution of prion genes in Chordata lineage. Numbers of LIV-1 ZIP and prion sequences in the selected organisms were extracted from gene data published by the Wellcome Trust Sanger Institute (TreeFam, http://www.treefam.org) and by multiple alignments of ZIP and prion protein sequences. For each organism, the number of the subset of sequences containing a cysteine-flanked core (CFC) domain is indicated in brackets.

Figure 4.3: The cysteine-flanked core within the prion-like domain of ZIP proteins is confined to metazoa. Multiple sequence alignment of the prion-like domain of select PrP and ZIP genes from metazoans to mammals. Baker’s yeast (Sc_YKE4), fission yeast (Sp_ZIP) and choanoflagellate (Mb_ZIP) sequences were included in this alignment in the interest of depicting a small number of representative LZT protein sequences outside of the metazoa realm. Their ectodomains appear, however, to lack a CFC domain based on (i) poor alignment, (ii) the

absence of a ‘CPALLY’ motif and (iii) the absence of conserved introns. Black squares (■) indicate the position of introns and asterisks (*) denote sequences for which complete intron/exon annotations were not available. Numbers in square brackets ([X]) indicate the length of a stretch of amino acids omitted for the purpose of clarity in a specific section of the alignment. Please note that Dm_ZIP (2) (marked with dagger symbol †) is the same protein sequence as that encoded by the Drosophila melanogaster fear-of-intimacy (foi) gene. Please see Table 4.1 for a complete list of scientific and common names of species referred to in this alignment with two-letter abbreviations.

4.3.2 ZIP genes of all evolutionary lineages are characterized by complex intron-exon structures not observed in prion gene sequences

The comparison of transcript structures of a set of related genes can sometimes shed light on the evolutionary history that links them to a common ancestor [327]. In particular, the number of exons and the relative position of intron-exon boundaries in relevant orthologous sequences can provide the basis for forming hypotheses regarding evolutionary relationships. To that end, we expanded upon the initial determination of introns flanking the CFC domain and investigated the intron-exon structure of the coding sequences of prion genes in vertebrates and of a representative subset of ZIP genes from diverse organisms (Figure 4.4). Species included in these analyses were selected with a view to (i) capture distant branches of the evolutionary tree, (ii) include PrP and ZIP gene sequences that are most similar (e.g. from pufferfish) or relatively distantly-related to each other (e.g. human sequences) according to our previous ZIP-prion evolutionary analyses [2], and (iii) extend the analysis of ZIP sequences to genomic lineages whose divergence predates the split of PrP and ZIP sequences and, thus, may be meaningful for deducing the gene structure of ZIP genes at the time when the prion gene emerged (Figure 4.1). Analyses relied on intron-exon genomic annotations provided by Ensembl and Entrez databases. Whenever annotations were ambiguous or conflicting, clarification was sought by comparing expressed sequence tag (EST) entries to the corresponding genomic sequences. In-depth analyses of prion genes in diverse organisms which preceded this work have repeatedly revealed a common gene structure composed of one or two short 5’ noncoding exons and a relatively long exon that codes for a short 5’ untranslated region (UTR), the entire open reading frame (ORF) and a 3’ untranslated region [323, 328-331]. Thus, the emphasis in this analysis was not on prion genes but on ZIP genes for which no detailed analyses had been undertaken.

Figure 4.4: Multiple introns observed in the coding regions of ZIP genes are missing from prion genes. Complex intron/exon arrangements of ZIP genes contrast the genomic organization of prion genes characterized by a coding sequence that is confined to one or, in rare instances, two exons. Only exons are depicted to scale. Black, hatched, grey or white fillings depict exons coding for the ectodomain, the cysteine-flanked core (CFC), the C-terminal multi-spanning transmembrane domain of ZIP transporters or non-coding segments of a given transcript, respectively. The black solid lines connecting exon boxes indicate introns. Untranslated regions (UTRs) are not depicted for a subset of sequences lacking reliable relevant annotation in the databases. For Tr_PrP1, part of the 5’ UTR, the ORF and the 3’ UTR are encoded in a single exon. Dm_FOI represents the ZIP ortholog in D. melanogaster with strong sequence similarity to mammalian ZIPs 5, 6 and 10. kbp, kilobase pairs.

Even at cursory inspection, the results revealed strikingly different intron-exon gene organizations of prion and ZIP genes. Consistent with data from the aforementioned studies, the ORFs of almost all vertebrate prion genes were confirmed to be contained in single exons. Exceptions to this genomic organization represent the prion gene homologs in stickleback and opossum, which apparently underwent genomic rearrangements that caused the coding sequences to be split into two exons. In contrast, the ORFs of all vertebrate ZIP genes we analyzed predict the splice-removal of multiple introns for generating the respective messenger RNAs. A closer look at ZIP genes revealed two categories of introns: (i) introns which display low positional conservation even amongst closely related members of the family, and (ii) introns that are highly conserved. For examples, two highly conserved introns which flank the CFC domain were observed in all LZT ZIPs that contain a prion-like ectodomain. In contrast to intron positions, intron lengths are known to change relatively rapidly in evolutionary time and are

therefore a poor indicator of sequence relationships. Consistent with this general observation, ZIP genes included in this analysis display a remarkable variation in the lengths of their corresponding introns, with human ZIP5 and ZIP12 serving as a pair of genes exhibiting multiple relatively short and long positionally-conserved introns, respectively (Figure 4.5). Significantly, the absence of introns flanking the CFC region in today’s prion sequences suggests that these introns disappeared shortly after or during the emergence of the prion gene founder from its ZIP ancestor.

Figure 4.5: Multiple introns observed in the coding regions of ZIP genes are missing from prion genes. Alternative presentation of data from intron/exon analysis shown in Figure 4.4 with both intron and exon lengths depicted to scale.

4.3.3 No shared genes in the genomic neighborhoods of ZIP and prion genes

Synteny analyses can be a powerful vehicle not only for establishing gene homology relationships, but also with respect to providing clues about the mechanistic origins of new genes. We therefore conducted an analysis of the genetic neighborhoods of select prion and ZIP genes. Specifically, the identity and relative genomic position of the three genes which map to genomic regions immediately adjacent to either side (5’ versus 3’) of the selected ZIP and PrP genes were recorded using Ensembl and Entrez genomic databases. In instances where the gene nomenclature did not readily reveal the identity of a gene, BLAST searches were conducted to establish possible relationships to other genes recorded in this manner. Consistent with previous

reports, organisms as distant to each other as pufferfish and humans exhibit synteny on the 3’ side of prion genes where prnd orthologs, rassf2 and slc23a2 genes are shared [24, 27]. Similarly, strong syntenic relations among ZIP gene orthologs are easily detected in all vertebrate sequences we scrutinized. For example, ZIP5 was flanked by rnf41 and ankrd52, and ZIP6 showed synteny with mocos, elp2, rprd1a, c18orf21 and galnt1 (Figure 4.6). The genes tmeff2, sdpr and stk17b were within the physical proximity of ZIP10 in human, chicken and fish genomes. More importantly, evidence for synteny could even be obtained for ZIP paralogs. Namely, the homologous variants of the obfc2b/a gene were detected in close proximity to human ZIP5 and ZIP10 genes, probably indicating an evolutionarily conserved linkage to the region that once hosted an ancestor of the subbranch of ZIP genes to which ZIP5 and ZIP10 belong. Notably though, no gene homologous to obfc2b/a was detected in proximity to ZIP6 genes, the third paralog in this ZIP subbranch, or the prion protein gene. And whereas two genes belonging to the ankyrin gene superfamily were located in spatial proximity to both zebrafish PrP (ankrd) and ZIP5 sequences from various organisms (ankrd52), a closer comparison of relevant sequences failed to reveal orthologous relationships for these genes and instead suggested ankrd and ankrd52 to be distant members of a large and diverse gene family.

100

Figure 4.6: Lack of shared genes in proximity of PrP and ZIP genes. The synteny analysis was restricted to three adjacent genes on either side of the relevant PrP and ZIP genes. Synteny was determined to be restricted to ortholog sequences, with paralogs of the gene obfc2 observed adjacent to both ZIP5 and ZIP10 genes serving as notable exceptions. Black boxes depict the genes of interests (PrP/ZIP), and colored boxes represent proximal genes. Please note that Tr_ZIP10 maps to the 5’ boundary of a genomic contig for which the adjacent genomic segment is not annotated. Black solid lines indicate interspersed non-coding regions. Mb, megabase pairs.

101

Taken together, no evidence for shared genomic context in proximity to PrP and ZIP genes emerged from this analysis, corroborating the impression that the homology of prion and ZIP genes may not extend beyond their respective coding regions.

4.3.4 Pseudogene analyses uncover instances of genomic insertions of spliced and reverse-transcribed ZIP transcripts in vertebrates

The data presented thus far suggested the intriguing possibility that a spliced and reverse- transcribed ZIP transcript may have served as an intermediate during the generation of the prion founder gene. We therefore wondered whether instances of retroposition of all or parts of a ZIP transcript harboring a prion-like domain can be traced in current genomes. We initially restricted our search to the human genome. A query of the PseudoGeneQuest online tool [325] with the human ZIP6 sequence returned multiple hits of which 10 were designated by the program as pseudogenes, 1 as a pseudogene fragment and 2 as interrupted processed pseudogenes. Similarly, 9 sequences which were flagged as pseudogenes, 2 as pseudogene fragments and 2 as interrupted processed pseudogenes were returned by the algorithm when queried with a human ZIP10 sequence (interrupted processed pseudogenes are search results possessing repeat content which is >50% of the length of the target) [332]. A subsequent closer analysis of these hits based on BLAST searches identified most of them as ZIP paralogs (of which there are 14 in the human genome) which had been misinterpreted by the algorithm to represent candidate pseudogenes. However, one interrupted processed pseudogene returned for both the ZIP6 and ZIP10 queries was confirmed by us to represent a bona fide ZIP pseudogene located on human chromosome 1 (residues 48,061 to 49,786, clone RP11-365D9, locus AL583844.11). In fact, the genomic Entrez/NCBI annotation of this clone already identified this sequence as a ZIP14 pseudogene. Alignment of this region with the human ZIP14 parent gene located on the short arm of chromosome 8 further refined the boundaries of the retroposed segment and revealed that the entire coding sequence of the ZIP14 parent gene is retained in this pseudogene sequence, but considerable sequence decay has accumulated since its formation (Figure 4.7A). A recent update to the human Ensembl/Vertebrate Genome Annotation (VEGA) genome database indicated the existence of a second ZIP pseudogene in humans. This pseudogene is located on chromosome 22 and derived from the ZIP1 parent gene (Figure 4.7B). However, given that ZIP1 does not contain a PL domain in its sequence, this pseudogene is of lesser relevance in the context investigated here.

102

Figure 4.7: Evidence for the existence of ZIP pseudogenes in the human genome. A. Human chromosome 1 contains a retrocopy of the human ZIP14 gene coded within the long arm of chromosome 8. The retrocopy is embedded within a relatively long intron of the guanine nucleotide binding protein gamma 4 (GNG4) gene. It exhibits telltale signs of sequence decay associated with pseudogenes such as an accumulation of multiple translation stop codons and the presence of more than a dozen predicted frameshifts relative to the predicted mRNA sequence of its parent ZIP14 gene. B. A relatively short ZIP1 pseudogene corresponding to a C-terminal segment of its ZIP1 parent gene coded within chromosome 1 can be identified on human chromosome 22. The pseudogene sequence features a translation stop codon and two predicted frameshifts.

We hypothesized that the development of the prion founder gene might have been accompanied by a loss of most of the C-terminal domain of its ancestral ZIP parent gene. Assuming that retroposition might have been the mechanism, we wondered whether such an event was a unique occurrence or whether it would be possible to find evidence that a similar ZIP retroposition paralleled by the loss of C-terminal transmembrane domains also occurred at a

103

different time. To address this question, we next searched the genomes of available chordate organisms for LIV-1 ZIP-like sequences that (i) were contiguous at conserved exon boundaries within segments of ZIP genes that code for their prion-like domain, and (ii) did not align to stretches of conserved C-terminal ZIP sequences. The objective was not to generate an exhaustive list of candidate sequences but to determine whether at least one such sequence could be found. Indeed, on chromosome 7 of the gray short-tailed opossum (Monodelphis domestica) a sequence was found which matched the filtering criteria (Figure 4.8). The respective pseudogene was flanked by a number of repetitive elements (Figure 4.8A) and aligned to exons 2, 3, 4 and parts of exons 1 and 5 of the opossum ZIP6 gene which maps to chromosome 3 (Figure 4.8B). This pseudogene features four stop codons but otherwise has experienced a low degree of sequence decay. Taken together, the pseudogene analyses uncovered specific examples of independent insertion events of spliced and reverse-transcribed ZIP transcripts in present-day vertebrate genomes.

104

Figure 4.8: Evidence for a C-terminally truncated ZIP6 pseudogene in the opossum genome. A. An N-terminal ZIP6-like pseudogene was identified on chromosome 7 of Monodelphis domestica (cont3.050765, GenBank: AAFR03050766.1), which consisted of exons 2, 3, 4 and parts of exons 1 and 5. Exon 3 which codes for the CFC is depicted in light green

105

color. Long terminal repeats (LTRs) and short interspersed nuclear elements (SINEs) in the vicinity of the pseudogene are marked. Please note that short direct repeats immediately flanking the pseudogene were not detected. Similarly, no evidence of a poly-A tail could be observed, consistent with the retroinsertion of a C-terminally truncated ZIP6 transcription product. B. Sequence alignment of opossum ZIP6 with the pseudogene and its flanking sequences, clearly demarcating the boundaries of retroinsertion. Identical base pairs in the two sequences are highlighted in yellow, and sequence features such as the CFC domain are marked at the amino acid level. kbp, kilobase pairs.

4.4 Discussion

The recently-proposed ZIP-prion evolutionary link [2] raised the possibility that a close examination of relevant genomic sequences may reveal insights into genomic rearrangements which precipitated the emergence of prion genes in the vertebrate lineage. In the following paragraphs we will present an argument based on the data from this report which proposes that the emergence of the prion founder gene depended on two genomic rearrangements which occurred hundreds of millions of years apart.

We will discuss that the first of these two events on the path to the prion founder gene may have involved the insertion of a CFC domain into a preexisting ZIP ancestor. This event likely occurred around the time when multicellular mobile metazoa emerged on the planet, possibly more than a billion years ago [333]. The second event, i.e., the actual formation of the prion founder gene, can be traced back to a time before the divergence of teleosts and tetrapods, approximately a half-billion years ago (Figure 4.9). The proverbial ‘smoking gun’ which would simplify the reconstruction of this genomic rearrangement (e.g., flanking short direct repeats and/or the presence of remnants of a poly-A tail) may no longer exist in the genomes of contemporary vertebrate species. Nonetheless, the cumulative data we presented relate a consistent story and suggest that the formation of the prion founder gene may have involved the genomic insertion of a reverse-transcribed ZIP transcript.

106

Figure 4.9: Two-step model of emergence of prion gene from a ZIP ancestor. At the time in evolution when early metazoa emerged, a CFC domain was inserted into an ancient ZIP transporter or evolved de novo. During early vertebrate speciation, a descendant of this ZIP ancestor, with ectodomain features resembling present-day ZIPs 5, 6 and 10, gave rise to a processed transcript which was reverse-transcribed and inserted into a genomic region that shares no synteny relationship with the parent gene. Through acquisition of a nearby 5’ promoter element, this retrocopy may have evolved into a fully functional retrogene – the first prion gene. An additional expansion of the subfamilies of LZT and prion genes occurred through gene duplication events. Genomic elements in this figure are not drawn to scale. The depiction of intron positions for the ZIP gene are based on the intron-exon structure of the Trichoplax adhaerens LIV-1 ZIP gene harboring a CFC domain described in this manuscript.

107

4.4.1 Emergence of prion-like ZIP ectodomain in early metazoa

Data presented in this manuscript established that ZIP sequences containing CFC domains can be identified in the genomes of metazoa with relatively primitive body plans, including the amoeba- like organism Trichoplax adhaerens (Ta) and cnidarians, but these domains seem to be absent in ZIP genes of all other branches of life. Thus, around the time when the metazoa lineage emerged, the CFC domain may have either gradually evolved or become inserted as a module into a preexisting ZIP gene (Figure 4.2). Multiple alignments of prion and ZIP sequences from a diverse selection of organisms undertaken for this work revealed a dichotomy in the degree of sequence conservation within the globular PL domain itself, i.e., sequences flanking the CFC domain are conspicuously enriched in charged residues but, in contrast to sequences within the CFC, show relatively little positional conservation (Figure 4.3). A number of alternative (and not necessarily mutually exclusive) explanations come to mind that may have limited the ability of the CFC to diversify: (i) the currently unknown function or molecular interactions of the CFC might have placed limitations on sequence variation; (ii) the predicted existence of a disulfide bridge formed between cysteine residues (so far only proven to exist in the CFC of tetrapod PrPC and Dpl) at its boundary may constitute a structural constraint that restricted sequence evolution; and/or (iii) a different rate of evolution might be the consequence of a genomic organization that causes this genome segment to evolve at a different pace than the surrounding sequences. Indeed, conserved exon/intron boundaries can be found immediately N-terminal and in close proximity to the C-terminal boundary of the CFC domain in the genomes of species ranging from Trichoplax to humans. It is therefore likely that the emergence of this domain was based on exon shuffling or an exonization of a preexisting intron, a process which can, for example, be triggered by intronic insertion of a retroelement providing novel splice acceptor motifs [334, 335]. Consistent with their ancient origins, the lengths of the positionally-conserved introns flanking the CFC vary widely from a few nucleotides to thousands of base pairs in LZT genes. The alternative model based on which the CFC domain was generated through gradual sequence evolution is less appealing because it fails to explain the concomitant emergence of the two highly-conserved flanking introns.

4.4.2 Generation of prion founder gene in vertebrates

The absence of introns flanking the CFC region in today’s prion sequences (Figure 4.4) suggests that these introns disappeared shortly after or during the emergence of the prion gene founder

108

from its ZIP ancestor. While these introns may indeed have disappeared after the actual gene duplication event and independent of it, this explanation neither represents the most parsimonious model nor does it suggest a satisfying answer for why intron loss in these positions occurred in the prion gene founder but is not observed in CFC-containing ZIP gene sequences. The literature surrounding intron loss and gain seems contradictory: on the one hand, forces of genome miniaturization have given rise to massive-scale intron loss in individual species [336], and on the other hand, intron loss has been described as a rare event relative to other types of genomic rearrangements [337, 338], a quality exploited in studies that use comparative intron mapping to determine deep evolutionary histories of gene families [339, 340]. A large-scale comparison of mouse and human genomes revealed, for example, that introns in these two species are only changed in 0.08% of positions, indicating a more than 1,000-fold higher level of conservation when compared with protein sequence changes [341]. Whenever intron loss is observed for two adjacent positionally-conserved introns, it appears to be the result of a reverse transcription of RNA intermediates [337]. Mechanistically, RNA intermediates play a role in two types of intron loss events: gene conversions by recombination with spliced transcripts from the affected gene, and retroposon-mediated gene transfers [338, 342]. Whereas the former mechanism converts the gene in its original genomic environment, the latter causes a transposition of a spliced copy of the original gene into a distant genomic acceptor site, generally assumed to represent transcriptionally active and open chromatin [343, 344]. Consequently, an important criterion for the designation of retrocopies is the loss of at least two positionally- conserved introns in regions that can be aligned to homologous parent genes [345, 346]. In the context discussed here, the application of this criterion suggests that the emergence of the prion gene founder may have been the result of an ancient germline retroposition event. The truncation of the prion gene founder could then have been the consequence of a frequently-observed shortening of mRNA sequences before or during reverse transcription [347], or could have occurred following the genomic insertion but prior to the divergence of PrP sequences as a result of speciation.

Following their genomic insertion, the majority of retrocopies turn into pseudogenes by falling transcriptionally silent and being subjected to relatively rapid genomic mutations, insertions and deletions that lead to sequence decay and can eventually cause the elimination of pseudogene sequences [348]. These rapid evolutionary changes to pseudogene sequences occur

109 because regulatory elements that could drive their expression and, consequently, help to realize a stabilizing selective advantage, are missing. Thus, for sustained survival of a retrocopy, it is critical that the genomic insertion event occurs in the vicinity of a preexisting promoter or proto- promoter that can be hijacked for transcriptional activity [342, 346]. When genomic insertions do not occur immediately proximal to a preexisting promoter but in some 3’ distance to it, retrocopies can adapt 5’ sequences and even untranslated exon/intron structures for gene regulatory purposes on their way to become transcriptionally-active retrogenes [349]. Thus, the existence of noncoding exons and introns in the 5’ untranslated region of today’s prion genes is not inconsistent with this model but represents a frequent occurrence in retrogenes [350]. Alternatively, retrogenes have been shown to acquire exons de novo during evolution. For example, a study of >1,000 retrocopies in the human genome revealed a surprisingly large percentage of retrogenes (27 out of a total of 120 retrocopies which had developed into bona fide genes) that had acquired untranslated exons in this manner [346]. In the absence of strong sequence conservation, the effects of genomic rearrangements and divergent sequence evolution which accumulate in a given pair of retro- and parent genes over time may mask the ability to recognize the origins of the former. For example, the gene encoding HNRPF, a protein involved in RNA processing, was not recognized as a retrogene until recently, possibly because it recruited three 5’ untranslated exons [346]. The most conspicuous lingering characteristic indicating retrogene origins might be the absence of introns within ORFs, a school of thought that provoked the proposition that many of the approximately 15% of genes in the human genome lacking introns in their ORF may have arisen by retroposition [351].

Because retroposition is accompanied by the loss of the surrounding genomic sequences, the absence of homology of promoter sequences and synteny relationships of a retrocopy and its parent gene constitute additional criteria routinely used for the distinction of retrocopies from segmentally duplicated genes. However, because of the extensive evolutionary time which has passed since the emergence of the prion founder gene and given the relatively rapid diversification of non-coding sequences, a comparison of promoter sequences seemed futile as no sequence conservation would be expected at this time, regardless of the mechanism of evolution. A promoter comparison we undertook for other purposes confirmed this prediction (Supplemental Figures 4.1 and 4.2).

110

In the case of prion gene family members, previous synteny analyses not only provided a framework for comparing the evolutionary links amongst prion-related genes, but also led to an intriguing model which posits that all prion genes known to date have emerged from a common prion gene founder [28]. The examinations of genetic neighborhoods undertaken in this work revealed robust synteny within ortholog comparisons of different ZIPs or prion genes (Figure 4.6), but failed to detect synteny across paralog boundaries with one notable exception: the gene obfc2b/a was found to be shared in proximity of both ZIP5 and ZIP10 genes of the ZIP LIV-1 subfamily. This is relevant as comparative genomic analyses we have undertaken suggest that the subbranch of the ZIP gene family populated by ZIPs 5, 6 and 10 may have undergone an expansion around the time when the prion founder gene emerged [2]. This is evident based on the existence of ZIPs 5, 6 and 10 genes in some teleost and tetrapod genomes but not in the early chordates. Thus, the synteny across ZIP5-ZIP10 paralog boundaries served as a positive control in this analysis. It documented that despite the approximately half-billion years which have passed since the divergence of the paralogous pair, this event can still be identified to have been mechanistically based on a duplication of a genomic segment containing a predecessor ZIP gene and its adjacent genes. Thus, it is conceivable that synteny between prion and ZIP genes could be observed in contemporary genomes if the emergence of the prion founder gene had relied on a similar genomic duplication event.

An important aspect in prion pathobiology which also relates to the emergence of the prion founder gene is the evolutionary time point at which the protein became capable of infection and aggregation. However, given that very little is known about prion disease outside of the mammalian clade, the characterization of the evolution of prion infectivity requires further research.

4.4.3 Other retropositional events

Just as inductive reasoning draws strength from specific repeated observations, the notion of retrocopy origins of the prion founder gene would be easier to embrace if other instances of retrocopy events involving ZIP genes could be traced in existing genomes. A non-exhaustive search in genomic databases we report in this manuscript revealed that independent retroinsertions of ZIP transcripts containing prion-like ectodomains have indeed occurred in the opossum and human genomes (Figures 4.7 and 4.8). Incidentally, the opossum ZIP pseudogene

111 was derived from a ZIP6 parent gene, a member of the very subbranch of ZIP family genes which we proposed to have given rise to the prion founder gene [2]. Remarkably, a comparison of gene boundaries of the pair of opossum ZIP6 retrocopy and parent gene revealed that the retrocopy lacks most of the C-terminal sequences coding for the multi-spanning transmembrane domain of its parent gene. Thus, the opossum with its ZIP6 pseudogene can be viewed as demonstrating a re-enactment of the ancient genomic rearrangement which may have caused the loss of C-terminal domains of the prion founder gene.

Is there a precedent of an unrelated gene family in which a phylogenetic subbranch originated from an ancient retroinsertion event? The gene family of glial cell-derived neurotrophic factor (GDNF) family ligand receptors (GFRα) may serve as an example: within the GFRA gene family, a majority of sequences share a common exon-intron structure. However, growth arrest-specific 1 (GAS1) genes – members of a GFRA gene family subbranch expressed in species as diverse as roundworms, honey bees and humans [352, 353] – lack all introns and have been proposed to have originated from an ancient retroinsertion event [354, 355] (Figure 4.10). Intriguingly, the parallels do not end there, as GAS1 (like the prion protein) is a glycosylphosphatidylinositol (GPI)-anchored member in a protein family that contains both transmembrane (e.g., GDNF family receptor alpha-like, GFRAL) and GPI-anchored proteins (GFRA 1-4) [353]. Thus, in both the GFRα family and the ZIP superfamily, a retrotransposition event may have given rise to a subbranch of C-terminally truncated genes. These observations are consistent with previous reports by other groups which established that a mere C-terminal truncation of genes coding for transmembrane proteins at a site adjacent to transmembrane- coding sequences can be sufficient to generate a signal sequence for the attachment of a GPI anchor [2, 200].

112

Figure 4.10: Precedent of retroposition event leading to a subbranch of GPI-anchored proteins within family of transmembrane proteins. Schematic representation of proposed mode of evolution of glial cell-derived neurotrophic factor (GDNF) family ligand receptor (GFRα) members. Horizontal and vertical cartoons depict modular gene and protein organization of GFRα members, respectively. Please note both the absence of introns and the emergence of the GPI membrane attachment mode in the Gas1 subbranch of GFRα proteins following a retroposition event which occurred early during metazoan speciation.

4.5 Conclusions

Taken together, our bioinformatic analyses of prion and ZIP genes and their genetic environments suggest that retroposition was the likely mode of emergence of prions from a LIV- 1 ZIP ancestor molecule. It is anticipated that this model can be further refined once additional genome sequences of species with relevance for elucidating pre-vertebrate evolution become available. Potentially more rewarding, however, might be to uncover (i) where the CFC domain within metazoan ZIP transporters originated from, and (ii) whether any molecular cousins of the prion protein exist which descended from the independent retroposition of ectodomain-coding sequences of ZIP transporters lacking a CFC domain.

113

Chapter 5 LIV-1 ZIP Ectodomain Shedding in Prion-Infected Mice Resembles Cellular Response to Transition Metal Starvation

Please note that all parts of this chapter, excluding minor modifications, were published in the following article [356]: Sepehr Ehsani*, Ashkan Salehzadeh*, Hairu Huo*, William Reginold, Cosmin L. Pocanschi, Hezhen Ren, Hansen Wang, Kelvin So, Christine Sato, Mohadeseh Mehrabian, Robert Strome, William S. Trimble, Lili-Naz Hazrati, Ekaterina Rogaeva, David Westaway, George A. Carlson, Gerold Schmitt-Ulms (2012) LIV-1 ZIP ectodomain shedding in prion-infected mice resembles cellular response to transition metal starvation. J Mol Biol 422(4):556‐74. [* equal contribution]

Candidate’s role: Contributed to biochemical characterization of ZIP10 in normal and prion- infected cells, manuscript assembly and editing

Summary: Here, we begin to address whether the study of LZTs can shed light on the biology of prion proteins in health and disease. Starting from an observation of an abnormal LZT immunoreactive band in prion-infected mice, subsequent cell biological analyses uncovered a surprisingly coordinated biology of ZIP10 and prion proteins that involves alterations to N- glycosylation and endoproteolysis in response to manipulations to the extracellular divalent cation milieu. Starving cells of manganese or zinc, but not copper, causes shedding of the N1 fragment of PrPC and of the ectodomain of ZIP10. For ZIP10, this posttranslational biology is influenced by an interaction between its PrP-like ectodomain and a conserved metal coordination site within its C-terminal multi-spanning transmembrane domain. The transition metal starvation-induced cleavage of ZIP10 can be differentiated by an immature N-glycosylation signature from a constitutive cleavage targeting the same site. Data from this work provide a first glimpse into a hitherto neglected molecular biology that ties PrP to its LZT cousins and suggest that manganese or zinc starvation may contribute to the etiology of prion disease in mice.

114

5.1 Introduction

In contrast to the vast literature surrounding the biology of PrP, little is known about the LZTs. This knowledge gap pertains to many aspects of ZIP biology, including their expression, posttranslational modifications, molecular structure, regulation and response to divalent cations. Assuming that some overlaps in the biology of LZT paralogs exist, a close examination of ZIP4 may be warranted in efforts to predict the biology of ZIPs 5, 6 and 10. ZIP4 has been genetically linked to an autosomal recessive zinc deficiency disease in children, termed acrodermatitis enteropathica (AE), and as a consequence has been subjected to closer scrutiny than other LZTs. Biochemical studies of ZIP4 suggest that its expression inversely correlates with extracellular zinc levels [136, 170]. Furthermore, it has been shown that a large part of what we termed the PrP-like ectodomain of ZIP4 can be shed by an endoproteolytic cleavage event in healthy individuals, but possibly not in a subset of individuals suffering from AE [194]. Interestingly, this shedding seems to occur in response to zinc depletion, suggesting that it may constitute a mechanism by which the cell can counteract undesired zinc starvation stresses. It should be noted, however, that the altered uptake kinetics due to ectodomain shedding are not clearly understood and active investigations are ongoing [357].

The objective of this work was to complement our previous interactome analyses, which identified LZTs as candidate PrPC interactors, and bioinformatic investigations of the relationship between prion and LZT genes, with experiments that explore a possible involvement of LZTs in the normal and disease biology of the prion protein. We document that in mice challenged with mouse-adapted Rocky Mountain Laboratory (RML) prions, the generation of a novel ZIP10 immunoreactive band can be observed from as early as 84 days after intracerebral (i.c.) inoculation. Subsequent experiments aimed at characterizing this ZIP10-derived signal uncovered a surprising degree of crosstalk in the biology of ZIP10 and prion proteins and shed light on a complex biology surrounding N-glycosylation and the generation of endoproteolytic cleavages within the PrP-like domains of ZIP10 and PrP that is tied to cellular divalent cation homeostasis. We discuss the relevance of our data to understanding the functional role of an alpha-site cleavage observed in PrPC and propose that prion disease in mice might phenocopy a transition metal starvation phenotype.

115

5.2 Materials and Methods

5.2.1 Cell viability assay

Mouse neuroblastoma Neuro-2a (N2a) cells were obtained from the American Type Culture Collection (ATCC; cat. no. CCL-131, Manassas, VA, USA). N2a cell viability was measured by MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) assay as reported previously [358]. Briefly, 5 mg/ml of MTT (Sigma-Aldrich, Oakville, ON, Canada) in phosphate buffered saline (PBS) was added to the culture medium of N2a cells grown on 96-well plates at a final concentration of 0.25 mg/ml. After incubation for 4 h, cells were solubilized in 70% isopropanol acidified to 0.1 N HCl. The absorbances were measured at a wavelength of 550 nm. The extent of MTT conversion in N2a cells exposed to the chelating reagent was expressed as a percentage of the control.

5.2.2 Confocal immunofluorescence analysis

N2a cells were grown on poly-d-lysine coated coverslips 24 h prior to and following transfection with Lipofectamine 2000 reagent (cat. no. 11668, Invitrogen, Carlsbad, CA, USA). Cells were then fixed with 4% formaldehyde in PBS for 10 min at 37°C and the reaction was quenched with 25 mM glycine and 25 mM ammonium chloride in PBS for 15 min at 37°C. Following permeabilization with 0.2% Triton X-100 in PBS and blocking with 0.1% saponin, 2% BSA, and 1% horse serum in PBS, cells were incubated for 1 h with primary antibodies. After five 3-min PBS washes, fluorescently-conjugated secondary antibodies were added and incubated for 1 h. Images were acquired using an inverted fluorescence microscope (DMIRE2; Leica Microsystems, Concord, ON, Canada) equipped with a back-thinned electron multiplier charge- coupled device (EM CCD) camera (Hamamatsu Photonics, Bridgewater, NJ, USA) and spinning disk confocal scan head. The unit was outfitted with four separate diode-pumped solid-state laser lines (405 nm, 491 nm, 561 nm and 638 nm; Spectral Applied Research, Richmond Hill, ON, Canada), an Improvision Piezo Focus Drive (PerkinElmer, Woodbridge, ON, Canada), a motorized xy stage (Applied Scientific Instrumentation, Eugene, OR, USA), and a 1.5× magnification lens (Spectral Applied Research). The equipment was controlled by Volocity acquisition software (PerkinElmer). Split Mander’s co-localization coefficients for comparing spatial signal overlaps of ZIPs and PrP were calculated in Volocity from at least 20 flattened Z-

116 stacks (encompassing 20-30 layers) of individual cells. The analyses were conducted in three technical replicates.

5.2.3 Immunohistochemical staining

5-micron thick sections of formalin-fixed paraffin-embedded tissues of wild-type mouse brains were dewaxed in five changes of xylenes and then brought down to water through graded alcohols. Endogenous peroxidase activities were blocked for 15 min using 3% aqueous solution of hydrogen peroxide before tissues were heat-retrieved with 10 mM citrate buffer at pH 6.0. Heat-induced epitope-retrieval was done using a Decloaking Chamber (Biocare Medical, Concord, CA, USA) at 120˚C for 2 min followed by cooling off to 90˚C before slides were removed from the Decloaking Chamber [359]. Tissues were then allowed to stay in the hot buffer for another 20 min at room temperature before washing well in running tap water. Sections were blocked for 30 min with 2.5% normal horse serum before incubating overnight with in-house rabbit polyclonal antibodies to ZIP10 at 1/100 dilution using an antibody diluting buffer containing Tris buffer and casein. Staining was finished using the ImmPRESS anti-rabbit Ig peroxidase kit (cat. no. MP-7401, Vector Laboratories, Burlingame, CA, USA). To validate the specificity of the stain, a negative control sample was generated by pre-saturating the polyclonal antibodies with the peptide antigen during overnight incubation. All incubations were done at room temperature inside a moist chamber and sections were washed thoroughly in three changes of Tris buffered saline before color development with a freshly prepared solution of NovaRed (cat. no. SK4800, Vector Laboratories). Finally, sections were lightly counterstained with Mayer’s hematoxylin (cat. no. MHS16-500ML, Sigma-Aldrich), washed well with water, dehydrated through graded alcohols, cleared in xylenes and then mounted in Permount (cat. no. SP15-500, Fisher Scientific, Nepean, ON, Canada).

5.2.4 Divalent metal ion measurements

An Optima 7300 spectrometer (PerkinElmer) was used for inductively coupled plasma atomic emission spectroscopy (ICP-AES)-based measurements of divalent metal ion concentrations in different cell media (ANALEST facility, University of Toronto). The device has a typical detection in the ppb (μg/L) to ppm (mg/L) range. Specifically, it has a quantifiable minimum detection of 10 ppb for zinc and 4 ppb for copper. Three reading wavelengths were selected for each element, and the spectrometer was adjusted to read for 5 to 10 seconds. A minimum of

117

three replicate analyses were carried out to determine transition metal concentrations in selected solutions. Calibration was performed using PlasmaCAL Multi-Element Standard QC-4 (SCP SCIENCE, Baie-d’Urfé, QC, Canada).

5.2.5 Glycosylation analysis

Cell lysates (20 μg of protein) were denatured in 0.5% SDS and 1% β-mercaptoethanol at 100°C for 10 min. NP-40 and sodium phosphate were added to final concentrations of 1% and 50 mM, respectively. 5 μL (2,500 units) of PNGase F (New England Biolabs, Ipswich, MA, USA) was incubated with the mixture at 37°C for between 3 h to overnight.

5.2.6 Metal chelation

The cell-permeable chelator N,N,N’,N’-tetrakis(2-pyridylmethyl)ethylenediamine (TPEN; cat. no. P4413, Sigma-Aldrich) was first dissolved in ethanol at 23.4 mM and subsequently diluted 100× in cell culture medium to produce a 234 μM stock. The cell-impermeable chelator diethylene triamine pentaacetic acid (DTPA; cat. no. D6518, Sigma-Aldrich) was dissolved in cell culture medium for a 1.2 mM stock solution. Both solutions were filter-sterilized before addition to cells, and incubated for 48 h, unless indicated otherwise.

5.2.7 Molecular Cloning

ZIP10 (Slc39a10 ORF in pCMV6-Entry) was obtained from OriGene (Rockville, MD, USA). A C-terminal hemagglutinin (HA) tag (YPYDVPDYA) was introduced in the plasmid following a glycine spacer residue. ZIP10-HA_d300 (deleted from residue 31 to 300), ZIP10-HA_d389 (deleted from residue 36 to 389), ZIP10-HA_N341A and ZIP10-HA-(H711A/H715A) were derived from the ZIP10-HA plasmid. The HotStar HiFidelity polymerase (Qiagen, Valencia, CA, USA) was used for site-directed mutagenesis. All plasmids were sequence-verified.

5.2.8 Polyclonal antibody generation

A KLH-conjugated peptide derived from ZIP10 (CNHDHSEQYEHNR) was synthesized and injected into Charles River Laboratories SPF rabbits (Division of Comparative Medicine, University of Toronto). Upon exsanguination, a portion of serum was purified using the SulfoLink Immobilization Kit for Peptides (cat. no. 44999, Pierce, Rockford, IL, USA) following instructions provided by the manufacturer.

118

5.2.9 Reverse transcription polymerase chain reaction (RT-PCR)

cDNAs were generated from the total RNA (5 µg) of N2a cells or mouse brain (CD-1) using the StrataScript First Strand synthesis kit (Stratagene, La Jolla, CA, USA). PCR primers were designed using Primer3 software (S. Rozen and H. Skaletsky, Whitehead Institute for Biomedical Research) as follows:

ZIP4: F: AGACTTGGTGGCAGAGGAGA, R: GGGATGGGATAGAGCTGTCA

ZIP5: F: CTCCCCTCTCTCTGCTGTTG, R: AGAAGCCGTCTGAGAAAGCA

ZIP6: F: GGATGTGGAGAGCAAGAAGC, R: CACAGCGACAGAGGTGCTTA

ZIP8: F: AGCAGTTGCTGTGTTTGGTG, R: GCGGAGAGGAAGTTGAACAG

ZIP10: F: TTACCTGGGTGTGGAGGAAG, R: TTGCGAAGATCCAGAGTGTG

ZIP14: F: TGCCTCCTTCACTGTGTCTG, R: AAGAGCTGCCTTTTCCATGA

A reaction volume of 50 µL containing 2 µL cDNA, 25 pmol of each primer, 5 µL PCR reaction buffer (Qiagen) and 2.5U Taq polymerase (Qiagen) was thermocycled for 35 cycles of 94°C for 30s, 60°C for 30s, and 72°C for 30s. The resulting PCR fragments were resolved on a 1.5% agarose gel.

5.2.10 Preparation of mouse brain homogenates

Scrapie inoculations with mouse-adapted Rocky Mountain Laboratory (RML) prions were carried out as previously described [360, 361]. Briefly, FVB/NCr prnp+/+ mice were intracerebrally inoculated in the left hemisphere with 20 µl of 1:10 diluted (w/v) brain homogenate in PBS, sourced from either clinically-ill RML-infected or normal mice. Following 84 or 131 (or 143) days p.i., the right hemispheres of mice were harvested and sliced into small sections. To homogenize the brain pieces, a tenfold excess (v/w) of Extraction Buffer (Tris-HCl, pH 7.5, 2% SDS and 100 mM dithiothreitol) was added, followed by a 5-min incubation at 90°C and 3 consecutive 1-min bead beating steps at ‘homogenization’ speed, with the steps interrupted by 1-min breaks (Mini Beadbeater-8, BioSpec Products, Bartlesville, OK, USA) in the presence of 1-mm zirconia beads (BioSpec Products). Samples were spun for 20 min at 38,000x g to remove insoluble debris and, subsequently, diluted in concentrated Laemmli sample buffer to

119

adjust to a final concentration of 2% SDS, 100 mM DTT, 30% glycerol, 60 mM Tris-HCl, pH 6.5, and 0.001% bromophenol blue.

5.2.11 SDS-PAGE and immunoblotting

For denaturing sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) analyses, protein mixtures were separated at 180 V in 4-12% or 12% NuPAGE Bis-Tris gels (Invitrogen) in MOPS running buffer. The transfer of proteins to polyvinyl fluoride (PVDF) membranes was accomplished within 1.5 hours at 40V. Subsequently, membranes were probed overnight with one of the following primary antibodies: anti-Drebrin (1:100; cat. no. D029-3, MBL, Woburn, MA, USA), anti-Munc18 (1:10,000; cat. no. 610336, BD Transduction, Mississauga, ON, Canada), anti-MnSOD (1:1,000; cat. no. SOD-110D, Enzo Life Sciences, Farmingdale, NY, USA), anti-Histone H4 (1:1,000; cat. no. 04-118, Millipore, Billerica, MA, USA), anti-TfR (1:500; cat. no. 13-6800, Invitrogen), anti-CCS (1:200; cat. no. sc-20141, Santa Cruz Biotechnology, Santa Cruz, CA, USA), anti-GAPDH (1:2,000; cat. no. G8795, Sigma- Aldrich), anti-ZIP10 (1:800), anti-HA.11 (1:1,000; cat. no. MMS-101P, Covance, Princeton, NJ, USA), anti-PrP rabbit monoclonal (1:5,000; cat. no. 2063-1, Epitomics, Burlingame, CA, USA) or anti-PrP Sha31 (1:5,000; cat. no. A03213, Medicorp, Montreal, QC, Canada). Membranes were then incubated with anti-mouse (1:10,000; cat. no. 170-6516, Bio-Rad, Mississauga, ON, Canada) or anti-rabbit (1:10,000; cat. no. 170-6515, Bio-Rad) secondary antibody, and visualized using ECL reagent (cat. no. RPN2106, GE Healthcare, Baie d’Urfe, QC, Canada). For band quantification, the histogram feature of Photoshop CS5 (Adobe Systems, San Jose, CA, USA) was used, and two-sample, two-tailed Student’s t-test assuming unequal variance was utilized for tests of statistical significance.

5.2.12 Proteinase K digestion

Brain extracts were digested with 12 μg/mL of proteinase K (PK) for 30 min at 37°C. The reaction was stopped with 2 mM phenylmethanesulfonylfluoride (PMSF) and incubated with 0.7% phosphotungstic acid (PTA) and 0.1% sodium lauroyl sarcosinate for 30 min at 37°C to precipitate PrPSc. The samples were subsequently pelleted, resuspended in sample buffer and heated at 90°C for 10 min.

120

5.2.13 Sucrose gradient isolation of lipid rafts

Lipid rafts (also referred to as caveolae-like membrane domains; CLDs) were isolated as described previously [362, 363]. Briefly, confluent cells were washed twice in ice-cold PBS (pH 7.2) and lysed in a Triton X-100-containing lysis buffer (25 mM Tris-HCl pH 7.5, 150 mM NaCl, 1% Triton X-100, 1× Roche Complete Protease Inhibitor Cocktail). After 20 min incubation on ice, cells were scraped and homogenized through successively higher gauge syringe needles. The lysate was diluted 1:1 in 80% sucrose and placed at the bottom of a 38.5 mL open-top thinwall ultra-clear ultracentrifuge tube (Beckman Coulter, Indianapolis, IN, USA). Two stepwise sucrose layers of (i) 25% and (ii) 5% were poured (at a rate <0.5 ml/min) above the lysate. All sucrose dilutions were prepared in lysis buffer lacking Triton X-100. Ultracentrifugation was performed for 21 h at 28,000 rpm (140,000 ×g) in a Beckman Coulter Optima L-100 XP ultracentrifuge and SW 32 Ti rotor. Upon completion of centrifugation, the 5%, 5%-25% transition (representing the lipid raft fraction), 25% and lysate layers were collected. The 5%-25% transition layer was diluted with lysis buffer and pelleted, and subsequently resuspended in 1× Laemmli sample buffer. All other layers were diluted 1:1 with 2× sample buffer. Lastly, the ultracentrifugation pellet was dissolved in 1× Laemmli sample buffer directly in the ultracentrifuge tube. All sample-buffer-containing samples were analyzed using Western blotting.

5.2.14 Transient overexpression and knockdown

Mouse neuroblastoma (N2a) cells were cultured in Minimum Essential Medium (MEM) (cat. no. 11095, Gibco, Burlington, ON, Canada) supplemented with 10% FBS (cat. no. 12484028, Gibco), 1% GlutaMAX (cat. no. 35050, Gibco) and 1% antibiotic/antimycotic solution (cat. no. 15250-062, Gibco). Lipofectamine 2000 transfection reagent (Invitrogen) was used as per the manufacturer’s instructions. For transient overexpressions, 5 mM sodium butyrate (cat. no. B5887, Sigma-Aldrich) was added to cells 24 h post-transfection from a 1.5 M stock solution in water. The transfection duration was 48 h. Cells were lysed in Lysis Buffer (50 mM Tris, pH 8.0, 150 mM NaCl, 0.5% NP-40, 0.5% deoxycholate, 1× Roche Complete Protease Inhibitor Cocktail), followed by BCA protein adjustment and 10-min heating at 45°C in Laemmli sample buffer before SDS-PAGE analysis.

121

5.3 Results

5.3.1 Prion disease in mice is accompanied by abnormal endoproteolysis of ZIP10

Whereas comprehensive time-course data exist which document changes to the global transcriptome in response to challenging mice with prions [290], a similar analysis at the proteome level is lacking. Of note, however, is a study which uncovered a robust and progressive reduction of protein but not mRNA levels of the prion paralog Shadoo during prion disease manifestation [143]. In light of these earlier observations we wondered whether protein levels of LZTs might be similarly modified in mice which had been intracerebrally inoculated with mouse-adapted prions derived from the RML prion strain. Because ZIP10 was predicted to be expressed in the brain [364] and was identified as a candidate interactor of PrPC in N2a cells, we set out to raise a polyclonal antibody directed against a paralog-specific epitope within the (N- terminal) ectodomain of endogenous mouse ZIP10 (anti-ZIP10_N). Subsequent immunoblot analyses following the separation of mouse brain extracts by denaturing SDS-PAGE confirmed the expression of ZIP10 in the brain (Figure 5.1). The band pattern revealed by the ZIP10- directed antibody, however, was complex, suggesting that the antibody may crossreact with other proteins and/or the respective endogenous ZIP protein may undergo complex posttranslational modifications (please see below).

Figure 5.1: Prion disease in mice causes appearance of novel protein band which can be detected with ZIP10-directed antibody. Mouse littermates were inoculated with 10% brain homogenates generated from the brains of naïve control mice or mice which had been infected

122

with mouse-adapted RML prions. 84 or 131 days following i.c. challenge with prions, the mice were sacrificed and their brain homogenates adjusted for equal protein concentration and side- by-side analyzed by Western blotting with antibodies directed against PrPC or ZIP10. Please note the appearance of a protein band which is barely visible in control mice but can be detected with the ZIP10-directed antibody in RML inoculated mice (marked with arrowhead). PrPSc infection in the 131d RML brain was confirmed by the presence of high-MW aggregates and relative proteinase K (PK) resistance. Based on the known epitope specificity of the anti-ZIP10 antibody (Figure 5.4A), the novel band can be inferred to comprise the respective PrP-like ectodomain of ZIP10. The apparent MWs of additional signals detected with the anti-ZIP10 antibody are indicated. Control Western blots with antibodies which can probe for GAPDH or cellular transporters of iron (transferrin receptor; TfR) and copper (copper chaperone for superoxide dismutase; CCS) revealed no differences in the signal levels of these proteins in brain homogenates derived from negative control or prion-challenged mice. Please see Figure 5.2 for quantitative analyses of levels of additional cell status reporters.

A comparison of the ZIP10 band pattern observed in control mice and littermates 84 days post-inoculation (p.i.) with RML prions appeared largely identical. Strikingly, however, the analysis of brain extracts of mice sacrificed 131 days p.i. with prions exhibited an additional band which migrated with an apparent molecular weight (MW) of 45 kDa (anti-ZIP10_N). In fact, this band was already detectable 84 days p.i. and was much less pronounced in age-matched control animals which had been inoculated with brain extracts from naïve mice. When probed with an antibody (Sha31) that recognizes endogenous PrP, the 131 p.i. RML-inoculated brain extract revealed the expected aggregation of PrP characteristic of prion-infected mice. Parallel comparisons we undertook with antibodies that recognize iron or copper transporters (TfR or CCS) or can probe for the expression of GAPDH documented no differences in protein levels or posttranslational modifications of these control proteins between non-inoculated and RML- inoculated mice. When we probed additional cell status reporter proteins which can inform about synaptic integrity (Drebrin and Munc18) or cellular stress (MnSOD and Histone H4) subtle age- related differences in their expression levels were observed but again no significant difference in age-matched RML-inoculated versus control brains were seen (Figure 5.2). A subsequent repetition of the entire experiment led to an identical result and thus ruled out that these observations represent mere idiosyncrasies of the first cohort of mice. Taken together, these experiments suggested that the manifestation of prion disease in mouse brains might be paralleled by alterations in the expression or posttranslational modification of the ZIP10 transporter which populates the evolutionarily-nearest phylogenetic subbranch of LZTs.

123

However, this experiment also raised many questions that might best be answered by moving to an experimental paradigm more amenable to experimental manipulation.

Figure 5.2: Quantitative comparisons of protein levels or posttranslational modifications of selected cell status reporter proteins reveal no significant differences between age-matched control or prion-infected mice. Control Western blots with antibodies which detect reporters of synapse integrity (Drebrin, Munc18), oxidative stress (MnSOD) or epigenetic modifications (Histone H4) revealed no new bands (not shown) or differences in the levels of expression in brain homogenates derived from age-matched negative control and prion challenged mice. Western blot analyses were repeated three times and relative intensities depicted in bar graphs, with error bars indicating standard error of the mean. To normalize band intensities across different Western blots for quantification, the intensities of the 84d control bands were read at the same level and used as base values for their respective blots. No statistically significant differences were found.

124

5.3.2 ZIP10 is one of many LZTs expressed in N2a cells that localizes to the plasma membrane and undergoes complex posttranslational maturation

Of the nine LZTs present in the genomes of mice and humans, up to seven may contain PrP-like ectodomains [1]. In preparation for studies in the mouse neuroblastoma cell model Neuro2a (N2a), the best-characterized cell model for the study of prion biology to date, reverse transcription polymerase chain reactions (RT-PCRs) comparing the transcriptional expression of relevant LZTs in N2a cells and brains of twelve-week-old CD-1 outbred mice were conducted. Data from this experiment suggested that a complex LZT biology may exist in N2a cells and that this cell model may constitute a suitable proxy to the brain for studying the biology of LZTs (Figure 5.3). Furthermore, it provided a straightforward explanation for why ZIP5, the LZT paralog which has retained the strongest sequence similarity to its molecular cousin PrPC, could not, unlike ZIPs 6 and 10, be identified as a PrPC candidate interactor in our interactome study which was also based on N2a cells [102].

Figure 5.3: Multiple LZTs but not ZIP5 are expressed in N2a cells. Expression analysis of LZT proteins by RT-PCR detected transcripts of ZIPs 4, 6, 8, 10 and 14 in N2a cells and brains of CD-1 outbred mice. Primer pairs were designed to amplify stretches of the respective mRNAs extending over 500-550 base pairs (bps). This experiment revealed the expected amplification products confirming expression of ZIPs 4, 6, 8, 10 and 14 but not ZIP5 in N2a cells and a generally overlapping transcriptional profile in CD-1 brains, with the notable exception of a minor signal of ZIP5 transcript present in brain tissue. In the absence of ZIP5 mRNA which could compete with ZIP5 genomic DNA during the primer annealing step, the RT-PCR resulted in the expected amplification of a genomic ZIP5 segment of 1,385 bps which served as an additional internal positive control, establishing the selectivity and sensitivity of the assay. Amplifications of genomic DNA are not expected in the presence of an excess of amplifiable RNA templates. Similar results to the ones shown here were observed for alternative pairs of RT- PCR primers for each of the ZIP sequences targeted in this analysis (data not shown). Expected

125

product sizes: ZIP4: 549 bp, ZIP5: 543 bp (1,385 bp with introns), ZIP6: 511 bp, ZIP8: 531 bp, ZIP10: 512 bp, ZIP14: 522 bp.

We next conducted confocal immunofluorescence analyses of HA-tagged ZIP10 in N2a cells and observed the expected predominant localization of this protein at the plasma membrane, with signals primarily appearing in the form of distinct puncta and occasionally extending into filamentous membrane protrusions (Figure 5.4B). Additional punctate intracellular signals but no nuclear signals were observed, consistent with a cell biology that involves passage through the secretory pathway and possible degradation in endolysosomal compartments. These data were in good agreement with immunohistochemical data we collected with the in-house generated antibody recognizing ZIP10 (Figure 5.4A and 5.4C). The latter, however, showed less intracellular staining, conceivably a consequence of a lower turnover rate of ZIP10 in mature brain relative to cells grown in culture. Next, co-immunofluorescence analyses of the same HA- tagged ZIP10 and endogenous PrPC were undertaken, revealing that, at most, a partial co- localization of PrP and ZIP10 may exist in cells (Figure 5.5A). Similarly, no ZIP10 was detected in the low density sucrose fraction (lipid rafts) populated by PrPC that is known to be enriched in cholesterol and sphingolipids and can be obtained following extraction of ZIP10-HA-expressing cells in cold Triton X-100 (Figure 5.5B). These data were consistent with the low ZIP10 sequence coverage observed in the PrPC interactome analysis which preceded this work [102], a possible indicator that the interaction with the PrPC bait was either relatively weak or that only a small population of the total cellular pools of PrPC and ZIP10 proteins engage in protein-protein interactions.

126

Figure 5.4: ZIP10 is localized at the plasma membrane and undergoes complex posttranslational modifications. Cellular distribution and posttranslational processing of ZIP10 in N2a cells. (A) Schematic diagram depicting targeted epitopes within ZIP10 as well as amino acid sequence of synthetic peptide antigens used for conjugation to KLH and for raising polyclonal rabbit antibodies. (B) N2a cells were transfected with a ZIP10-HA expression vector and the distribution of the heterologously expressed protein analyzed by immunofluorescence. A comparison of bright field, extended focus and confocal views revealed predominant localization of ZIP10 at the plasma membrane and its association with punctate structures seen within the cell

127

and within neuritic membrane protrusions (see inset). Please note that a negligible background stain was observed in non-transfected cells (not shown). (C) Immunohistochemical analysis of endogenous ZIP10 on formaldehyde crosslinked and paraffin-embedded mouse brain sections revealed predominant punctate staining of plasma membranes of cortical neurons. Immunohistochemical staining generated with ZIP10-directed polyclonal antibodies which had been pre-saturated with their cognate peptide served as negative controls. (D) Transient transfection of ZIP10-HA followed by detection with anti-HA antibody recognizing C-terminal HA-tag or with anti-ZIP10_N antibody. A comparison of signals revealed the SDS-stable ZIP10 dimer to probably represent an overexpression artifact, and identified two C-terminal cleavage products which were designated as C1 and C_ecto on the basis of evidence revealed in experiments described below. The asterisk indicates an unspecific band. (E) Cellular extracts of ZIP10-HA transfected cells digested with PNGase F were analyzed next to undigested extracts. A comparison of band patterns indicated that both full-length ZIP10 and the C-terminal fragment C1, but not C_ecto, were N-glycosylated.

128

Figure 5.5: ZIP10 exhibits low level of co-localization with PrPC in N2a cells. (A) Representative co-immunofluorescence data generated with antibodies directed against endogenous PrPC and the C-terminal HA-tag present on heterologously expressed ZIP10-HA. Note the largely distinct localizations of PrPC and ZIP10. (B) N2a cell extracts were generated in the presence of cold Triton X-100. Following their adjustment to 40% sucrose, extracts were overlaid with layers of 25% and 5% sucrose and subject to separation by ultracentrifugation. No ZIP10 was detected in the low density sucrose fraction populated by PrPC known to be enriched in cholesterol and sphingolipids.

129

A wealth of data exists characterizing the molecular biology of PrPC and its posttranslational processing in N2a cells. To begin to understand the significance of the ZIP10- PrP relationship at the molecular level, a biochemical characterization of ZIP10 in N2a cells might be informative. When analyzed by denaturing SDS-PAGE, the heterologous expression of ZIP10 in N2a cells caused the appearance of multiple bands (Figure 5.4D). Prominent signals detected migrated with apparent MWs of 80-90 kDa, consistent with the expected size of full- length ZIP10. Additional lower MW bands were observed in the range of 40-60 kDa. Initially, the ZIP10-HA expression construct was separately analyzed with a monoclonal antibody that detects the HA-tag at the C-terminus of ZIP10-HA or the aforementioned in-house generated polyclonal antibody (used in Figure 5.1) that recognizes an epitope present within the ectodomain of both endogenous ZIP10 and overexpressed ZIP10-HA (Figure 5.4A). This analysis revealed a band pattern for endogenous ZIP10 in the non-transfected (empty plasmid) control that was qualitatively identical to the one observed following ZIP10 overexpression, thereby establishing that the appearance of the band duplet, which we interpreted to represent full-length ZIP10, did not constitute an overexpression artifact. On the basis of published work documenting N-glycosylation of a subset of LZTs [131, 173] and the presence of NxT consensus motifs within the amino acid sequences of ZIPs 4, 5 and 10 (Figure 5.6), it seemed likely that N- glycosylation also contributed to the complexity of the band patterns observed in these experiments. Indeed, PNGase F digestion of cell extracts harvested following ZIP10-HA transfection caused the full-length ZIP10-HA band duplet to be replaced with a single band (Figure 5.4E). These data established ZIP10 to be partially posttranslationally modified by N- glycosylation and to be subject to endoproteolysis. On the other hand, alternative splicing was unlikely to be a major factor contributing to ZIP10 band complexity because the expression of ZIP10 from a heterologous cassette vector which lacked the genomic intron/exon structure gave rise to all major bands seen with endogenous ZIP10.

130

131

Figure 5.6: N-glycosylation acceptor sites within ZIP10. The primary structure of ZIP10 contains multiple ‘Nx(S/T)’ sequons which can be classified as strong, medium or weak consensus motifs for N-glycosylation. The only sequon that is positionally conserved between ZIP10, ZIP5 and PrPC resides within the cysteine flanked core (CFC) domain. Data from this study establish that ZIP10 can, like PrPC, be N-glycosylated at this site and further suggest that up to three additional acceptor sites N-terminal of its first transmembrane domain (TMD1) can be occupied in mature ZIP10. No occupancy of putative ZIP10 N-glycan acceptor sites was observed within the C-terminal transmembrane domains forming the transition metal channel.

5.3.3 ZIP10 is N-glycosylated at ‘NxT’ acceptor site that it shares with PrPC

To determine whether N-glycosylation of ZIP10 occurs at the ‘NxT’ acceptor site conserved between PrP and ZIP10 [2] (Figure 5.6), a ZIP10-HA_N341A expression construct was generated which replaced the critical asparagine within the N-glycosylation consensus motif with an alanine. Additional ZIP10-HA-derived truncation constructs were generated to define endoproteolysis sites (Figure 5.7A). The rationale for choosing truncation sites followed the anticipation that the cysteine-flanked core (CFC) domain (the most conserved stretch of sequence in multiple alignments of PrP and LZT sequences [2]) was unlikely to be targeted by endoproteolysis and that instead cleavage sites may flank this domain. Heterologous expression of these constructs in N2a cells followed by immunoblot analyses and PNGase F treatment (Figure 5.7B) allowed several conclusions which established that (i) the N341 N-glycan acceptor site (positionally shared with PrP) is indeed quantitatively N-glycosylated; (ii) additional N- glycosylations occur within the ectodomain but not within the C-terminal transmembrane domain; and (iii) endoproteolytic cleavages indeed occur at sites which flank the CFC and give rise to C-terminal fragments with approximate N-terminal boundaries near amino acids 300 and 400 (numbering based on wild-type mouse ZIP10), hereafter referred to as ZIP10 C1 and ZIP10 C_ecto, respectively.

132

Figure 5.7: ZIP10 is N-glycosylated at ‘NxT’ acceptor site it shares with PrPC and is partially shed from the membrane by at least two distinct cleavages. (A) Schematic presentation of ZIP10-HA-derived expression constructs used in subsequent experiments. Colors and symbols are explained in the legend to the right of the schematic. (B) Transient expression analysis of ZIP10 expression constructs introduced in panel A with and without prior digestion of extracts with PNGase F. A comparison of bands revealed the N341 residue within the ‘NxT’ acceptor site to be quantitatively N-glycosylated. Please note, however, that additional N- glycosylation sites must exist since digestion with PNGase F leads to a 3- to 4-fold larger shift in the apparent molecular weight of the slow-migrating full-length ZIP10 signal. Please also note the similarity in apparent molecular weight of C-terminal cleavage products C1 and C_ecto to the length of ZIP10 expression constructs with N-terminal truncations that extend to amino acids 300 and 389, respectively. Extract amounts analyzed were adjusted for total protein levels.

5.3.4 Appearance of truncated and N-glycosylated variants of ZIP10 depends on extracellular divalent metal ion status

In preparation for experiments addressing whether metal-induced cleavages of LZTs contribute to the appearance of faster-migrating bands observed following inoculation of mice with RML

133 prions and/or during SDS-PAGE analysis of LZTs in N2a cells, the quantity of divalent metals in the N2a growth medium was measured by inductively coupled plasma atomic emission spectroscopy (ICP-AES). This analysis revealed low micromolar zinc levels, ten-fold lower copper levels and negligible manganese levels in the fetal bovine serum-supplemented minimal essential medium (MEM) used for culturing N2a cells (Table 5.1). Next, levels of zinc, copper or manganese were experimentally augmented or suppressed in the growth medium by adding the respective salts or including the transition-metal chelator N,N,N’,N’-tetrakis (2- pyridylmethyl) ethylenediamine (TPEN) (Figure 5.8A). Interestingly, the individual bands within the full-length ZIP10 band duplet were affected by these treatments differently, i.e., when transition metals were replete, levels of the N-glycosylated derivative of the full-length ZIP10 band were slightly reduced. In contrast, and reminiscent of what had been described for ZIP4 [194], starving cells of transition metals by chelation resulted in an increase in the intensity of the signal interpreted to represent N-glycosylated full-length ZIP10 and the appearance of a smaller ZIP10 product which migrated at an apparent molecular weight of 45 kDa. Interestingly, the chelation of transition metals with TPEN also caused changes in the way PrPC-reactive signals were presented when Western blots were immunoprobed with the Sha31 antibody. More specifically, a considerable proportion of full-length N-glycosylated PrPC had been replaced with Sha31-reactive bands which migrated at lower apparent molecular weights, including signals at 19 kDa, and as such lower than the expected molecular weight of unglycosylated full-length PrPC. This result suggested that treatment of cells with TPEN causes profound changes to posttranslational modifications of PrPC that include its endoproteolysis. Thus, endogenous ZIP10 and PrPC undergo proteolysis in the presence of TPEN; and both the inoculation of mice with mouse-adapted RML prions and TPEN-based chelation can generate a product of endogenous mouse ZIP10 that migrates at 45 kDa.

134

Table 5.1: Inductively coupled plasma atomic emission spectroscopy analysis of transition metals in relevant media and solutions

Sample Zn2+ (μM) Cu2+ (μM) Fe2+ (μM) Mn2+ (μM) MEM 0.3 0.1 0.1 0.0 MEM + 10%FBS 3.3 0.3 2.6 0.0 FBS (stock) 28.5 2.0 27.3 0.5 DMEM 0.4 0.0 0.2 0.0 Opti-MEM 0.9 0.1 0.4 0.0

Notes: 1. Measurement wavelengths: Zn = 213.857 nm Cu = 324.752 nm Fe = 239.562 nm Mn = 257.610 nm

2. Molecular weights: Zn MW = 65.382 g/mol Cu MW = 63.546 g/mol Fe MW = 55.845 g/mol Mn MW = 54.938 g/mol

135

136

Figure 5.8: Depletion of zinc and manganese but not copper from the cell culture medium causes shedding of the ZIP10 ectodomain and cleavage of PrPC in N2a cells. N2a or GT1 cells were grown for two days in cell culture media which were either supplemented with defined concentrations of individual transition metals or functionally depleted of transition metals by the addition of chelation agents. Subsequently, the posttranslational processing of PrPC and ZIP10 was characterized by immunoblotting of cellular extracts. (A) Whereas both ZIP10 and PrPC demonstrate only subtle changes in their overall full-length expression levels in the presence of an excess of transition metals, both proteins undergo endoproteolysis when transition metals were depleted by the cell-permeable transition metal chelator TPEN. (B) To determine the cellular tolerance toward the use of transition metal chelators in the cell culture medium, an MTT assay analysis was conducted with a range of TPEN or DTPA concentrations. (C) DTPA- induced ZIP10 cleavage causes the majority of N-terminal ZIP10 fragments to be shed into the cell culture medium. On the basis of calculations which consider the relative amounts of cell culture volume and cell pellet analyzed, the ratio of shed versus cell-bound ZIP10 fragments can be estimated to exceed 95%. (D) A phenotypic rescue experiment was conducted to determine the identity of the transition metal(s) which can rescue the DTPA-mediated endoproteolysis of PrP and ZIP10. The experimental steps were as for panel A, except that individual transition metals were added to the cell culture medium concomitantly with DTPA at the concentrations

137 indicated and the data are presented in composite Western blots. (E) No endoproteolysis of PrP or ZIP10 was observed when analogous experiments were conducted with GT1 cells. Of note, however, are the opposite effects of DTPA or DTPA plus 50 µM zinc on full-length PrPC and ZIP10 protein levels which echoed observations made in N2a cells. Note that the absence of the cleavage products may reflect that endoproteolytic activities are less pronounced in these cells or proteolytic fragments are more effectively cleared.

5.3.5 Appearance of truncated ZIP10 can be rescued by manganese and zinc but not copper replenishment

Given that TPEN has been shown to pass the cellular membrane [365] and, in our hands, was toxic to N2a cells after extended incubation periods (even at low micromolar concentrations), we replaced TPEN in subsequent experiments with the cell-impermeable transition-metal chelator diethylene triamine pentaacetic acid (DTPA) which exhibited lower cytotoxicity than TPEN (Figure 5.8B). Importantly, the replacement of TPEN with DTPA did not abolish the generation of the 45 kDa endogenous mouse ZIP10 product, or the formation of the 17 kDa PrP-derived signal, indicating that the appearance of these bands is indeed most likely a consequence of N2a cells sensing an extracellular environment deprived of transition metals rather than a result of TPEN exerting this effect independent of its metal-chelation properties or on the basis of its capacity to permeate into cells. To explore whether the 45 kDa fragment is generated as a byproduct of an intracellular ZIP10 degradation pathway or is shed/secreted into the medium, the experiment was repeated with N2a cells grown in serum-free medium (Figure 5.8C). The side- by-side Western blot analysis of an aliquot of the cell culture supernatant of cells grown in the presence of DTPA next to cellular extract fractions from the same cells, followed by densitometric analysis of signal intensities, revealed that approximately 95% of the 45 kDa product was indeed detected in the medium. Since neither TPEN nor DTPA possess exquisite selectivity for a particular transition metal, experiments up to this point had not answered which of the transition metals, when depleted, caused the appearance of the 45 kDa ZIP10 band. Rescue experiments conducted by adding DTPA together with zinc, copper or manganese to the growth medium established that 25 μM concentrations of zinc or manganese but not copper can prevent the formation of the 45 kDa ZIP10 band (Figure 5.8D). Of note, the functional depletion of transition metals from the medium by the addition of DTPA was again paralleled by a diminution of full-length PrPC levels. Strikingly, the activity profile of transition metals which enabled

138 rescue of this PrPC effect was also identical to the profile we had observed for rescuing ZIP10 endoproteolysis, i.e., the addition of copper had no influence on the DTPA-mediated diminution of full-length PrPC signals, despite the ability of DTPA to also chelate copper [191] and a well- documented preferential association of PrPC with copper. Finally, to assess whether another well- established cell model for the study of prion biology responds similarly to the starvation of transition metals, GT1 cells (immortalized murine hypothalamic neurons [366]) were grown in the presence of DTPA and cellular extracts were again analyzed by immunoblotting with anti- ZIP10 or anti-PrP antibodies (Figure 5.8E). The analysis gave rise to less complex immunoblot band patterns for both ZIP10 and PrPC, consistent with the interpretation that endoproteolytic activities are less pronounced in these cells or proteolytic fragments are more effectively cleared. However, the overall trends with which full-length ZIP10 or PrPC bands were altered in response to DTPA as well as the ability of zinc to rescue this phenotype was not only retained but appeared even more pronounced. Taken together, these experiments established that starving the cell culture medium of zinc or manganese not only triggers an increase in full-length N- glycosylated ZIP10 levels and the shedding of a 45 kDa ZIP10 ectodomain fragment, but also causes a diminution of full-length N-glycosylated PrPC, possibly through an endoproteolysis/degradation step. Experiments up to this point, however, failed to reveal the identity of the respective ZIP10 or PrPC cleavage products and did not address whether a putative zinc metalloprotease active center present in LZTs may contribute to endoproteolytic cleavages observed.

5.3.6 A zinc coordination site within TMD5 regulates N-glycosylation and endoproteolysis of ZIP10

Sequence alignments of LZTs revealed a conserved ‘(C/L)(H/E)E(L/I//V/F)PHE(L/I/V/M)GDF’ sequence motif within their predicted transmembrane domain 5 (TMD5) (Figure 5.9A). The conservation of this motif had also been noted by others [11], and two separate roles of histidine residues embedded within this motif were considered: (i) a function in the gating of transition metals; or (ii) a contribution to the assembly of a zinc metalloproteinase (MP) active center with resemblance to metzincin or deformylase group MPs [367]. Hypothesizing that the presence of TPEN or DTPA may have interfered with the transition metal occupancy of this coordination center and that one may gain more insight into the identity of the 45 kDa cleavage band by close examination of the role of these histidines, site-directed mutagenesis was employed to generate

139 alanine replacement mutants. The comparison of transiently expressed wild-type ZIP10-HA and mutant ZIP10-HA-(H711A/H715A) by immunoblotting revealed two striking differences: (i) a larger proportion of full-length ZIP10-HA-(H711A/H715A) migrated at the level of N- glycosylated wild-type ZIP10-HA, suggesting that point mutations either altered SDS-PAGE separation characteristics or that the mutant is indeed more readily glycosylated as it traverses the secretory pathway; and (ii) expression of ZIP10-HA-(H711A/H715A) caused the appearance of conspicuous faster-migrating bands. These bands were detected with the polyclonal antibody which targets the ZIP10 ectodomain and migrated in SDS-PAGE analyses with apparent MWs of 50 and 52 kDa (Figure 5.9B).

140

Figure 5.9: A conserved zinc coordination site within TMD5 influences N-glycosylation and shedding of the ZIP10 ectodomain. Altered processing of ZIP10, in which histidines within a putative metalloprotease catalytic center characterized by the ‘CHELPHELGDF’ motif were mutated to alanines. (A) Multiple sequence alignment of mouse ZIP paralogs in the region surrounding a putative zinc metalloprotease catalytic center present within the predicted transmembrane domain 5 (TMD5) in a majority of sequences belonging to the LZT branch of ZIP transporters. (B) N2a cells were transiently transfected with wild-type ZIP10-HA or a mutant ZIP10-HA-(H711A/H715A) and, following 48 hours of cell growth, cell extracts were

141 analyzed by immunoblotting. Please note the differences in the apparent molecular weights of a subset of ZIP10 wild-type and ZIP10 mutant expression products. The numbering 1 to 4 shown next to the N_ecto label indicates the number of N-glycans present within individual ZIP10 ectodomains observed (see also panels C-E for evidence in support of this interpretation). (C) When cells expressing mutant ZIP10-HA were grown in the presence of DTPA, the generation of the ZIP10 N_ecto cleavage product modified by a single N-glycan takes precedence over the generation of more fully N-glycosylated N_ecto cleavage products. Please note that the molecular weight ladder indicates protein weights in both Western blots. (D) Cellular extracts obtained from naïve N2a cells or N2a cells grown in the presence of DTPA were subjected to digestion with PNGase F at a low or high enzymatic dose. All ZIP10 cleavage products merged into a single band when fully digested with PNGase F. (E) Similarly, cleavage products of heterologously expressed ZIP10-HA-(H711A/H715A) all converted to a band of the same apparent molecular weight when digested with a high dose of PNGase F.

We next investigated the relationship between the faster-migrating bands observed with either TPEN-/DTPA-treated endogenous ZIP10 or with mutant ZIP10-HA-(H711A/H715A). Whereas the former appeared to migrate at 45 kDa and the mechanism of its generation was not immediately apparent, the latter was observed as a duplet of 50 and 52 kDa and had to be an endoproteolysis product because it was generated from a plasmid-encoded parent molecule. Interestingly, the DTPA-mediated starvation of transition metals in the culture medium of cells expressing ZIP10-HA-(H711A/H715A) caused these cells to override the molecular biology which led to the 50/52 kDa band duplet and made the ZIP10-HA-(H711A/H715A)-derived signals look as if they had originated from cells transfected with wild-type ZIP10-HA (Figure 5.9C). Thus, the 45 kDa fragment generated by DTPA chelation must itself originate from an endoproteolysis event which either supersedes cleavages giving rise to the 50/52 kDa band duplet or precedes the formation of the latter within the cellular life cycle of a ZIP10 molecule. Given that the absence of the transition metal coordination center caused a more pronounced N- glycosylation of the parent molecule, it then seemed plausible that the slower migration of the 50/52 kDa band duplet (relative to the 45 kDa band seen with TPEN or DTPA) merely reflected a higher level of N-glycosylation. Indeed, the side-by-side removal of N-linked glycans caused the fast-migrating 50/52 kDa band duplet to be replaced by a single band of 38 kDa, identical in size to the deglycosylated derivative of the 45 kDa band seen when cells were grown in transition-metal-depleted growth medium (Figure 5.9D).

142

Finally, by further analyzing ZIP10-HA-(H711A/H715A) following the removal of N- linked glycans, it became apparent that the point mutations had indeed caused a larger population of ZIP10-HA-(H711A/H715A) to be N-glycosylated (Figure 5.9E). This result was counterintuitive given that N-linked glycans had been mapped to the ectodomain (Figure 5.7B), not the C-terminal TMDs which form the membrane channel of which the mutated sequence motif is a part. This result was illuminating, however, as it suggested that the ectodomain might engage in an interaction that relies on the integrity of the conserved histidines within TMD5. Taken together, experiments in this section revealed the identity of multiple ectodomain fragments observed with different apparent molecular weights in response to transition metal starvation or mutation of conserved TMD5 histidine residues. It showed that these fragments constitute products of ZIP10 cleavage at a site near ZIP10 amino acid 400 (giving rise to C_ecto and N_ecto products) which merely differ with regard to the level of occupancy of up to four N- glycan acceptor sites present within the ectodomain of ZIP10 (Figures 5.6 and 5.7A).

5.3.7 Transition metal starvation or RML infection causes PrPC to shed distinct N1 or N2 fragments but triggers shedding of identical N_ecto fragments from immature ZIP10

We next turned to characterizing the underlying cause for the diminution of PrPC levels under manganese or zinc starvation and to understanding the relationship of cleavages observed following infection with prions versus starvation of transition metals. It has repeatedly been shown that the prion protein can be cleaved at a so-called α-site bond connecting amino acids 109 and 110 (in mouse) [368], causing the release of an N-terminal fragment N1 into the medium and the cellular retention of a C-terminal C1 fragment. A second scissile bond within PrP, termed the β-site, has been identified in proximity to the C-terminal boundary of the octapeptide repeat region [190]. Cleavages at the β-site, giving rise to fragments designated as N2 and C2, had been shown to be considerably increased upon prion infection [177]. Both C1 and C2 are routinely identified by their characteristic apparent MWs following enzymatic removal of N-glycans [369]. In this study, the immunoblot analysis of cellular fractions obtained following transition metal starvation versus infection with prions demonstrated the former treatment to increase the amount of C1, and confirmed the anticipated increase of C2 in prion-infected cells or mouse brain (Figure 5.10A). In contrast, a side-by-side comparison of these extracts with the antibody reactive toward the ZIP10 N-terminus revealed that both transition metal starvation of cells or

143

RML inoculation of mice led to the selective appearance of a ZIP10 fragment that migrates with an identical apparent MW of 45 kDa which can be reduced to a band migrating at 38 kDa when deglycosylated (Figures 5.9D and 5.10B). On the basis of the analyses presented above, it can be deduced that this band represents a ZIP10 ectodomain (N_ecto) fragment in which only one N- glycan acceptor site is occupied. Taken together, these experiments (i) established that α-site cleavages in PrP can be triggered by transition metal starvation; and (ii) identified a subset of ZIP10 expression products in prion-infected mice to bear a molecular signature of cellular transition metal starvation.

144

Figure 5.10: The novel ZIP10 reactive band observed in prion-infected mice phenocopies ZIP10 ectodomain shedding in response to zinc or manganese depletion and is distinct from constitutive ZIP10 ectodomain shedding. A comparison of PrP and ZIP10 cleavage products observed upon chelation of transition metals from the medium versus RML inoculation is depicted. (A) Both transition metal chelation and prion infection led to PrPC cleavage products which could be detected with a PrP-directed antibody (Sha31) that binds to the respective C- terminal PrPC stubs retained on the plasma membrane. Upon PNGase F digestion, these PrPC cleavage products were revealed to be distinct. Whereas the PrPC C1 fragment appeared to have been accumulated in transition metal depleted cells (identified by its characteristic deglycosylated apparent molecular weight; please note that a small amount of partially deglycosylated C1 migrating slightly above can also be seen), the PrPC C2 fragment accumulated in prion infected cells or brain (depicted in composite Western blots). (B) In contrast, transition metal depletion or prion infection caused the appearance of a ZIP10 endoproteolytic cleavage product of the same apparent molecular weight. On the basis of data in Figure 5.9, this cleavage product was identified as the ZIP10 N_ecto cleavage product carrying one N-glycan. (C) Schematic drawing summarizing epitopes of antibody binding sites relative to endoproteolytic cleavages and N-glycosylation sites observed in this study. (D) Cartoon summarizing a mechanistic model consistent with data presented herein. According to this model, (i) wild-type ZIP10 undergoes complex maturation processes which lead to the attachment of up to four N- glycans to its ectodomain but not the channel domain. The ectodomain may make contact with the conserved CHELPHEL sequence embedded in the putative TM domain 5 of LZT ZIP transporters. (ii) Disruption of this ectodomain-TMD5 interface by the introduction of point mutants which eliminate a transition metal coordination site or by depletion of Mn2+ or Zn2+ from the extracellular milieu, causes shedding of the ZIP10 ectodomain (N_ecto). (iii) Whereas this shedding event follows ZIP10 maturation (as evidenced by the presence of 2-4 N-glycans on the N_ecto cleavage product) in cells expressing the ZIP10-(H711A/H715A) mutant, endoproteolysis at the same (or a very nearby site) occurs on immature ZIP10 (as evidenced by

145 the presence of only a single N-glycan) in Mn2+ or Zn2+ depleted medium or in mouse brains inoculated with RML prions.

5.4 Discussion

Data presented in this manuscript constitute the first steps towards elucidating the functional significance of the evolutionary relationship between the prion gene and the gene family of ZIP metal ion transporters [1, 2, 370]. Because the primary motivation to conduct this study was to explore whether members of the LZT subfamily of ZIP transporters play a role in prion disease, an observation of abnormal LZT-derived expression products in mice challenged with prion disease served as a starting point for this project. During the course of this work interesting facets of a hitherto largely unknown biology of LZTs were uncovered, including: (i) the broad expression of many LZTs in N2a cells, (ii) the predominant localization of ZIP10 to the cell surface, (iii) the demonstration of occupancy of up to four N-glycosylation acceptor sites within the ectodomain of ZIP10, (iv) the approximate location of scissile bonds flanking the CFC sequence segment within ZIP10, (v) the finding that zinc or manganese but not copper starvation, and not overabundance of transition metals, causes shedding of the ectodomain of immature ZIP10 into the medium, (vi) evidence that the replacement of conserved histidine residues within TMD5 but not the addition of a C-terminal HA tag interferes with the activity of ZIP10, and (vii) the observation that both the growing of cells in medium starved of transition metals or the expression of the ZIP10 construct harboring TMD5 histidine replacement mutations cause cleavage of the same ZIP10 scissile bond, with the consequence that the bulk of ZIP10 PrP-like ectodomains are released into the medium (Figures 5.10C and D). Taken together, these data combine to a model whereby cells grown in standard medium express primarily full-length ZIP10 which is gradually N-glycosylated, reaching a maximum occupancy when all four NxT ectodomain acceptor sites carry N-glycans. Expression levels of ZIP10 are posttranslationally controlled by the cell, possibly to avoid toxic intracellular transition metal levels from accumulating through ZIP10-dependent passive diffusion (it should be noted that regulation at the transcriptional level seems implausible because ZIP10-HA or ZIP10-HA-(H711A/715A) were shown in this work to express at very different levels despite their transcription being driven from identical heterologous expression plasmids) (Figure 5.9C). Removal of a highly conserved zinc coordination center in TMD5 generates a non-gating ZIP10 molecule which is no

146 longer subject to the same posttranslational degradation processes and therefore accumulates to higher protein levels. A higher N-glycosylation level and increased ectodomain shedding of this ZIP10 mutant suggests that the ZIP10 ectodomain which contains both the four NxT acceptor sites and the scissile bond for endoproteolysis might engage in a transition metal-dependent interaction with TMD5 in wild-type ZIP10 but not in ZIP10-(H711A/H715A). Transmembrane mutants of this kind are often observed to accumulate in early compartments of the secretory pathway through non-productive aggregation pathways [371, 372], a phenomenon well- documented for the cystic fibrosis transmembrane conductance regulator (CFTR) [373]. The ZIP10 TMD5-mutant’s increased N-glycosylation relative to wild-type ZIP10 and lack of observations indicative of its aggregation suggest that the protein is not predominantly retained or aggregated in the ER. In addition, our recent data document the ZIP10-(H711A/H715A) mutant to retain activity as an importer of extracellular zinc (unpublished), consistent with the interpretation that at least a subpool of its expression products reach the plasma membrane. Planned quantitative immunolocalization studies will resolve this question and will help to refine our understanding about its trafficking characteristics relative to wild-type ZIP10. When the cell culture medium is starved of transition metals, the first line of response by the cells is to slow the degradation of full-length N-glycosylated ZIP10 which leads to a stronger signal not only for the corresponding Western blot signal derived from endogenous ZIP10, but also causes the heterologous expression of wild-type ZIP10-HA to proceed to higher steady-state levels of the mature ZIP10 protein. The low transition metal influx which is likely experienced by cells starved of manganese or zinc further causes newly-made ZIP10 ectodomains to be shed at a time when only one N-glycan has been transferred to its four eligible NxT acceptor sites. The i.c. challenge of mice with mouse-adapted RML prions causes a subpopulation of ZIP10 to acquire the molecular characteristics of ZIP10 molecules grown in cell medium starved of transition metals.

With the current data it cannot be resolved whether the shedding of ZIP10 ectodomains and N-terminal PrPC fragments constitute independent responses to the same extracellular stimulus or rely on more immediate crosstalk between LZTs and PrPC. In particular, the transition metal starvation data presented in this work could either represent the manifestation of independent cellular pathways which share a reliance on a transition metal starvation stimulus, or may reflect direct crosstalk and a shared molecular biology of LZTs and PrPC. Favoring the latter

147 interpretation are observations which documented that: (i) ZIPs and PrPC can co- immunoprecipitate, demonstrated in work which preceded this study for ZIP6/ZIP10 [102], (ii) RML-infection of mice causes the abnormal appearance of ZIP10-derived products as early as 84 days following i.c. challenge, i.e., at a time during disease progression when other molecular defects are not yet pronounced and no overt symptoms can be detected in mice, and (iii) a striking correlation exists in the response of ZIP10 or PrPC to transition metal starvation. This correlation not only pertains to the timing of cleavages and the strength of the response, but also extends to the identity of transition metals which, when withdrawn, caused (manganese or zinc) or failed to cause (copper) the shedding of N-terminal fragments from ZIP10 or PrPC.

5.4.1 Functional significance of ZIP10 ectodomain shedding

The function of what we had termed the PrP-like ectodomain within ZIP10 had not been specifically investigated prior to this study. However, data on the not-so-distant LZT protein family member ZIP4 suggested that its ectodomain may serve a role as a negative regulator or sensor of transition metal uptake. It had also been observed that under conditions of transition metal starvation, the ZIP4 ectodomain can be shed to overcome the inhibitory influence it exerts on cation uptake [194]. Data generated in this work support this view and suggest that the proposed inhibitory role may be a more common feature of LZT ectodomains in general. If the ectodomains of LZTs indeed serve an inhibitory function toward the uptake of transition metals, it then follows that the prion founder gene itself, which was evolutionarily descended from an ectodomain segment of an ancestral LZT gene, might have had a biological function in the sensing and negative regulation of transition metal uptake. Is there any hint in the data presented here that could substantiate the possibility that contemporary prion proteins have retained facets of such an inhibitory function? One of the most striking observations made in this study was the diametrical nature in which full-length N-glycosylated ZIP10 and PrPC signals responded to starving the cell culture medium of GT1 cells of transition metals (Figure 5.8E). One way to interpret these observations would be to assume that both the increase in steady-state N- glycosylated full-length ZIP10 levels and the dramatic decrease in mature PrPC observed under these conditions served a common objective, i.e., to respond to the shortage of transition metals in the medium by adjusting the cellular proteome in ways that facilitate transition metal uptake. It can then be deduced that the cellular expression of PrPC might indeed represent a hindrance to

148 transition metal uptake. Enthusiasm for such a model might be tempered by a body of literature which documents preferential binding of mammalian PrPC to copper, and only to a much lesser extent to zinc or manganese. If the ability of PrPC to coordinate transition metals with the known binding affinities had any connection to its putative inhibitory function, then it would seem counterintuitive that starving cells of manganese or zinc but not copper leads to a diminution of PrPC levels. One way to reconcile these seemingly paradoxical observations is to assume that the cell’s response to zinc or manganese starvation with regard to PrPC levels constitutes a remnant from an earlier evolutionary time. According to this model, the transition metal binding profile of ancient prion proteins was dominated by zinc coordination sites. Only gradually, and most pronounced in the tetrapod lineage, prion protein sequences diverged in ways that led to a bias for copper binding. In support of this model, a subset of prion proteins coded for by prion gene orthologs in contemporary fish species still display strong affinity for zinc [223]. There is also the possibility that PrPC may exert an inhibitory effect on zinc or manganese uptake independent of its own transition metal binding profile but on the basis of its ability to bind to a subset of LZTs which themselves may play a role in zinc or manganese uptake. The demonstration that PrPC co-localizes and can be immunoprecipitated with a subset of LZTs, together with the observation that highly conserved amino acids within the putative transmembrane domain of LZTs (shown in this work for ZIP10) can influence cleavages and N-glycosylation of their PrP- like ectodomain, strengthen this scenario. It is noteworthy that it is not necessary to invoke that PrPC represents a critical functional interactor of ZIPs, as it would be sufficient for contemporary PrPC molecules to have retained a rudimentary ability to engage in contacts with the transmembrane domains of a subset of LZTs. Similarly, the α-site cleavage of PrPC might in this scenario not constitute a ‘useful’ endoproteolysis event in the same sense in which the shedding of the ZIP10 ectodomain from the C-terminal ZIP10 TMD would be expected to open the passive diffusion channel of this ion transporter for the import of zinc or manganese. Instead, the cleavage of PrPC at the α-site scissile bond may represent a bystander effect inadvertently directed at PrPC molecules which, due to their affinity for a subset of LZTs, can be found enriched within their molecular environment. For this to be a plausible course of events, PrPC would merely need to have kept enough of its original LZT ectodomain features to still serve as a substrate for an endoprotease which gets recruited to LZTs to rid them of their inhibitory ectodomains whenever cells are starved of zinc or manganese. Interestingly, a similar course of

149 events might also underlie the posttranslational degradation of the prion family member Shadoo, observed in various strains of mice during the progression of prion disease [374]. The cellular compartment in which the LZT ectodomain shedding event occurs is currently not known. With regard to ZIP4 ectodomain shedding, it had been proposed that recycling endosomes may host this cleavage event. Data presented here now suggest that at least for ZIP10, the shedding of its ectodomain in response to transition metal starvation may selectively affect immature ZIP10 ectodomains. It was further demonstrated that the appearance of this band supersedes the generation of the mature ectodomain fragment observed when ZIP10-HA-(H711A/H715A) is expressed. Taken together, the most parsimonious explanation for these observations appears to be that the shedding of the ZIP10 ectodomain occurs while ZIP10 traverses the secretory pathway. This timing would fit well with recent data which document that cleavage of PrPC at its α-site might also occur during maturation within the secretory pathway [185].

5.4.2 Alpha-site cleavage of PrPC

If indeed the endoproteolysis of ZIP10 and PrPC observed in response to manganese or zinc starvation was the work of the same endoproteolytic activity, why would this shedding cause the release of the entire ectodomain in ZIP10 versus the release of only the α-site N-proximal fragment N1 from PrPC? One angle from which to explore this question might be to compare the minimal requirements for this cleavage. Whereas previous data are missing for ZIP10 in this regard, it was recently shown that α-site cleavage of PrPC requires the integrity of the hydrophobic domain which is flanking this cleavage site on the C-terminal site of the scissile bond (mapped to murine residues 109/110) [368]. More specifically, cleavage at the α-site was size-dependently impaired by deletions of the hydrophobic domain spanning residues 106-119 but was surprisingly tolerant toward changes within the amino acid sequence. This description is consistent with substrate determinants of sheddases which target cleavage sites N-proximal to transmembrane domains. Thus, from the perspective of a sheddase dispatched to shed the ectodomain of LZTs in response to transition metal starvation, the cleavage of a nearby PrPC molecule at its α-site might not seem so different. Although it has been shown that protein kinase C agonists can stimulate α-site cleavage [179], most reports refer to the cleavage of PrPC at this site as the ‘normal’ or constitutive cleavage [185]. The robust stimulation of PrPC cleavage at this site in response to manganese or zinc starvation observed in this study, therefore, came as a

150 surprise, particularly in light of a literature which seemed to have left no stone unturned in efforts to understand the transition metal biology of PrPC. Considerable interest in the proteolytic activity responsible for α-site cleavage arises from the critical role ascribed to the hydrophobic domain for the disease-associated conversion of the prion protein. The ability to manipulate cleavages at this site therefore may provide an angle for disease intervention strategies. From the data presented in this work, it cannot currently be deduced whether the ZIP10 ectodomain cleavage observed in response to the challenge of mice with mouse-adjusted RML prions indicates that transition metal starvation exists in these mice or whether a molecular phenocopy of transition metal starvation is observed. Work by other researchers which attempted to address whether transition metal imbalances exist in prion-infected mice has been largely inconclusive or at times led to contradictory results. Recent improvements to imaging methods which now allow interrogation of the neuroanatomy of metal distributions in the brain attest to a surprising heterogeneity in the distribution of transition metals and suggest that PrPC may indeed contribute to this biology [375].

Finally, as additional molecular biological tools, which are currently limiting, are developed for mammalian ZIP proteins, it would be essential to compare and complement the results obtained here for ZIP10 vis-à-vis its relationship to PrP biology with those for other LZT members, particularly ZIP6 and ZIP5.

5.5 Conclusions

Data presented in this study provided multiple novel angles from which to explore the biology of the prion protein in health and disease, but also raised many questions which need to be answered. More work will be needed to (i) understand whether a new look at the transition metal biology of the prion protein, with emphasis on its relationship to LZTs, may provide avenues for therapeutic intervention, (ii) determine whether binding of prion proteins to LZTs contributes to differences in posttranslational signatures of the prion protein underlying well-established strain phenomena in prion disease, and (iii) explore whether the LZT expression profiles of cells or the ectodomain shedding of LZTs contribute to a poorly-understood biology surrounding the susceptibility of cells to prion infection or the spread of prion disease.

151

Chapter 6 Future Directions of the ZIP-Prion Connection

Please note that all parts of this chapter were published in the following article [370]: Sepehr Ehsani, Mohadeseh Mehrabian, Cosmin L. Pocanschi, Gerold Schmitt-Ulms (2012) The ZIP-Prion connection. Prion 6(4):317‐21.

Candidate’s role: Co-wrote manuscript

Summary: This chapter provides a discussion on the possible significance of the reunion of ZIP and prion gene subfamilies for understanding the biology of the prion protein in health and disease. The emphasis is to comment on experimental paradigms, ongoing developments and challenges.

It may be appropriate to start a discussion on future directions in this research area with the question of what significance, if any, does the discovery of the evolutionary origins of the prion protein have? In the prion literature one can find references to ‘kryptonite’, ‘the Borg’, and ‘martians’ [376, 377]. Continuing on the science fiction theme, we would like to use the analogy of an alien, newly arrived on Earth, who finds a doorknob on the ground. It is apparent that no amount of isolated investigations of this doorknob would reveal the fact that it is meant to be attached to a door, nor would it allow the alien to discover what door it used to be attached to, and why. In this analogy the prion protein is, so to speak, descended from an ancient ‘doorknob’. The connection to ZIP proteins links the prion protein ‘doorknob’ to a whole family of ‘intact doors’ (the ZIP transporters) which have retained their doorknobs (the N-terminal extracellular domains of ZIP transporters). Finding the doors should be highly useful for this extraterrestrial observer intent on understanding the doorknob. It provides the ability to study the mechanisms of attached doorknobs and explore the spaces in front of and behind the doors they separate (e.g., upstream and downstream signaling pathways). As time passes, it may emerge that the detached doorknob (prion protein) no longer functions quite like the attached doorknobs (N-terminal domain of ZIP transporters). However, the fact that PrP co-purified with ZIPs 6 and 10 already

152

suggests that the prion protein may at least have retained some of its original ‘doorknob’ properties [102] (Chapter 2).

At first glance it may seem counterintuitive that proteins encoded by genes which are linked through evolution in the way proposed for LZTs and prion genes would bind to each other. However, it is not difficult to see how a physical encounter of the gene products of LZTs and the prion founder gene would have occurred. Firstly, given that LZTs appear to be expressed in every eukaryotic cell within the metazoan lineage, so long as the retroinsertion of the prion founder gene occurred downstream from any active promoter, the premise that an LZT was co- translated within the same cell (a minimum requirement for a physical encounter) was fulfilled. It should be noted that the prion founder gene retained both an N-terminal signal sequence required for directing the nascent protein to the secretory pathway, and a means by which it could be inserted into the lipid bilayer (a GPI anchor), preserving its orientation and relative distance to the membrane [2] (Chapter 3). Moreover, because the prion founder retrogene would have initially coded for an expression product identical to the ectodomain of its LZT ancestor (Chapter 4), it can be envisaged how these shared attributes would have directed both gene products to the same molecular environment, raising the possibility that they may interact with an overlapping set of binding partners. Our preliminary investigation of the localization of ZIPs has revealed that a small subpopulation within the cellular pool of LZTs may indeed associate with raft-like membrane structures which can be purified on the basis of their buoyancy during sucrose gradient centrifugations and which are known to be strongly enriched in cholesterol, sphingolipids and a subset of GPI-anchored membrane proteins, including PrP (unpublished data).

A direct interaction between members of the prion protein family and LZTs can also seem plausible if one assumes a scenario in which the LZT ancestor of the prion founder gene would have assembled into dimers or oligomers to fulfill its function as a cation importer. Thus, one can deduce that the ability of LZTs to assemble into functional dimers or oligomers could have translated into an ability of LZT predecessor and PrP founder gene products to bind to each other, if elements of the LZT ectodomain were contributing to the assembly of dimer or oligomer structures. To date, not much is known about the molecular organization of a majority of ZIP transporters, including the subbranch of LZTs comprising ZIPs 5, 6 and 10 from which the prion founder gene was descended. However, recent observations made with knockdown and metal

153 chelation experiments, which queried the expression, stability and maturation of a subset of LZTs, were suggestive of a functional heterodimerization involving ZIP6 and ZIP10 (unpublished data). The aforementioned scenario further gains strength from the observation that ZIP13, a more distant LZT member of the ZIP family, appears to assemble into functional dimers [378]. In light of the cumulative sequence divergence which can be observed amongst contemporary members of the prion gene subfamily and LZTs, the physiological significance of their shared origins has likely been fading with evolutionary time. However, if confirmed, the aforementioned model may not only serve as an explanation for weak interactions observed between contemporary prion proteins and LZTs, but may also further our understanding of why members of the mammalian prion protein family – PrP, Sho and Doppel – can be shown to interact [102, 379].

It is to be expected that more conclusive insights into this aspect of ZIP protein biology will emerge from detailed structural investigations of LZTs. To date, no high-resolution structural model exists for any ZIP protein. This contrasts an abundance of NMR and X-ray data available for a large range of prion protein orthologs and their derivatives, including recombinant constructs featuring point mutations or larger deletions [167, 380]. Incidentally, however, no PrP ortholog of fish has so far yielded a structure, despite considerable similarity and an undisputed phylogenetic linkage of fish and mammalian prion protein sequences [27, 381]. The failure to obtain a high-resolution structure may be a reflection of fish prion proteins differing in some way from their mammalian cousins that renders them refractory to tried-and-true in vitro folding strategies [382]. This piece of information is relevant and might serve as a harbinger indicating that attempts to solve the structure of the PrP-like ectodomains of LZTs may not be a trivial undertaking, because the branch of LZTs which aligns best to contemporary prion sequences exhibits the closest sequence similarity to fish rather than mammalian prion proteins [2] (Chapter 3).

If LZT ectodomains do indeed acquire a PrP-like fold, then it will be valuable to investigate the purpose of this fold in relation to their cation channel activity. Our recent work presented in Chapter 5 suggests that a subset of LZT ectodomains can be shed into the medium in response to divalent cation starvation [356]. These observations are consistent with earlier data indicating that a rare zinc deficiency disease, acrodermatitis enteropathica, characterized by an inability of affected individuals to absorb zinc when dietary levels of this metal are limited, is

154

linked to mutations in the gene coding for ZIP4, which appear to interfere with the zinc starvation-induced ectodomain shedding of the ZIP4 cellular cation import channel [194]. Taken together, these data suggest that the PrP-like ectodomain present in a subset of LZTs serves a function as a negative regulator of divalent cation uptake [1]. There is robust evidence for the ability of the prion protein to coordinate divalent cations, with copper exhibiting the strongest binding, and levels of available cations influencing the coordination geometry observed [210]. Attempts to uncover the functional significance of the metal binding characteristics of the prion protein have been less successful. Our preliminary observations suggest that the prion protein may, similar to LZTs, undergo endoproteolysis in response to divalent metal starvation (Chapter 5). More specifically, when N2a mouse neuroblastoma cells are grown in medium starved of transition metals for a prolonged time period, a diminution of full-length PrPC levels can be observed. Strikingly, this effect can be rescued by the addition of zinc but not copper [356]. In light of the evolutionary relationship of ZIPs and PrP, a research direction to pursue could be the determination of whether this zinc starvation-induced cleavage of PrP (i) occurs in the context of its ability to interact with LZTs and, (ii) similar to LZT ectodomain shedding, serves a function in cellular cation homeostasis.

Due to the molecular complexity and heterogeneity of the LIV-1 subfamily of ZIP transporters, investigations into the functional relationship of LZTs and prion proteins may prove to be far from straightforward. To date, knockout studies of the most closely related LZTs are not available; yet, if work on other ZIPs is indicative of the results to be expected, it is likely that crosstalk, functional redundancies, spatial overlap and complex posttranslational modifications will complicate their investigation [91]. That this is not merely a theoretical scenario is evident from our observation that N2a cells actively transcribe at least five out of six LZTs which were probed [356] (Chapter 5). To understand the functional relationship of LZTs and the prion protein, we have begun to investigate the phylogenetic subbranch most closely related to PrP comprising ZIPs 5, 6 and 10. Interestingly, when expressed at similar levels and with identical C- terminal tags in neuroblastoma cells, the degree to which members within this LZT subbranch co-localize with the endogenous prion protein differs, with heterologously expressed ZIP5 demonstrating the most pronounced overlap in its cellular distribution (unpublished data). Assuming that this co-localization is reflective of direct binding of the prion protein to LZTs, it may not only raise the possibility to map the molecular interface between these proteins but may

155

also help to investigate whether the interaction alters the divalent cation uptake biology of the respective LZTs. If so, one may find that despite PrP’s own preference for binding copper over zinc, its physiological cation biology may bear the cumulative signature of the ZIP transporters it interacts with, and as such may primarily affect zinc rather than copper influx to the cell.

From a prion research perspective, the most pressing question remains whether LZTs play a role in the molecular etiology underlying prion diseases. On the basis of data which did not detect an influence of the brain-expressed PrP paralog Sho on PrP conversion [374, 383] (despite a selective and progressive reduction of Sho protein levels in prion disease [143]), one may anticipate that other members of the prion/ZIP gene family similarly play no critical role in prion disease. However, given that Sho lacks both the cysteine-flanked core and C-terminal channel domains present in LZTs, the Sho data available to date may not be a suitable proxy for predicting whether LZTs play a role in the disease-associated conversion and/or the cell-to-cell spread of prion disease. It will be interesting to investigate whether there is an aspect to the physiological function of the PrP-like ectodomain of LZTs that may rely on an intrinsic ability to populate alternative folding states. In light of the likely function of the ectodomain in the gating and regulation of LZTs, it is, for example, conceivable that this domain can acquire alternative folds as part of an adaptive response to altering cation levels, similar to an observation made for the disordered domain within the prion protein [384]. If so, the study of ZIPs may help to define the cellular triggers and molecular constraints which promote conformational changes, which may in turn further our understanding of the profound changes in the physicochemical properties of PrP in prion diseases. Studies which employ hybrid PrP-LZT constructs may represent an experimental approach that would be informative in this context.

Recent biochemical data suggest that both the occurrence of prion strains and the species barrier phenomenon observed in prion diseases may relate to a conformational selection process which can be simulated with defined components in vitro [385-387]. It is likely that this selection process would be influenced in vivo by PrP’s physiological milieu and its protein-protein interactions. In fact, it has repeatedly been speculated that other cellular factors which bind to the prion protein or reside in its proximity during its passage through the secretory pathway may contribute to these phenomena in vivo. In particular, the endoproteolytically-shed PrP-like LZT ectodomains may be plausible candidates in this context and could also act as possible modulators of prion disease spread; not only because they contain exclusively structural elements

156

present in PrPC, but also because their lack of transmembrane domains would be expected to confer added mobility.

Finally, if PrPC modulates or contributes to a cation biology dictated by LZTs, changes to its expression level and/or fold, as seen in prion disease, would be expected to affect such a physiological role. This effect could manifest either as (i) an ablation of such a function, for example if PrPSc would no longer be able to interact with LZTs, or as (ii) a poisoning, if PrPSc’s interaction with LZTs was retained but acquired a non-physiological quality. Needless to say that both scenarios could contribute to the symptoms of prion diseases and their impact would be expected to correlate with the severity of the underlying physiological perturbation.

157

References

1. Ehsani, S., et al. Family reunion--the ZIP/prion gene family. Progress in neurobiology, 2011. 93(3): p. 405-20.

2. Schmitt-Ulms, G., et al. Evolutionary descent of prion genes from the ZIP family of metal ion transporters. PLoS One, 2009. 4(9): p. e7208.

3. Eide, D., et al. A novel iron-regulated metal transporter from plants identified by functional expression in yeast. Proc Natl Acad Sci U S A, 1996. 93(11): p. 5624-8.

4. Zhao, H. and D. Eide. The yeast ZRT1 gene encodes the zinc transporter protein of a high-affinity uptake system induced by zinc limitation. Proc Natl Acad Sci U S A, 1996. 93(6): p. 2454-8.

5. Zhao, H. and D. Eide. The ZRT2 gene encodes the low affinity zinc transporter in Saccharomyces cerevisiae. J Biol Chem, 1996. 271(38): p. 23203-10.

6. Manning, D.L., et al. Oestrogen-regulated genes in breast cancer: association of pLIV1 with lymph node involvement. Eur J Cancer, 1994. 30A(5): p. 675-8.

7. Masiakowski, P., et al. Cloning of cDNA sequences of hormone-regulated genes from the MCF-7 human breast cancer cell line. Nucleic Acids Res, 1982. 10(24): p. 7895-903.

8. Aguzzi, A., F. Baumann, and J. Bremer. The Prion's Elusive Reason for Being. Annu Rev Neurosci, 2008. 31: p. 439-477.

9. Prusiner, S.B. Prions. Proc Natl Acad Sci U S A, 1998. 117: p. 421-434.

10. Taylor, K.M., et al. The emerging role of the LIV-1 subfamily of zinc transporters in breast cancer. Mol Med, 2007. 13(7-8): p. 396-406.

11. Taylor, K.M. and R.I. Nicholson. The LZT proteins; the LIV-1 subfamily of zinc transporters. Biochim Biophys Acta, 2003. 1611(1-2): p. 16-30.

12. Kury, S., et al. Identification of SLC39A4, a gene involved in acrodermatitis enteropathica. Nat Genet, 2002. 31(3): p. 239-40.

13. Wang, K., et al. A novel member of a zinc transporter family is defective in acrodermatitis enteropathica. Am J Hum Genet, 2002. 71(1): p. 66-73.

14. Moynahan, E.J. Letter: Acrodermatitis enteropathica: a lethal inherited human zinc- deficiency disorder. Lancet, 1974. 2(7877): p. 399-400.

15. Gloriam, D.E., R. Fredriksson, and H.B. Schioth. The G protein-coupled receptor subset of the rat genome. BMC Genomics, 2007. 8: p. 338.

158

16. Povey, S., et al. The HUGO Gene Nomenclature Committee (HGNC). Hum Genet, 2001. 109(6): p. 678-80.

17. Hediger, M.A., et al. The ABCs of solute carriers: physiological, pathological and therapeutic implications of human membrane transport proteinsIntroduction. Pflugers Arch, 2004. 447(5): p. 465-8.

18. Fredriksson, R., et al. The solute carrier (SLC) complement of the human genome: phylogenetic classification reveals four major families. FEBS Lett, 2008. 582(27): p. 3811-6.

19. Schlessinger, A., et al. Comparison of human solute carriers. Protein Sci, 2010. 19(3): p. 412-28.

20. Guerinot, M.L. The ZIP family of metal transporters. Biochim Biophys Acta, 2000. 1465(1-2): p. 190-8.

21. Gaither, L.A. and D.J. Eide. Eukaryotic zinc transporters and their regulation. Biometals, 2001. 14(3-4): p. 251-70.

22. Feeney, G.P., et al. The phylogeny of teleost ZIP and ZnT zinc transporters and their tissue specific expression and response to zinc in zebrafish. Biochim Biophys Acta, 2005. 1732(1-3): p. 88-95.

23. Premzl, M., et al. Shadoo, a new protein highly conserved from fish to mammals and with similarity to prion protein. Gene, 2003. 314: p. 89-102.

24. Rivera-Milla, E., et al. Disparate evolution of prion protein domains and the distinct origin of Doppel- and prion-related loci revealed by fish-to-mammal comparisons. Faseb J, 2006. 20(2): p. 317-9.

25. Rivera-Milla, E., C.A.O. Stuermer, and E. Malaga-Trillo. An evolutionary basis for scrapie disease: identification of a fish prion mRNA. Trends Genetics, 2003. 19: p. 72-75.

26. Cotto, E., et al. Molecular characterization, phylogenetic relationships, and developmental expression patterns of prion genes in zebrafish (Danio rerio). FEBS J, 2005. 272(2): p. 500-13.

27. Premzl, M. and V. Gamulin. Comparative genomic analysis of prion genes. BMC Genomics, 2007. 8: p. 1.

28. Premzl, M., et al. Evolution of vertebrate genes related to prion and shadoo proteins - clues from comparative genomic analysis. Mol Biol Evol, 2004. 21: p. 2210-2231.

29. Kim, Y., J. Lee, and C. Lee. In silico comparative analysis of DNA and amino acid sequences for prion protein gene. Transbound Emerg Dis, 2008. 55(2): p. 105-14.

30. Brown, D.R. Brain proteins that mind metals: a neurodegenerative perspective. Dalton Trans, 2009(21): p. 4069-76.

159

31. Caughey, B., R.E. Race, and B. Chesebro. Detection of prion protein mRNA in normal and scrapie-infected tissues and cell lines. J Gen Virol, 1988. 69 ( Pt 3): p. 711-6.

32. Collinge, J. Prion diseases of humans and animals: their causes and molecular basis. Annu Rev Neurosci, 2001. 24: p. 519-50.

33. Collinge, J. and A.R. Clarke. A general model of prion strains and their pathogenicity. Science, 2007. 318(5852): p. 930-6.

34. Hu, W., et al. Prion proteins: physiological functions and role in neurological disorders. J Neurol Sci, 2008. 264(1-2): p. 1-8.

35. Jeong, B.H., et al. Polymorphism at 3' UTR +28 of the prion-like protein gene is associated with sporadic Creutzfeldt-Jakob disease. Eur J Hum Genet, 2005. 13(9): p. 1094-7.

36. Makrinou, E., J. Collinge, and M. Antoniou. Genomic characterization of the human prion protein (PrP) gene locus. Mamm Genome, 2002. 13(12): p. 696-703.

37. Mead, S., et al. Examination of the human prion protein-like gene doppel for genetic susceptibility to sporadic and variant Creutzfeldt-Jakob disease. Neurosci Lett, 2000. 290(2): p. 117-20.

38. Peoc'h, K., et al. First report of polymorphisms in the prion-like protein gene (PRND): implications for human prion diseases. Neurosci Lett, 2000. 286(2): p. 144-8.

39. Uboldi, C., et al. Cloning of the bovine prion-like Shadoo (SPRN) gene by comparative analysis of the predicted genomic locus. Mamm Genome, 2006. 17(11): p. 1130-9.

40. Watts, J.C. and D. Westaway. The prion protein family: diversity, rivalry, and dysfunction. Biochim Biophys Acta, 2007. 1772: p. 654-672.

41. Cousins, R.J., J.P. Liuzzi, and L.A. Lichten. Mammalian zinc transport, trafficking, and signals. J Biol Chem, 2006. 281(34): p. 24085-9.

42. Eide, D.J. The SLC39 family of metal ion transporters. Pflugers Arch, 2004. 447(5): p. 796-800.

43. Eide, D.J. Zinc transporters and the cellular trafficking of zinc. Biochim Biophys Acta, 2006. 1763(7): p. 711-22.

44. Kambe, T., et al. Overview of mammalian zinc transporters. Cell Mol Life Sci, 2004. 61(1): p. 49-68.

45. Liuzzi, J.P. and R.J. Cousins. Mammalian zinc transporters. Annu Rev Nutr, 2004. 24: p. 151-72.

160

46. Mocchegiani, E., R. Giacconi, and M. Malavolta. Zinc signalling and subcellular distribution: emerging targets in type 2 diabetes. Trends Mol Med, 2008. 14(10): p. 419- 28.

47. Franklin, R.B., et al. hZIP1 zinc uptake transporter down regulation and zinc depletion in prostate cancer. Mol Cancer, 2005. 4: p. 32.

48. Kambe, T., et al. Slc39a1 to 3 (subfamily II) Zip genes in mice have unique cell-specific functions during adaptation to zinc deficiency. Am J Physiol Regul Integr Comp Physiol, 2008. 294(5): p. R1474-81.

49. Desouki, M.M., et al. hZip2 and hZip3 zinc transporters are down regulated in human prostate adenocarcinomatous glands. Mol Cancer, 2007. 6: p. 37.

50. Peters, J.L., et al. Targeting of the mouse Slc39a2 (Zip2) gene reveals highly cell-specific patterns of expression, and unique functions in zinc, iron, and calcium homeostasis. Genesis, 2007. 45(6): p. 339-52.

51. Mathews, W.R., et al. Drosophila fear of intimacy encodes a Zrt/IRT-like protein (ZIP) family zinc transporter functionally related to mammalian ZIP proteins. J Biol Chem, 2005. 280(1): p. 787-95.

52. Weaver, B.P., et al. Zip4 (Slc39a4) expression is activated in hepatocellular carcinomas and functions to repress apoptosis, enhance cell cycle and increase migration. PLoS One, 2010. 5(10).

53. Li, M., et al. Down-regulation of ZIP4 by RNA interference inhibits pancreatic cancer growth and increases the survival of nude mice with pancreatic cancer xenografts. Clin Cancer Res, 2009. 15(19): p. 5993-6001.

54. Li, M., et al. Aberrant expression of zinc transporter ZIP4 (SLC39A4) significantly contributes to human pancreatic cancer pathogenesis and progression. Proc Natl Acad Sci U S A, 2007. 104(47): p. 18636-41.

55. Zhang, Y., et al. ZIP4 regulates pancreatic cancer cell growth by activating IL-6/STAT3 pathway through zinc finger transcription factor CREB. Clin Cancer Res, 2010. 16(5): p. 1423-30.

56. Zhang, Y., et al. ZIP4 upregulates the expression of neuropilin-1, vascular endothelial growth factor, and matrix metalloproteases in pancreatic cancer cell lines and xenografts. Cancer Biol Ther, 2010. 9(3): p. 236-42.

57. Geiser, J., et al. A mouse model of acrodermatitis enteropathica: loss of intestine zinc transporter ZIP4 (Slc39a4) disrupts the stem cell niche and intestine integrity. PLoS Genet, 2012. 8(6): p. e1002766.

58. Chowanadisai, W., S.L. Kelleher, and B. Lonnerdal. Zinc deficiency is associated with increased brain zinc import and LIV-1 expression and decreased ZnT-1 expression in neonatal rats. J Nutr, 2005. 135(5): p. 1002-7.

161

59. Taylor, K.M. A distinct role in breast cancer for two LIV-1 family zinc transporters. Biochem Soc Trans, 2008. 36(Pt 6): p. 1247-51.

60. Hogstrand, C., et al. Zinc transporters and cancer: a potential role for ZIP7 as a hub for tyrosine kinase activation. Trends Mol Med, 2009. 15(3): p. 101-11.

61. Waterworth, D.M., et al. Genetic variants influencing circulating lipid levels and risk of coronary artery disease. Arterioscler Thromb Vasc Biol, 2010. 30(11): p. 2264-76.

62. Dalton, T.P., et al. Identification of mouse SLC39A8 as the transporter responsible for cadmium-induced toxicity in the testis. Proc Natl Acad Sci U S A, 2005. 102(9): p. 3401- 6.

63. Besecker, B., et al. The human zinc transporter SLC39A8 (Zip8) is critical in zinc- mediated cytoprotection in lung epithelia. Am J Physiol Lung Cell Mol Physiol, 2008. 294(6): p. L1127-36.

64. Galvez-Peralta, M., et al. ZIP8 zinc transporter: indispensable role for both multiple- organ organogenesis and hematopoiesis in utero. PLoS One, 2012. 7(5): p. e36055.

65. Matsuura, W., et al. SLC39A9 (ZIP9) regulates zinc homeostasis in the secretory pathway: characterization of the ZIP subfamily I protein in vertebrate cells. Biosci Biotechnol Biochem, 2009. 73(5): p. 1142-8.

66. Kaler, P. and R. Prasad. Molecular cloning and functional characterization of novel zinc transporter rZip10 (Slc39a10) involved in zinc uptake across rat renal brush-border membrane. Am J Physiol Renal Physiol, 2007. 292(1): p. F217-29.

67. Kagara, N., et al. Zinc and its transporter ZIP10 are involved in invasive behavior of breast cancer cells. Cancer Sci, 2007. 98(5): p. 692-7.

68. Fukada, T., et al. The zinc transporter SLC39A13/ZIP13 is required for connective tissue development; its involvement in BMP/TGF-beta signaling pathways. PLoS One, 2008. 3(11): p. e3642.

69. Liuzzi, J.P., et al. Interleukin-6 regulates the zinc transporter Zip14 in liver and contributes to the hypozincemia of the acute-phase response. Proc Natl Acad Sci U S A, 2005. 102(19): p. 6843-8.

70. Thorsen, K., et al. Alternative splicing of SLC39A14 in colorectal cancer is regulated by the Wnt pathway. Mol Cell Proteomics, 2011. 10(1): p. M110 002998.

71. Zhao, N., et al. ZRT/IRT-like protein 14 (ZIP14) promotes the cellular assimilation of iron from transferrin. J Biol Chem, 2010. 285(42): p. 32141-50.

72. He, L., et al. Discovery of ZIP transporters that participate in cadmium damage to testis and kidney. Toxicol Appl Pharmacol, 2009. 238(3): p. 250-7.

162

73. Aydemir, T.B., H.S. Sitren, and R.J. Cousins. The zinc transporter Zip14 influences c- Met phosphorylation and hepatocyte proliferation during liver regeneration in mice. Gastroenterology, 2012. 142(7): p. 1536-46 e5.

74. Hu, W., R.N. Rosenberg, and O. Stuve. Prion proteins: a biological role beyond prion diseases. Acta Neurol Scand, 2007. 116(2): p. 75-82.

75. Flechsig, E. and C. Weissmann. The role of PrP in health and disease. Curr Mol Med, 2004. 4(4): p. 337-53.

76. Brandner, S., et al. Neuroinvasion of prions: insights from mouse models. Exp Physiol, 2000. 85(6): p. 705-12.

77. Sakudo, A., et al. Recent advances in clarifying prion protein functions using knockout mice and derived cell lines. Mini Rev Med Chem, 2006. 6(5): p. 589-601.

78. Bremer, J., et al. Axonal prion protein is required for peripheral myelin maintenance. Nat Neurosci, 2010. 13(3): p. 310-8.

79. Le Pichon, C.E., et al. Olfactory behavior and physiology are disrupted in prion protein knockout mice. Nat Neurosci, 2009. 12(1): p. 60-9.

80. Steele, A.D., et al. Prion protein (PrPc) positively regulates neural precursor proliferation during developmental and adult mammalian neurogenesis. Proc Natl Acad Sci U S A, 2006. 103(9): p. 3416-21.

81. Zhang, C.C., et al. Prion protein is expressed on long-term repopulating hematopoietic stem cells and is important for their self-renewal. Proc Natl Acad Sci U S A, 2006. 103(7): p. 2184-9.

82. Fasano, C., V. Campana, and C. Zurzolo. Prions: protein only or something more? Overview of potential prion cofactors. J Mol Neurosci, 2006. 29(3): p. 195-214.

83. Lasmezas, C.I. Putative functions of PrP(C). Br Med Bull, 2003. 66: p. 61-70.

84. Vana, K., et al. Novel aspects of prions, their receptor molecules, and innovative approaches for TSE therapy. Cell Mol Neurobiol, 2007. 27(1): p. 107-28.

85. Gondo, Y., et al. Next-generation gene targeting in the mouse for functional genomics. BMB Rep, 2009. 42(6): p. 315-23.

86. Kambe, T., et al. Sequence similarity and functional relationship among eukaryotic ZIP and CDF transporters. Genomics Proteomics Bioinformatics, 2006. 4(1): p. 1-9.

87. Lin, S.J. and V.C. Culotta. Suppression of oxidative damage by Saccharomyces cerevisiae ATX2, which encodes a manganese-trafficking protein that localizes to Golgi- like vesicles. Mol Cell Biol, 1996. 16(11): p. 6303-12.

163

88. MacDiarmid, C.W., L.A. Gaither, and D. Eide. Zinc transporters that regulate vacuolar zinc storage in Saccharomyces cerevisiae. EMBO J, 2000. 19(12): p. 2845-55.

89. Kumanovics, A., et al. YKE4 (YIL023C) encodes a bidirectional zinc transporter in the endoplasmic reticulum of Saccharomyces cerevisiae. J Biol Chem, 2006. 281(32): p. 22566-74.

90. Mastrangelo, P. and D. Westaway. The prion gene complex encoding PrP(C) and Doppel: insights from mutational analysis. Gene, 2001. 275(1): p. 1-18.

91. Dufner-Beattie, J., et al. Mouse ZIP1 and ZIP3 genes together are essential for adaptation to dietary zinc deficiency during pregnancy. Genesis, 2006. 44(5): p. 239-51.

92. Dufner-Beattie, J., et al. The mouse acrodermatitis enteropathica gene Slc39a4 (Zip4) is essential for early development and heterozygosity causes hypersensitivity to zinc deficiency. Hum Mol Genet, 2007. 16(12): p. 1391-9.

93. Wang, B., et al. Generation of a Slc39a8 hypomorph mouse: markedly decreased ZIP8 Zn(2)(+)/(HCO(3)(-))(2) transporter expression. Biochem Biophys Res Commun, 2011. 410(2): p. 289-94.

94. Vazquez, F. and A. Grider. The effect of the acrodermatitis enteropathica mutation on zinc uptake in human fibroblasts. Biol Trace Elem Res, 1995. 50(2): p. 109-17.

95. Lin, W., et al. Selective electrodiffusion of zinc ions in a Zrt-, Irt-like protein, ZIPB. J Biol Chem, 2010.

96. Chao, Y. and D. Fu. Thermodynamic studies of the mechanism of metal binding to the Escherichia coli zinc transporter YiiP. J Biol Chem, 2004. 279(17): p. 17173-80.

97. Grass, G., et al. FieF (YiiP) from Escherichia coli mediates decreased cellular accumulation of iron and relieves iron stress. Arch Microbiol, 2005. 183(1): p. 9-18.

98. Lu, M. and D. Fu. Structure of the zinc transporter YiiP. Science, 2007. 317(5845): p. 1746-8.

99. Cherezov, V., et al. Insights into the mode of action of a putative zinc transporter CzrB in Thermus thermophilus. Structure, 2008. 16(9): p. 1378-88.

100. Hofer, N., et al. Crystallization and preliminary X-ray diffraction analysis of a soluble domain of the putative zinc transporter CzrB from Thermus thermophilus. Acta Crystallogr Sect F Struct Biol Cryst Commun, 2007. 63(Pt 8): p. 673-7.

101. Lu, M., J. Chai, and D. Fu. Structural basis for autoregulation of the zinc transporter YiiP. Nat Struct Mol Biol, 2009. 16(10): p. 1063-7.

102. Watts, J.C., et al. Interactome analyses identify ties of PrP and its mammalian paralogs to oligomannosidic N-glycans and endoplasmic reticulum-derived chaperones. PLoS Pathog, 2009. 5(10): p. e1000608.

164

103. Qin, K., et al. Doppel-induced apoptosis and counteraction by cellular prion protein in neuroblastoma and astrocytes. Neuroscience, 2006. 141(3): p. 1375-88.

104. Rambold, A.S., et al. Stress-protective signalling of prion protein is corrupted by scrapie prions. EMBO J, 2008. 27(14): p. 1974-84.

105. Priola, S.A., et al. A 60-kDa prion protein (PrP) with properties of both the normal and scrapie-associated forms of PrP. J Biol Chem, 1995. 270(7): p. 3299-305.

106. Li, A., et al. Prion protein with an octapeptide insertion has impaired neuroprotective activity in transgenic mice. EMBO J, 2007. 26(11): p. 2777-85.

107. Baumann, F., et al. Lethal recessive myelin toxicity of prion protein lacking its central domain. EMBO J, 2007. 26(2): p. 538-47.

108. Jiayu, W., et al. Mapping the interaction site of prion protein and Sho. Mol Biol Rep, 2009.

109. Yamasaki, S., et al. Zinc is a novel intracellular second messenger. J Cell Biol, 2007. 177(4): p. 637-45.

110. Frederickson, C.J., J.Y. Koh, and A.I. Bush. The neurobiology of zinc in health and disease. Nat Rev Neurosci, 2005. 6(6): p. 449-62.

111. Devirgiliis, C., et al. Zinc fluxes and zinc transporter genes in chronic diseases. Mutat Res, 2007. 622(1-2): p. 84-93.

112. Wilson, M., C. Hogstrand, and W. Maret. Picomolar concentrations of free zinc(II) ions regulate receptor protein-tyrosine phosphatase beta activity. J Biol Chem, 2012. 287(12): p. 9322-6.

113. Lye, J.C., et al. Systematic functional characterization of putative zinc transport genes and identification of zinc toxicosis phenotypes in Drosophila melanogaster. J Exp Biol, 2012. 215(Pt 18): p. 3254-65.

114. Van Doren, M., et al. fear of intimacy encodes a novel transmembrane protein required for gonad morphogenesis in Drosophila. Development, 2003. 130(11): p. 2355-64.

115. Taylor, K. and R.I. Nicholson. Fear of intimacy--a close LIV-1 acquaintancy? Development, 2006. 133(16): p. 3053.

116. Moore, L.A., et al. Identification of genes controlling germ cell migration and embryonic gonad formation in Drosophila. Development, 1998. 125(4): p. 667-78.

117. Mathews, W.R., et al. Zinc transport activity of Fear of Intimacy is essential for proper gonad morphogenesis and DE-cadherin expression. Development, 2006. 133(6): p. 1143- 53.

165

118. Yamashita, S., et al. Zinc transporter LIVI controls epithelial-mesenchymal transition in zebrafish gastrula organizer. Nature, 2004. 429(6989): p. 298-302.

119. Malaga-Trillo, E., et al. Regulation of embryonic cell adhesion by the prion protein. PLoS Biol, 2009. 7(3): p. e55.

120. Thiery, J.P., et al. Epithelial-mesenchymal transitions in development and disease. Cell, 2009. 139(5): p. 871-90.

121. Tozlu, S., et al. Identification of novel genes that co-cluster with estrogen receptor alpha in breast tumor biopsy specimens, using a large-scale real-time reverse transcription-PCR approach. Endocr Relat Cancer, 2006. 13(4): p. 1109-20.

122. Schneider, J., et al. Identification and meta-analysis of a small gene expression signature for the diagnosis of estrogen receptor status in invasive ductal breast cancer. Int J Cancer, 2006. 119(12): p. 2974-9.

123. Taylor, K.M., S. Hiscox, and R.I. Nicholson. Zinc transporter LIV-1: a link between cellular development and cancer progression. Trends Endocrinol Metab, 2004. 15(10): p. 461-3.

124. Shen, H., H. Qin, and J. Guo. Concordant correlation of LIV-1 and E-cadherin expression in human breast cancer cell MCF-7. Mol Biol Rep, 2009. 36(4): p. 653-9.

125. Zhau, H.E., et al. Epithelial to mesenchymal transition (EMT) in human prostate cancer: lessons learned from ARCaP model. Clin Exp Metastasis, 2008. 25(6): p. 601-10.

126. Unno, J., et al. LIV-1 enhances the aggressive phenotype through the induction of epithelial to mesenchymal transition in human pancreatic carcinoma cells. Int J Oncol, 2009. 35(4): p. 813-21.

127. Taylor, K.M., et al. Structure-function analysis of LIV-1, the breast cancer-associated protein that belongs to a new subfamily of zinc transporters. Biochem J, 2003. 375(Pt 1): p. 51-9.

128. Suzuki, A. and T. Endo. Ermelin, an endoplasmic reticulum transmembrane protein, contains the novel HELP domain conserved in eukaryotes. Gene, 2002. 284(1-2): p. 31- 40.

129. Huang, L., et al. The ZIP7 gene (Slc39a7) encodes a zinc transporter involved in zinc homeostasis of the Golgi apparatus. J Biol Chem, 2005. 280(15): p. 15456-63.

130. Dufner-Beattie, J., et al. The adaptive response to dietary zinc in mice involves the differential cellular localization and zinc regulation of the zinc transporters ZIP4 and ZIP5. J Biol Chem, 2004. 279(47): p. 49082-90.

131. Wang, F., et al. The mammalian Zip5 protein is a zinc transporter that localizes to the basolateral surface of polarized cells. J Biol Chem, 2004. 279(49): p. 51433-41.

166

132. Lyons, T.J., et al. Genome-wide characterization of the Zap1p zinc-responsive regulon in yeast. Proc Natl Acad Sci U S A, 2000. 97(14): p. 7957-62.

133. Gitan, R.S. and D.J. Eide. Zinc-regulated ubiquitin conjugation signals endocytosis of the yeast ZRT1 zinc transporter. Biochem J, 2000. 346 Pt 2: p. 329-36.

134. Gitan, R.S., et al. Zinc-induced inactivation of the yeast ZRT1 zinc transporter occurs through endocytosis and vacuolar degradation. J Biol Chem, 1998. 273(44): p. 28617-24.

135. Ryu, M.S., et al. Zinc transporters ZnT1 (Slc30a1), Zip8 (Slc39a8), and Zip10 (Slc39a10) in mouse red blood cells are differentially regulated during erythroid development and by dietary zinc deficiency. J Nutr, 2008. 138(11): p. 2076-83.

136. Weaver, B.P., et al. Novel zinc-responsive post-transcriptional mechanisms reciprocally regulate expression of the mouse Slc39a4 and Slc39a5 zinc transporters (Zip4 and Zip5). Biol Chem, 2007. 388(12): p. 1301-12.

137. Choi, C.J., et al. Interaction of metals with prion protein: possible role of divalent cations in the pathogenesis of prion diseases. Neurotoxicology, 2006. 27(5): p. 777-87.

138. Rana, A., et al. Prion metal interaction: is prion pathogenesis a cause or a consequence of metal imbalance? Chem Biol Interact, 2009. 181(3): p. 282-91.

139. Toni, M., et al. Extracellular copper ions regulate cellular prion protein (PrPC) expression and metabolism in neuronal cells. FEBS Lett, 2005. 579(3): p. 741-4.

140. Qin, K., et al. ATM-mediated transcriptional elevation of prion in response to copper- induced oxidative stress. J Biol Chem, 2009. 284(7): p. 4582-93.

141. Armendariz, A.D., et al. Gene expression profiling in chronic copper overload reveals upregulation of Prnp and App. Physiol Genomics, 2004. 20(1): p. 45-54.

142. Varela-Nallar, L., et al. Induction of cellular prion protein gene expression by copper in neurons. Am J Physiol Cell Physiol, 2006. 290(1): p. C271-81.

143. Watts, J.C., et al. The CNS glycoprotein Shadoo has PrP(C)-like protective properties and displays reduced levels in prion infections. EMBO J, 2007. 26: p. 4038-4050.

144. Zheng, D., et al. Regulation of ZIP and ZnT zinc transporters in zebrafish gill: zinc repression of ZIP10 transcription by an intronic MRE cluster. Physiol Genomics, 2008. 34(2): p. 205-14.

145. Wimmer, U., et al. Two major branches of anti-cadmium defense in the mouse: MTF- 1/metallothioneins and glutathione. Nucleic Acids Res, 2005. 33(18): p. 5715-27.

146. Weaver, B.P. and G.K. Andrews. Regulation of zinc-responsive Slc39a5 (Zip5) translation is mediated by conserved elements in the 3'-untranslated region. Biometals, 2012. 25(2): p. 319-35.

167

147. Bellingham, S.A., et al. Regulation of prion gene expression by transcription factors SP1 and metal transcription factor-1. J Biol Chem, 2009. 284(2): p. 1291-301.

148. Funke-Kaiser, H., et al. Functional characterization of the human prion protein promoter in neuronal and endothelial cells. J Mol Med, 2001. 79(9): p. 529-35.

149. Mahal, S.P., et al. Isolation and functional characterisation of the promoter region of the human prion protein gene. Gene, 2001. 268(1-2): p. 105-14.

150. Riechmann, V., et al. The genetic control of the distinction between fat body and gonadal mesoderm in Drosophila. Development, 1998. 125(4): p. 713-23.

151. Fu, Y., et al. Vertebrate tinman homologues XNkx2-3 and XNkx2-5 are required for heart formation in a functionally redundant manner. Development, 1998. 125(22): p. 4439-49.

152. Cano, A., et al. The transcription factor snail controls epithelial-mesenchymal transitions by repressing E-cadherin expression. Nat Cell Biol, 2000. 2(2): p. 76-83.

153. Tanaka-Matakatsu, M., et al. Cadherin-mediated cell adhesion and cell motility in Drosophila trachea regulated by the transcription factor Escargot. Development, 1996. 122(12): p. 3697-705.

154. Nieto, M.A. The snail superfamily of zinc-finger transcription factors. Nat Rev Mol Cell Biol, 2002. 3(3): p. 155-66.

155. Inghirami, G., et al. New and old functions of STAT3: a pivotal target for individualized treatment of cancer. Cell Cycle, 2005. 4(9): p. 1131-3.

156. Groner, B., P. Lucks, and C. Borghouts. The function of Stat3 in tumor cells and their microenvironment. Semin Cell Dev Biol, 2008. 19(4): p. 341-50.

157. Micalizzi, D.S. and H.L. Ford. Epithelial-mesenchymal transition in development and cancer. Future Oncol, 2009. 5(8): p. 1129-43.

158. van Zijl, F., et al. Epithelial-mesenchymal transition in hepatocellular carcinoma. Future Oncol, 2009. 5(8): p. 1169-79.

159. Zhao, L., et al. LIV-1 suppression inhibits HeLa cell invasion by targeting ERK1/2- Snail/Slug pathway. Biochem Biophys Res Commun, 2007. 363(1): p. 82-8.

160. Kitamura, H., et al. Toll-like receptor-mediated regulation of zinc homeostasis influences dendritic cell function. Nat Immunol, 2006. 7(9): p. 971-7.

161. Lopez, V. and S.L. Kelleher. Zip6-attenuation promotes epithelial-to-mesenchymal transition in ductal breast tumor (T47D) cells. Exp Cell Res, 2010. 316(3): p. 366-75.

162. Watt, N.T. and N.M. Hooper. The prion protein and neuronal zinc homeostasis. Trends Biochem Sci, 2003. 28: p. 406-410.

168

163. Schmitt-Ulms, G., et al. Binding of neural cell adhesion molecules (N-CAMs) to the cellular prion protein. J Mol Biol, 2001. 314: p. 1209-1225.

164. Santuccione, A., et al. Prion protein recruits its neuronal receptor NCAM to lipid rafts to activate p59fyn and to enhance neurite outgrowth. J Cell Biol, 2005. 169(2): p. 341-54.

165. Lehembre, F., et al. NCAM-induced focal adhesion assembly: a functional switch upon loss of E-cadherin. EMBO J, 2008. 27(19): p. 2603-15.

166. Frame, M.C. and G.J. Inman. NCAM is at the heart of reciprocal regulation of E- cadherin- and integrin-mediated adhesions via signaling modulation. Dev Cell, 2008. 15(4): p. 494-6.

167. Wuthrich, K. and R. Riek. Three-dimensional structures of prion proteins. Adv Protein Chem, 2001. 57: p. 55-82.

168. Zahn, R., et al. NMR structure of a variant human prion protein with two disulfide bridges. J Mol Biol, 2003. 326(1): p. 225-34.

169. Silverman, G.L., et al. Doppel is an N-glycosylated, glycosylphosphatidylinositol- anchored protein. Expression in testis and ectopic production in the brains of Prnp(0/0) mice predisposed to Purkinje cell loss. J Biol Chem, 2000. 275(35): p. 26834-41.

170. Wang, F., et al. Acrodermatitis enteropathica mutations affect transport activity, localization and zinc-responsive trafficking of the mouse ZIP4 zinc transporter. Hum Mol Genet, 2004. 13(5): p. 563-71.

171. He, L., et al. ZIP8, member of the solute-carrier-39 (SLC39) metal-transporter family: characterization of transporter properties. Mol Pharmacol, 2006. 70(1): p. 171-80.

172. Taylor, K.M., et al. Structure-function analysis of a novel member of the LIV-1 subfamily of zinc transporters, ZIP14. FEBS Lett, 2005. 579(2): p. 427-32.

173. Girijashanker, K., et al. Slc39a14 gene encodes ZIP14, a metal/bicarbonate symporter: similarities to the ZIP8 transporter. Mol Pharmacol, 2008. 73(5): p. 1413-23.

174. Wang, F., et al. Zinc-stimulated endocytosis controls activity of the mouse ZIP1 and ZIP3 zinc uptake transporters. J Biol Chem, 2004. 279(23): p. 24631-9.

175. Pan, K.M., N. Stahl, and S.B. Prusiner. Purification and properties of the cellular prion protein from Syrian hamster brain. Protein Sci, 1992. 1(10): p. 1343-52.

176. Shyng, S.L., M.T. Huber, and D.A. Harris. A prion protein cycles between the cell surface and an endocytic compartment in cultured neuroblastoma cells. J Biol Chem, 1993. 268(21): p. 15922-8.

177. Chen, S.G., et al. Truncated forms of the human prion protein in normal brain and in prion diseases. J Biol Chem, 1995. 270(32): p. 19173-80.

169

178. Jimenez-Huete, A., et al. Endogenous proteolytic cleavage of normal and disease- associated isoforms of the human prion protein in neural and non-neural tissues. Am J Pathol, 1998. 153(5): p. 1561-72.

179. Vincent, B., et al. Phorbol ester-regulated cleavage of normal prion protein in HEK293 human cells and murine neurons. J Biol Chem, 2000. 275(45): p. 35612-6.

180. Guillot-Sestier, M.V., et al. The alpha-secretase-derived N-terminal product of cellular prion, N1, displays neuroprotective function in vitro and in vivo. J Biol Chem, 2009. 284(51): p. 35973-86.

181. Haigh, C.L., et al. PrPC-related signal transduction is influenced by copper, membrane integrity and the alpha cleavage site. Cell Res, 2009. 19(9): p. 1062-78.

182. Vincent, B., et al. The disintegrins ADAM10 and TACE contribute to the constitutive and phorbol ester-regulated normal cleavage of the cellular prion protein. J Biol Chem, 2001. 276(41): p. 37743-6.

183. Alfa Cisse, M., et al. Isoform-specific contribution of protein kinase C to prion processing. Mol Cell Neurosci, 2008. 39(3): p. 400-10.

184. Cisse, M.A., et al. The disintegrin ADAM9 indirectly contributes to the physiological processing of cellular prion by modulating ADAM10 activity. J Biol Chem, 2005. 280(49): p. 40624-31.

185. Walmsley, A.R., et al. alpha-cleavage of the prion protein occurs in a late compartment of the secretory pathway and is independent of lipid rafts. Mol Cell Neurosci, 2009. 40(2): p. 242-8.

186. Taylor, D.R., et al. Role of ADAMs in the ectodomain shedding and conformational conversion of the prion protein. J Biol Chem, 2009. 284(34): p. 22590-600.

187. Borchelt, D.R., et al. Release of the cellular prion protein from cultured cells after loss of its glycoinositol phospholipid anchor. Glycobiology, 1993. 3(4): p. 319-29.

188. Harris, D.A., et al. Processing of a cellular prion protein: identification of N- and C- terminal cleavage sites. Biochemistry, 1993. 32(4): p. 1009-16.

189. Parkin, E.T., et al. Dual mechanisms for shedding of the cellular prion protein. J Biol Chem, 2004. 279(12): p. 11170-8.

190. Mange, A., et al. Alpha- and beta- cleavages of the amino-terminus of the cellular prion protein. Biol Cell, 2004. 96(2): p. 125-32.

191. McMahon, H.E., et al. Cleavage of the amino terminus of the prion protein by reactive oxygen species. J Biol Chem, 2001. 276(3): p. 2286-91.

192. Watt, N.T., et al. Reactive oxygen species-mediated beta-cleavage of the prion protein in the cellular response to oxidative stress. J Biol Chem, 2005. 280(43): p. 35914-21.

170

193. Dron, M., et al. Endogenous proteolytic cleavage of disease-associated prion protein to produce C2 fragments is strongly cell- and tissue-dependent. J Biol Chem, 2010. 285(14): p. 10252-64.

194. Kambe, T. and G.K. Andrews. Novel proteolytic processing of the ectodomain of the zinc transporter ZIP4 (SLC39A4) during zinc deficiency is inhibited by acrodermatitis enteropathica mutations. Mol Cell Biol, 2009. 29(1): p. 129-39.

195. Mao, X., et al. A histidine-rich cluster mediates the ubiquitination and degradation of the human zinc transporter, hZIP4, and protects against zinc cytotoxicity. J Biol Chem, 2007. 282(10): p. 6992-7000.

196. Taylor, K.M. LIV-1 breast cancer protein belongs to new family of histidine-rich membrane proteins with potential to control intracellular Zn2+ homeostasis. IUBMB Life, 2000. 49(4): p. 249-53.

197. Jongeneel, C.V., J. Bouvier, and A. Bairoch. A unique signature identifies a family of zinc-dependent metallopeptidases. FEBS Lett, 1989. 242(2): p. 211-4.

198. Rawlings, N.D. and A.J. Barrett. Evolutionary families of metallopeptidases. Methods Enzymol, 1995. 248: p. 183-228.

199. Feng, L., et al. Structure of a site-2 protease family intramembrane metalloprotease. Science, 2007. 318(5856): p. 1608-12.

200. Bell, L.M., et al. Cytoplasmic tail deletion of T cell receptor (TCR) β-chain results in its surface expression as glycosylphosphatidylinositol-anchored polypeptide on mature T cells in the absence of TCR-α. J Biol Chem, 1994. 269: p. 22758-22763.

201. Lanier, L., et al. Membrane anchoring of a human IgG Fc receptor (CD16) determined by a single amino acid. Science, 1989. 246: p. 1611-1613.

202. Naghibalhossaini, F. and C.P. Stanners. Minimal mutations are required to effect a radical change in function in CEA family members of the Ig superfamily. J Cell Sci, 2004. 117: p. 761-769.

203. Rogers, E.E., D.J. Eide, and M.L. Guerinot. Altered selectivity in an Arabidopsis metal transporter. Proc Natl Acad Sci U S A, 2000. 97(22): p. 12356-60.

204. Grass, G., et al. The metal permease ZupT from Escherichia coli is a transporter with a broad substrate spectrum. J Bacteriol, 2005. 187(5): p. 1604-11.

205. Taudte, N. and G. Grass. Point mutations change specificity and kinetics of metal uptake by ZupT from Escherichia coli. Biometals, 2010. 23(4): p. 643-56.

206. Tominaga, K., et al. SLC39A14, a LZT protein, is induced in adipogenesis and transports zinc. FEBS J, 2005. 272(7): p. 1590-9.

171

207. Himeno, S., T. Yanagiya, and H. Fujishiro. The role of zinc transporters in cadmium and manganese transport in mammalian cells. Biochimie, 2009. 91(10): p. 1218-22.

208. Liuzzi, J.P., et al. Zip14 (Slc39a14) mediates non-transferrin-bound iron uptake into cells. Proc Natl Acad Sci U S A, 2006. 103(37): p. 13612-7.

209. Eng, B.H., et al. Sequence analyses and phylogenetic characterization of the ZIP family of metal ion transport proteins. J Membr Biol, 1998. 166(1): p. 1-7.

210. Millhauser, G.L. Copper and the prion protein: methods, structures, function, and disease. Annu Rev Phys Chem, 2007. 58: p. 299-320.

211. Pauly, P.C. and D.A. Harris. Copper stimulates endocytosis of the prion protein. J Biol Chem, 1998. 273(50): p. 33107-10.

212. Perera, W.S. and N.M. Hooper. Ablation of the metal ion-induced endocytosis of the prion protein by disease-associated mutation of the octarepeat region. Curr Biol, 2001. 11(7): p. 519-23.

213. Kenward, A.G., L.J. Bartolotti, and C.S. Burns. Copper and zinc promote interactions between membrane-anchored peptides of the metal binding domain of the prion protein. Biochemistry, 2007. 46(14): p. 4261-71.

214. Morante, S., et al. Inter- and intra-octarepeat Cu(II) site geometries in the prion protein: implications in Cu(II) binding cooperativity and Cu(II)-mediated assemblies. J Biol Chem, 2004. 279(12): p. 11753-9.

215. Zahn, R. The octapeptide repeats in mammalian prion protein constitute a pH-dependent folding and aggregation site. J Mol Biol, 2003. 334(3): p. 477-88.

216. Walter, E.D., et al. The prion protein is a combined zinc and copper binding protein: Zn2+ alters the distribution of Cu2+ coordination modes. J Am Chem Soc, 2007. 129(50): p. 15440-1.

217. Qin, K., et al. The PrP-like protein Doppel binds copper. J Biol Chem, 2003. 278(11): p. 8888-96.

218. Cereghetti, G.M., et al. Copper(II) binding to the human Doppel protein may mark its functional diversity from the prion protein. J Biol Chem, 2004. 279(35): p. 36497-503.

219. Behrens, A., et al. Absence of the prion protein homologue Doppel causes male sterility. EMBO J, 2002. 21(14): p. 3652-8.

220. Paisley, D., et al. Male infertility and DNA damage in Doppel knockout and prion protein/Doppel double-knockout mice. Am J Pathol, 2004. 164(6): p. 2279-88.

221. Yamaguchi, S., et al. Zinc is an essential trace element for spermatogenesis. Proc Natl Acad Sci U S A, 2009. 106(26): p. 10859-64.

172

222. Camponeschi, F., et al. Structural features of the Zn(2+) complex with the single repeat region of "prion related protein" (PrP-rel-2) of zebrafish zPrP63-70 fragment. Dalton Trans, 2009(24): p. 4643-5.

223. Szyrwiel, L., et al. Zn(II) ions bind very efficiently to tandem repeat region of "prion related protein" (PrP-rel-2) of zebra-fish. MS and potentiometric evidence. Dalton Trans, 2008. 44: p. 6117-6120.

224. Taylor, D.R. and N.M. Hooper. The prion protein and lipid rafts. Mol Membr Biol, 2006. 23: p. 89-99.

225. Supattapone, S., N.R. Deleault, and J.R. Rees. Amplification of purified prions in vitro. Methods Mol Biol, 2008. 459: p. 117-130.

226. Ben-Zaken, O., et al. Cellular heparan sulfate participates in the metabolism of prions. J Biol Chem, 2003. 278(41): p. 40041-9.

227. Horonchik, L., et al. Heparan sulfate is a cellular receptor for purified infectious prions. J Biol Chem, 2005. 280(17): p. 17062-7.

228. Fischer, M.B., et al. Binding of disease-associated prion protein to plasminogen. Nature, 2000. 408(6811): p. 479-83.

229. Nelson, J., et al. The 67 kDa laminin receptor: structure, function and role in disease. Biosci Rep, 2008. 28: p. 33-48.

230. Rieger, R., et al. The human 37-kDa laminin receptor precursor interacts with the prion protein in eukaryotic cells. Nat Med, 1997. 3: p. 1383-1388.

231. Schmitt-Ulms, G., et al. Binding of neural cell adhesion molecules (N-CAMs) to the cellular prion protein. J. Mol. Biol., 2001. 314: p. 1209-1225.

232. Hundt, C., et al. Identification of interaction domains of the prion protein with its 37- kDa/67-kDa laminin receptor. EMBO J, 2001. 20(21): p. 5876-86.

233. Warner, R.G., et al. Identification of the heparan sulfate binding sites in the cellular prion protein. J Biol Chem, 2002. 277(21): p. 18421-30.

234. Gauczynski, S., et al. The 37-kDa/67-kDa laminin receptor acts as the cell-surface receptor for the cellular prion protein. EMBO J, 2001. 20: p. 5863-5875.

235. Mouillet-Richard, S., et al. Signal transduction through prion protein. Science, 2000. 2000: p. 1925-1928.

236. Santuccione, A., et al. Prion protein recruits its neuronal receptor NCAM to lipid rafts to activate p59fyn and to enhance neurite outgrowth. J Cell Biol, 2005. 169: p. 341-354.

237. Schmitt-Ulms, G., et al. Time controlled transcardiac perfusion crosslinking for the study of protein interactions in complex tissues. Nat Biotechnol, 2004. 22: p. 724-731.

173

238. Race, R.E., L.H. Fadness, and B. Chesebro. Characterization of scrapie infection in mouse neuroblastoma cells. J Gen Virol, 1987. 68: p. 1391-1399.

239. Telling, G.C., et al. N-terminally tagged prion protein supports prion propagation in transgenic mice. Protein Science, 1997. 6: p. 825-833.

240. Einhauer, A. and A. Jungbauer. The FLAG peptide, a versatile fusion tag for the purification of recombinant proteins. J Biochem Biophys Methods, 2001. 49: p. 455-465.

241. Chalkley, R.J., et al. Comprehensive Analysis of a Multidimensional Liquid Chromatography Mass Spetrometry Dataset Acquired on a Quadrupole Selecting Quadrupole Collision Cell, Time-of-flight Mass Spectrometer. II. New Developments in Protein Prospector Allow for Reliable and Comprehensive Automatic Analysis of Large Datasets. Mol Cell Prot, 2005. 4: p. 1194-1204.

242. Shannon, P., et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res, 2003. 13(11): p. 2498-504.

243. Zieske, L.R. A perspective on the use of iTRAQ reagent technology for protein complex and profiling studies. J Exp Bot, 2006. 57: p. 1501-1508.

244. Elias, J.E. and S.P. Gygi. Target-decoy search strategy for increased confidence in large- scale protein identifications by mass spectrometry. Nat Methods, 2007. 4: p. 207-214.

245. Moore, R.C., et al. Ataxia in prion protein (PrP)-deficient mice is associated with upregulation of the novel PrP-like protein doppel. J Mol Biol, 1999. 292(4): p. 797-817.

246. Altruda, F., et al. Cloning of cDNA for a novel mouse membrane glycoprotein (gp42): shared identity to histocompatibility antigens, immunoglobulins and neural-cell adhesion molecules. Gene, 1989. 85: p. 445-452.

247. Heller, M., et al. The immunoglobulin-superfamily molecule basigin is a binding protein for oligomannosidic carbohydrates: an anti-idiotypic approach. J Neurochem, 2003. 84: p. 557-565.

248. Horstkorte, R., et al. The fourth immunoglobulin-like domain of NCAM contains a carbohydrate recognition domain for oligomannosidic glycans implicated in association with L1 and neurite outgrowth. J Cell Biol, 1993. 121: p. 1409-1421.

249. Heiland, P.C., et al. Tyrosine and serine phosphorylation of the neural cell adhesion molecule L1 is implicated in its oligomannosidic glycan dependent association with NCAM and neurite outgrowth. Eur J Cell Biol, 1998. 75: p. 97-106.

250. Orberger, G., et al. Structure of the N-linked oligosaccharides of the human transferrin receptor. Eur J Biochem, 1992. 205: p. 257-267.

251. Schmitz, B., et al. Monoclonal antibodies raised against membrane glycoproteins from mouse brain recognize N-linked oligomannosidic glycans. Glycobiology, 1993. 3: p. 609- 617.

174

252. Ohe, Y., et al. Characterization of nucleotide pyrophosphatase-5 as an oligomannosidic glycoprotein in rat brain. Biochem Biophys Res Com, 2003. 308: p. 719-725.

253. Pesheva, P., A.F. Horwitz, and M. Schachner. Integrin, the cell surface receptor for fibronectin and laminin, expresses the L2/HNK-1 and L3 carbohydrate structures shared by adhesion molecules. Neurosci Let, 1987. 83: p. 303-306.

254. Borchelt, D.R., A. Taraboulos, and S.B. Prusiner. Evidence for synthesis of scrapie prion proteins in the endocytic pathway. J Biol Chem, 1992. 267: p. 16188-16199.

255. Caughey, B. and G.J. Raymond. The scrapie-associated form of PrP is made from a cell surface precursor that is both protease- and phospholipase-sensitive. J Biol Chem, 1991. 266: p. 18217-18223.

256. Mandel, R., et al. Inhibition of a reductive function of the plasma membrane by bacitracin and antibodies against protein disulfide-isomerase. Proc Natl Acad Sci U S A, 1993. 90: p. 4112-4116.

257. Caughey, B., et al. N-terminal truncation of the scrapie-associated form of PrP by lysosomal protease(s): implications regarding the site of conversion of PrP to the protease-resistant state. J Virol, 1991. 65(12): p. 6597-603.

258. Taraboulos, A., et al. Synthesis and trafficking of prion proteins in cultured cells. Mol Biol Cell, 1992. 3(8): p. 851-63.

259. Moore, R.C., et al. Doppel-induced cerebellar degeneration in transgenic mice. Proc Natl Acad Sci U S A, 2001. 98: p. 15288-15293.

260. Rutishauser, D., et al. The comprehensive native interactome of a fully functional tagged prion protein. PLoS One, 2009. 4: p. 1-13.

261. Fasano, C., V. Campana, and Z. C. Protein Only or Something More? Overview of Potential Prion Cofactors. J Mol Neurosci, 2006. 29: p. 195-214.

262. Hetz, C., et al. The disulfide isomerase Grp58 is a protective factor against prion neurotoxicity. J Neurosci, 2005. 25: p. 2793-2802.

263. Turano, C., et al. Proteins of the PDI family: unpredicted non-ER locations and functions. J Cell Physiol, 2002. 193: p. 154-163.

264. Lucassen, R., K. Nishina, and S. Supattapone. In vitro amplification of protease-resistant prion protein requires free sulfhydryl groups. Biochemistry, 2003. 42: p. 4127-4135.

265. Ellgaard, L. and L.W. Ruddock. The human protein disulphide isomerase family: substrate interactions and functional properties. EMBO reports, 2005. 6: p. 28-32.

266. Weissmann, C. The state of the prion. Nat Rev Microbiol, 2004. 2: p. 861-871.

175

267. Gardai, S.J., et al. Cell-surface calreticulin initiates clearance of viable or apoptotic cells through trans-activation of LRP on the phagocyte. Cell, 2005. 123: p. 321-334.

268. Westergard, L., H.M. Christensen, and D.A. Harris. The cellular prion protein (PrPC): its physiological function and role in disease. Biochim Biophys Acta, 2007. 1772: p. 629- 644.

269. McDonnell, J.M., et al. Calreticulin binding affinity for glycosylated laminin. J Biol Chem, 1996. 271: p. 7891-7894.

270. White, T.K., Q. Zhu, and M.L. Tanzer. Cell surface calreticulin is a putative mannoside lectin which triggers mouse melanoma cell spreading. J Biol Chem, 1995. 270: p. 15926- 15929.

271. Kleene, R., et al. Prion protein regulates glutamate-dependent lactate transport of astrocytes. J Neurosci, 2007. 27: p. 12331-12340.

272. Hemler, M.E., et al. Multiple very late antigen (VLA) heterodimers on platelets. Evidence for distinct VLA-2, VLA-5 (fibronectin receptor), and VLA-6 structures. J Biol Chem, 1988. 263: p. 7660-7665.

273. Zhu, Q., et al. Calreticulin-integrin bidirectional signaling complex. Biochem Biophys Res Com, 1997. 232: p. 354-358.

274. Canfield, S.M. and A.Y. Khakoo. The nonintegrin laminin binding protein (p67 LBP) is expressed on a subset of activated human T lymphocytes and, together with the integrin very late activation antigen-6, mediates avid cellular adherence to laminin. J Immunol, 1999. 163: p. 3430-3440.

275. Vana, K., et al. LRP/LR as an alternative promising target in therapy of prion diseases, Alzheimer's disease and cancer. Infect Disord Drug Targets, 2009. 9(1): p. 69-80.

276. Zuber, C., et al. Delivery of single-chain antibodies (scFvs) directed against the 37/67 kDa laminin receptor into mice via recombinant adeno-associated viral vectors for prion disease gene therapy. J Gen Virol, 2008. 89(Pt 8): p. 2055-61.

277. Pflanz, H., et al. Scrapie-infected transgenic mice expressing a laminin receptor decoy mutant reveal a prolonged incubation time associated with low levels of PrPres. J Mol Biol, 2009. 388(4): p. 721-9.

278. Hundt, C. and S. Weiss. The prion-like protein Doppel fails to interact with itself, the prion protein and the 37 kDa/67 kDa laminin receptor in the yeast two-hybrid system. Biochim Biophys Acta, 2004. 1689(1): p. 1-5.

279. Yin, S.M., et al. Interaction of Doppel with the full-length laminin receptor precursor protein. Arch Biochem Biophys, 2004. 428(2): p. 165-9.

176

280. Sato, M., et al. Ribosome-Associated Protein LBP/p40 binds to S21 protein of 40S ribosome: Analysis using a yeast Two-Hybrid System. Biochem Biophys Res Com, 1999. 256: p. 385-390.

281. Stelzl, U., et al. A human protein-protein interaction network - a resource for annotating the proteome. Cell, 2005. 122: p. 957-968.

282. Cai, S., et al. CD98 modulates integrin b1 function in polarized epithelial cells. J Cell Sci, 2004. 118: p. 889-899.

283. Prager, G.W., et al. CD98hc (SLC3A2) interaction with the integrin beta subunit cytoplasmic domain mediates adhesive signaling. J Biol Chem, 2007. 282: p. 24477- 24484.

284. Cho, J.Y., et al. The functional interactions between CD98, beta-1integrins, and CD147 in the induction of U937 homotypic aggregation. Blood, 2001. 98: p. 374-382.

285. Xu, D. and M.E. Hemler. Metabolic activation-related CD147-CD98 complex. Mol Cell Proteomics, 2005. 4: p. 1061-1071.

286. Mori, K., et al. The functional interaction between CD98 and CD147 in regulation of virus-induced cell fusion and osteoclast formation. Med Microbiol Immunol, 2004. 193: p. 155-162.

287. Takahashi, Y., et al. Sequence-specific interaction between the disintegrin domain of mouse ADAM3 and murine eggs: role of beta1 integrin-associated proteins CD9, CD81 and CD98. Mol Biol Cell, 2001. 12: p. 809-820.

288. Saxenam, D.K., et al. Behaviour of a sperm surface transmembrane glycoprotein basigin during epididymal maturation and its role in fertilization in mice. Reproduction, 2002. 123: p. 435-444.

289. Behrens, A., et al. Absence of the prion protein homologue Doppel causes male sterility. EMBO J, 2002. 21: p. 3652-3658.

290. Hwang, D., et al. A systems approach to prion disease. Mol Syst Biol, 2009. 5: p. 252.

291. Caughey, B. and G.S. Baron. Prions and their partners in crime. Nature, 2006. 443: p. 803-810.

292. Rivera-Milla, E., et al. Disparate evolution of prion protein domains and the distinct origin of Doppel- and prion-related loci revealed by fish-to-mammal comparisons. Faseb J, 2005. 20: p. 317-319.

293. Date, S.V. Estimating protein function using protein-protein relationships. Methods Mol Biol, 2007. 408: p. 109-127.

177

294. Rappsilber, J., Y. Ishihama, and M. Mann. Stop and go extraction tips for matrix-assisted laser desorption/ionization, nanoelectrospray, and LC/MS sample pretreatment in proteomics. Anal Chem, 2003. 75: p. 663-670.

295. Bai, Y., et al. The in vivo brain interactome of the amyloid precursor protein. Mol Cell Proteomics, 2008. 7(1): p. 15-34.

296. Perkins, D.N., et al. Probability-based protein identification by searching sequence databases using mass spectrometry data. Electrophoresis, 1999. 20: p. 3551-3567.

297. Mann, M. and M. Wilm. Error-tolerant identification of peptides in sequence datases by peptides sequence tags. Anal Chem, 1994. 66: p. 4390-4399.

298. Katoh, K., et al. MAFFT: a novel method for rapid multiple sequence alignment based on fast Fourier transform. Nucleic Acids Res, 2002. 30: p. 3059-3066.

299. Larkin, M.A., et al. Clustal W and Clustal X version 2.0. Bioinformatics, 2007. 23: p. 2947-2948.

300. Lu, G. and E.N. Moriyama. Vector NTI, a balanced all-in-one sequence analysis suite. Brief Bioinform, 2004. 5: p. 378-388.

301. Sadreyev, R.I. and N.V. Grishin. Accurate statistical model of comparison between multiple sequence alignments. Nucleic Acids Res, 2008. 36: p. 2240-2248.

302. Jaroszewski, L., et al. FFAS03: a server for profile-profile sequence alignments. Nucleic Acids Res, 2005. 33: p. W284-288.

303. Riek, R., et al. NMR structure of the mouse prion protein domain PrP(121-321). Nature, 1996. 382: p. 180-182.

304. Mo, H., et al. Two different neurodegenerative diseases caused by proteins with similar structures. Proc Natl Acad Sci U S A, 2001. 98: p. 2352-2357.

305. Calzolai, L., et al. Prion protein NMR structures of chickens, turtles, and frogs. Proc Natl Acad Sci U S A, 2005. 102(3): p. 651-5.

306. Huang, C.C., et al. Chimera: An Extensible Molecular Modeling Application Constructed Using Standard Components. Pacific Symposium on Biocomputing, 1996. 1: p. 724.

307. Gingras, A.C., et al. Analysis of protein complexes using mass spectrometry. Nat Rev Mol Cell Biol, 2007. 8: p. 645-654.

308. Lo Conte, L., et al. SCOP: a Structural Classification of Proteins database. Nucleic Acids Research, 2000. 28: p. 257-259.

309. Gough, J., et al. Assignment of homology to genome sequences using a library of hidden Markov models that represent all proteins of known structure. J Mol Biol, 2001. 313: p. 903-919.

178

310. Sprenger, J., et al. LOCATE: a mammalian protein subcellular localization database. Nucleic Acids Res, 2008. 36: p. D230-233.

311. Holm, L. and J. Park. DaliLite workbench for protein structure comparison. Bioinformatics, 2000. 16: p. 566-567.

312. Dunn, C.W., et al. Broad phylogenomic sampling improves resolution of the animal tree of life. Nature, 2008. 452(7188): p. 745-9.

313. Davies, P. and D.R. Brown. The chemistry of copper binding to PrP: is there sufficient evidence to elucidate a role for copper in protein function? Biochem J, 2008. 410: p. 237- 244.

314. Rachidi, W., et al. Prion protein protects against zinc-mediated cytotoxicity by modifying intracellular exchangeable zinc and inducing metallothionein expression. J Trace Elem Med Biol, 2009. 23(3): p. 214-23.

315. Lichten, L.A. and R.J. Cousins. Mammalian Zinc Transporters: Nutritional and Physiologic Regulation. Annu Rev Nutr, 2009.

316. Kraut, J., et al. The aromatic substrate binding site in subtilisin BPN' and its resemblance to chymotrypsin. Cold Spring Harb Symp Quant Biol, 1972. 36: p. 117-123.

317. Kornegay, J.R. Molecular genetics and evolution of stomach and nonstomach lysozymes in the hoatzin. J Mol Evol, 1996. 42: p. 676-684.

318. Shmerling, D., et al. Expression of amino-terminally truncated PrP in the mouse leading to ataxia and specific cerebellar lesions. Cell, 1998. 93(2): p. 203-14.

319. Rost, B. Twilight zone of protein sequence alignments. Prot Engineering, 1999. 12: p. 85-94.

320. Baldwin, M.A. Analysis of glycosylphosphatidylinositol protein anchors: the prion protein. Methods Enzymol, 2005. 405: p. 172-187.

321. Rojas, M., et al. Radical differences in functions of closely related members of the human carcinoembryonic antigen gene family. Cell Growth Differ, 1996. 7: p. 655-662.

322. Angst, B.D., C. Marcozzi, and A.I. Magee. The cadherin superfamily: diversity in form and function. J Cell Sci, 2001. 114: p. 629-641.

323. Lee, I.Y., et al. Complete genomic sequence and analysis of the prion protein gene region from three mammalian species. Genome Res, 1998. 8(10): p. 1022-37.

324. Ehsani, S., et al. Evidence for retrogene origins of the prion gene family. PLoS One, 2011. 6(10): p. e26800.

179

325. Ortutay, C., M. Siermala, and M. Vihinen. Molecular characterization of the immune system: emergence of proteins, processes, and domains. Immunogenetics, 2007. 59(5): p. 333-48.

326. Tarailo-Graovac, M. and N. Chen. Using RepeatMasker to identify repetitive elements in genomic sequences. Curr Protoc Bioinformatics, 2009. Chapter 4: p. Unit 4 10.

327. Long, M. Evolution of novel genes. Curr Opin Genet Dev, 2001. 11(6): p. 673-80.

328. Basler, K., et al. Scrapie and cellular PrP isoforms are encoded by the same chromosomal gene. Cell, 1986. 46(3): p. 417-28.

329. Simonic, T., et al. cDNA cloning of turtle prion protein. FEBS Lett, 2000. 469(1): p. 33- 8.

330. Suzuki, T., et al. cDNA sequence and tissue expression of Fugu rubripes prion protein- like: a candidate for the teleost orthologue of tetrapod PrPs. Biochem Biophys Res Commun, 2002. 294(4): p. 912-7.

331. Saeki, K., et al. Three-exon structure of the gene encoding the rat prion protein and its expression in tissues. Virus Genes, 1996. 12(1): p. 15-20.

332. Ortutay, C. and M. Vihinen. PseudoGeneQuest - service for identification of different pseudogene types in the human genome. BMC bioinformatics, 2008. 9: p. 299.

333. Seilacher, A., P.K. Bose, and F. Pfluger. Triploblastic animals more than 1 billion years ago: trace fossil evidence from india. Science, 1998. 282(5386): p. 80-3.

334. Gal-Mark, N., et al. The pivotal roles of TIA proteins in 5' splice-site selection of alu exons and across evolution. PLoS Genet, 2009. 5(11): p. e1000717.

335. Sorek, R. When new exons are born. Heredity, 2009. 103(4): p. 279-80.

336. Roy, S.W. Intron-rich ancestors. Trends Genet, 2006. 22(9): p. 468-71.

337. Roy, S.W. and D. Penny. Large-scale intron conservation and order-of-magnitude variation in intron loss/gain rates in apicomplexan evolution. Genome Res, 2006. 16(10): p. 1270-5.

338. Coulombe-Huntington, J. and J. Majewski. Characterization of intron loss events in mammals. Genome Res, 2007. 17(1): p. 23-32.

339. Irimia, M. and S.W. Roy. Spliceosomal introns as tools for genomic and evolutionary analysis. Nucleic Acids Res, 2008. 36(5): p. 1703-12.

340. Rogozin, I.B., et al. Remarkable interkingdom conservation of intron positions and massive, lineage-specific intron loss and gain in eukaryotic evolution. Curr Biol, 2003. 13(17): p. 1512-7.

180

341. Waterston, R.H., et al. Initial sequencing and comparative analysis of the mouse genome. Nature, 2002. 420(6915): p. 520-62.

342. Kaessmann, H., N. Vinckenbosch, and M. Long. RNA-based gene duplication: mechanistic and evolutionary insights. Nat Rev Genet, 2009. 10(1): p. 19-31.

343. Cereseto, A. and M. Giacca. Integration site selection by retroviruses. AIDS Rev, 2004. 6(1): p. 13-21.

344. Sandmeyer, S.B., L.J. Hansen, and D.L. Chalker. Integration specificity of retrotransposons and retroviruses. Annu Rev Genet, 1990. 24: p. 491-518.

345. Marques, A.C., et al. Emergence of young human genes after a burst of retroposition in primates. PLoS Biol, 2005. 3(11): p. e357.

346. Vinckenbosch, N., I. Dupanloup, and H. Kaessmann. Evolutionary fate of retroposed gene copies in the human genome. Proc Natl Acad Sci U S A, 2006. 103(9): p. 3220-5.

347. Terai, G., et al. Discovery of short pseudogenes derived from messenger RNAs. Nucleic Acids Res, 2010. 38(4): p. 1163-71.

348. Vanin, E.F. Processed pseudogenes. Characteristics and evolution. Biochimica et biophysica acta, 1984. 782(3): p. 231-41.

349. Brosius, J. and S.J. Gould. On "genomenclature": a comprehensive (and respectful) taxonomy for pseudogenes and other "junk DNA". Proceedings of the National Academy of Sciences of the United States of America, 1992. 89(22): p. 10706-10.

350. Fablet, M., et al. Evolutionary origin and functions of retrogene introns. Mol Biol Evol, 2009. 26(9): p. 2147-56.

351. Brosius, J. Echoes from the past--are we still in an RNP world? Cytogenet Genome Res, 2005. 110(1-4): p. 8-24.

352. Schueler-Furman, O., et al. Is GAS1 a co-receptor for the GDNF family of ligands? Trends Pharmacol Sci, 2006. 27(2): p. 72-7.

353. Cabrera, J.R., et al. Gas1 is related to the glial cell-derived neurotrophic factor family receptors alpha and regulates Ret signaling. J Biol Chem, 2006. 281(20): p. 14330-9.

354. Airaksinen, M.S., L. Holm, and T. Hatinen. Evolution of the GDNF family ligands and receptors. Brain Behav Evol, 2006. 68(3): p. 181-90.

355. Hatinen, T., L. Holm, and M.S. Airaksinen. Loss of neurturin in frog--comparative genomics study of GDNF family ligand-receptor pairs. Mol Cell Neurosci, 2007. 34(2): p. 155-67.

356. Ehsani, S., et al. LIV-1 ZIP ectodomain shedding in prion-infected mice resembles cellular response to transition metal starvation. J Mol Biol, 2012. 422(4): p. 556‐74.

181

357. Antala, S. and R.E. Dempski. The human ZIP4 transporter has two distinct binding affinities and mediates transport of multiple transition metals. Biochemistry, 2012. 51(5): p. 963-73.

358. Wang, H., et al. Genetic evidence for adenylyl cyclase 1 as a target for preventing neuronal excitotoxicity mediated by N-methyl-D-aspartate receptors. J Biol Chem, 2007. 282(2): p. 1507-17.

359. St Bernard, R., et al. Fibroblast growth factor receptors as molecular targets in thyroid carcinoma. Endocrinology, 2005. 146(3): p. 1145-53.

360. Carlson, G.A., et al. Prion isolate specified allotypic interactions between the cellular and scrapie prion proteins in congenic and transgenic mice. Proc Natl Acad Sci U S A, 1994. 91(12): p. 5690-4.

361. Kumar, R., et al. Cholesterol transporter ATP-binding cassette A1 (ABCA1) is elevated in prion disease and affects PrPC and PrPSc concentrations in cultured cells. J Gen Virol, 2008. 89(Pt 6): p. 1525-32.

362. Brown, D.A. and J.K. Rose. Sorting of GPI-anchored proteins to glycolipid-enriched membrane subdomains during transport to the apical cell surface. Cell, 1992. 68(3): p. 533-44.

363. Sargiacomo, M., et al. Signal transducing molecules and glycosyl-phosphatidylinositol- linked proteins form a caveolin-rich insoluble complex in MDCK cells. The Journal of cell biology, 1993. 122(4): p. 789-807.

364. Lichten, L.A., et al. MTF-1-mediated repression of the zinc transporter Zip10 is alleviated by zinc restriction. PLoS One, 2011. 6(6): p. e21526.

365. Arslan, P., et al. Cytosolic Ca2+ homeostasis in Ehrlich and Yoshida carcinomas. A new, membrane-permeant chelator of heavy metals reveals that these ascites tumor cell lines have normal cytosolic free Ca2+. J Biol Chem, 1985. 260(5): p. 2719-27.

366. Mellon, P.L., et al. Immortalization of hypothalamic GnRH neurons by genetically targeted tumorigenesis. Neuron, 1990. 5(1): p. 1-10.

367. Taylor, K.M., et al. Structure-function analysis of a novel member of the LIV-1 subfamily of zinc transporters, ZIP14. FEBS letters, 2005. 579(2): p. 427-32.

368. Oliveira-Martins, J.B., et al. Unexpected Tolerance of alpha-Cleavage of the Prion Protein to Sequence Variations. PLoS One, 2010. 5(2): p. e9107.

369. Yadavalli, R., et al. Calpain-dependent endoproteolytic cleavage of PrPSc modulates scrapie prion propagation. J Biol Chem, 2004. 279(21): p. 21948-56.

370. Ehsani, S., et al. The ZIP-Prion connection. Prion, 2012. 6(4): p. 317-21.

182

371. Leube, R.E. The topogenic fate of the polytopic transmembrane proteins, synaptophysin and connexin, is determined by their membrane-spanning domains. J Cell Sci, 1995. 108 ( Pt 3): p. 883-94.

372. Gallagher, M.J., et al. The GABAA receptor alpha1 subunit epilepsy mutation A322D inhibits transmembrane helix formation and causes proteasomal degradation. Proc Natl Acad Sci U S A, 2007. 104(32): p. 12999-3004.

373. Cormet-Boyaka, E., et al. Rescuing cystic fibrosis transmembrane conductance regulator (CFTR)-processing mutants by transcomplementation. Proc Natl Acad Sci U S A, 2004. 101(21): p. 8221-6.

374. Westaway, D., et al. Down-regulation of Shadoo in prion infections traces a pre-clinical event inversely related to PrP(Sc) accumulation. PLoS Pathog, 2011. 7(11): p. e1002391.

375. Pushie, M.J., et al. Prion protein expression level alters regional copper, iron and zinc content in the mouse brain. Metallomics : integrated biometal science, 2011. 3(2): p. 206- 14.

376. Aguzzi, A. and L. Rajendran. The transcellular spread of cytosolic amyloids, prions, and prionoids. Neuron, 2009. 64(6): p. 783-90.

377. Prusiner, S.B. The prion diseases. Sci Am, 1995. 272(1): p. 48-51, 54-7.

378. Bin, B.H., et al. Biochemical characterization of human ZIP13 protein: a homo-dimerized zinc transporter involved in the spondylocheiro dysplastic Ehlers-Danlos syndrome. J Biol Chem, 2011. 286(46): p. 40255-65.

379. Jiayu, W., et al. Mapping the interaction site of prion protein and Sho. Mol Biol Rep, 2010. 37(5): p. 2295-300.

380. Hornemann, S., et al. Prion protein library of recombinant constructs for structural biology. FEBS J, 2009. 276(8): p. 2359-67.

381. Premzl, M., et al. Evolution of vertebrate genes related to prion and Shadoo proteins-- clues from comparative genomic analysis. Mol Biol Evol, 2004. 21(12): p. 2210-31.

382. Christen, B., K. Wuthrich, and S. Hornemann. Putative prion protein from Fugu (Takifugu rubripes). The FEBS journal, 2008. 275(2): p. 263-70.

383. Watts, J.C., et al. Protease-resistant prions selectively decrease Shadoo protein. PLoS Pathog, 2011. 7(11): p. e1002382.

384. Chattopadhyay, M., et al. The octarepeat domain of the prion protein binds Cu(II) with three distinct coordination modes at pH 7.4. Journal of the American Chemical Society, 2005. 127(36): p. 12647-56.

385. Colby, D.W. and S.B. Prusiner. De novo generation of prion strains. Nature reviews. Microbiology, 2011. 9(11): p. 771-7.

183

386. Sweeting, B., et al. Structural factors underlying the species barrier and susceptibility to infection in prion disease. Biochemistry and cell biology = Biochimie et biologie cellulaire, 2010. 88(2): p. 195-202.

387. Weissmann, C., et al. Prions on the move. EMBO reports, 2011. 12(11): p. 1109-17.

184

Appendices

Supplemental Table 2.1: Complete quantitative interactome dataset of prion protein family in mouse neuroblastoma cells.

Supplemental Figure 3.1: Multiple full-length sequence alignment of selected mammalian and teleost ZIP and prion gene sequences.

Supplemental Table 3.1: Abridged quantitative interactome dataset from Supplemental Table 2.1 pertaining to prion protein family members, ZIP6, ZIP10 and actin.

Supplemental Figure 4.1: Weak similarities in the composition but not the order of regulatory elements in PrP and ZIP promoters. 1,000-bp regulatory sequences of prion and ZIP family members, consisting of 800 bps upstream (5’) and 200 bps downstream (3’) of the transcription start site, were pairwise compared. Depicted are the positions of TRANSFAC motifs which were shared by the respective promoter pairs within 250 bps of each other, with the transcription start site (indicated with a black vertical bar) serving as a reference for positional alignments. All numbers represent motif designators from the TRANSFAC database (the letter ‘R’ at the beginning of all TRANSFAC designators has been omitted). Please note that shared TRANSFAC elements occurring in identical order around the transcription start site can only be seen in the pairwise comparison of Sho and PrP promoter sequences.

Supplemental Figure 4.2: Visualization of TRANSFAC promoter motif analyses (Supplemental Figure 4.1) with nucleotide sequences.

185

Chapter 1: Progress in Neurobiology (Elsevier); License #: 2931171133736 (17 June 2012)

Chapter 2: PLoS Pathogens (Public Library of Science); “Open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.”

Chapter 3: PLoS ONE (Public Library of Science); “Open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.”

Chapter 4: PLoS ONE (Public Library of Science); “Open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.”

Chapter 5: Journal of Molecular Biology (Elsevier); License #: 2945401364374 (10 July 2012)

Chapter 6: Prion (Landes Bioscience); “Ownership of the copyright remains with the Authors, and provided that, when reproducing all or part of the Contribution, the Authors first acknowledge and reference publication in the Journal, the Authors retain the following non- exclusive rights: (a) Use of all or part of the Contribution (after publication in the Journal) in any book, article, thesis or dissertation written by the Authors without first requiring permission from Landes.”