Elucidating the Molecular Mechanisms Involved in Assembly/Folding and Targeting V-Atpase A-Subunit Isoforms to Their Functional Destinations

Elucidating the Molecular Mechanisms Involved in Assembly/Folding and Targeting V-ATPase a-subunit Isoforms to their Functional Destinations

Sally Esmail

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Faculty of Dentistry University of Toronto

Elucidating the Molecular Mechanisms Involved in Assembly/Folding and Targeting V-ATPase a-subunit Isoforms to their Functional Destinations

Sally Esmail Doctor of Philosophy

Faculty of Dentistry University of Toronto

2017 Abstract

Vacuolar H+-ATPases (V-ATPases) are proton pumps distributed across membranes of specialized cells and luminal compartments. V-ATPases in the plasma membrane of osteoclasts are responsible for acidifying the surface of bone, essential for bone resorption. V-ATPases in the plasma membrane of metastatic cells acidify the extracellular space to facilitate invasion. The

V-ATPase a subunit has four isoforms (a1-a4) that localize to distinct compartments. In invasive cancer, plasma membrane expression of a3 and a4 are required for metastasis while a3 is specific for the plasma membrane of osteoclasts. Thus, both isoforms are potential therapeutic targets. Sequences of a isoforms reveal putative N-glycosylation sites within extracellular loop II

(ELII). Upon PNGase F and Endo H treatment, all a isoforms showed faster mobility on SDS-

PAGE indicating the presence of N-linked oligosaccharide. Using site-directed mutagenesis, I showed that deglycosylated a1–a4 had shorter half-lifes, more rapid proteasomal degradation, endoplasmic reticulum (ER) retention, defective Golgi trafficking, and an inability to associate with ER assembly factor, VMA21. In addition, deglycosylated a4 showed defective cell-surface iii

expression and assembly. Cutis laxa type II, osteopetrosis and distal renal tubular acidosis

(dRTA) result from mutations within the a2, a3 and a4 subunits, respectively. To further map critical domains essential for V-ATPase structure and function, I studied human disease-causing missense mutations that affect conserved residues of a isoforms, specifically: a2P405L, a4R449H and a4G820R. a4R449H and a2P405L were unstable and degraded by the proteasomal pathway. The data also indicated that a2-P405 is required for Golgi trafficking while a4-R449 is essential for ER exit and cell-surface expression. a4R449H shows increased association with the assembly factor,

VMA21. Molecular modeling of a4 predicts a4G820R would interfere with proton translocation through the cytoplasmic half channel formed by the a subunit. This work enhances our knowledge of a isoforms structure and informs possible therapeutic interventions against cancer metastasis and lytic bone diseases. iv

Acknowledgments

First, I would like to express my sincere gratitude to my supervisor, Prof. Morris F Manolson for giving me the chance to join his lab and the tremendous support, motivation and guidance during the course of my PhD program. His guidance and scientific expertise greatly helped me in all aspects of my PhD research and writing of this thesis. I have been extremely lucky to have a supervisor who cared so much about my research. I would also like to extend my gratitude to my co-supervisor Prof. Reinhart A.F. Reithmeier; his expertise and deep knowledge of the research subject inspired me. He helped me to conceptualize and analyze all the data presented in this thesis. His careful and critical editing positively impacted the production of this thesis. I feel exceedingly lucky and could not have imagined better advisors.

Completing this work would have been more difficult were it not for the support of Dr. Norbert Kartner. His critical thinking and guidance helped me in designing all the experiments and in interpretation of data. I am deeply thankful to his scientific expertise and professional academic writing skills, without which I would not have publishable research.

I would like to thank my thesis committee member Prof. Boris Hinz for his constructive criticism, insightful comments and encouragement, but also for the hard questions which encouraged me to expand my horizons from various research perspectives.

I must also thank all lab members of the Manolson’s lab and the Reithmeier’s lab, with special thanks to Dr. Yeqi Yao and Jing Li for all their help and technical support.

Finally, I would like to thank my family and friends who always believed in me and encouraged me to come over to Canada to pursue my research path and remained patient and positive. v

Table of Contents

Acknowledgments...... iv

Table of Contents ...... iv

Original Contribution by author ...... ix

List of Figures and Tables...... xi

Abbreviations ...... xiii

1. Introduction–Thesis Rationale ...... 1

2. Literature Review...... 3

2.1. Recent Advances in the Understanding of Membrane Protein N-glycosylation Structure, Function, and Regulation in Health and Disease ...... 3

2.1.1. Biosynthesis and elongation of mammalian N-linked glycan ...... 3

2.1.2. Role of N-glycans in protein folding, stability, and quality control in the ER ...... 6

2.1.3. Regulation of mammalian N-glycosylation ...... 9

2.1.4. N-glycosylation in cell biology, signal transduction, and immunity ...... 10

2.1.5. N-glycosylation in diseases...... 11

2.1.6. N-glycosylation as a potential therapeutic target ...... 12

2.1.7. Future prospectives: Advancing the knowledge of N-glycan 3D structure ...... 13

2.2. V-ATPase structure, function, regulation and drug targeting ...... 14

2.2.1. V-ATPase Function ...... 14

2.2.2 V-ATPases Structure ...... 16

2.2.3 V-ATPases subunits isoforms ...... 17

2.2.4. Regulation of V-ATPase assembly and trafficking ...... 20

2.2.5 Involvement of V-ATPases in human diseases ...... 20

2.2.6 V-ATPase a subunit topology and atomic models ...... 21

2.2.7 V-ATPase inhibitors and their limitation ...... 27 vi

2.2.8 V-ATPase a subunit as a potential therapeutic target ...... 28

3. Research Objectives and Hypotheses ...... 29

Hypotheses...... 29

Central Objective ...... 30

Aim 1 ...... 30

Aim 2 ...... 30

Significance...... 30

+ 4. N-Linked Glycosylation Is Required for Vacuolar H -ATPase (V-ATPase) a4 Subunit Stability, Assembly , and Cell Surface Expression ...... 31

4.1 Abstract ...... 32

4.2. Introduction ...... 32

4.3. Materials and Methods ...... 34

4.4. Results ...... 39

4.4.1. The Human V-ATPase a4 subunit is N-glycosylated...... 39

4.4.2. N-glycosylation is required for a4 stability ...... 40

4.4.3. Un-glycosylated a4 is degraded in proteasomal and lysosomal pathways ...... 41

4.4.4. Un-glycosylated a4 is mostly retained in the ER ...... 42

4.4.5. Un-glycosylated a4N489D is unable to assemble into the V-ATPase complex...... 42

4.4.6. Defective plasma membrane trafficking of un-gycosylated a4 ...... 43

4.5. Discussion ...... 44

4.5.1. The V-ATPase a4 subunit isoform is N-glycosylated ...... 44

4.5.2. U-glycosylated a4N489D is retained in the ER and is not assembled into V1V0 ....45

4.5.3. Un-glycosylated a4N489D undergoes both proteasomal and lysosomal degradation ....45

5. N-linked Glycosylation of a Subunit Isoforms is Critical for Vertebrate Vacuolar H+- ATPase (V-ATPase) Biosynthesis ...... 56 vii

5.1 Abstract ...... 57

5.2. Introduction ...... 58

5.3. Materials and Methods ...... 59

5.4. Results ...... 64

5.4.1. N-glycosylation of V-ATPase a1, a2 and a3 subunits ...... 64

5.4.2. N-glycosylation requirement for stability of a subunits ...... 66

5.4.3. N-glycosylation requirement for ER exit of a subunits ...... 67

5.4.4. N-glycosylation requirement for association of a subunits with ER-resident V- ATPase assembly factor, VMA21 ...... 68

5.5. Discussion ...... 69

5.5.1. All human V-ATPase a subunit isoforms are N-glycosylated ...... 69

5.5.2. N-glycosylation of a subunits contributes to their stability ...... 71

5.5.3. N-glycosylation of a subunits is required for ER exit ...... 72

5.5.4. N-glycosylation is required for association with assembly factor, VMA21 ...... 72

5.6. Conclusion ...... 73

6. Molecular Mechanisms of Cutis Laxa and Distal Renal Tubular Acidosis-Causing Mutations in V-ATPase a subunits, ATP6V0A2 and ATP6V0A4 ...... 87

6.1. Abstract ...... 88

6.2. Introduction ...... 88

6.3. Materials and Methods ...... 90

6.4. Results ...... 96

6.4.1. Amino acid residues a2 P405, a4 R449 AND a4 G820 are highly conserved ...... 96

6.4.2. Glycosylation and stability of cutis laxa mutant, a2P405L ...... 96

6.4.3. Glycosylation and stability of dRTA mutants, a4R449H and a4G820R ...... 97

6.4.4. Unstable a2P405L is degraded in proteasome and lysosome while a4R449H is degraded in proteasome ...... 98

6.4.5. a2 P405 is required for Golgi trafficking, and a4 R449 for ER exit ...... 98 viii

6.4.6. Defective cell surface expression of a4R449H ...... 99

6.4.7. a4R449H shows increased association with VMA21 ...... 100

6.4.8. a4 G820 resides within the putative proton pathway...... 101

6.5. Discussion ...... 102

6.5.1. a2 P405, a4 R449 and a4 G820 are conserved and crucial for function ...... 102

6.5.2. Human a2P405L and a4R449H are N-glycosylated, but are unstable and degraded in the proteasomal pathway ...... 103

6.5.3. a4 R449 is required for ER exit and a2 P405 is required for Golgi trafficking ....103

6.5.4. a4 R449H is crucial for cell surface expression and a–V0 association...... 104

6.5.5. a4 G820 is a functional residues and resides in the putative proton pathway ...... 104

6.6. Conclusion and Future Perspective ...... 105

7. General Discussion and Future Prespectives ...... 119

7.1. General discussion ...... 119

7.1.1 Role of a subunit N-glycosylation ...... 119

7.1.2. Disease causing mutations pinpoint functional domains within a isoforms ...... 121

7.1.3. Refining the mechanism of Human V-ATPase assembly and membrane targeting ...... 122

7.2. Future perspectives...... 123

7.2.1. Advance the knowledge of mammalian V-ATPase assembly ...... 123

7.2.2. Role of N-glycosylation on V-ATPase activity ...... 125

7.2.3. High throughput screening for plasma membrane V-ATPases specific inhibitor(s)...... 126

8. References ...... 127

ORIGINAL CONTRIBUTION BY AUTHOR

Publications and submitted manuscripts resulting from this PhD thesis: 1. Esmail S, Yao Y, Kartner N, Li J, Reithmeier R.A.F., and Manolson M.F. N-Linked

Glycosylation Is Required for Vacuolar H+-ATPase (V-ATPase) a4 Subunit Stability,

Assembly, and Cell Surface Expression. The Journal of Cellular Biochemistry,

117(12):2757-276, 2016.

Esmail S performed all experiments and prepared constructs used in the study. Yao Y.

helped in designing constructs used in the study and provided technical expertise. Li J.

helped in designing initial transfection and western blotting experiments and provided

technical expertise. Esmail S., Kartner N., Reithmeier R.A.F. and Manolson M.F.

conceptualized, planned and analyzed experimental work. Esmail S. and Kartner N. wrote

the manuscript and prepared figures.

2. Esmail S, Kartner N, Yao Y, Kim J.W., Reithmeier R.A.F., and Manolson M.F. N-linked

glycosylation of a subunit isoforms is critical for vertebrate vacuolar H+-ATPase (V-

ATPase) biosynthesis.

The Journal of Cellular Biochemistry, 2017 Jun 29. doi: 10.1002/jcb.26250. [Epub ahead of

print]

Esmail S. performed all experiments except those represented in Fig. 5.2A and B, which

were performed by Joo W.K. Yao Y. prepared constructs used in the study and provided

technical expertise. Esmail S., Kartner N., Reithmeier R.A.F and Manolson M.F. x

conceptualized, planned and analyzed experimental work. Esmail S. and Kartner N. wrote

the manuscript and prepared figures.

3. Esmail S, Kartner N, Yao Y, Kim J.W., Reithmeier R.A.F., and Manolson M.F. Molecular

Mechanisms of Cutis Laxa and Distal Renal Tubular Acidosis-Causing Mutations in V-

ATPase a subunits, ATP6V0A2 and ATP6V0A4

Manuscript in preparation and will be submitted to Journal of Biological Chemistry

Esmail S., Kartner N., Reithmeier R.A.F. and Manolson M.F. conceptualized, planned and

analyzed experimental work. Esmail S. performed all experiments except those represented

in Fig. 6.1C, which were performed by Joo W.K. Esmail S. and Yao Y. prepared constructs

used in the study and provided technical expertise. Esmail S. and Kartner N. wrote the

manuscript and prepared figures.

4. Esmail S and Manolson M.F. Recent Advances in the Understanding of Membrane Protein

N-glycosylation Structure, Function, and Regulation in Health and Disease.

Review will be submitted to Trends in Biomedical Sciences Journal

LIST OF FIGURES AND TABLES

FIGURES

Figure 2.1: N-linked Protein Glycosylation in the ER...... 4

Figure 2.2: Processing of N-glycans in Golgi ...... 8

Figure 2.3: Structure of yeast V-ATPase ...... 16

Figure 2.4: Human and mouse CTa of a subunit, and yeast Vph1p, alignments ………….22

Figure 2.5: Cartoon of secondary structure of yeast a subunit...... 27

Figure 4.1: Human a4 is N-glycosylated...... 47

Figure 4.2: Glycosylation is required for a4 subunit stability ...... 48

Figure 4.3: a4 is degraded in both proteasomal and lysosomal degradation pathways ...... 49

Figure 4.4: Localization of WT a4 and a4N489D after transient transfection………………………………………………………………………..………..50

Figure 4.5: Un-glycosylated a4 is mostly retained in the ER ...... 51

Figure 4.6: N-glycosylation is required for V1–V0 assembly ...... 52

Figure 4.7: Glycosylation is required for a4 cell-surface expression ...... 53

Figure 4.8: Surface biotinylation of HEK 293 cells transfected with a4 or un-glycosylated a4N489D………………………………..…………………………………………………….54

Figure 5.1: N-glycosylation sequons of a subunit isoforms from various species ...... 74

Figure 5.2: Human a1, a2 and a3 subunit isoforms are N-glycosylated ...... 75

Figure 5.3: Glycosylation is required for a1, a2 and a3 stability ...... 76 xii

Figure 5.4: Subunits a1, a2 and a3 are degraded in the proteasomal pathway……………………………………………………………………………………..78

Figure 5.5: Unglycosylated a1, a2 and a3 are mostly retained in the ER ...... 79

Figure 5.6: WT a1, a2 and a3 traffic to Golgi ...... 80

Figure 5.7: N-glycosylation is required for a subunit association with V0 assembly factor, VMA21...... 81

Figure 6.1: Glycosylation and stability of a2P405L, a4R449H and a4G820R ...... 106

Figure 6.2: a2P405L and a4R449H are degraded in the proteasomal pathway ...... 108

Figure 6.3: Localization of mutant a subunit proteins in the secretory pathway ...... 109

Figure 6.4: Defective cell-surface expression of a4R449H …………………………………110

Figure 6.5: Association of a subunit with V-ATPase assembly chaperone, VMA21, and V0 marker, ATP6V0B1 ...... 112

Figure 6.6: a4 G820 resides within the putative proton translocation pathway ...... 113

Figure 6.7: Model for human a4 trafficking in the secretory pathway and to the plasma membrane ...... 115

TABLES

Table 2.1: Mammalian V-ATPase subunit isoforms and their functions ...... 18

Table 2.2: Commonly used V-ATPases inhibitors ...... 28

Table 5.1: Significance of half-life comparisons for data in figures 5.3 and 5.4 ..... 82

Table 6.1: Significance of half-life comparisons for data in figures 6.1 and 6.2 ... 116

xiii

ABBREVIATIONS

ASBT Apical Sodium–Bile acid Transporter

ATP Adenosine Triphosphate

BRET bio-luminescent resonance energy transfer

Cav3 calcium channel

CDG congenital disorder of glycosylation

CFTR Cystic Fibrosis Transmembrane-conductance Regulator

CHX Cycloheximide

CNS Central Nervous System

Cryo-EM Cryo-electron microscopy

CTa C-terminal domain of a subunit

CTD C-Terminal domain dRTA distal Renal tubular acidosis

EL2 or ELII External cellular loop 2

ELISA enzyme-linked immunosorbent assay

Endo H Endoglycosidase H

ER Endoplasmic Reticulum

ERAD Endoplasmic Reticulum Associated Degradation

FRET fluorescent resonance energy transfer

GTP Guanosine triphosphate

HA tag Human influenza hemagglutinin

H-K-ATPase Hydrogen potassium ATPase mTORC1 mammalian target of rapamycin complex 1

NTa N-terminal domain of a subunit xiv

OST Oligosaccharyltransferase H-K-ATPase

PA1b Pea albumin 1b

PNGase F Peptide N-Glycosidase F

RA Rheumatoid arthritis

RANKL Receptor activator of nuclear factor kappa-B ligand

1. Introduction–Thesis Rationale

Maintenance of cytosolic pH, the pH within cellular organelles and the extracellular micro- environment is essential for an optimal array of cellular and physiological functions (1). Indeed, lack of cellular pH homeostasis is a key characteristic of several genetic and chronic diseases (2). V-ATPases, ATP-dependent multi-subunit proton pumps, are a major player in acidifying the lumen of cellular organelles and extracellular matrix of some specialized cells in both physiological and pathological conditions (3,4).

The V-ATPase complex is composed of at least 14 protein subunits with new subunits still being discovered and disputed (5,6). In this thesis, I focus on one V-ATPase subunit, the 100 kDa a subunit. In mammals, the a subunit has four isoforms (a1–a4), with the different isoforms thought to be responsible for targeting the V-ATPase isocomplex to different functional destinations. The a1 and a2 isoforms target V-ATPases to intracellular compartments, a3 targets V-ATPases to the osteoclast ruffled border, while a4 targets V-ATPases to apical membrane in renal intercalated cells (3). Point mutations in human a2, a3 and a4 are associated with cutis laxa, osteopetrosis, and distal renal tubular acidosis (dRTA), respectively. In cancer cells, increased plasma membrane expression of both a3 and a4 have been linked to creation of a more acidic micro-environment and an increase in metastasis (7). Compounds that specifically target these a isoforms and their interactions have been selected and tested for their therapeutic ability to prevent bone loss, cancer metastasis and even as potential pesticides (6,8-10). Despite the clinical importance of the V-ATPase a isoforms, little is known about their structural details. My thesis seeks to address this issue. Here I pinpointed some of the functional domains within human a subunit isoforms with the ultimate goal of informing drug discovery to uniquely target the V-ATPases implicated in diseases without altering their normal physiological functions.

A major focus of my research is on the role of N-glycosylation of a subunits on the functional expression of the V-ATPase. Based on sequence analysis I found that putative N-glycosylation sites within the V-ATPase a subunits are highly conserved in all four isoforms. This observation led me to hypothesize that N-glycosylation of the a subunit is important for proper protein folding and the assembly and/or targeting of the V-ATPase complex to their functional destination. To test this hypothesis, I have determined whether all four V-ATPase a subunit isoforms are N-glycosylated when expressed in HEK-293 cells. Also, I have determined the 2

effect of eliminating the glycosylation site(s) on subunit stability, targeting, assembly and cell surface expression. The findings of these studies are described in chapters four and five.

As mentioned above, human point mutations in a2, a3 and a4 are associated with cutis laxa, osteopetrosis, and distal renal tubular acidosis (dRTA), respectively. This led me to hypothesize that human disease-causing missense mutations within a subunits will identify critical domains essential for V-ATPase targeting, activity and/or regulation. To test this hypothesis, I have determined the effect of the human a subunit missense mutation: a2P405L (Cutis laxa), a4R449H and a4G820R (renal tubular acidosis) with respect to subunit stability, glycosylation, assembly, localization in the secretory pathway and cell surface expression using transfected cells. The findings of this study are described in chapter six.

The literature review is divided into two major sections. The first section is a general background about the role of N-glycosylation in membrane protein structure and function, and N- glycosylation as a potential drug target. The second section is a review of recent advances in V- ATPases structure and function in health and disease.

This thesis ends with a general discussion (Chapter seven) that illustrates how my research has improved our understanding of V-ATPase a subunit structure and how it will potentially inform drug discovery of novel V-ATPases inhibitors. Finally, the thesis will end with a brief discussion on the future directions for research into V-ATPases. 3

2. Literature Review 2.1. Recent Advances in the Understanding of Membrane Protein N-glycosylation Structure, Function, and Regulation in Health and Disease

Asparagine (N)-linked glycosylation of membrane proteins is a common co-translational modification that is formed by the covalent attachment of a common high mannose oligosaccharide onto particular asparagine residues of polypeptide chains, followed by post- translational processing to complex structures (11). This protein modification is conserved from prokaryotes to eukaryotes (12-14). N-glycosylation can be summarized in three steps: 1, the formation of the lipid-linked oligosaccharide donor (LLO); 2, the co-translational transfer of the glycan onto Asn-X-Ser/Thr (where X represents any amino acid except Pro) nascent polypeptide

chain and 3, processing of the Glc3Man9GlcNAc2 oligosaccharide chain in the endoplasmic reticulum (ER) and Golgi. This review discusses factors that influence the occupancy of N- glycosylation sites in membrane glycoproteins and on the involvement of mammalian N-glycans in the molecular and cellular mechanisms that control health and disease.

1.1.1 2.1.1. Biosynthesis and elongation of mammalian N-linked glycan

2.1.1.1 N-glycans biosynthesis in the ER

The mechanism of N-glycosylation in mammals is summarized in Figure 2.1. Briefly, N- glycosylation starts with the attachment of glycan to LLO on the cytoplasmic side of the ER.

This step is followed by the synthesis of DolPPGlcNAc2-Man5 heptasaccharide (M5-DLO) that is formed by the transfer of N-acetylglucosamine (GlcNAc)-1-phosphate from uridine diphosphate N-acetylglucosamine (UDP-GlcNAc) onto a dolichol phosphate (DolP) carrier in a reaction catalyzed by the asparagine-linked glycosylation phosphotransferase-7 (Alg7) to yield DolPP-GlcNAc. Within the following reactions, five GTP-activated mannoses are added in a sequential manner by three specific glycotransferases to build the branched heptasaccharide. These reactions all occur on the cytosolic side of the ER membrane. In the second part of the assembly process, specific ER membrane proteins, flippases, facilitate flip-flop or translocation 4

of lipid M5-DLO from the cytoplasmic side to luminal side of the ER (12). Once it faces the ER lumen, the flipped lipid-linked heptasaccharide is further elongated by four mannose and three glucose residues, forming the final donor Glc3Man9GlcNAc2-P-P-dolichol (12). The oligosaccharyltransferase (OST) complex then transfers the completely assembled

Glc3Man9GlcNAc2-P-P-dolichol glycan onto the N-glycosylation sequons “N-X-T/S” of a nascent polypeptide as it translocates into the ER lumen. Factors that modulate N-X-T/S sequon occupancy are discussed in the next section.

Figure 2.1: N-linked Protein Glycosylation in the ER. The synthesis of M5-DLO is initiated by the transfer of GlcNAc2-Man5 heptasaccharide to the lipid anchor, dolichol phosphate, on the cytosolic side of the ER. The M5-DLO then flips from the cytosolic side to the lumenal side, a process that mediated by a flippase enzyme. Sequentially, the Glc3Man9GlcNAc2-P-P-dolichol was formed by further addition of four mannose and three glucose residues. Finally, the OST complex catalyzes the transfer and assembly of Glc3Man9GlcNAc2-P-P-dolichol onto the N-glycosylation sequon N-X-S/T of the nascent polypeptide chain. Green circles represent mannose; blue squares, N-acetylglucosamine (GlcNAc) and red hexagons, glucose residues. Modified from (15).

2.1.1.2 N-glycosylation occupancy is modulated by local amino acid sequences

In order to increase the affinity between asparagine’s side chain and Glc3Man9GlcNAc2 oligosaccharide; a specific conformation of the N-glycosylation acceptor peptide is required (14). Two studies have estimated that 2/3 of all putative N-glycosylation sites on glycoproteins are occupied (16,17). These estimates were based on analysis of 506 and 749 well-characterized glycoproteins in the Protein Data Bank crystallographic database and in the SWISS-PROT database, respectively. Analyses showed a greater preference in glycan occupancy for N-X-T sites over N-X-S sites (16,18,19). Furthermore, studies reveal that the second position (X) in the N-glycosylation sequon is an important determinant of glycosylation efficiency (20). Site directed mutagenesis studies of the X amino acid revealed a varied degree of glycosylation efficiency (21); interestingly, amino acids like Asp, Glu, Trp and Leu showed a significant reduction in glycosylation, whereas proline resulted in complete inhibition of N-glycosylation (19-22). Other studies examining the effect of amino acids adjacent to utilized sequons revealed an increase in aromatic and hydrophobic groups and a decrease in the presence of basic groups (16,18,23). Given the hydrophilic nature of oligosaccharides, these studies suggested that N- linked glycans could be involved in covering/stabilizing hydrophobic domains on protein surfaces.

2.1.1.3. Secondary structure around the glycosylated N-X-S/T sequon

N-glycosylation can occur on all types of secondary structures, but favors turns and bends (16). The tendency of N-glycosylation occurring at or near the point of change in the peptide secondary structure is high (14,23) suggesting that N-glycosylation may promote protein folding and stabilize peptide conformation. There is also a correlation between the degrees of accessibility of potential N-glycosylation sites on the protein surface, such as in turns, and the occupancy of such sites. It is less likely for N-glycosylation to occur at buried or very exposed N-X-S/T sequon (16,19).

2.1.2. Role of N-glycans in protein folding, stability, and quality control in the ER

2.1.2.1. N-glycans enhance protein stability

N-linked glycans play both direct and indirect roles in newly-synthesized glycoproteins folding and assembly in the secretory pathway to help ensure that only properly-folded proteins are delivered from the ER. The properties of the nascent polypeptide chain are clearly modified by the addition of hydrophilic glycan residues (24). N-glycans also stabilize the conformation of the secondary structure of glycoproteins by modulating the folding of residues around the N-X-S/T sequon. An interaction of N-glycan with the polypeptide chain has been shown to induce a β-turn structure (25). Several studies have shown how N-glycosylation is required for protein stability and folding (26-28). Our own work in chapter four has shown that the glycosylated Vacuolar (V) H+-ATPases a4-subunit is less stable when the glycosylation site is eliminated through site- directed mutagenesis (27). Furthermore, N-glycosylation enhanced overall protein stability and resistance to proteolysis when compared to the un-glycosylated counterparts (27,28). Another study showed that the glycosylated gastric H-K-ATPase β-subunit of the apical pump is more resistant to proteolysis by both trypsin and proteinase K than its N-glycosylation-deficient mutants (29). Similar results were obtained with the bile acid transporter, ASBT, and the − − Cl /HCO3 exchanger, SLC26A3, where their un-glycosylated forms showed both defective function as well an increased sensitivity to proteolysis (30,31).

2.1.2.2. N-glycans promote folding and quality control in the ER

N-glycans also assist folding and quality control in the ER by serving as recognition or sorting signals that allow glycoproteins to interact with a variety of chaperones, glycosidases, and glycosyltransferases (24). Briefly, newly formed N-linked glycoproteins in the ER may bind to either calnexin or calreticulin, lectin-binding chaperones, which in turn facilitate the binding with co-chaperones that assist protein folding, disulfide bond formation and ER exit (32,33). If the protein is properly folded, it leaves the ER and traffics to the Golgi (32). Misfolded glycoproteins can be reglycosylated by a luminal ER glucosyltransferase; the misfolded glycosylated proteins may bind again to lectin chaperones and other co-chaperones to achieve a 7

proper folded state. Misfolded proteins are retained in the ER and activate the ER-associated degradation (ERAD) pathway to prevent accumulation of defective proteins (33-36). The misfolded proteins are retrograde translocated into the cytosol where they are degraded by the proteasome.

In glycoproteins with multiple N-glycosylation sites, individual N-glycans are not always of equal importance in mediating folding, ER retention or exit. Our work on the V-ATPases a subunit isoforms (see chapter five) have revealed two N-glycosylation sites within the second extracellular loop domain of the a2 and a3 isoforms at Asn484+Asn505 and Asn483+Asn503, respectively. Using site-directed mutagenesis, we were able to demonstrate that the a2-Asn484 and a3-Asn483 glycosylation sites were more essential than the a2-Asn505 and a3-Asn503 with respect to both ER exit and protein stability (see chapter five). The β-subunit of the multi-subunit H-K-ATPase expressed in LLC-PK1 cells has seven N-glycosylation sites and mutation of some glycosylation sites alter its exit from the ER, while mutations of others sites have little to no impact. The glycosylated β-subunit is normally targeted to the plasma membrane but when all seven N-glycosylation were removed, the result was ER retention (29). These findings provide evidence that N-glycosylation of multimeric proteins can be important for subunit folding, assembly and ER exit. Once a polypeptide is glycosylated within the ER, it traffics through the Golgi apparatus prior to arriving at its functional destination. During this passage, the N-linked oligosaccharide may be subjected to further modification by the actions of various glycosidases and/or glycosyltransferases as discussed in the next section (37).

2.1.2.3. Processing and elongation of N-glycan in the Golgi

A few glycoproteins containing high-mannose N-glycans traffic through Golgi to the plasma membrane without further processing (see Figure 2.2), however, most N-glycans are modified further in the Golgi. Golgi mannosidases and multiple glycosyltransferases catalyze elongation and branching of the glycan, producing hybrid or complex N-glycans (37). Thus, individual branches of the mature N-glycan can vary in both their length and their carbohydrate composition based on their proposed function (38). Individual branches are elongated by the addition of galactose and terminated by sialic acid residues catalyzed by galactosyltransferases 8

and sialyltransferases, respectively (39). Processing and branching are modulated by the location with certain N-glycans having more accessibility to Golgi glycosyltransferases and therefore having a greater probability to be processed to complex N-glycans than others. As a result, particular glycosylation sites in a protein may be occupied by complex N-glycans and other sites by high-mannose or hybrid N-glycans (40,41). As described in the next section, the processing of N-glycans is also dependent on the expression levels of enzymes in different cell types.

Figure 2.2: Processing of N-glycans in Golgi. Golgi mannosidases and glycosyltransferases catalyze branching and elongation of high mannose producing diverse carbohydrate structures, which can be categorized as hybrid or complex N-glycans. Green circles represent mannose residues, blue squares represent N-acetylglucosamine (GlcNAc), yellow spheres represent galactose residues and magenta hexagons represent N-acetylneuraminic (sialic) acid residues.

2.1.3. Regulation of mammalian N-glycosylation

N-glycosylation elongation, branching and glycosidase enzymes can regulate glycan variation. Structural variations in glycans at the cell surface produce numerous biomarkers, some of which correlate with differentiation, recognition, cell activation, and diseases (42,43). Regulation of glycosylation in the secretory pathway is controlled by modulating glycosyltransferase or glycosidase expression profile and their function as well as their accessibility to substrates in a cell-specific manner. Gene transcription of glycosyltransferases and glycosidases are cell-type specific and their levels are regulated based on cellular needs. Modulations of RNA expression profiles of these glycosylation enzymes have an impact on glycan formation. Large-scale microarray data on the transcription levels of various glycosylation enzymes has started to provide information about types and levels of various glycans in different cells (44-46). Profiling of N-glycosylation enzymes has the potential to pinpoint the role of cellular glycosylation in health and disease.

More information on the regulation of glycosyltransferase and glycosidase gene expression (the cellular glycome) is needed to fully understand the impact on their regulation on various cellular process (47). As mentioned above, the regulation of glycosylation enzymes, glycosyltransferases and glycosidases, could be done on a transcription level. In addition, there is both posttranscriptional and post-translational regulation. For instance, some of the glycosylation enzymes themselves must be glycosylated or phosphorylated to be properly targeted and fully active (48,49). Major alteration in the glycome is mediated by the loss of some chaperones and/or other factors that alter glycosyltransferase trafficking between the ER and Golgi (50). Availability of glycosyltransferases and glycosidases and/or availability of their specific substrate will modulate the glycosylation profile. Towards a therapeutic strategy, controlling substrate/enzyme accessibility could modulate glycosylation status in diseases where normal N-glycosylation profile is altered.

2.1.4. N-glycosylation in cell biology, signal transduction, and immunity

1.1.1.1 2.1.4.1. Cell adhesion and self/nonself recognition

It has been reported that cell-cell adhesion and aggregation induced by lectin binding may have contributed to cell-cell recognition in the evolution of the earliest metazoans (51). This cell adhesion system is regulated by specific cell surfaces, including the endothelium of the vasculature and on most leukocytes, thereby contributing to leukocyte trafficking responses essential in immune-system homeostasis, hematopoiesis, and inflammation (52). N-glycosylation also modulates cell-cell adhesion in early mammalian embryos with the loss of specific glycans disrupting fertilization (53).

In mammalian cells, lectins are used for self/non-self recognition; a process mediated by activation of Toll-like receptors upon binding of the invading organism. The ability of mammalian lectins to recognize glycans from divergent organisms such as bacteria, yeast, and

mammals highlights a mechanism of self/non-self recognition (54). Lectins are highly expressed on mammalian cells involved in innate immunity, and several bind to glycans specifically expressed among phylogenetically older organisms. Thus, defective glycosylation can alter self/nonself recognition and could lead to autoimmune diseases. Mammalian glycans serve as self-antigens and facilitate the development of cellular immune response to infection, and participate in binding to major histocompatibility complexes (54).

2.1.4.2 Receptor activation, endocytosis and drug delivery

N-glycosylation also modulates interactions of receptors and ligands, regulatory molecules, and distinct membrane domains of intact cells, thereby altering signal transduction. For example, fibroblast growth-factor receptors bind to specific heparan sulfate glycosaminoglycans on some proteoglycans, and this interaction facilitates the co-presentation of ligand monomers to achieve receptor dimerization and activation (55). Multiple receptor systems appear to employ glycosaminoglycans on proteoglycans to modulate receptor activation in regulating morphogenesis and organogenesis signaling during early ontogeny. Studies have found that endocytosis of cell-surface glycoproteins is modulated by glycoproteins produced in Golgi, thereby controlling cell surface expression of receptor and hence thresholds for cell signaling 11

(56). Different cell types appear to use distinct glycans to alter rates of endocytosis (57,58). Additionally, glycans have gained attention as a target for drug delivery using specific glycan recognition mechanisms. Glycosylation has applied to selective deliver therapies to specific tissues based on receptor-mediated endocytosis (59,60).

2.1.5. N-glycosylation in diseases

2.1.5.1. Disorder in protein N-glycosylation leads to neurological abnormalities

Congenital disorder of glycosylation syndrome (CDGs) is a rare but severe disorder associated with abnormalities in either the assembly or the processing of both N- and O-glycosylation. CDGs due to defects in O-glycosylation will not be reviewed here. While CDGs are multisystem disorders, the central nervous system (CNS) is often the main site of clinical involvement (61). Classic disorders of N-glycosylation are due to enzyme defects either in the assembly or subsequent trimming stage of the oligosaccharides. Defects in the N-glycosylation pathway that results in hypo-glycosylation lead to CDG-Type I (62). CDG-Type I is a severe autosomal recessive disease, often with onset in early infancy and significant involvement of the CNS that has been recognized as a multisystem disorder (62). Neurological symptoms include seizures, cognitive impairment and developmental disabilities (63).

2.1.5.2. Change in protein N-glycosylation: Potential tumors markers

Glycomics and glycoproteomics are promising fields for the discovery of sensitive and specific glycan biomarkers for early detection of cancer, evaluation of therapeutic efficacy of cancer treatment, and for assessment of prognosis (64,65). Ligand-receptor interactions may be affected by aberrant expression of glycans and/or any change/truncation in glycan structure. Changes in the expression of glycans might interfere with the regulation of cell proliferation, adhesion and migration (66,67). Large N-glycans are a unique feature in many cancerous cells, thus changes in aberrant glycosylation as well as the altered expression of glycosyltransferases and glycosidases 12

are used as cancer biomarkers (68,69). In most cancers, sialylation and fucosylation are modified (70,71).

2.1.5.3. N-glycosylation changes in rheumatoid diseases

Rheumatoid arthritis (RA) is the most common autoimmune arthritis and is associated with a defect in a galactosyltransferase (GTase) resulting in changes to the galactosylation of immunoglobulin G (72). It is hypothesized that disruptions in glycosylation homeostasis is associated with the pathogenesis of RA possibly involving a generation of unique GTase isoenzymes that differ from those in healthy individuals (72,73). These changes are not unique to RA, but there may be a unique signature of sugar changes associated with a number of other rheumatic diseases. This change is referred to as “sugar printing the rheumatic diseases” (74), a concept that may be useful diagnostically and therapeutically.

2.1.6. N-glycosylation as a potential therapeutic target

As described above, abnormal glycoproteins may be involved in CGDs, cancer progression and metastasis, and RA. Therefore, modification of the glycosylation pathway could be a strategy for the prevention of metastasis and RA. Numbers of inhibitors with broad specificity have been used to eliminate N-glycosylation. For example, Tunicamycin is used to inhibit the assembly of dolichol-pp-GlcNAC while plant alkaloids are used to inhibit glycosidases and glycotransferases (reviewed in (75)). However, specific inhibitors would be required to modulate N-glycosylation enzymes which are implicated in pathological conditions (76,77). Such inhibitors might be found through screening chemical libraries, followed by multiple in vitro validations of these inhibitors and finally, elucidating the in vivo effect of the potential inhibitors on inhibiting the growth of solid tumors. The potential of using the same approach for RA should be explored using animal models. Promising inhibitors must be specific for N-glycosylation in pathological conditions without altering normal physiological function – a continuing challenge. Targeting glycoproteins on the surface of viruses using carbohydrate binding agents (CBAs) have led to the development of powerful anti-virals (78). On the other hand, researchers are investigating possible enzyme replacement therapy for the treatment of CDG (79-81). 13

2.1.7. Future prospectives: Advancing the knowledge of N-glycan 3D structure

Resolving 3D structures of disease-implicated glycans is a necessity to target them in pathological conditions. However, as a result of the heterogeneity of chemical structure, mobility and conformation of glycans, characterization of glycoproteins structure by X-ray crystallography is challenging. Nevertheless, technical advances in glycoprotein crystallography have increased 3D structural information (19,82). The glycan sequence at each glycosylation site should be analyzed in detail by mass spectrometry, coupled with liquid chromatography (83). Other techniques, such as molecular dynamic simulation and NMR, are needed to help join the discontinuous snapshots derived from X-ray studies (84,85). In addition, statistical analysis of “snapshots” of glycoproteins can provide clues to understanding their structure and dynamics. Tailoring technical procedures to resolve the 3D structure of glycan will directly enhance the understanding of glycan structure/function relationships, and in turn, will advance the knowledge of the physiological and/or pathological role of protein N-glycosylation.

Research on the function, structure, and regulation of mammalian glycosylation has demonstrated that glycans are involved in multiple disciplines spanning cell biology, immunology, and neurobiology, and are linked to a number of genetic diseases. It is now clear that cell-surface glycans are used to organize plasma membrane receptors and to control the recruitment of intracellular signal transduction mediators. In some cases, different modifications of glycans are the only source of variation between identical proteins produced within the same cell. Clearly, N-glycosylation can regulate function, stability and quality control of glycoproteins. Nevertheless, simply knowing the specific type of glycan modification does not allow one to predict what effect it will have on the glycoprotein. Anticipating biological function of N-glycan on proteins is a challenging process and more information on how different glycan structures affect protein structure and function is required. To cure diseases where glycosylation is altered, there is a now a pressing need for high throughput screening of structural variation among glycans through 3D structural analysis to identify therapeutic targets. 14

2.2. V-ATPase structure, function, regulation and drug targeting

2.2.1. V-ATPase Function

1.1.1.2 2.2.1.1. Endomembrane V-ATPases

Vacuolar H+ ATPases (V-ATPases) are multimeric proton nano-motors found on endomembranes of acidic organelles and on the plasma membrane of some specialized cells. In acidic endomembrane organelles, V-ATPases are responsible for acidification (86). The functions of intracellular compartments are largely dependent on their luminal pH thus tight regulation of pH by V-ATPases is a crucial aspect for cellular homeostasis (1). V-ATPases on endosomes is essential for receptor-mediated endocytosis as acidification mediates the release of internalized receptors (87). In phagocytosis, the acidic pH facilitates pathogen uptake and subsequent fusion to lysosomes for degradation and activation of the innate immune system (88). Lysosome acidification is crucial for protein degradation and enzyme activation. Acidic pH is important for various Golgi functions such as vesicular trafficking, glycosylation of proteins and maintaining the integrity of Golgi morphology (89). V-ATPases also have other non-canonical functions listed in the following section.

2.2.1.2 Non-canonical functions of endomembrane V-ATPases

V-ATPases can act as a pH sensor to both regulate membrane fusion (90) and control recycling of internalized cholesterol from endosome back to the plasma membrane (91). V-ATPases also play a role in viral and bacterial virulence, where they can facilitate the formation of fusion pore between viral envelopes of influenza and Ebola virus and endosomes allowing viruses to invade host cells (reviewed in (3)). In pathogenic bacteria such as Bacillus anthracis and

Corynebacterium diphtheriae, V-ATPases permit the entrance of their toxins to host cells (3). Additionally, V-ATPases are involved in pH dependent sorting of cargo in the secretory pathway (92), amino acid sensor in the lysosome leading to activation of mTORC1 signaling (93,94) and regulation of the final stage of autophagy, a process by which the cell breaks down aggregated proteins in the autolysosome (95,96), where the application of V-ATPase inhibitors suppress autophagy (95). V-ATPases act as scaffolds in protein-protein interaction that can in turn modulate V-ATPase function and regulation. For instance, in activated osteoclast, V-ATPases 15

interact with actin cytoskeleton and this interaction thought to be involved in vesicular sorting and trafficking (97). V-ATPases also interact with the metabolic enzymes phosphofructokinase-1 and aldolases to modulate V-ATPase activity through the reversible dissociation of V0-V1 (98,99).

2.2.1.3 Plasma membrane V-ATPases

V-ATPases are also found in the plasma membranes of specialized cells, where they pump protons to the extracellular matrix to maintain cellular acid-base homeostasis and coupled transport (87). Plasma membrane V-ATPases in intercalated cells of the kidney pump protons into the urine, allowing for bicarbonate re-absorption into the blood (100). In osteoclasts, bone resorbing cells, V-ATPases pump protons onto the bone surface, necessary for bone resorption (101,102). V-ATPase acidification in the apical membrane of epididymal clear cells is crucial for maturation and activation of sperm (103). In activated neutrophil and macrophages, plasma membrane V-ATPases maintain a neutral cytosolic pH by continuously pumping protons to the extracellular matrix (104). In osteoclasts, plasma membrane V-ATPases are essential for ruffled border formation, osteoclast maturation and activation (105,106). In synaptic vesicles of nerve cells, V-ATPases maintain an electrochemical gradient crucial for neurotransmitter uptake (107). In midgut cells of growing insects, plasma membrane V-ATPases pump protons into the gut creating an electrochemical gradient, which in turn activates the H+/2K+ antiporters that allow protons back to the midgut cells in exchange for K+, a process that is up-regulated by the insects’ appetite during growth, but is inhibited during molting (108). In addition to normal physiological functions, it has been reported that plasma membrane V-ATPases are critical for metastasis of both highly invasive MB-231 and MDA-MB-231 cells breast cancer cell lines (6,109). It is hypothesized that acidification of the extracellular space allows optimal activity of proteolytic enzymes thus facilitating the cells ability to metastasize through the degraded matrix. Blocking the acidification may be a therapy beyond the classic inhibition of proteolytic enzyme activity.

2.2.2 V-ATPases Structure

V-ATPases are multi-subunit ATP-powered rotary pumps that translocate protons against their concentration gradient driven by ATP hydrolysis. V-ATPases are composed of two domains (see figure 2.3); a soluble cytoplasmic domain V1 and membrane-embedded domain V0, which combine together to form a functional V1V0 complex. The V1 domain catalyzes ATP hydrolysis and couples it to proton translocation through the V0 membrane domain. The subunit composition of V1 sub-complex is A3B3CDE3FG3H, whereas the subunits composition of V0 sub-complex is ac5c”de. Yeast has additional subunits c’ and f (87,110) while the V0 sector of higher eukaryotes have accessory subunits Ac45 and M8-9 (87). Each subunit or combination of subunits forms a unique structural feature for the V-ATPase complex, which will be detailed below.

Figure 2.3: Structure of yeast V-ATPase. V-ATPases are composed of at least 14 subunits organized in two domains; a cytoplasmic V1 and a membrane-embedded V0. The catalytic hexamer A3B3 hydrolyze

ATP and provides the energy required for proton translocation through V0.

2.2.2.1 Structural components and function of V-ATPase V1 sector

The V1 sector is composed of three functional domains; catalytic hexamer, peripheral stalks and the central stalk. The catalytic hexamer, A3B3, is composed of three copies of subunit A and B 17

(111). The function of the catalytic hexamer is to hydrolyze ATP and couple the released energy to proton translocation (112). In yeast, high resolution structural and biochemical evidence suggest that V-ATPases have three peripheral stalks; each stalk is composed of “EG” subunit hetero-dimers (113-115). Structural analyses of V-ATPases showed that the V1 and V0 sectors are separated by a 100 Å central stalk composed of subunits D, F and d. The d subunit, at the base of the central stalk (113,114,116), couples the rotation at the interface between the V1 and

V0 (117).

2.2.2.2 Structural components of V-ATPase V0 domain.

The yeast membrane-embedded V0 functions as a proton pathway and is composed of six subunits: a, d, e, c, c´ and c´´. Mammalian V-ATPases are lacking c´, but have two accessory proteins Ac45 and M8-9. To date the number of total c subunits in the proteolipid c-ring is still controversial. In yeast based on genetic and quantitative amino acid analyses, a stoichiometry of 4:1:1 (c: c´:c´´) was predicted (118-120). In contrast, the latest high resolution cryo-EM model of

V0 showed that the c-ring was composed of 10 c subunits in a stoichiometry of 8:1:1 instead of 4:1:1 and this c-ring was in close proximity to subunits a and e (121). Additionally, cryo-EM maps of the yeast V-ATPase at 11 and 6.4 Å showed that the C-terminal and proton translocation domains of the a subunit are in direct contact with the c-ring, a finding that sheds light on a mechanistic insight into how protons might translocate from the highly conserved glutamic acid residue (E137) within the c subunits to proton translocation residues within the a subunit cytoplasmic and luminal half-channels (116,121).

2.2.3 V-ATPases subunits isoforms

Many V-ATPases subunits have multiple isoforms, and it is suggested that these isoforms target V-ATPase complexes to their functional destination. For instance, a3 and a4 subunits target V- ATPases to the plasma membrane of osteoclasts and kidney intercalated cells respectively (109). In addition, a recent study has shown that over expression of a3 in a non-invasive breast epithelial cell line leads to an increase in V-ATPase expression on the cell surface as well as 18

increase in cell invasion (122). V-ATPases subunits isoforms, their function, expression and stoichiometry has been summarized in table 1 (87,109,123).

Table 2.1: Mammalian V-ATPase subunit isoforms and their functions

Subunit Sector Human Stoichiometry Mr Function Expression

Isoforms Genes (kDa)

A ATP6V1A 3 70 catalytic/stator Ubiquitous

B1 ATP6V1B1 3 58 catalytic/regulatory/stator B1: kidney and ear

56 B2: ubiquitous B2 ATP6V1B2

C1 ATP6V1C1 42 C1: ubiquitous

1 48 regulatory/stator C2a: lung C2 ATP6V1C2 V1 C2b:Kidney

D ATP6V1D 1 34 rotor Ubiquitous

E1 ATP6V1E1 3 31 Stator E1: acrosomal lumen

E2: ubiquitous E2 ATP6V1E2

F ATP6V1F 1 14 rotor Ubiquitous

G1 ATP6V1G1 G1: ubiquitous 19

3 13 Stator G2: Brain G2 ATP6V1G2 G3: kidney

G3 ATP6V1G3

H ATP6V1H 1 56 regulatory/stator Ubiquitous

a1 ATP6V0A1 a1: ubiquitous

a2: ubiquitous a2 ATP6V0A2 1 100– stator/proton a3: ubiquitous, but

3 116 enriched in osteoclasts a3 ATP6V0A3 a4:kidney, epididymal

clear cells a4 ATP6V0A4

c” ATP6V0B 1 21 proton translocation/ rotor Ubiquitous V0

C ATP6V0C 5 16 proton translocation/ rotor Ubiquitous

d1 ATP6V0D1 1 38 Rotor d1:ubiquitous

d2: kidney d2 ATP6V0D2

e1 ATP6V0E1 1 9 Stator e1: ubiquitous

e2: enriched in e2 ATP6V0E2 osteoclast

Ac45 ATP6AP1 1 40 accessory /stator

M8-9 ATP6AP2 1 9 accessory /stator 20

2.2.4. Regulation of V-ATPase assembly and trafficking

In mammals, the assembly of the V1V0 domains is not well understood; however it is better studied in yeast. In yeast, two assembly mechanisms have been experimentally supported. In the independent pathway, the fully assembled V1 domain couples with a fully assembled V0 domain; while in the concerted pathway, partially assembled V1 and V0 subunits combine in a sequential manner to form the V1V0 complex (87,124,125). Yeast assembly chaperones Vma12p, Vma21p, and Vma22p are essential for V0 domain assembly; reviewed in (87). VMA21 is the human homolog of yeast Vma21p, an essential assembly chaperone of V-ATPases (126). Mutations in human VMA21 results in X-linked Myopathy with Excessive Autophagy (XMEA). In XMEA, there is an increase in lysosomal pH and a subsequent decrease in lysosomal ability to degrade proteins, which in turn leads to reduction in free amino acids required for cell growth. This reduction in amino acid leads to the down-regulation of mTORC1 pathway and increased macro- autophagy (126).

Yeast V-ATPase assembly is regulated by amino acid and glucose levels; amino acid starvation increases V1V0 assembly, while glucose depletion induces V1V0 reversible disassembly (109,127). Additionally, V-ATPase assembly is also regulated by extracellular and cytoplasmic pH as well as by modulating the expression levels of the fully assembled V-ATPase complex on biological membranes (93,128,129). Targeting of V-ATPases to biological membranes is not well characterized in mammals; however, in yeast the N-terminal domains of the a subunit (NTa) of the two a subunit isoforms encode targeting signals that direct Vph1p to the vacuole and Stv1p to Golgi apparatus (130,131). The putative membrane targeting domains within four mammalian a subunit isoforms would be excellent targets for therapeutic intervention, but to date, little is known about their location.

2.2.5 Involvement of V-ATPases in human diseases

Functional abnormalities within V-ATPases are associated with several human diseases. Human mutations in the B1 subunit were the first to be reported to result in both distal renal tubular acidosis (dRTA) and sensorineural deafness (132). Loss of the mouse d2 subunit isoform results in mild osteopetrosis, but there is no equivalent mutations reported in human. Yet, many disease- 21

causing mutations were mapped in V-ATPase subunit genes; however, most these mutations involve a subunit isoforms. As my thesis is focused on mapping functional domains of the mammalian a subunit, the next section will focus on the clinical implication of this particular subunit and its four isoforms.

2.2.5.1 Clinical implication of V-ATPase a subunit.

As mentioned before, the a1 and a2 isoforms are primarily localized to intracellular compartments, a3 is localized to the osteoclast plasma membrane, and a4 is expressed in the kidney in intercalated cells of the distal nephron. Knock down of a1 is reported to be embryonically-lethal in Drosophila (133), but loss of function mutations in a2 causes cutis laxa, (wrinkle skin syndrome), which is thought to be caused by defect in Golgi glycosylation (134,135). Loss of a3 function causes autosomal recessive osteopetrosis, characterized by high bone density due to defective bone resorption (136,137), and a4 loss of function causes dRTA (138). In contrast, excessive V-ATPase activity can also lead to disease: studies have suggested that plasma membrane V-ATPases activities are important for cancer metastasis. In invasive human breast cancer cells, a3 and a4 are highly expressed while siRNA knockdown of both V- ATPase a subunit isoforms abrogates surface expression of V-ATPases and inhibits invasion of cancer cells (109).

2.2.6 V-ATPase a subunit topology and atomic models

The a subunit is conserved from yeast to human (See alignment figure 2.4) and is thought to be involved in proton translocation, V1V0 coupling, and signaling (87,139).

The ~50 kDa N-terminal of a “NTa” interacts with the V1 sector to serve as a stabilizing structure or stator (5,116). The crystal structure of the N-terminal cytosolic domain from the Meiothermus ruber subunit “I” homolog of subunit a has been characterized and shown to be composed of “a curved long central α-helix bundle capped on both ends by two lobes with similar α/β architecture” (140). The crystal structure and model fit well into reconstructions from electron microscopy of prokaryotic and eukaryotic vacuolar H+-ATPases and show that the N- 22

terminal is parallel to the cytoplasmic surface of the membrane with close proximity to the N- terminal domain of H subunit (116,140).

The ~50 kDa C-terminal domain “CTa” is proposed to serve as the proton pathway by forming the membrane-embedded proton channel (5,139). The precise topology of the a subunit is controversial as the crystal structures for the C-terminus has yet to be obtained (reviewed in (141)). A topology model of the a subunit was created based on an alignment between CTa of all the yeast and mammalian a subunits) which suggested eight transmembrane domains (139). This model is based on the consensus of seven topology prediction algorithms and was validated experimentally using epitope tagging, green fluorescent protein fusion, and protease accessibility analysis in purified yeast vacuoles (139). The number of TM domains and orientation of the C- and N-terminal domains represented in this topology model are supported by the recent atomic models of the CTa domain from several rotary ATPases (see figure 2.5); models that were obtained were based on multiple structural analyses, molecular covariance and molecular flexible fitting of the atomic models in the cryo-EM maps (5,110). Cryo-EM shows several TMs tilted in the membrane with TM7 and TM8 both highly tilted, contacting the c-ring and almost parallel to the lipid bilayer.

Alignments of CTa domain >>

______TM1_____

H1 SIQFALRRGTEHSGSTVPSILNRMQTNQTPPTYNKTNKFTYGFQNIVDAYGIGTYREINPAPYTIITFPF 401

H2 DLRRALEEGSRESGATIPSFMNIIPTKETPPTRIRTNKFTEGFQNIVDAYGVGSYREVNPALFTIITFPF 406

H3 ALQEALRDSSMEEG--VSAVAHRIPCRDMPPTLIRTNRFTASFQGIVDAYGVGRYQEVNPAPYTIITFPF 398

H4 RIKRALEQGMELSGSSMAPIMTTVQSKTAPPTFNRTNKFTAGFQNIVDAYGVGSYREINPAPYTIITFPF 403

M1 SIQFALRRGTEHSGSTVPSILNRMQTNQTPPTYNKTNKFTHGFQNIVDAYGIGTYREINPAPYTVITFPF 401

M2 GLRRALEEGSRESGATIPSFMNTIPTKETPPTLIRTNKFTEGFQNIVDAYGVGSYREVNPALFTIITFPF 406

M3 TVQQALQSGSSEEG--VSAVAHRIPCQDMPPTLIRTNRFTSSFQGIVDAYGVGRYREVNPAPYTIITFPF 399 23

M4 HIKKALEQGMELSGSSMIPIMTEVETKTDPPTFNRTNKFTAGFQNIVDAYGVGSYREINPAPYTIITFPF 403

YV TLQARLGEMIARLGIDVPSIIQVLDTNHTPPTFHRTNKFTAGFQSICDCYGIAQYREINAGLPTIVTFPF 417

______EL1___TM2______CL1______TM3______EL2______

H1 LFAVMFGDFGHGILMTLFAVWMVLRESRILSQKNENEMFSTVFSGRYIILLMGVFSMYTGLIYNDCFSKS 471

H2 LFAVMFGDFGHGFVMFLFALLLVLNENHPRLNQSQ-EIMRMFFNGRYILLLMGLFSVYTGLIYNDCFSKS 475

H3 LFAVMFGDVGHGLLMFLFALAMVLAENRPAVKAAQNEIWQTFFRGRYLLLLMGLFSIYTGFIYNECFSRA 468

H4 LFAVMFGDCGHGTVMLLAALWMILNERRLLSQKTDNEIWNTFFHGRYLILLMGIFSIYTGLIYNDCFSKS 473

M1 LFAVMFGDFGHGILMTLFAVWMVLRESRILSQKHENEMFSMVFSGRYIILLMGLFSIYTGLIYNDCFSKS 471

M2 LFAVMFGDFGHGFVMFLFALLLVLNENHPRLSQSQ-EILRMFFDGRYILLLMGLFSVYTGLIYNDCFSKS 475

M3 LFAVMFGDVGHGLLMFLFALAMVLTENRPAVKAAQNEIWQTFFGGRYLLLLMGLFSVYTGFIYNECFSRA 469

M4 LFAVMFGDCGHGMVMLMAALWMVLNERHLLAQKSTNEMWNIFFNGRYLILLMGIFSIYTGLIYNDCFSKS 473

YV MFAIMFGDMGHGFLMTLAALSLVLNEKKINKMKRG-EIFDMAFTGRYIILLMGVFSMYTGFLYNDIFSKT 486

______

H1 LNIFGSSWSVRPMFTY------NWTEETLRGNPVLQLNPALPGVFGG-PYPFGIDPIWNIATNKL 529

H2 VNLFGSGWNVSAMYSSSHPPAEHKKMVLWNDSVVRHNSILQLDPSIPGVFRG-PYPLGIDPIWNLATNRL 544

H3 TSIFPSGWSVAAMANQSG------WSDAFLAQHTMLTLDPNVTGVFLG-PYPFGIDPIWSLAANHL 527

H4 LNIFGSSWSVQPMFRNGT------WNTHVMEESLYLQLDPAIPGVYFGNPYPFGIDPIWNLASNKL 533

M1 LNIFGSSWSVRPMFTQG------NWTEETLLGSSVLQLNPAIPGVFGG-PYPFGIDPIWNIATNKL 530

M2 VNLFGSGWNVSAMYSSSHSPEEQRKMVLWNDSTIRHSRTLQLDPNIPGVFRG-PYPFGIDPIWNLATNRL 544

M3 TTIFPSGWSVAAMANQSG------WSDEYLSQHSMLTLNPNITGVFLG-PYPFGIDPIWSLATNHL 528

M4 FNIFGSSWSVQPMFRNGT------WNTHIVENSPYLQLDPAIPGVYSGNPYPFGIDPIWNLASNKL 533 24

YV MTIFKSGWKWPDHWKK------GE------SITATSVG-TYPIGLDWAWHGTENAL 529

_____TM4______CL2______TM5______EL3___

H1 TFLNSFKMKMSVILGIIHMLFGVSLSLFNHIYFKKPLNIYFGFIPEIIFMTSLFGYLVILIFYKWTAYDA 599

H2 TFLNSFKMKMSVILGIIHMTFGVILGIFNHLHFRKKFNIYLVSIPELLFMLCIFGYLIFMIFYKWLVFSA 614

H3 SFLNSFKMKMSVILGVVHMAFGVVLGVFNHVHFGQRHRLLLETLPELTFLLGLFGYLVFLVIYKWLCVWA 597

H4 TFLNSYKMKMSVILGIVQMVFGVILSLFNHIYFRRTLNIILQFIPEMIFILCLFGYLVFMIIFKWCCFDV 603

M1 TFLNSFKMKMSVILGIIHMLFGVSLSLFNHIYFKKPLNIYFGFIPEIIFMSSLFGYLVILIFYKWTAYDA 600

M2 TFLNSFKMKMSVILGIFHMTFGVVLGIFNHLHFRKKFNVYLVSVPEILFMLCIFGYLIFMIIYKWLAYSA 614

M3 SFLNSFKMKMSVILGVTHMAFGVFLSIFNHVHFGQSHRLLLETLPELIFLLGLFGYLVFLIVYKWVNVSA 598

M4 TFLNSYKMKMSVILGIAHMIFGVILSLFNHIYFRRTLNIILQFIPEMIFMLSLFGYLVFMIIFKWCRYDA 603

YV LFSNSYKMKLSILMGFIHMTYSYFFSLANHLYFNSMIDIIGNFIPGLLFMQGIFGYLSVCIVYKWAVDWV 599

______TM6______CL3______

H1 HTSENAPSLLIHFINMFLFSYPESGYSMLYSGQKGIQCFLVVVALLCVPWMLLFKPLVLRRQYLRRKHLG 669

H2 ETSRVAPSILIEFINMFLFPASKT-SG-LYTGQEYVQRVLLVVTALSVPVLFLGKPLFLLWLHNGRSCFG 682

H3 ARAASAPSILIHFINMFLFSHSPS-NRLLYPRQEVVQATLVVLALAMVPILLLGTPLHLL--HRHR---- 661

H4 HVSQHAPSILIHFINMFLFNYSDSSNAPLYKHQQEVQSFFVVMALISVPWMLLIKPFILRASHR-KSQLQ 672

M1 HSSRNAPSLLIHFINMFLFSYPESGNAMLYSGQKGIQCFLIVVAMLCVPWMLLFKPLILRHQYLRKKHLG 670

M2 ETSREAPSILIEFINMFLFPTSKT-HG-LYPGQAHVQRVLVALTVLAVPVLFLGKPLFLLWLHNGRNCFG 682

M3 ASASSAPSILIHFINMFLFSQNPT-NHLLFHGQEVVQYVLVVLALATVPILLLGTPLYLLRQHRHR---- 664 25

M4 HTSRKAPSILIHFIGMFLFDYDDSSNAPLYGHQQEVQTFFVIIALVSVPWMLLIKPFVLRAKHQ-KSQLQ 672

YV KDGKPAPGLLNMLINMFLSPG--TIDDELYPHQAKVQVFLLLMALVCIPWLLLVKPLHFKFTHKKKSHEP 667

______M7______

H1 TLNFGGIRVGNGPTEEDAEIIQHDQL------STHSEDADEPSEDEVFDFGDTMVHQAIHTIEYCLGC 731

H2 VNRSGYTLIRKDSEEEVSLLGSQDIE------EGNHQVEDGCREMACEEFNFGEILMTQVIHSIEYCLGC 746

H3 R-RLRRRPADRQEENKAGLLDLPDAS--VNGWSSDEEKAGGLDDEEEAELVPSEVLMHQAIHTIEFCLGC 727

H4 ASRIQEDATENIEGDSSSPSSRSGQR------TSADTHGALDDHGEEFNFGDVFVHQAIHTIEYCLGC 734

M1 TLNFGGIRVGNGPTEEDAEIIQHDQL------STHSEDAEEPTEDEVFDFGDTMVHQAIHTIEYCLGC 732

M2 MSRSGYTLVRKDSEEEVSLLGNQDIE------EGNSRMEEGCREVTCEEFNFGEILMTQAIHSIEYCLGC 746

M3 R-NTQRRPAGQQDEDTDKLLASPDASTLENSWSPDEEKAGSPGDEE-TEFVPSEIFMHQAIHTIEFCLGC 731

M4 SFTIHEDA---VEGDHSGHSSK---K------T-AGAHGMKDGHEEEFNFGDIFVHQAIHTIEYCLGC 727

YV LPSTEADAS-SEDLEAQQLISAMDAD------DAEEEEVGSGSHGEDF--GDIMIHQVIHTIEFCLNC 726

______EL4_TM8______

H1 ISNTASYLRLWALSLAHAQLSEVLWTMVIHIGLSVKSLAG--GLVLFFFFTAFATLTVAILLIMEGLSAF 799

H2 ISNTASYLRLWALSLAHAQLSDVLWAMLMRVGLRVDTTYG--VLLLLPVIALFAVLTIFILLIMEGLSAF 814

H3 VSNTASYLRLWALSLAHAQLSEVLWAMVMRIGLGLGREVGVAAVVLVPIFAAFAVMTVAILLVMEGLSAF 797

H4 ISNTASYLRLWALSLAHAQLSEVLWTMVMNSGLQTRGWGG--IVGVFIIFAVFAVLTVAILLIMEGLSAF 802

M1 ISNTASYLRLWALSLAHAQLSEVLWTMVIHIGLHVRSLAG--GLGLFFIFAAFATLTVAILLIMEGLSAF 800

M2 ISNTASYLRLWALSLAHAQLSDVLWAMLMRVGLRVDTTYG--VLLLLPVMAFFAVLTIFILLVMEGLSAF 814

M3 ISNTASYLRLWALSLAHAQLSEVLWAMVMRIGLGMGREIGVAAVVLVPVFAAFAVLTVAILLVMEGLSAF 801 26

M4 ISNTASYLRLWALSLAHAELSEVLWTMVMSIGLRLQGWAG--LVGVFIIFAVFAVLTVAILLVMEGLSAF 795

YV VSHTASYLRLWALSLAHAQLSSVLWTMTIQIAFGFRGFVG--VFMTVALFAMWFALTCAVLVLMEGTSAM 794

___CTD______

H1 LHALRLHWVEFQNKFYSGTGFKFLPFSFEHIREGKFEE 837

H2 LHAIRLHWVEFQNKFYVGAGTKFVPFSFSLLSSKFNNDDSVA 856

H3 LHALRLHWVEFQNKFYSGTGYKLSPFTFAATDD 830

H4 LHALRLHWVEFQNKFYVGDGYKFSPFSFKHILDGTAEE 840

M1 LHALRLHWVEFQNKFYTGTGFKFLPFSFEHIREGKFDE 838

M2 LHAIRLHWVEFQNKFYVGAGTKFVPFSFSLLSSKFSNDDSIA 856

M3 LHALRLHWVEFQNKFYSGTGYKLSPFTFTVDSD 834

M4 LHALRLHWVEFQNKFYEGAGSKFSPFSFKHVLEGTAEE 833

YV LHSLRLHWVESMSKFFVGEGLPYEPFAFEYKDMEVAVASASSSASS 840

Figure 2.4: Human and mouse CTa of a subunit, and yeast Vph1p, alignments. Domain assignments are based on a computational topology for Vph1p (139) and the two recently published atomic models of yeast Vph1p subunit [PDB: 5I1M and 5JT5] (5,110). Cyan highlights indicate TMs of the CTa domain. Magenta highlights indicate glycosylation sites and blue highlights indicate second glycosylation sites in EL2. Domains are indicated above alignments. Red font in the YV line indicates TM boundaries proposed by (142). Underlining indicates hydrophobic sequences that may represent a membrane-inserted reentrant loop. Experimental validation of the Topology model used is detailed in (139). H1–4, human ATP6V0A1–4; M1–4, mouse Atp6v0a1–4; YV, yeast Vph1p. Alignments kindly provided by N. Kartner. 27

Figure 2.5: Cartoon of secondary structure of yeast a subunit. V-ATPase a subunit is composed of eight transmembrane domains and at least four extracellular loops (EL). The largest EL2 is glycosylated in all human a subunit isoforms (27,143). The CTa and the NTa domains are cytoplasmic. Cartoon kindly generated by N. Kartner.

2.2.7 V-ATPase inhibitors and their limitation

Developing specific therapeutic approaches to target V-ATPases to prevent pathological bone loss has been the center of attention for some time. However, the current V-ATPase inhibitors have not been clinically approved due to their broad specificity to all V-ATPases and anticipated side effects. Table 2 shows summary of current V-ATPase inhibitors.

Table 2.2: Commonly used V-ATPases inhibitors

Inhibitor Binding site; mechanism of action Source References

Bafilomycins Bind to the b/c subunits; disrupt c-ring Streptomyces (144,145) rotation and alter proton pumping of V- Plecomacrolides ATPases

Concanamycin Binds to c-subunit; disrubt c-ring rotation Streptomyces (145,146) and alter proton pumping of V-ATPases

Archazolid Binds to c-subunit; not clear mechanism Archangium (147,148) gephyra Compete with concanmycin binding site

Benzolactone Enamides Unidentified Variety (148,149)

Pea albumin 1 subunit b (PA1b) Unidentified; active only against insects’ Peas (150) V-ATPases

2.2.8 V-ATPase a subunit as a potential therapeutic target

Designing specific inhibitors directed against specific a subunit isoforms will require more knowledge of the plasma membrane targeting domains. As reviewed before, a3 and a4 are capable of targeting V-ATPases to the plasma membrane of some specialized cells although the precise molecular mechanism is still unclear. Understanding how a subunit isoforms target V- ATPases complexes will pinpoint novel therapeutic approaches to target plasma membrane V- ATPases without disrupting normal physiological functions. Our group has characterized the interaction between a3-B2 subunits to be specific for osteoclast plasma membrane V-ATPases using yeast two-hybrid and fusion proteins. Small molecules were screened for their ability to disrupt a3/B2 binding, while not affecting B2 interactions with others a subunit isoforms (8). The 29

screening resulted in the discovery of a novel inhibitor, benzohydrazide that both disrupted a3- B2 binding and reduced the ability of osteoclasts to resorb bones in vitro (8). The same strategy was taken to screen for inhibitor that disrupt the osteoclast specific a3-d2 interaction, resulting in the discovery of luteolin that is able to both inhibit a3-d2 binding as well as osteoclast resorption ability (9).

3. Research Objectives and Hypotheses

V-ATPase a isoforms are viable targets for therapeutic interventions towards a number of diseases. Understanding isoform-specific interactions, including protein–protein interaction motifs and structural domains affecting the a subunit folding, trafficking, membrane targeting, function and regulation, will enhance our ability to target specialized V-ATPase isocomplexes (i.e. the osteoclast plasma membrane V-ATPase or the plasma membrane of metastatic cells), without altering normal physiological functions of V-ATPases. Specific a subunit isoforms are associated with specific disease processes, and these could be uniquely targeted if enough were known about their structural and functional differences and how they are targeting to specific cellular organelles. Thus my research hypotheses and objectives are as follow:

Hypotheses

1. Human a subunit isoforms are N-glycosylated and glycosylation is important for their folding and assembly and/or targeting V-ATPase complexes to the plasma membrane. 2. Human disease-causing missense mutations within a subunit isoforms will identify critical domains essential for V-ATPase targeting, activity and/or regulation.

Central Objective

Elucidate specific features of a subunit isoforms which govern structure, function and membrane targeting of V-ATPases, in order to facilitate the development of therapeutics to prevent cancer metastasis and pathological bone loss.

Aim 1

Characterize a subunit N-glycosylation and its role in stability, assembly, trafficking and plasma membrane targeting.

Aim 2

Study the molecular repercussions of the Cutis laxa causing mutation; a2P405L and the renal tubular acidosis causing mutations; a4R449H and a4G820R with respect to subunit glycosylation, localization, stability and assembly.

Significance

Studying human mutations and N-glycosylation of a isoforms could pinpoint critical domains that could be specifically targeted in pathological conditions 31

4. N-Linked Glycosylation Is Required for Vacuolar H+- ATPase (V-ATPase) a4 Subunit Stability, Assembly , and Cell Surface Expression

Sally Esmail,1 Yeqi Yao,1 Norbert Kartner,1 Jing Li,2 Reinhart A. F. Reithmeier,2 and Morris F. Manolson1,2

1Faculty of Dentistry, Dental Research Institute, University of Toronto,

Toronto, Ontario, Canada M5G 1G6

2Department of Biochemistry, University of Toronto, Toronto, Ontario,

Canada M5S 1A8

Journal of Cellular Biochemistry, 117(12):2757-2768 (2016)

4.1 Abstract

The a subunit is the largest of 14 different subunits that make up the V-ATPase complex. In mammalian species this membrane protein has four paralogous isoforms, a1–a4. Clinically, a subunit isoforms are implicated in diverse diseases; however, little is known about their structure and function. The subunit has conserved, predicted N-glycosylation sites, and the a3 isoform has been directly shown to be N-glycosylated. Here we ask if human a4 (ATP6V0A4) is N- glycosylated at the predicted site, Asn489. We transfected HEK 293 cells, using the pCDNA3.1 expression-vector system, to express cDNA constructs of epitope-tagged human a4 subunit, with or without mutations to eliminate the putative glycosylation site. Glycosylation was characterized also by treatment with endoglycosidases; expression and localization were assessed by immunoblotting and immunofluorescence. Endoglycosidase-treated wild type (WT) a4 showed increased relative mobility on immunoblots, compared with untreated WT a4. This relative mobility was identical to that of unglycosylated mutant a4N489D, demonstrating that the a4 subunit is glycosylated. Cycloheximide pulse-chase experiments showed that the un- glycosylated subunit degraded at a higher rate than the N-glycosylated form. Un-glycosylated a4 was degraded mostly in the proteasomal pathway, but also, in part, through the lysosomal pathway. Immunofluorescence colocalization data showed that un-glycosylated a4 was mostly retained in the ER, and that plasma membrane trafficking was defective. Co- N489D immunoprecipitation studies suggested that a4 does not assemble with the V-ATPase V1 domain. Taken together, these data show that N-glycosylation plays a crucial role in a4 stability, and in V-ATPase assembly and trafficking to the plasma membrane.

4.2. Introduction

V-ATPases are highly-conserved, ATP-dependent molecular motors that play vital roles in transporting protons across various endomembranes to acidify intracellular compartments, or in translocating protons across the plasma membrane and out of the cell (151-153). They are multi- subunit complexes with a bipartite structure consisting of a cytoplasmic V1 sector that is catalytic and has the subunit composition, A3B3CDE3FG3H, and an integral membrane V0 sector that is involved in proton translocation and has the subunit composition ac5c”de. Some V-ATPase 33

subunits have multiple isoforms, and isoform composition may account in part for targeting of the V-ATPase holocomplex to the specific intracellular destination where its function is required

(154,155). In mammalian V-ATPases, the V0 a subunit can be one of four isoforms, a1–a4. The a1–a3 isoforms are ubiquitously expressed, whereas a4 is predominantly expressed in apical membranes of renal intercalated cells (156). Clinically, a subunits are implicated in diverse V- ATPase-related, isoform-specific disease mechanisms such as cancer metastasis (a3 and a4), kidney disease (a4), cutis laxa (a2) and both sclerotic and lytic bone disease (a3) (151,156-158). This provides compelling incentives for systematically studying their structures and their cellular and molecular biology.

The a subunit consists of two major structural domains, the cytoplasmic N-terminal half and the integral membrane C-terminal half. The N-terminal domain interacts with the V-ATPase V1 sector and serves as a stator, to stabilize the complex during ATP-driven proton transport (159- 161). The C-terminal domain forms the proton channel and facilitates proton translocation across the membrane bilayer (142). Despite its functional importance, and the prominence of the a subunit within the V-ATPase holoenzyme, little is yet known about the details of its structure and function. In previous work, we showed that the a3 subunit is a glycoprotein that is N- glycosylated at two sites on the second extracellular loop, EL2, within the C-terminal domain (162).

N-linked glycosylation is one of the most common post-translational modifications to occur in membrane protein biosynthesis. It starts in the ER with the co-translational transfer of

Glc3Man9GlcNAc2-PP-dolichol to the nascent polypeptide at Asn-X-Ser/Thr residues, where X represents any amino acid except proline (163,164). N-glycosylation of a membrane protein can be crucial for its biological function and also for the quality control of its folding, structural stability, and subsequent protection against proteolysis (165-168).

For the different a subunit isoforms, predicted glycosylation sites appear to be highly conserved; however, with the recent exception of a3 (161,162), there has been no direct biochemical evidence to identify glycosylation sites for a subunit isoforms. Here we show that the kidney- specific, plasma membrane a4 subunit is N-glycosylated, and that disruption of its glycosylation has functional consequences for its stability and trafficking to the plasma membrane. 34

4.3. Materials and Methods

REAGENTS, ANTIBODIES AND ENZYMES: Protease inhibitor cocktail, phenylmethylsulfonyl fluoride (PMSF) and proteasome inhibitor MG132, N- (benzyloxycarbonyl) leucinylleucinylleucinal (catalog nos. P8340, P7626 and C2211, respectively), were from Sigma-Aldrich (Oakville, Canada). Octaethylene glycol mono-n- dodecyl ether (C12E8) was from NIKKO Chemicals (Barnet Products, Englewood Cliffs, NJ). Cycloheximide (CHX; CYC003) was from BioShop (Burlington, Canada). Phosphate buffered saline (PBS; 10010.023), Dulbecco’s phosphate buffered saline (DPBS; 1404182), heat- inactivated fetal bovine serum (FBS; 16140071), Dulbecco’s modified Eagle’s medium (DMEM (1X); 11965092) and penicillin/streptomycin mix (15140-122), were from Gibco, Fisher Scientific (Whitby, Canada). Bradford protein assay reagent (500-0006) was from Bio-Rad (Mississauga, Canada). ProLong gold antifade reagent (P36930) was from Molecular Probes, Fisher Scientific. Bovine serum albumin (BSA; ALB001.50) was from BioShop Canada (Burlington, Canada). Novex ECL horseradish peroxidase (HRP) chemiluminescent substrate reagent kit (WP20005) was from Fisher Scientific. Endoglycosidases, endo- -N- acetylglucosaminidase H (Endo H; P0702S) and peptide N-glycosidase F (PNGase F; P0704S), were from New England Biolabs (Whitby, Canada). Antibodies used in this study were: rabbit polyclonal anti-FLAG (ab1162) and mouse monoclonal anti-syntaxin 6 (ab56656) from Abcam (Cedarlane; Burlington, Canada); mouse monoclonal anti-calnexin, IgG (3H4A7; sc-130059), mouse monoclonal anti-LAMP-2 (H4B4; sc-18822), mouse monoclonal anti-glyceraldehyde-3- phosphate dehydrogenase (anti-GAPDH 0411; sc-47724), rabbit polyclonal anti-HA, IgG (4-11; sc-805), HRP-conjugated goat anti-rabbit IgG (sc-2004), and HRP-conjugated goat anti-mouse IgG (sc-2005) were from Santa Cruz (Dallas, TX). Alexa Fluor 488-conjugated goat anti-rabbit IgG (H+L; A11034), Alexa Fluor 568-conjugated goat anti-mouse IgG (H+L; A11004), Alexa Fluor 568-conjugated goat anti-rabbit IgG (H+L; A11011), and Alexa Fluor 488-conjugated goat anti-mouse IgG (H+L; A11001) were from Molecular Probes, Fisher Scientific. DAPI (10236276001) was from Roche Diagnostics (Mississauga, Canada). Site-directed mutagenesis was performed using the QuikChange II XL Site-Directed Mutagenesis Kit (200521) from Agilent Technologies (Mississauga, Canada). 35

HUMAN a4-SPECIFIC ANTIBODY: The human a4-homologous synthetic peptide, 500- GTWNTHVMEESLYLQLDPA-518, the sequence of which is located in the EL2 domain (161) was synthesized and used to immunize rabbits. Immune sera containing the polyclonal a4 antibody were affinity-purified against the immobilized immunizing peptide (Cedarlane custom antibody services; Burlington, Canada).

CONSTRUCTION OF WT AND UNGLYCOSYLATED HUMAN a4 MUTANTS: The cDNA coding for human a4, supplied in the pGEM-T vector, was a kind gift from Dr. Fiona E. Karet (University of Cambridge, Cambridge, UK). The a4 coding region was fused with tandem C- terminal FLAG epitope tags (2×DYKDDDDK; designated 2FLAG) and sub-cloned between EcoRI and ApaI sites in the pcDNA3.1(+) vector. Unglycosylated mutants (a4N489D, a4N489Q and a4T491A) were created using the QuikChange II XL Site-Directed Mutagenesis Kit (200521) from Agilent Technologies, by mutating the N-glycosylation acceptor Asn to Asp or Gln (within the N-X-S/T signal site), or by mutating Thr to Ala, which also disrupts N-glycosylation (169). Constructs were confirmed by DNA sequencing. All primer synthesis and DNA sequencing was performed by ACGT Corp. (Toronto, Canada).

CELL SURFACE EPITOPE TAGGING OF a4: Triple, in-tandem HA epitope tags (3HA) were introduced into a poorly conserved region within the predicted human a4 second extracellular loop (EL2; specifically between E498 and E499) to serve as a plasma membrane marker that is accessible on the surface of intact cells. Briefly, 3HA (3×YPYDVPDYA) was inserted into EL2 of WT a4, which was simultaneously fused with a C-terminal 2FLAG cytoplasmic epitope tag, within the pcDNA3.1(+) vector construct. The same procedure was repeated for the unglycosylated mutant a4N489D. The final epitope-tagged constructs were the wild type, WT a4- 3HA-2FLAG, and the unglycosylated mutant, a4N849D-3HA-2FLAG.

CELL CULTURE: Human embryonic kidney cells, HEK 293 (CRL-1573TM; American Type Culture Collection (ATCC), Manassas, VA) were grown in DMEM (1X) supplemented with

10% FBS, 0.5% penicillin/streptomycin mix (50 units/ml penicillin, 50 g/ml streptomycin, final), in a humidified 5% CO2 atmosphere at 37 °C. 36

TRANSFECTION AND EXPRESSION: HEK 293 cells were transiently transfected using Lipofectamine LTX with Plus Reagent (15338100; Invitrogen, Life Technologies). Briefly, 70– 80% confluent HEK 293 cells growing in six-well plates were transfected with 1 µg/well plasmid DNA. Cells were harvested 24–48 h post-transfection. Transfection efficiency, as judged by expression of markers, was routinely approximately 40% (unless otherwise noted in figure legend). No significant difference was observed in amounts of wild type and mutant a4 constructs expressed, and expression of those constructs resulted in only marginal changes in endogenous a4 expression (data not shown). Specifically, endogenous expression was unchanged after transfection of a4, and after a4N489D transfection it was marginally lowered (by approx. 20%; p = 0.033). Exogenous levels of expression were approximately 2-fold higher than endogenous levels; however, since this was observed in extracts from unselected transfected cells, taking transfection efficiency (40%) into account, exogenous levels are estimated to be 5- fold greater than endogenous levels.

WHOLE-CELL PROTEIN EXTRACTION, DEGLYCOSYLATION AND IMMUNOBLOTTING: Whole-cell lysates were prepared according to Li et al. (170), with modification. Briefly, transfected HEK 293 cells were grown in six-well plates for 24–48 h, then were collected in 0.3 ml/well lysis buffer (PBS containing 1% C12E8, protease inhibitor cocktail (1:100, v/v) and 1 mM PMSF) and incubated 40 min on ice, followed by centrifugation at 15,000g for 30 min at 4 °C. Supernatant was collected and total protein concentration assayed. Deglycosylation of native a4 and a4-2FLAG was performed by adding PNGase F or Endo H according to the manufacturer’s instructions. Briefly, 20 µg of whole cell lysate protein was combined with 2 µl of 10X glycoprotein denaturation buffer, and the final volume was adjusted to 10 µl using dH2O. Samples were then denatured by heating at 65 °C for 10 min, followed by addition of 0.1 volume each of 10X GlycoBuffer and 10% NP-40, and then 1,000 units of

PNGase F or Endo H. The final reaction volume was adjusted to 20 µl by adding dH2O, and the reactions were incubated for 1 h at 37 °C, followed by mixing with an equal volume of 2X SDS- PAGE sample-loading buffer. Proteins were analyzed by 7% SDS-PAGE and transferred onto nitrocellulose membrane. FLAG-tagged proteins were detected with rabbit polyclonal anti- FLAG antibody (1:3,000 dilution). Native a4 proteins were detected with rabbit polyclonal anti- a4 antibody (1:1,000 dilution), followed by HRP-labeled anti-rabbit IgG secondary antibody (1:5,000 dilution). As a protein loading control, blots were also probed with mouse monoclonal 37

anti-GAPDH (1:10,000 dilution), followed by HRP-labeled anti-mouse IgG secondary antibody (1:5,000 dilution).

CYCLOHEXIMIDE (CHX) PULSE-CHASE PROTEIN STABILITY ASSAY: Transfected HEK 293 cells were grown in six-well plates for 24 h, then were incubated with 10 µg/ml CHX, with or without either proteasome inhibitor (10 µM MG132), or lysosome inhibitor (25 mM

NH4Cl), for 0, 4, 8, 12 and 20 h. The cells were then collected and whole cell lysates were prepared, and proteins were analyzed by SDS-PAGE and immunoblotted with anti-FLAG antibody.

QUANTITATIVE PROTEIN BAND ANALYSIS: Immunoblot protein bands of the CHX pulse- chase assays were quantified using Bio-Rad Quantity One 4.6.9 software. Background subtraction was performed using the rolling disc method and the significance of difference with respect to controls was evaluated by using two-tailed Student’s t tests. Briefly, all band intensities were normalized to GAPDH in the same lane, as a protein loading control, and then intensities at indicated hours were normalized to intensities at zero time. To accurately quantify protein levels, band intensities were compared after treating all WT a4-2FLAG samples with PNGase F. This ensured that all bands were uniformly sharp (rather than having some with the typical diffuse appearance of glycoproteins), allowing a more accurate comparison of protein levels between WT and unglycosylated subunits.

IMMUNOFLUORESCENCE: Transiently transfected HEK 293 cells expressing WT a4-2FLAG and a4N89D-2FLAG were grown on glass coverslips for 24 h. Cells were washed with DPBS and fixed with 4% (w/v) paraformaldehyde for 10 min at room temperature. Cells were then permeabilized using DPBS containing 0.2% Triton X-100 at room temperature for 15 min, and blocked with DPBS containing 5% BSA for 30 h at room temperature. Cells were immunostained with anti-FLAG (1:1,000), anti-Calnexin (1:500), anti-Syntaxin 6 (1:500), or anti-LAMP-2 (1:500) antibodies in DPBS containing 3% BSA for 45 min at room temperature. This was followed by washing with DPBS and incubation with fluorescent-labeled second antibodies (1:500) for 45 min at room temperature, and finally nuclear staining with 0.1 µg/ml DAPI in PBS for 10 min. To detect protein expression on the cell surface, the same protocol was used, except that the cells were non-permeabilized. Anti-HA antibody was used to specifically detect HA-tagged a4 protein on the cell surface. This was incubated for 45 min at room 38

temperature, followed by washing and incubation with secondary fluorescent antibody (1:1,000) for 45 min. Immunolabeled cells were preserved with ProLong Gold Antifade Reagent. Images were obtained with either a Leica DMIRE2 inverted fluorescence microscope equipped with a Hamamatsu Back-Thinned EM-CCD camera, or a Quorum Spinning Disk Confocal System equipped with a Hamamatsu C9100-13 EM-CCD, Yokogawa CSU X1 scan head, and Improvision Piezo focus drive (Imaging Facility, Hospital for Sick Children, Toronto, Canada).

CONFOCAL MICROSCOPY QUANTITATIVE ANALYSIS: Quantitative analysis of confocal images was performed using PerkinElmer Volocity V6.3 3D image analysis software (PerkinElmer, Woodbridge, Canada). Quantification results were expressed as Pearson’s correlation coefficients (r). Data were compared to controls, and the significance of differences were evaluated by using two-tailed Student’s t tests.

CELL SURFACE BIOTINYLATION: Cell surface biotinylation was performed using EZ-Link NHS-SS-Biotin reagent (Pierce 21328; Fisher Scientific) according to manufacturer’s instructions, with some modification. Briefly, transfected HEK 293 cells were grown in six-well plates for 24 h, then washed with ice-cold PBS and incubated with 1 mg/ml freshly prepared EZ- Link NHS-SS-Biotin for 1 h at 4 °C, with gentle agitation. This was followed by addition of ice- cold quenching buffer (25 mM Tris, pH 8.3 at 25 °C, 192 mM glycine) to remove unreacted biotin. Cells then were collected in 0.5 ml ice-cold RIPA buffer (10 mM Tris, pH 7.5 at 25 °C, 1% Triton X-100, 150 mM NaCl, 1% sodium deoxycholate, 1 mM EDTA, and 0.1% SDS) containing protease inhibitor cocktail (1:100 v/v) and 1 mM PMSF, and incubated for 30 min on ice, then centrifuged at 15,000g for 30 min at 4 °C. The supernatant was collected and total protein concentration was assayed. In order to purify biotinylated cell-surface protein, the whole- cell lysate was incubated with 100 µl of 50% slurry of streptavidin agarose (Pierce 20347; Fisher Scientific) for 2 h at 4 °C. The biotinylated proteins were then eluted, using SDS-PAGE sample buffer containing 4% SDS and 5% 2-mercaptoethanol, for 1 h at room temperature. Eluted proteins were resolved on 7% SDS-PAGE, followed by immunoblotting with anti-FLAG antibody (1:3,000 dilution), followed by HRP-labeled anti-rabbit IgG secondary antibody (1:5,000 dilution).

CO-IMMUNOPRECIPITATION AND V1–V0 ASSEMBLY: Immunoprecipitations of WT a4- 2FLAG and un-glycosylated mutant a4N489D-2FLAG were performed using rabbit polyclonal 39

anti-FLAG antibody. Briefly, transiently transfected HEK 293 cells expressing WT a4-2FLAG and a4N89D-2FLAG were grown for 24 h. Cells were washed with ice-cold PBS and lysed in IP buffer (25 mM Tris HCl, 150 mM NaCl, pH 7.2 at 25 °C) containing 1% C12E8, 1:100 (v/v) mammalian protease inhibitor cocktail and 1 mM PMSF. After 30 min incubation on ice, the lysate was centrifuged at 17,000g for 30 min at 4 °C, and the supernatant was collected. Total protein concentrations were measured using the Bradford assay. Aliquots of 50 µg protein from the whole-cell lysates of WT a4-2FLAG or a4N489D-2FLAG were incubated with 5 µg anti- FLAG antibody overnight, with gentle mixing at 4 °C. Immunoprecipitation was performed by adding 100 µl of protein A agarose beads (50% by volume) to the antigen-antibody complex and incubated with gentle mixing for 2 h at room temperature. The beads were then washed with IP buffer and centrifuged at 2,500g for 3 min, the supernatant was discarded and the immunocomplex was eluted by incubating the beads with SDS-PAGE sample buffer for 5 min at 95 °C. After centrifugation at 2,500g, the supernatant was collected and analyzed by 7% SDS- PAGE and transferred onto nitrocellulose membrane. Blots were probed with anti-FLAG and anti-B1 antibodies to assess V1V0 complex assembly.

4.4. Results

4.4.1. The Human V-ATPase a4 subunit is N-glycosylated

Amino acid sequence alignment of a4 subunit cDNA from different species identified a single highly conserved, putative N-glycosylation site within the EL2 domain (at Asn489 in human a4), as shown in Fig. 4.1A. In the current study we characterized a4 glycosylation by using two enzymes: the amidase, PNGase F, which cleaves between the polypeptide Asn and the GlcNAc of high mannose, hybrid and complex carbohydrate moieties and is used to determine whether a protein is N-glycosylated (171); and the endoglycosidase, Endo H, which cleaves the carbohydrate from the proximal GlcNac residue of core glycosylated, high-mannose and hybrid glycoproteins that have not been processed by α-mannosidase II in the Golgi (172,173). Fig. 4.1B shows immunoblots of FLAG-tagged a4 subunit constructs, WT a4-2FLAG, and un- glycosylated mutant protein, a4N489D-2FLAG, expressed transiently in HEK 293 cells, with and without PNGase F or Endo H treatment. WT a4-2FLAG was observed as a 105-kDa band, and 40

upon PNGase F and Endo H treatment the band size was reduced to 98 kDa (un-glycosylated a4- 2FLAG). The latter size corresponds well to the predicted molecular mass of 98.5 kDa (96.4 kDa for human a4, plus the 2.1-kDa 2FLAG tag). Similar results were seen for native a4, as shown in Fig. 4.1C. Here, the native, endogenous a4 protein from untransfected HEK 293 cells was immunoblotted and visualized using polyclonal anti-a4 antibody, showing an intense band at 103 kDa. Upon PNGase F and Endo H treatment, a deglycosylated band was observed at the predicted polypeptide size of 96 kDa. This analysis shows that the endogenous protein, like the protein expressed in HEK 293 cells, contains a high-mannose oligosaccharide and is not processed to a complex form.

4.4.2. N-glycosylation is required for a4 stability

CHX pulse-chase experiments were done to investigate a4 subunit stability. After inhibiting the de novo synthesis of proteins with CHX for up to 20 h, the treated cells were harvested and immunoblotted. The resultant time course is shown in Fig. 4.2. Here, the rates of turnover of un- glycosylated mutant a4 and WT are compared. To rule out the possibility of destabilizing a4 structure when introducing the point mutation to eliminate glycosylation, three different un- glycosylated mutants, N489D, T491A, and N489Q, were independently introduced into the WT a4-2FLAG construct. Fig. 4.2A, B and C show immunoblots of the CHX pulse-chase assay, comparing the stabilities of these three constructs, respectively. All the blots were visualized by staining with anti-FLAG antibody, and all samples were treated with PNGase F prior to blotting and analysis to remove the carbohydrate moiety. Quantitative band analyses were performed and signals were normalized to the time zero control (see Experimental Procedures). These results (Fig. 4.2D) showed quantitatively that unglycosylated mutants were significantly less stable, with half-lives of approximately 7.5 h, compared with the WT protein that had a half-life > 20 h. Since the three mutants had similar decreases in stability, with the a4N489D mutant being somewhat more stable than the a4N489Q and a4T491A constructs, we used a4N489D exclusively for further studies.

4.4.3. Un-glycosylated a4 is degraded in proteasomal and lysosomal pathways

There are two major degradation pathways for membrane proteins, the proteasomal ERAD pathway, and the plasma membrane endocytosis and lysosomal degradation pathway. The proteasome inhibitor MG132 is commonly used to distinguish the involvement of the proteasomal pathway in protein degradation in mammalian cells (174). It has been reported that treating cells with 1–50 µM MG132 for 10–24 h effectively inhibits proteasomal degradation without impacting cell viability (175,176). We used 10 µM MG132 to assess the role of the proteasomal pathway in degradation of a4N489D. As shown in Fig. 4.3A, proteasomal inhibition resulted in significant enhancement of steady state levels of un-glycosylated a4N489D protein. MG132 treatment of cells slightly increased the stability of WT a4, but had a dramatic effect on the stability of the un-glycosylated mutant, increasing its half-life to nearly that of WT. Thus, the un-glycosylated mutant is subject to degradation by the proteasomal pathway.

It has been reported in several studies that clearing unfolded protein from the ER through the proteasomal pathway alone may be insufficient, resulting in ER stress, which in turn activates the lysosomal degradation pathway (176,177). Thus, we assessed the contribution of lysosomal degradation by treating cells with the weak base, NH4Cl, which is known to increase lysosomal pH, inhibiting lysosomal proteases. As shown in Fig. 4.3B, NH4Cl treatment resulted in measurable enhancement of un-glycosylated a4N489D protein level, but to a much lesser extent than was obtained by inhibition of the proteasomal pathway. This is confirmed in Fig. 4.3C and D, by quantification of the immunoblots shown in Fig. 4.3A and B, respectively, where the half- life of the un-glycosylated protein is increased twofold to approximately 15 h.

Both proteasomal and lysosomal inhibitors decreased the rate of degradation of the un- glycosylated a4, by approximately 80% and 10%, respectively. Though only a small fraction of the protein appeared to be degraded in the lysosome, we sought to confirm its presence in that organelle. Fig. 4.4A shows the photomicrographic colocalization of a4N489D with the lysosomal marker, LAMP-2, with quantitative correlation data shown in Fig. 4.4B. These data show that a4N489D localizes significantly to lysosomes in comparison with WT (p < 0.05). All together, these results confirm that a4N489D is degraded predominantly in the proteasomal pathway, but also to a lesser extent in the lysosomal protein degradation pathway. 42

4.4.4. Un-glycosylated a4 is mostly retained in the ER

The a4 subunit is a plasma membrane isoform that would typically be processed and trafficked through the secretory pathway (178). Given the instability and apparent degradation of un- glycosylated a4, it was of interest to compare the fates of WT and un-glycosylated a4 in the secretory pathway. Fig. 4.5A shows representative confocal fluorescence micrographs of untransfected HEK 293 cells and cells transiently transfected with WT a4-2FLAG, or a4N489D- 2FLAG, and immunostained with antibodies directed toward the ER marker, calnexin (red) and anti-FLAG antibody (green). Although transfection efficiency was only 10% in this set of experiments, the data revealed a perinuclear localization of the un-glycosylated a4N489D-2FLAG in cells where it was expressed, whereas WT a4-2FLAG was more evenly distributed throughout the cell. Similarly, we looked at colocalization with Golgi, in Fig. 4.5B, where the cells were immunostained with antibodies directed toward the Golgi marker, syntaxin 6 (red), and with anti-FLAG antibody. These experiments showed that WT a4-2FLAG colocalized with the Golgi marker. In contrast, however, the un-glycosylated a4N489D-2FLAG did not colocalize with syntaxin 6. Quantification of either WT a4-2FLAG or a4N489D-2FLAG colocalization with calnexin (Fig. 4.5C) showed that there was a significantly higher association (p < 0.001) of un- glycosylated a4N489D-2FLAG with the ER (r > 0.7). These data suggested that un-glycosylated a4-2FLAG is mostly retained in the ER. Quantification of colocalization of a4-2FLAG or a4N489D-2FLAG with syntaxin 6 showed a significantly higher degree of association of WT a4- 2FLAG with Golgi (r > 0.5; p < 0.001), indicating a moderate degree of colocalization; however, correlation of colocalization of the un-glycosylated protein with Golgi was very low. These data support the finding of Fig. 4.1, suggesting that WT a4 traffics through the Golgi apparatus, even though its oligosaccharide moiety is not processed to a complex form, and they further support the notion that the un-glycosylated protein is largely retained in the ER.

4.4.5. Un-glycosylated a4N489D is unable to assemble into the V-ATPase complex

V-ATPase is a multi-subunit complex that is composed of two functional domains, the V1 and V0 sub-complexes. V-ATPase complexes can have unique tissue-specific subunit isoform compositions. For example a4 and B1 are subunit isoforms that occur together in the kidney 43

(179). In the present work we used a4 (a component of V0) and B1 (a component of V1) as markers to assess V1–V0 assembly; when the V1V0 complex is intact, a4 and B1 can be co- immunoprecipitated. Fig. 4.6A shows immunoprecipitates of WT a4-2FLAG and un- glycosylated a4N489D-2FLAG from whole cell lysates of transfected HEK 293 cells, using anti- FLAG antibodies. These immunoprecipitates were blotted and probed with anti-FLAG and anti- B1 antibodies, and the results revealed an association between B1 and WT a4-2FLAG. In contrast, there was no association between the B1 subunit and un-glycosylated a4N489D. Whole cell lysates, without immunoprecipitation, were also immunoblotted with anti-B1 antibody to show that B1 was expressed at similar levels, independent of the transfected construct (Fig. 4.6B).

4.4.6. Defective plasma membrane trafficking of un-gycosylated a4

To assess the effect of preventing glycosylation on the plasma membrane expression of a4, we constructed WT and un-glycosylated a4 with HA epitope tags within the luminal EL2 domain (see Experimental Procedures), to serve as a cell surface marker. These constructs retained also the C-terminal tandem FLAG tag of previously described a4 constructs. The new constructs were designated WT a4-3HA-2FLAG and un-glycosylated a4N489D-3HA-2FLAG. Fig. 4.7A–D show surface expression in non-permeabilized HEK 293 cells transfected with either WT a4-3HA- 2FLAG or un-glycosylated a4N489D-3HA-2FLAG constructs. The transfected cells were immunostained with anti-HA antibody (green). Comparing with the negative control (Fig. 4.7B), immunofluorescence data showed surface expression on the WT a4-transfected cells, but no signal was observed for the un-glycosylated a4N489D (Fig. 4.7C and D, respectively). As an additional control (Fig. 4.7E–G), fluorescent anti-HA staining was done on non-permeabilized HEK 293 cells expressing WT or un-glycosylated a4, then cells were permeabilized and counter- stained with fluorescent anti-FLAG antibody. This verified expression of a4N489D-3HA-2FLAG, but only the wild type showed surface HA staining.

To further confirm this finding we performed cell-surface protein biotinylation on intact cells. Fig. 4.8 shows that the glycosylated WT a4 was expressed at the cell surface, but there was no cell surface labeling, implying a lack of cell surface expression of the un-glycosylated a4N489D. 44

These data show directly that a4 N-glycosylation is required for cell surface expression of V- ATPases incorporating the a4 subunit isoform.

4.5. Discussion

4.5.1. The V-ATPase a4 subunit isoform is N-glycosylated

The goal of the present work was to assess V-ATPase a4 subunit glycosylation and its role in protein stability and in V-ATPase assembly and trafficking to the plasma membrane. We have shown here that the kidney-specific, plasma membrane-targeted human a4 isoform is N- glycosylated at one site, Asn489, situated on the second extracellular (luminal) loop, EL2. Epitope-tagged human a4, expressed in HEK 293 cells, was detected in immunoblots as an Endo H-sensitive, 105-kDa, N-glycosylated band, suggesting that a4 has a high-mannose or hybrid type carbohydrate moiety. Our data (Figs. 4.5–4.7) also support the notion that a4 is normally synthesized in the ER, followed by assembly into the V-ATPase complex and trafficking to the Golgi apparatus, and is ultimately localized to the plasma membrane. The endogenous a4 protein also retains a high mannose oligosaccharide, indicating that the protein is not processed to a complex form, yet is localized to the plasma membrane. This is in agreement with our previous findings for the glycosylation and processing (at two sites, Asn484 and Asn504, also on EL2) of the mouse a3 subunit, which is also targeted to the plasma membrane in specialized cells (161). With regard to the possible effects that epitope tagging might have on the function and trafficking of the a4 subunit, it is worth noting that in yeast we have reconstituted Vph1p (yeast V-ATPase a subunit) into Vph1p deletion strains and fully restored V-ATPase activity, even when the introduced Vph1p was C-terminally tagged with green fluorescent protein (GFP), a 238 amino acid (26.9 kDa) polypeptide. In this case, assembly, activity and localization were essentially identical to what was seen for endogenously expressed Vph1p (161).

4.5.2. U-glycosylated a4N489D is retained in the ER and is not assembled into V1V0

In yeast, several studies have shown that V0 sector assembly occurs in the ER, by a mechanism that requires multiple assembly factors. The chaperone proteins, Vma21p, Vma12p and Vma22p, facilitate the assembly of V0 in the ER, which is followed by the trafficking of the V0 sector to the Golgi apparatus for assembly with the V1 sector (151). When the yeast a subunit ortholog,

Vph1p, fails to assemble into V0 in the ER, it is rapidly degraded (180,181). There is a high degree of homology between a4 and Vph1p, and consequently it is likely that a similar mechanism accounts for a4 assembly into the V0 complex in the mammalian ER. As confirmed by immunofluorescence colocalization of glycosylated a4 with syntaxin 6 (Fig. 4.5), WT a4 does traffic to the Golgi. In contrast, our observations showed predominant colocalization of un- glycosylated a4N489D with the ER marker, calnexin, suggesting that un-glycosylated a4 is mostly retained in the ER and is unable to traffic to the Golgi. Fig. 4.6 showed that, unlike WT a4, un- N489D glycosylated a4 was not associated with the B1 subunit, indicating defective V1–V0 assembly. Taking these data into account, it seems that the carbohydrate moiety might serve as a recognition signal for V0 assembly chaperones, so that un-glycosylated a4 fails to assemble into

V0 and is retained in the ER for degradation; however the precise mechanism of un-glycosylated a4 ER retention remains unclear and will require further investigation.

4.5.3. Un-glycosylated a4N489D undergoes both proteasomal and lysosomal degradation

Others have shown that there appears to be a general association between protein stability and glycosylation of glycoproteins (182-184). It was of interest, therefore, to elucidate the role of N- glycosylation in a4 protein folding or stability. Our data indicate a significantly lower stability for un-glycosylated a4N489D relative to glycosylated WT a4 (Fig. 4.2). To rule out the possibility that a stability artifact might be induced by insertion of a point mutation into the protein primary sequence, we introduced two other mutations to eliminate N-glycosylation (N489Q and T491A in the NGT glycosylation signal sequence) (169). The a4N489Q and a4T491A constructs were significantly less stable than WT a4, but a4N489D had a relatively higher degree of stability, so it 46

was used for our investigations. As shown in Fig. 4.3, a4N489D protein stability was largely restored upon inhibition of proteasomal degradation, and to a lesser extent upon inhibition of lysosomal degradation. Quantification of colocalization of either WT a4 or un-glycosylated a4N489D with LAMP-2 (Fig. 4.5C) showed that there was a significantly higher (p < 0.001) association of un-glycosylated a4N489D with the lysosome. The correlation coefficient, r > 0.5, suggested a moderate degree of colocalization, indicating at least a partial role for the lysosomal pathway in un-glycosylated a4 degradation (Fig. 4.4A and B).

Based on the above observations in the high-expression HEK 293 cell system, we speculate that degradation of the un-glycosylated a4 protein by ER-associated degradation (ERAD) might be activated through the unfolded protein response (UPR) and ER stress (185). In a similar example, Fujita et al. (186) have shown that misfolded mutant dysferlin is degraded by two pathways, ERAD(I), the retrotranslocon-mediated ubiquitin/proteasome pathway, and ERAD(II), the alternate autophagy/lysosomal pathway that clears aggregated protein that cannot be retrotranslocated to the cytoplasm. Our data (Fig. 4.7 and 4.8) show that no a4N489D is present at the cell surface, so it seems unlikely that the endocytotic lysosomal pathway of degradation comes into play. Rather, it appears that the un-glycosylated a4N489D is largely retained in the ER as a result of protein misfolding, activating the ERAD(I) pathway and targeting the protein for proteasomal degradation. In the HEK 293 over-expression system, however, some fraction of the un-glycosylated a4 construct is likely aggregated, resulting in ER stress and activation of the autophagy/lysosomal degradation pathway of ERAD(II). Whether both these pathways would be utilized in a more native expression system is unclear, however the ultimate consequence of degradation of un-glycosylated a4 in either pathway is the absence of V1–V0 assembly and subsequent membrane trafficking of a4-containing V-ATPase, as was shown in Fig. 4.6 and 4.7. Furthermore, it has been demonstrated in several studies that proteins generally require at least one glycosylation site for effective trafficking to the plasma membrane (187,188), and our observations on a4 appear to support this generalization.

Further studies will be needed to establish more precisely the structure and the functional roles of the carbohydrate moiety of the V-ATPase a4 subunit. A better understanding of structure and function of a4 will ultimately impact the development of targeted therapeutics for the treatment 47

of some types of distal renal tubular acidosis (dRTA), and for the prevention of metastasis of malignant tumors.

Figure 4.1: Human a4 is N-glycosylated. A: Alignment of putative EL2 polypeptide sequences from 1 2 3 4 5 human (Hsa), mouse (Mmu), bovine (Bta), chicken (Gga), and Xenopus (Xtr) a4 subunits. Black highlighted residues are the highly conserved glycosylation signals, gray highlights indicate identical conserved residues; number above the first line indicates position of glycosylated Asn residues in human a4 coding sequence. Numbers at ends of lines indicate first and last residues of segment, according to predicted position in polypeptide sequence. B: HEK 293 cells were transiently transfected with C-terminal FLAG-tagged constructs, WT a4-2FLAG (indicated as WT a4), or un-glycosylated mutant a4N489D- 2FLAG (indicated as a4N489D), and protein extracts (20 µg protein per lane) were immunoblotted with anti-FLAG antibody after incubation with (+) and without (−) PNGase F or Endo H. Lane 1, WT a4- 2FLAG without endoglycosidase treatment. Lane 2, after Endo H treatment. Lane 3, after PNGase F treatment. Lanes 4 and 5, a4N489D-2FLAG with and without PNGase F treatment. C: native, endogenous 103-kDa a4 from untransfected HEK 293 cells and results after incubation with (+) and without (−) PNGase F or Endo H Data shown are representative of three independent experiments.

Figure 4.2: Glycosylation is required for a4 subunit stability. HEK 293 cells were transiently transfected with a4 (WT or un-glycosylated mutants), and expression was allowed for 24 h. Cells were then treated with CHX (10 µg/ml) for the indicated time, then harvested and immunoblotted (all samples were treated with PNGase F, indicated by ‘/PF’). A: results for a4N489D mutant; blot probed with anti- FLAG antibody. GAPDH was used as a loading control. B: same as panel A, but showing results for a4T491A mutant. C: same as panel A, but showing results for a4N489Q mutant. D: plots of band intensities quantified from panels A–C. Data were normalized to GAPDH signal and zero time control. Error bars indicate ± SD (three independent experiments). 49

Figure 4.3: a4 is degraded in both proteasomal and lysosomal degradation pathways. A: representative Western blot for a4N489D-transfected (left panels) and WT a4-transfected (right panels) HEK 293 cells treated with CHX (10 µg/ml), with and without proteasome inhibitor (10 µM MG132), for the time indicated. Western blots were probed with anti-FLAG antibody. GAPDH served as a loading control. MG132 treatment is designated ‘/MG’. B: same as panel A, but cells were treated with lysosomal inhibitor, NH4Cl (25 mM). NH4Cl treatment is designated ‘/Am’. C: plots of quantified bands from panel A. D, plots of quantified bands from panel B. Error bars indicate ± SD (three independent experiments).

Figure 4.4: Localization of WT a4 and a4N489D after transient transfection. A: representative confocal fluorescence images of WT a4-2FLAG (center panel) and un-glycosylated a4N489D-2FLAG (right panel) expressed in transiently transfected HEK 293 cells; immunostaining for a4 was with anti-FLAG antibody (green) and lysosomal localization was determined using anti-LAMP-2 antibody (red). LAMP-2 control is shown in left panel. Nuclei were stained with DAPI (blue). All panels are same scale (scale bar, 5 µm, in lower right of right panel). Images are representative of 15 each from three independent experiments. B: quantitative colocalization analysis of confocal microscopy images in panel E. Ordinate is Pearson’s correlation coefficient (r). Data shown are representative of three independent experiments (15 cells each). Error bars indicate ± SD.

Figure 4.5: Un-glycosylated a4 is mostly retained in the ER. A: representative confocal fluorescence B images of control (empty-vector transfected) HEK 293 cells (left panel), or cells transiently transfected \ with WT a4-2FLAG construct (center panel), or a4N489D-2FLAG construct (right panel). All panels show cells stained with anti-calnexin (red) and anti-FLAG (green) antibodies. Nuclei are counter-stained with DAPI (blue). B: same as in panel A, but stained with anti-syntaxin 6 antibody (red) instead of anti- calnexin. Scale bar in bottom right corner is 5 µm and applies to all panels. C: quantitative colocalization analysis of data in panels A and B. Ordinate is Pearson’s correlation coefficient (r). Results show that 52

a4N489D is mostly retained in the ER (p < 0.05). Data shown are representative of three independent experiments (20 cells each). Transfection efficiencies for these experiments were approximately 10–20%. Error bars indicate ± SD.

Figure 4.6: N-glycosylation is required for V1–V0 assembly. HEK 293 cells were transfected with empty vector (Control), or with WT a4-2FLAG, or a4N488D-2FLAG constructs. Cell lysates were subjected to immunoprecipitation, using anti-FLAG antibody. A: immunoprecipitates were blotted and probed with anti-FLAG, or anti-B1 antibodies, as indicated. B: whole cell lysate blotted before immunoprecipitation, probed with anti-B1 antibody. Blots in this figure are representative of three independent experiments.

Figure 4.7: Glycosylation is required for a4 cell-surface expression. In each of the rows, A–D, the left panel shows DAPI (blue) stained nuclei of HEK 293 cells, the center panel shows fluorescent antibody staining (as indicated) of the same field, and the right panel shows the merged images. For all images, 54

cells are intact and non-permeabilized. A: empty-vector transfected (Control) HEK 293 cells stained with anti-calnexin antibody (red). Absence of calnexin staining shows that cells are impermeable to antibody. B: same as row A, except cells were immunostained with anti-HA antibody (green). C: same as row B, but cells were transfected for transient expression of WT a4-3HA-2FLAG. Surface staining with anti-HA antibody (green) indicates surface accessibility of epitope tag. D: same as row C, except cells were transiently transfected with un-glycosylated a4N489D-3HA-2FLAG. Scale bar at top left of right panel is 5 µm; images in all panels of rows A–D in this figure are of the same magnification. For rows E–G the left panel shows DAPI (blue) stained nuclei of HEK 293 cells, the second from left panel, and third from left panels show fluorescent double antibody staining, first with anti-HA antibody on non-permeabilized cells, and then with anti-FLAG antibody after cell permeabilization, respectively (identical fields). The right panel shows the merge of all images to the left. E: empty-vector transfected (Control) HEK 293 cells stained with anti-FLAG (green) and anti-HA (red). F: same as row E, except cells were transfected for transient expression of WT a4-3HA-2FLAG. G: same as row F, but cells were transfected for transient expression of WT a4N489D-3HA-2FLAG. Cell surface staining with anti-HA antibody (red) indicates surface accessibility of epitope tag. Scale bar at top left of bottom right panel is 5 µm; all panels in rows E–G are of the same magnification. Each panel is representative of 15 images from 3 independent experiments.

Figure 4.8: Surface biotinylation of HEK 293 cells transiently transfected with WT a4-3HA-2FLAG (a4) or un-glycosylated a4N489D-3HA-2FLAG (a4N489D). Lane 1: whole cell lysate from WT a4-3HA- 2FLAG transfected HEK 293 cells; lane 2: surface protein biotinylation, showing surface protein fraction from cells transfected with WT a4-3HA-2FLAG; lane 3: same as lane 1 except cells were transfected with a4N489D-3HA-2FLAG; lane 4: same as lane 2, except cells were transfected with a4N489D-3HA- 55

2FLAG. Surface proteins on the transfected cells were subjected to biotinylation followed by streptavidin purification prior to immunoblotting. The blot was probed with anti-HA antibody and is representative of three independent experiments. Abbreviations: a.p., affinity purified (biotin-streptavidin); glyc., glycosylated; unglyc., un-glycosylated.

5. N-linked Glycosylation of a Subunit Isoforms is Critical for Vertebrate Vacuolar H+-ATPase (V-ATPase) Biosynthesis

Sally Esmail‡, Norbert Kartner‡, Yeqi Yao‡, Joo Wan Kim‡, Reinhart A. F.

Reithmeier§ and Morris F. Manolson‡§1

From the ‡Dental Research Institute, Faculty of Dentistry, University of Toronto,

Toronto, Ontario M5G 1G6, Canada, and §Department of Biochemistry, University

of Toronto, Toronto, Ontario M5S 1A8, Canada

Journal of Cellular Biochemistry, 2017 Jun 29. doi: 10.1002/jcb.26250. [Epub ahead of print]

5.1 Abstract

The a subunit of the V0 membrane-integrated sector of human V-ATPase has four isoforms, a1– a4, with diverse and crucial functions in health and disease. They are encoded by four conserved paralogous genes, and their vertebrate orthologs have positionally conserved N-glycosylation sequons within the second extracellular loop, EL2, of the a subunit membrane domain. Previously, we have shown directly that the predicted sequon for the a4 isoform is indeed N- glycosylated. Here we extend our investigation to the other isoforms by transiently transfecting HEK 293 cells to express cDNA constructs of epitope-tagged human a1–a3 subunits, with or without mutations that convert the acceptor Asn to Gln at putative N-glycosylation sites. Expression and N-glycosylation were characterized by immunoblotting and mobility shifts after enzymatic deglycosylation, and intracellular localization was determined using immunofluorescence microscopy. All un-glycosylated mutants, where predicted N-glycosylation sites had been eliminated by sequon mutagenesis, showed increased relative mobility on immunoblots, identical to what was seen for wild type a subunits after enzymatic deglycosylation. Cycloheximide pulse-chase experiments showed that un-glycosylated subunits were turned over at a higher rate than N-glycosylated forms by degradation in the proteasomal pathway. Immunofluorescence colocalization analysis showed that un-glycosylated a subunits were retained in the ER, and co-immunoprecipitation studies showed that they were unable to associate with the V-ATPase assembly chaperone, VMA21. Taken together with our previous a4 subunit studies, these observations show that N-glycosylation is crucial in all four human V- ATPase a subunit isoforms for protein stability and ultimately for functional incorporation into V-ATPase complexes via interaction with VMA21.

5.2. Introduction

V-ATPases are highly conserved, heteromultimeric nanomotors that utilize ATP hydrolysis to pump protons across membranes (151-153,189). They are composed of a multi-subunit cytoplasmic sector, V1, which hydrolyzes ATP to drive proton translocation, and a multi-subunit integral membrane sector, V0, that includes the crucial proton channel-forming a subunit (142). In mammalian V-ATPases, each complex has one copy of one of four paralogous a subunit isoforms (a1–a4) that are differentially expressed in tissues and may encode isoform-specific signals that direct assembled V-ATPase complexes to localize to distinct cellular organelles, as has been demonstrated for yeast orthologs (154,155). The a subunit has a bipartite structure (predicted size c. 95 kDa), with the N-terminal half being hydrophilic and cytoplasmically oriented, whereas the C-terminal half is integral to the membrane, consisting of 8 predicted transmembrane α-helices (TMs), separated by cytoplasmic and luminal loops, and a cytoplasmic C-terminal tail (161). Because of the complexity and membrane association of both the V- ATPase holoenzyme and its a subunit, their structures had remained poorly described until recently. By fitting X-ray crystal structures determined for some subunits, taking into account homology with the better known F-ATPase structure, and using computational methods including 3D-reconstructive cryo-electron microscopy imaging and evolutionary covariance of mutation analysis, informative structural models have emerged for both the holoenzyme and the a subunit (190,191). Clinically, a subunit malfunction is implicated in diverse V-ATPase-related isoform-specific diseases, like cutis laxa (a2), diseases of bone homeostasis (a3), cancer metastasis (a3 and a4), and renal acidosis (a4) (151,156,157,189). These clinical manifestations have recently provided strong incentives for more completely characterizing the structure and biosynthesis of the a subunit and its incorporation into the V-ATPase complex.

N-glycosylation of proteins occurs in the lumen of the endoplasmic reticulum (ER) and consists initially of the en bloc co-translational transfer of a Glc3Man9GlcNAc2 oligosaccharide chain from a dolichol pyrophosphate donor to the Asn located in an Asn-X-Ser/Thr (X≠Pro) sequon (164). The process is catalyzed by the oligosaccharyltransferase (OST) complex, but multiple cellular factors can influence its efficiency, including translation rate, oligosaccharide donor availability, and OST expression. Furthermore, the size of the OST complex requires the sequon 59

to be at least 12 amino acids from an upstream flanking transmembrane α-helix and at least 14 amino acids from a downstream flanking transmembrane α-helix for N-glycosylation to proceed (169,192). Additionally, specific amino acids at the X position (other than Pro), and at positions immediately preceding and following the sequon, may influence the efficiency of N- glycosylation, particularly for Asn-X-Ser sequons (193-196). Thus, N-glycosylation is a highly regulated process and not all potential glycosylation sites are necessarily occupied, as has been shown directly by analysis of over 3,800 glycoproteins in the SWISS-PROT and Protein Data Bank databases where only two-thirds of available N-glycosylation sites were found to be occupied (197,198). Nevertheless, N-glycosylation of a membrane protein is often crucial for quality control of its folding, for stability and protection against proteolysis, and for biological function (165-168). For a better understanding of V-ATPase structure and function, and potentially its therapeutic targeting, it is therefore important to determine directly whether putative N-glycosylation sites are indeed occupied and what role they play.

Previously, we showed that mouse a3 and human a4 are N-glycosylated (161,162,199), and we also showed that, for human a4, N-glycosylation is required for ER exit, stability, V-ATPase complex assembly and cell-surface expression (199). It remained to be determined, however, whether N-glycosylation has any biological function in other a subunit isoforms. Here we extend our characterization and understanding of the sites and biological implications for N- glycosylation on human V-ATPase a1, a2 and a3 subunit isoforms.

5.3. Materials and Methods

ENZYMES, ANTIBODIES AND REAGENTS: Restriction enzymes, peptide N-glycosidase F (PNGase F; catalog no. P0704S), and endo-β-N-acetylglucosaminidase H (Endo H; P0702S), were from New England Biolabs (Whitby, Canada). Antibodies used in the presented study were: rabbit polyclonal IgG anti-FLAG (ab1162) and mouse monoclonal IgG1 anti-syntaxin 6 (ab56656) from Abcam (Cedarlane; Burlington, Canada); rabbit IgG anti-VMA21 antibody (HPA010972) from Sigma-Aldrich (Oakville, Canada); mouse monoclonal IgG2b anti-calnexin (3H4A7; sc-130059), mouse monoclonal IgG1 anti-glyceraldehyde-3-phosphate dehydrogenase (anti-GAPDH, 0411; sc-47724), horseradish peroxidase (HRP)-conjugated goat polyclonal IgG 60

anti-mouse IgG (sc-2005) and HRP-conjugated goat polyclonal IgG anti-rabbit IgG (sc-2004) from Santa Cruz (Dallas, TX); Alexa Fluor 568-conjugated goat anti-mouse IgG (A-11004), Alexa Fluor 488-conjugated goat polyclonal IgG anti-rabbit IgG (A-11034), Alexa Fluor 488- conjugated goat polyclonal IgG anti-mouse IgG (A11001), and Alexa Fluor 568-conjugated goat polyclonal IgG anti-rabbit IgG (A-11011) from Molecular Probes, Fisher Scientific (Whitby, Canada). Phenylmethylsulfonyl fluoride (PMSF; P7626), the proteasome inhibitor N- (benzyloxycarbonyl)leucinylleucinylleucinal (MG132; C2211), and Protease Inhibitor Cocktail (P8340) were from Sigma-Aldrich. Bradford protein assay reagent (500-0006) was from Bio-Rad (Mississauga, Canada), cycloheximide (CHX; CYC003) was from BioShop (Burlington, Canada), and 4′,6-diamidine-2′-phenylindole dihydrochloride (DAPI) (10236276001) was from Roche Diagnostics (Mississauga, Canada). Octaethylene glycol mono-n-dodecyl ether (C12E8) was from NIKKO Chemicals (Barnet Products, Englewood Cliffs, NJ).

PLASMIDS AND cDNA CONSTRUCTS: All synthetic DNA fragments (a1 coding sequence, mutation inserts, and 2FLAG epitope tags) were obtained from GeneArt Gene Synthesis, Invitrogen (Thermo Fisher Scientific; Carlsbad, CA). Sequences coding for C-terminal 2FLAG (tandem DYKDDDDK) epitope tags were added to all constructs used in this study. The a1 coding sequence (ATP6V0A1; NCBI Reference Sequence NM_001130020) was synthesized with the C-terminal 2FLAG sequence by GeneArt and supplied in the pMK vector. WT a1-2FLAG was excised and inserted into the NotI restriction site of the pCMV-XL4 expression vector. The a1N495Q-2FLAG mutant construct was prepared by inserting the synthetic fragment bearing the N495Q mutation between PshAI/EcoRI restriction sites in the wild type pCMV-XL4 construct. The a2 coding sequence (ATP6V0A2; NM_012463) was obtained from the commercially available construct supplied in pCMV6-XL4 (SC115366; Origene). This construct was modified by inserting the 2FLAG coding fragment between XbaI/HindIII restriction sites to obtain the WT a2-2FLAG construct, which was then transferred to the pCMV6-XL4 vector between ApaI/HindIII restriction sites. Mutant constructs for a2N484Q-2FLAG, a2N505Q-2FLAG and a2N484Q+N505Q-2FLAG were prepared by independently inserting fragments bearing the mutations between BmgBI/Pf1FI restriction sites in the wild type construct in the pCMV-XL4 expression vector. The DNA insert for WT a3 (TCIRG1; NM_006019) was excised from the commercially available construct in the pEnter vector (CH899064; ViGene Biosciences; Rockville, MD) and inserted into the pCDNA3.1(+) expression vector between KpnI/EcoRV restriction sites. This 61

construct contained a single FLAG tag and a second was added in tandem using PCR primers incorporating the coding sequence for the second tag. The a3N483Q-2FLAG, a3N503Q-2FLAG and a3N483Q+N503Q-2FLAG constructs were prepared using synthetic DNA bearing each of these mutations, inserted into BglII/AgeI restriction sites in the wild type a3 pCDNA3.1 construct. WT a4-2FLAG and a4N489D-2FLAG were prepared as previously described (199). All constructs were confirmed by DNA sequencing.

CELL CULTURE, TRANSFECTION, AND EXPRESSION OF HUMAN V-ATPASE A SUBUNIT ISOFORMS: Human embryonic kidney cells (HEK 293; CRL-1573TM) were from the American Type Culture Collection (ATCC; Manassas, VA). Cell culture media and reagents, Dulbecco’s modified Eagle’s medium (DMEM; 11965092), heat-inactivated fetal bovine serum (FBS; 16140071), penicillin/streptomycin mix (15140122), phosphate buffered saline (PBS; 10010023), Dulbecco’s phosphate buffered saline (DPBS; 1404182), and 1X Trypsin/EDTA (25200056), were from Gibco (Fisher Scientific). GenJet In Vitro DNA Transfection Reagent (SL100488) was from SignaGen Laboratories (Rockville, MD).

Liquid nitrogen-stored HEK 293 cells were revived by rapid thawing in a water bath at 37 °C and subsequently incubated in 75 cm2 tissue culture flasks (containing 17 ml DMEM supplemented with 10% FBS and 1% penicillin/streptomycin mix) in a humidified 5% CO2 incubator for 3 days at 37 °C. Cells were trypsinized at 70–80% confluence, using 1 ml of 1X Trypsin/EDTA per drained flask, and then seeded into 6-well plates at densities of 4–7 × 105 cells/well. For transient expression of constructs, after 24 h of growth, the cells were transfected with 1 µg/well of plasmid construct in a transfection complex. The transfection complex consisted of GenJet in Vitro DNA Transfection Reagent and plasmid DNA in the ratio 3:1 (µl GenJet to µg DNA), and was diluted to 200 µl with serum-free DMEM and then incubated for 10 min prior to addition to the drained cell monolayers (transfection time zero). Cells were incubated as usual and were harvested for analysis 24 h post-transfection.

WHOLE-CELL LYSATES PREPARATION: Whole-cell lysates were prepared as previously described (199). Briefly, cells were harvested 24 h post-transfection in 0.2 ml/well lysis buffer (PBS containing 1% C12E8, 1 mM PMSF and 1:100 (v/v) Protease Inhibitor Cocktail) and incubated on ice for 30 min. Lysates were then centrifuged at 15,000g for 30 min at 4 °C. Protein concentrations of collected supernatants were assayed using the Bradford protein assay. 62

DEGLYCOSYLATION OF HUMAN a SUBUNIT ISOFORMS: HEK 293 cells were transfected, cultured for 24 h, and then whole-cell lysates were prepared, followed by deglycosylation analysis with Endo H or PNGase F. Briefly, 3 µl of 10X Glycoprotein Denaturation Buffer (5% sodium dodecyl sulfate, 0.4 M dithiothreitol; New England Biolabs) was added to 30 µg of whole-cell lysate protein and the final volume was adjusted to 20 µl with distilled H2 M sodium phosphate, pH 7.4 at 25 °C, for PNGase F, or 0.5 M sodium citrate, pH 7.5 at 25 °C, for Endo H; New England Biolabs), 2 µl 10% (w/v) NP-40 (New England Biolabs), and 2,000 u

PNGase F, or Endo H, were added. The final volume was adjusted to 40 µl with distilled H2O, incubated for 1 h at 37 °C, and then analyzed by SDS-PAGE/immunoblotting.

IMMUNOBLOTTING ANALYSIS AND PROTEIN STABILITY ASSAY: Protein expression analysis was performed by immunoblotting. Briefly, 20 µg of whole-cell lysate proteins, in equal volumes of 2X SDS sample loading buffer, were loaded per lane. Samples were run on 8% SDS- PAGE and transferred onto nitrocellulose membrane. The membrane was immunostained with anti-FLAG antibody (1:3,000 dilution), followed by HRP-labeled second antibody (1:5,000 dilution), to detect FLAG-tagged proteins. The V0 assembly factor, VMA21, was detected using anti-VMA21 antibody (1:2,000 dilution). To provide loading controls, blots were also probed with anti-GAPDH antibody (1:5,000 dilution) and HRP-labeled second antibody.

Protein stability was assessed using the cycloheximide (CHX) pulse-chase assay. Briefly, 24 h post-transfection, HEK 293 cells were exposed to 10 µg/ml CHX with and without lysosomal inhibitor (25 mM NH4Cl), or proteasomal inhibitor (10 µM MG132), for 0, 4, 8 and 12 h. The cells were then harvested at the indicated time points, whole-cell lysates were prepared, and immunoblots were probed with anti-FLAG antibodies, and anti-GAPDH antibodies as a loading control.

QUANTITATIVE PROTEIN BAND ANALYSIS: Protein band intensities of CHX immunoblots were estimated using Bio-Rad Quantity One 4.6.9 software. The rolling-disc method of band quantification was used, and background was subtracted from protein band signals. The significance of differences in quantified band intensities was determined using the two-tailed Student’s t-test. Relative protein levels were calculated by adjustment of intensities of bands of interest to GAPDH loading controls and then normalizing to zero time controls. 63

Protein-half lives (h) were calculated by GraphPad Prism 5 software. Briefely, curve fitting equation were modified either as one phase decay [Y= (Y0) *exp (-K*X)] or plateau followed by one phase decay [Y= IF (X

Glycosylated proteins typically run in SDS-PAGE as diffuse bands, possibly due to carbohydrate moiety heterogeneity, and this confounds quantification results. Therefore, for accurate band intensity estimation, WT samples were treated with PNGase F prior to immunoblotting, yielding sharper protein bands for more accurate analysis.

IMMUNOFLUORESCENCE of a SUBUNIT ISOFORMS AND COLOCALIZATION ANALYSIS: HEK 293 cells grown on glass coverslips and transiently expressing WT and un- glycosylated mutant human a1, a2 or a3 were washed with DPBS and fixed with 3.7% (w/v) paraformaldehyde for 15 min at room temperature. Cells were then permeabilized with DPBS containing 0.2% Triton X-100 at room temperature for 15 min. Cells were blocked with DPBS containing 5% bovine serum albumin for 30 h at room temperature, followed by immunostaining with anti-FLAG (1:1,000), anti-syntaxin 6 (1:500), or anti-calnexin (1:500) antibodies in DPBS containing 3% bovine serum albumin for 45 min at room temperature. Cells then were washed with DPBS and immunostained with fluorescent-labeled second antibodies (1:500) for 45 min at room temperature, and nuclei were stained with 0.1 mg/ml DAPI in DPBS for 10 min. Finally, cells were mounted with ProLong Gold Antifade Reagent (Fisher Scientific). Images were obtained using a Quorum Spinning Disk Confocal System equipped with a Hamamatsu C9100- 13 EM-CCD, Yokogawa CSU X1 scan head, and Improvision Piezo focus drive (Imaging Facility, Hospital for Sick Children, Toronto, Canada).

Colocalization analyses of 15 micrographs from three independent experiments were performed for each a subunit construct, using Volocity v6.3 3D image analysis software (PerkinElmer, Woodbridge, Canada). Co-associations of two fluorescent signals (red and green) were quantified and expressed as Pearson’s correlation coefficients (r). Significance of differences between WT and mutants were estimated using two-tailed Student’s t tests.

Co-IMMUNOPRECIPITATION: Whole-cell lysates of HEK 293 cells expressing WT and un- glycosylated mutant constructs, 24 h post-transfection, were prepared in IP Buffer (150 mM 64

NaCl, 25 mM Tris HCl, pH 7.2 at 25 °C, containing 1% C12E8, 1:100 (v/v) Protease Inhibitor Cocktail and 1 mM PMSF), as previously described (199). Immunoprecipitations were performed by incubating 50 µg whole-cell lysate protein with 5 µg of anti-FLAG antibody overnight at 4 °C with agitation. Antigen-anti-FLAG antibody complexes were affinity purified by incubation with 100 µl (50% packed volume) of protein A agarose beads for 2 h at room temperature with gentle agitation. Antigens were eluted by incubation of beads with SDS-PAGE sample buffer for 5 min at 95 °C, and supernatants were collected after centrifugation at 2,500g for 3 min. These were analyzed by immunoblotting with anti-FLAG and anti-VMA21 antibodies to compare the assembly of WT and un-glycosylated a subunit isoforms.

5.4. Results

5.4.1. N-glycosylation of V-ATPase a1, a2 and a3 subunits

There had been no direct biochemical evidence that human a1, a2 and a3 subunits are N- glycosylated, but the high degree of sequon conservation in orthologous a subunit isoforms strongly suggests that this must be the case, as had been demonstrated directly for a4 (199). Fig. 5.1A shows polypeptide sequence alignments of the EL2 domains of the four paralogous human a subunit isoforms with their orthologs in different species. There are four isoforms in each of the tetrapod vertebrates examined, all of which have positionally conserved N-glycosylation sequons. The two yeast isoforms of the V-ATPase a subunit, Vph1p and Stv1p, are shown for comparison. The yeast orthologs remain highly conserved in their polypeptide sequence, but unlike the vertebrate isoforms, they do not have conserved sequons and are not N-glycosylated. In Fig. 5.1B, a dendrogram indicates that human a1 and a4 polypeptides are most closely related to each other, and that a2 and a3 are similarly related. Also, all of the paralogous a subunit isoforms are closely related to each other, and their orthologs are highly conserved, so they presumably originated from an ancestral a subunit gene that underwent gene duplication events prior to early vertebrate evolution. Most membrane proteins are N-glycosylated in single extracellular loops, typically near the beginning of the membrane spanning region (168) and the a subunits follow this pattern, utilizing the first extracellular loop large enough to support N- glycosylation. Single, positionally conserved, sequons are found within the EL2 extracellular 65

loop (161) of the C-terminal, integral membrane domain of the related a1 and a4 subunit isoforms, with N-glycosylation sites at Asn495 and Asn489, respectively, while two similarly conserved sequons are found within EL2 of a2 and a3, with N-glycosylation sites at Asn484+Asn505 and Asn483+Asn503, respectively. Further implications of the data shown in Fig. 5.1 are addressed in the Discussion section.

In the present study, we assessed a subunit glycosylation using the same methods previously reported by us in the characterization of the human a4 isoform (199). Two enzymes were used as probes of glycosylation: the glycoamidase, PNGase F, which removes high mannose, hybrid and complex carbohydrate moieties from the polypeptide by cleavage between the proximal GlcNAc and Asn, and is used to determine whether a protein is N-glycosylated (200), and the endoglycosidase, Endo H, which cleaves (leaving the proximal GlcNAc attached to the polypeptide) core, high-mannose and mannose-rich hybrid glycans that have not been processed by Golgi α-mannosidase II (172,173). Fig. 5.2A shows immunoblots of FLAG-tagged a1 subunit constructs, WT a1-2FLAG, and the mutant protein, a1N495Q-2FLAG (where the putative sequon Asn acceptor was disrupted by conversion to Gln), expressed transiently in HEK 293 cells, with and without PNGase F or Endo H treatment of cell lysates. WT a1-2FLAG was observed as a 105-kDa band and, upon PNGase F and Endo H treatment, its relative mobility was reduced to 103 kDa, representing the deglycosylated a1-2FLAG. Fig. 5.2B shows immunoblots of FLAG- tagged a2 subunit constructs, WT a2-2FLAG, and the mutant protein, a2N484Q+N505Q-2FLAG (where both putative sequons were disrupted), expressed transiently in HEK 293 cells, with and without PNGase F or Endo H treatment. WT a2-2FLAG was observed as a 110-kDa band and, upon PNGase F or Endo H treatment, its relative mobility was reduced to 105 kDa, representing the deglycosylated a2-2FLAG. Fig. 5.2C shows immunoblots of FLAG-tagged a2 subunit constructs, WT a2-2FLAG, and the mutant proteins, a2N484Q 2FLAG or a2N505Q-2FLAG (where the two putative sequons were each independently disrupted), expressed transiently in HEK 293 cells, with and without PNGase F or Endo H treatment. Both a2N484Q 2FLAG and a2N505Q- 2FLAG were observed as 107-kDa bands, and upon PNGase F or Endo H treatment the band sizes were reduced to 105 kDa, consistent with the size of the deglycosylated WT a2-2FLAG. Fig. 5.2D shows immunoblots of FLAG-tagged a3 subunit constructs, WT a3-2FLAG, single- un-glycosylated mutant proteins, a3N483Q-2FLAG and a3N503Q-2FLAG, and the un-glycosylated double mutant protein, a3N483Q+N503Q-2FLAG, expressed transiently in HEK 293 cells, with and 66

without PNGase F treatment. WT a3-2FLAG was observed as a 110-kDa band and, upon PNGase F treatment, the band size was reduced to (for single disruption of sequons) 107 kDa and (for double disruption of sequons) 105 kDa. Similar results were seen for human a3 treated with Endo H, as shown in Fig. 5.2E. Taken together, these results demonstrated that all of the predicted sequons are, in fact, N-glycosylated.

5.4.2. N-glycosylation requirement for stability of a subunits

Our previous work with the human V-ATPase a4 subunit showed that its stability is dependent on N-glycosylation (199). Furthermore, the latter work and previous evidence showed that C- terminal domain tagging of the V-ATPase a subunit does not affect its stability or intracellular trafficking (161,162). Thus, C-terminally FLAG-tagged WT a1, a2 and a3, or their un- glycosylated mutant constructs (a1N495Q, a2N484Q+N505Q, a3N483Q+N503Q), were transfected into HEK 293 cells to provide insights into the role that N-glycosylation plays in stability and localization of human a1, a2 and a3 subunit isoforms. Initial observations, consistent with those of previous work, using human WT a4-2FLAG, or un-glycosylated a4N489D-2FLAG (199), showed that the un-glycosylated mutants a1N495Q, a2N484Q+N505Q and a3N483Q+N503Q were expressed at a lower level than their wild type counterparts after transient expression in HEK 293 cells (results not shown).

As with our previous work on a4, the reduction in expression of the mutant a subunit isoforms prompted us to compare the stabilities of wild type and un-glycosylated mutant proteins. CHX pulse-chase experiments were performed for this purpose. Whole-cell lysates were prepared from cells that were transfected with WT a1, WT a2 and WT a3, or their un-glycosylated mutants (a1N495Q, a2N484Q+N505Q, a3N483Q+N503Q) and pretreated with 10 µg/ml CHX for up to 12 h. All samples were treated with PNGase F prior to immunoblotting and analysis to remove the glycan moieties. Fig. 5.3 shows quantitative band analysis of these immunoblots. All band intensities were normalized to GAPDH loading controls and to the zero time controls. Fig. 5.3A–C and table 5.1 show quantitatively that un-glycosylated mutants were significantly less stable than the corresponding WT a subunits. Fig. 5.3A shows that a1N495Q-2FLAG had a half-life of 4.8± 0.41 h compared to 13.45± 1.89 h for WT-a1-2FLAG, Fig. 5.3B that a2N484Q+N505Q-2FLAG, a2N484Q- 2FLAG and a2N505Q-2FLAG had half-lives of 7.81± 0.45, 10.77± 0.42 and 15.75± 0.83 h, respectively, compared to 32.21± 1.64 h for WT-a2-2FLAG, and Fig. 5.3C that a3N483Q+N503Q- 67

2FLAG, a3N483Q-2FLAG and a3N503Q-2FLAG had half-lives of 5.59± 0.187, 7.79± 0.023 and 10.87± 0.375 h, respectively, compared to 20.79± 2.2 h for WT-a3-2FLAG (p<0.005).

To determine whether the higher turnover rates observed in the cells transfected with un- glycosylated constructs were due to degradation in proteasomal or lysosomal pathways, these activities were inhibited with MG132 or NH4Cl, respectively, as previously described (199,201). Data presented in Fig. 5.4A showed that the degradation of un-glycosylated a1N495Q-2FLAG was restored to the level observed for WT-a1-2FLAG upon proteasomal pathway inhibition with MG132 (p<0.0001). Similarly, the degradation of un-glycosylated a2N484Q+N505Q-2FLAG and a3N483Q+N503Q-2FLAG was diminished to the levels observed for their corresponding wild type proteins (p<0.05), as shown in Figs. 5.4C and E. Upon inhibition of the lysosomal pathway, however, there was no significant reduction in rates of degradation of un-glycosylated proteins (p>0.05), as shown in Figs. 5.4B, D, F and table 5.1. Collectively, these findings demonstrated that N-glycosylation is critical for the stability of human V-ATPase a1, a2 and a3 subunit isoforms, and in its absence these proteins are turned over at a higher rate through ER-associated protein degradation (ERAD) in the proteasomal pathway.

5.4.3. N-glycosylation requirement for ER exit of a subunits

Degradation of un-glycosylated a1, a2 and a3 subunits in the proteasomal pathway strongly suggested that human V-ATPase a subunits require N-glycosylation to pass ER quality control in order to exit from the ER. To test this, the fates of un-glycosylated a1N495Q-2FLAG, a2N484Q+N505Q-2FLAG and a3N483Q+N503Q-2FLAG in the secretory pathway were investigated by fluorescence microscopy. Fig. 5.5A shows representative fluorescence micrographs of HEK 293 cells that transiently expressed either WT-a1-2FLAG or a1N495Q-2FLAG (anti-FLAG stain; green). The ER-specific marker, calnexin, was immunostained (red) in all panels to show ER distribution in control empty-vector transfected cells (left panel), cells expressing wild type a subunit (middle panel), and cells expressing un-glycosylated mutant a subunit (right panel). Little colocalization of calnexin and WT a subunit was seen, but a strong perinuclear colocalization of a1N495Q-2FLAG with calnexin-positive compartments was observed. Similarly, un-glycosylated a2N484Q+N505Q-2FLAG and a3N483Q+N503Q-2FLAG also showed significant 68

perinuclear colocalization with calnexin (Figs. 5.5B and C). Quantitative colocalization analysis of all WT a subunit isoforms (a1, a2 and a3) and their corresponding un-glycosylated mutants with calnexin was performed and charted, as shown in Fig. 5.5D. This analysis showed that there was a significantly higher association of un-glycosylated a1N495Q-2FLAG, a2N484Q+N505Q-2FLAG and a3N483Q+N503Q -2FLAG with the ER, in comparison with wild type proteins (p < 0.001; r = 0.5–0.8). This suggested that un-glycosylated a1N495Q-2FLAG, a2N484Q+N505Q-2FLAG and a3N483Q+N503Q -2FLAG were mostly retained in the ER.

We also looked at colocalization with the Golgi-specific marker, syntaxin 6. Fig. 5.6A–C, shows cells immunostained with anti-syntaxin 6 (red) and anti-FLAG antibodies (green). Merged images revealed that WT a1-2FLAG, WT a2-2FLAG and WT a3-2FLAG colocalized with the Golgi marker. In contrast, the un-glycosylated a1N495Q-2FLAG, a2N484Q+N505Q-2FLAG and a3N483Q+N503Q-2FLAG did not colocalize with syntaxin 6. Quantification of colocalization of all WT a subunit isoforms (a1, a2 and a3), and their corresponding un-glycosylated mutants, with syntaxin 6 (Fig. 5.6D), showed that there was a significantly higher association of WT a1- 2FLAG, WT a2-2FLAG and WT a3-2FLAG with Golgi in comparison with mutant proteins (p < 0.001; r = 0.4–0.6). Colocalization of un-glycosylated protein with Golgi was, in fact, very low. These data suggested that the WT proteins, but not their un-glycosylated counterparts, were capable of intracellular trafficking to Golgi.

5.4.4. N-glycosylation requirement for association of a subunits with ER- resident V-ATPase assembly factor, VMA21

Mammalian V-ATPase V1–V0 assembly remains poorly understood, but VMA21 has been identified as the mammalian ortholog of the ER-resident yeast V-ATPase assembly factor,

Vma21p (202). VMA21 binding to the a subunit is required for V0 assembly; therefore, we investigated whether N-linked glycosylation of human V-ATPase a subunits plays a role in this interaction. Fig. 5.7A shows immunoprecipitates of WT a1-2FLAG, WT a2-2FLAG, WT a3- 2FLAG and WT a4-2FLAG, and also un-glycosylated a1N495Q-2FLAG, a2N484Q+N505Q-2FLAG a3N483Q+N503Q-2FLAG and a4N489D-2FLAG, pulled down with anti-FLAG antibodies from whole- cell lysates of transfected HEK 293 cells. These immunoprecipitates were blotted and probed 69

with anti-FLAG and anti-VMA21 antibodies, and the results revealed co-immunoprecipitation and, thus, a presumptive association of VMA21 with all of the WT a subunit isoforms. In contrast, there was no apparent association between any of the un-glycosylated mutant a subunit isoforms and VMA21. As a control, whole-cell lysates, without immunoprecipitation, were immunoblotted with anti-VMA21 antibody to show that VMA21 was expressed at similar levels for all transfected constructs (Fig. 5.7B). These data supported the notion that N-glycosylation may be a critical factor for a subunit–VMA21 association in the ER, and consequently for subsequent V-ATPase assembly.

5.5. Discussion

5.5.1. All human V-ATPase a subunit isoforms are N-glycosylated

Amino acid alignments of the four mammalian V-ATPase a subunit isoforms with orthologs from various species, revealed positionally highly conserved N-glycosylation sequons in EL2 of the membrane-integrated C-terminal domain of the protein. In the polypeptide sequences of other tetrapod vertebrate species examined, all of which had four a subunit isoforms, a1 and a4 had single putative N-glycosylation sites, whereas a2 and a3 had two sites each (Fig. 5.1A). A dendrogram comparing a subunit isoform homologies suggested that the polypeptide sequence of the a1 isoform is most closely related to a4, and a2 to a3 (Fig 5.1B). These comparisons, and the high degree of homology among all a subunit isoforms in all vertebrate species, further suggested that a gene duplication event involving a single ancestral a subunit gene gave rise to a3 and a1/2/4 ancestors. Subsequent gene duplications then gave rise to a2 and a1/4 ancestors, and more recently to a1 and a4, but all of these events likely predated the appearance of tetrapod vertebrates c. 400 million years ago.

The observation of positional conservation of the N-glycosylation sequons suggested that they have functional importance, which in turn requires that they actually are glycosylated. Glycosylation is of importance in organisms for cell surface recognition and signaling functions in embryogenesis and tissue differentiation, but N-glycosylation can be important also in protein biosynthesis for determining protein stability, folding, sorting, and trafficking (166,182,188,199). Nevertheless, there is evidence that not all N-glycosylation sequons are necessarily occupied by N-linked oligosaccharide and that, where there are multiples sites, their actual glycosylation is 70

not of equal importance for determining successful glycoprotein biosynthesis and function (203). Thus, in spite of strong predictive evidence, N-glycosylation must be confirmed by direct experimental means. In the present work, N-glycosylation of all of the predicted sequons was confirmed directly for each of the human a subunit isoforms at: a1, Asn495; a2, Asn484+505; a3, Asn483+503; and (in previous work) a4, Asn489 (see Fig. 5.2, and previous work (199)). All of the epitope-tagged human a subunit isoforms expressed in HEK 293 cells were detected in immunoblots as both PNGase F- and Endo H-sensitive bands, suggesting that all have mannose- rich oligosaccharide moieties, unprocessed by Golgi α-mannosidase II. However, our data also confirm that a subunit glycoproteins, after synthesis in the ER, undergo trafficking to the Golgi apparatus (Figs. 5.5 and 5.6) as part of the V-ATPase complex.

Efficiency of N-glycosylation of Asn-X-Ser sequons has been shown to be sensitive to amino acid composition at the X position and at the positions immediately preceding (X-2) and following (Y) the sequon (193-196). In in vitro systems, Trp at the X-2 position is detrimental to N-glycosylation (c. 32% efficiency (195)) and Asp at the X position is also detrimental (c. 20% efficiency (194)). The former detrimental sequence occurs at both sequons in the human a2 subunit, and the latter additionally at the second sequon, suggesting the likelihood of a very low additive N-glycosylation efficiency at this site (Fig. 5.1A). More recently, however, Malaby et al. have shown that experimental cell-based systems for N-glycosylation are less sensitive to amino acid variations at these positions than in vitro (e.g. N-glycosylation where Asp is at the X position is 83% efficient in cells (193), compared with c. 20% in vitro (194)). In the present work, we did not observe any secondary un-glycosylated bands for any of the wild type subunits or for the single-glycosylated a2 and a3 mutant constructs. This suggests a high efficiency of N- glycosylation at all human a subunit sequons in the HEK 293 exogenous expression system that we used, more consistent with the cell-based observations of Malaby et al. (193).

On further examination of the polypeptide alignments shown in Fig. 5.1A, it is of interest to note that, while N-glycosylation sequons of the vertebrate a subunit isoforms appear to be highly conserved in position and number, the yeast a subunit isoforms, Vph1p or Stv1p, have no conserved sequons; moreover, there is no experimental evidence for their N-glycosylation. Also, while the flanking regions within EL2 are highly conserved for all isoforms in all species, including yeast, the c. 33 amino acid central sequence, where the vertebrate N-glycosylation 71

sequons are located, is relatively poorly conserved, and a large part of this same region is also missing in the yeast orthologs (Fig. 5.1A). These observations suggested that ancestral pre- vertebrate species acquired what might be a “spacer” or “platform”, the exact amino acid sequence of which is unimportant, having the linear dimension to support and separate two N- linked oligosaccharides present on the ancestral a subunit glycoprotein. Interestingly, this platform is present (positionally and dimensionally, not by sequence conservation) in a1 and a4, even though these isoforms have only a single N-glycosylation sequon. This might suggest that the a subunit lineage leading to a1 and a4 had two glycosylation sites and that one of these was lost as a1 and a4 evolved functional identities different from those of a2 and a3 (as alluded to in Fig. 5.1B). It is of interest also that, whereas the position of the first sequon was retained in most vertebrate a1 and a4 orthologs, the position of the second sequon was retained in Xenopus a1, supporting the notion that there were likely two sequons prior to divergence of a1 and a4, and while the first sequon was favored over the second in a1, evolutionarily, there is apparently an example where this is reversed.

5.5.2. N-glycosylation of a subunits contributes to their stability

There are numerous examples where glycoprotein stability is dependent on N-glycosylation (182-184), and our CHX-chase data similarly indicated a significantly higher turnover rate for un-glycosylated a subunits relative to the glycosylated wild type isoforms (Fig. 5.3, table 5.1 and (199)). It is interesting to note also that for a2 and a3, the single-glycosylated mutants had intermediate stability between wild type and fully un-glycosylated mutants. This suggested that some glycosylation is better than none for the stability of these isoforms, but also that single glycosylation cannot fully rescue the stability of the protein after one oligosaccharide is lost, possibly because the glycans of the two sites are different, with somewhat different functions that impact protein stability. Furthermore, as shown in Fig. 5.4, a subunit protein turnover rates were largely restored to wild type levels upon inhibition of proteasomal degradation with

MG132, but not upon inhibition of lysosomal degradation with NH4Cl, indicating that the un- glycosylated a subunits were degraded primarily through the ERAD pathway, which is the fate typically observed for misfolded proteins (185). 72

5.5.3. N-glycosylation of a subunits is required for ER exit

Quantification of colocalization of either WT or un-glycosylated a subunits with calnexin (Fig. 5.5) showed that there was a significantly higher (p < 0.001) association of un-glycosylated a subunit with the ER compartment, suggesting that there was ER retention of the un-glycosylated protein, likely as part of the quality control mechanism. As was confirmed by immunofluorescence, the exogenously expressed glycosylated wild type a subunit colocalized with syntaxin 6 (Fig. 5.6), indicating that it was fully capable of trafficking to the Golgi. As we have shown for a4 in previous work (199), it appeared that the un-glycosylated a subunits, in general, were largely retained in the ER as a result of protein misfolding, culminating in their being targeted for ubiquitin-mediated proteasomal degradation. The ultimate consequence of retention and degradation of un-glycosylated a subunits was their absence for V1–V0 assembly and subsequent further membrane trafficking of the V-ATPase holoenzyme, in comparison with wild type a subunits, as was shown in Figs. 5.5 and 5.6.

5.5.4. N-glycosylation is required for association with assembly factor, VMA21

Assembly of the V1V0 complex is incompletely understood, but studies in yeast have demonstrated that assembly of the V0 sector requires the ER-resident chaperones, Vma12p,

Vma21p, and Vma22p. These chaperones facilitate the assembly of V0 in the ER, after which membrane trafficking takes it to the Golgi apparatus for further assembly with the V1 sector

(151). In yeast, if the a subunit ortholog, Vph1p, fails to assemble into V0 in the ER, it is rapidly degraded (180,181). There is a high degree of homology between human a subunits and the yeast equivalent, Vph1p, and consequently a similar mechanism likely accounts for a subunit assembly into the V0 complex in human ER. VMA21 is the human homolog of the yeast Vma21p, and is the only essential chaperone identified for human V-ATPase assembly, to date (202). We showed by co-immunoprecipitation experiments (Fig. 5.7 and (199)) that, in contrast to wild type subunit isoforms, all un-glycosylated a subunit isoforms were unable to bind VMA21. Thus, in the absence of glycosylation, this essential assembly chaperone was functionally unavailable to the a subunit to ensure its inclusion in V0 assembly. 73

5.6. Conclusion

Taking all our observations into account, it seems that the oligosaccharide moieties of the human

V-ATPase a subunits might serve as recognition signals for V0 assembly chaperones, such as

VMA21, so that un-glycosylated a subunit fails to assemble into V0 and is retained in the ER, ultimately destined for degradation; however, the precise mechanism of un-glycosylated a subunit ER retention remains unclear and will require further investigation. Further studies will also be needed to establish more precisely the nature of the oligosaccharide moieties of the V- ATPase a subunits. A better understanding of their structure and function could ultimately impact the development of targeted therapeutics for treatment of the diverse diseases in which they are implicated, like cutis laxa, malignant osteopetrosis, distal renal tubular acidosis, and tumor metastasis.

Figure 5.1: N-glycosylation sequons of a subunit isoforms from various species. A: alignments of V- ATPase EL2 domain polypeptide sequences from various vertebrate species, grouped according to isoform, a1–a4, and compared also with yeast Vph1p and Stv1p, which are not N-glycosylated. These sequences are flanked on the left by the transmembrane α-helix TM3 and on the right by TM4 (TM sequences not shown). Abbreviations: Bta, Bos taurus (cattle; mammalian); Hsa, Homo sapiens (human; mammalian); Gga, Gallus gallus (chicken; avian); Mmu, Mus musculus (mouse; mammalian); Stv, Stv1p of Saccharomyces cerevisiae (baker’s yeast; fungal); Vph, Vph1p of S. cerevisiae; Xtr, Xenopus tropicalis (Western clawed frog; amphibian). Numbers above lines are Asn positions in N-glycosylation sequons of human subunit sequences (Xenopus in brackets, for a1). Highlights: red, identity throughout species and isoforms (conserved substitutions in magenta); dark grey, identity throughout vertebrate 75

species and isoforms (conserved substitutions in light grey); black, N-glycosylation sequons. Asterisk after human a2 subunit position 507 shows the point of insertion of sequences in the right margin (positions 508–517). These are mammalian a2-specific insert sequences, shown separately to avoid large gaps in alignments. B: dendrogram showing a subunit isoform divergence; putative gene duplication events and number of N-glycosylation sites in EL2 domains of present a subunit isoforms are indicated (and speculative numbers for ancestral proto-isoforms; see Discussion).

Figure 5.2: Human a1, a2 and a3 subunit isoforms are N-glycosylated. HEK 293 cells were transiently transfected with C-terminal FLAG-tagged constructs of a subunit isoforms, then whole-cell lysates were collected after 24 h expression, treated with PNGase F or Endo H, and immunoblotted (20 µg protein per lane) with anti-FLAG antibody, as follows: A: WT a1-2FLAG (WT a1), or un- glycosylated mutant a1N495Q-2FLAG (a1N495Q), treated with (+) or without (-) PNGase F or Endo H. For a1 isoform constructs, glycosylated bands were observed at c. 105 kDa (upper arrow), deglycosylated and un-glycosylated bands at c. 103 kDa (lower arrow). B: WT a2-2FLAG (WT a2), or un-glycosylated mutant a2N484Q+N505Q-2FLAG (a2N484Q+N505Q). For a2 isoform constructs, glycosylated bands were 76

observed at 110 kDa (upper arrow), deglycosylated and un-glycosylated bands at 105 kDa (lower arrow). C: single-glycosylated mutants a2N484Q-2FLAG (a2N484Q) or a2N505Q-2FLAG (a2N505Q). Single-glycosylated a2 isoform bands were observed at 107 kDa (upper arrow), deglycosylated bands were observed at 105 kDa (lower arrow). D: WT a3-2FLAG (WT a3), single-glycosylated mutants a3N483Q-2FLAG (a3N483Q), a3N503Q-2FLAG (a3N503Q) or un-glycosylated mutant a3N483Q+N503Q-2FLAG (a3N483Q+N503Q). Fully glycosylated bands were observed at 110 kDa (upper arrow), single-glycosylated bands at 107 kDa (middle arrow) and deglycosylated bands at 105 kDa (lower arrow). This panel shows results only with or without PNGase F treatment. E: same as panel D, but showing results only for Endo H treatment. In all cases, Endo H treatment resulted in the same mobility shift that was seen with PNGase F. Blots in all panels are representative of three independent experiments.

Figure 5.3: Glycosylation is required for a1, a2 and a3 stability. HEK 293 cells were transfected with a1, a2 or a3 constructs (FLAG-tagged WT and un-glycosylated mutants, as indicated), and expression 77

was allowed for 24 h. Cells were subsequently treated with CHX (10 μg/ml) for the indicated time, and then whole-cell lysates were collected and separated by SDS-PAGE and immunoblotted with anti-FLAG antibody (GAPDH was used as a loading control). A-C: plots of band intensities quantified from immunoblots. Data were normalized to GAPDH signal and zero time control. Error bars indicate ± S.D.

(three independent experiments). 78

Figure 5.4: Subunits a1, a2 and a3 are degraded in the proteasomal pathway. A: plots of quantified bands from immunoblots of WT a1 and un-glycosylated mutant a1N495Q, transfected into HEK 293 cells subsequently treated with CHX (10 μg/ml), with and without proteasome inhibitor (10 µM MG132) for 79

the time indicated. Immunoblots were probed with anti-FLAG antibody. Probing with anti-GAPDH antibody served as a loading control. MG132 treatment is designated ‘+MG’. B: same as panel A, but

cells were treated with lysosomal inhibitor, NH4Cl (25 mM). NH4Cl treatment is designated ‘+Am’. C, same as panel A, except quantified bands were from immunoblots of WT a2 and un-glycosylated mutant a2N484Q+N505Q. D, same as panel C: but cells were treated with lysosomal inhibitor. E and F: same as panels C and D, respectively, but quantified bands were from immunoblots of WT a3 and un- glycosylated a3N483+N503Q. Proteasomal, but not lysosomal, inhibitors restored the rates of degradation of the un-glycosylated a1, a2 and a3 to those seen for wild type proteins. Data were normalized to GAPDH signal and zero time control. Error bars indicate ± S.D. (three independent experiments).

Figure 5.5: Un-glycosylated a1, a2 and a3 are mostly retained in the ER. A: representative confocal B \ fluorescence images of control (empty-vector transfected) HEK 293 cells (left panel), or cells transiently transfected with WT a1-2FLAG construct (middle panel), or a1N495Q-2FLAG construct (right panel). All 80

panels show cells stained with anti-calnexin (red) and anti-FLAG (green) antibodies. Nuclei are counter- stained with DAPI (blue). B: same as in panel A, but cells were transfected with WT a2-2FLAG (middle panel), or un-glycosylated a2N484Q+N505Q-2FLAG (right panel). C: same as in panel A, but cells were transfected with WT a3-2FLAG (middle panel), or un-glycosylated a3N483Q+N503Q-2FLAG (right panel). All panels are the same magnification; scale bar in lower right panel is 5 µm. D: quantitative colocalization analysis of data in panels A, B and C. Ordinate is Pearson’s correlation coefficient (r). Results show that un-glycosylated a1N495Q, a2N484Q+N505Q and a3N483+N503 are mostly retained in the ER (p < 0.001). Images are representative of 15 each, from three independent experiments.

Figure 5.6: WT a1, a2 and a3 traffic to Golgi. A: representative confocal fluorescence images of control B \ (empty-vector transfected) HEK 293 cells (left panel), or cells transiently transfected with WT a1-2FLAG 81

construct (middle panel), or a1N495Q-2FLAG construct (right panel). All panels show cells stained with anti-syntaxin 6 (red) and anti-FLAG (green) antibodies. Nuclei are counter-stained with DAPI (blue). B: same as in panel A, but cells were transfected with WT a2-2FLAG or un-glycosylated a2N484Q+N505Q- 2FLAG, as indicated. C: same as in panel A, but cells were transfected with WT a3-2FLAG or un- glycosylated a3N483Q+N503Q-2FLAG, as indicated. Magnification is the same in all panels; scale bar in lower right panel is 5 µm. D: quantitative colocalization analysis of data in panels A, B and C. Ordinate is Pearson’s correlation coefficient (r). Results show that WT a1, WT a2 and WT a3 can all localize to Golgi (p < 0.001). Images are representative of 15 each, from three independent experiments.

Figure 5.7: N-glycosylation is required for a subunit association with V0 assembly factor, VMA21. HEK 293 cells were transfected with empty vector (control), or with constructs WT a1-2FLAG (a1), un- glycosylated mutant a1N495Q-2FLAG (indicated as a1–), WT a2-2FLAG (a2), un-glycosylated mutant a2N484Q+N505Q-2FLAG (a2=), WT a3-2FLAG (a3), un-glycosylated mutant a3N483Q+N503Q-2FLAG (a3=), WT a4-2FLAG (a4), or un-glycosylated mutant a4N489D-2FLAG (a4-). Whole-cell lysates were immunoprecipitated using anti-FLAG antibody. A: immunoprecipitates blotted and probed with anti- FLAG antibodies, or with anti-VMA21 antibodies to detect co-immunoprecipitated VMA21, as indicated. B: whole-cell lysates, blotted prior to immunoprecipitation, probed with anti-VMA21 antibody. Blots are representative of three independent experiments. 82

Table 5.1: Significance of half-life comparisons for data in figures 5.3 and 5.4

(P-values derived from unpaired 2-tailed t-tests; NS > 0.05, * ≤ 0.05, ** ≤ 0.01, *** ≤ 0.001)

Construct 1 vs. Construct 2 Half-life 1±SD Half-life 2±SD P-value

WT a1 WT a1 + MG 13.45 ± 1.89 12.34 ± 1.25 0.4507 NS

WT a1 WT a1 + Am 13.45 ± 1.89 10.75 ± 0.81 0.1169 NS

WT a1 a1N495Q 13.45 ± 1.89 4.80 ± 0.41 0.0125 *

WT a1 a1N495Q + MG 13.45 ± 1.89 12.54 ± 0.28 0.4930 NS

WT a1 a1N495Q + Am 13.45 ± 1.89 4.56 ±0.097 0.0145 *

WT a1 + MG a1N495Q 12.34 ± 1.25 4.80 ± 0.41 0.0051 **

WT a1 + MG a1N495Q + MG 12.34 ± 1.25 12.54 ± 0.28 0.8101 NS

WT a1 + Am a1N495Q 10.75 ± 0.81 4.80 ± 0.41 0.0016 **

WT a1 + Am a1N495Q + Am 10.75 ± 0.81 4.56 ±0.097 0.0051 ** 83

a1N495Q a1N495Q + MG 4.80 ± 0.41 12.54 ± 0.28 <0.0001 ***

a1N495Q a1N495Q + Am 4.80 ± 0.41 4.56 ±0.097 0.4188 NS

WT a2 WT a2 + MG 32.21 ± 1.64 52.61 ± 16.3 0.1612 NS

WT a2 WT a2 + Am 32.21 ± 1.64 23.46 ± 2.43 0.0094 **

WT a2 a2N484Q 32.21 ± 1.64 10.77 ± 0.42 0.0011 **

WT a2 a2N505Q 32.21 ± 1.64 15.75 ± 0.83 0.0006 ***

WT a2 a2N484Q+N505Q 32.21 ± 1.64 7.81 ± 0.45 0.0008 ***

WT a2 a2N484Q+N505Q + MG 32.21 ± 1.64 27.60 ±6.45 0.3409 NS

WT a2 a2N484Q+N505Q + Am 32.21 ± 1.64 8.94 ± 0.99 0.0001 ***

WT a2 + MG a2N484Q+N505Q 52.61 ± 16.3 7.81 ± 0.45 0.0413 * 84

WT a2 + MG a2N484Q+N505Q + MG 52.61 ± 16.3 27.60 ±6.45 0.1026 NS

WT a2 + Am a2N484Q+N505Q 23.46 ± 2.43 7.81 ± 0.45 0.0065 **

WT a2 + Am a2N484Q+N505Q + Am 23.46 ± 2.43 8.94 ± 0.99 0.0040 **

a2N484Q a2N505Q 10.77 ± 0.42 15.75 ± 0.83 0.0002 ***

a2N484Q a2N484Q+N505Q 10.77 ± 0.42 7.81 ± 0.45 0.0012 **

a2N505Q a2N484Q+N505Q 15.75 ± 0.83 7.81 ± 0.45 0.0006 ***

a2N484Q+N505Q a2N484Q+N505Q + MG 7.81 ± 0.45 27.60 ±6.45 0.0331 *

a2N484Q+N505Q a2N484Q+N505Q + Am 7.81 ± 0.45 8.94 ± 0.99 0.1765 NS

WT a3 WT a3 + MG 20.79 ± 2.21 29.59 ± 1.52 0.0068 **

WT a3 WT a3 + Am 20.79 ± 2.21 23.48 ± 4.77 0.4445 NS

WT a3 a3N483Q 20.79 ± 2.21 7.79 ± 0.023 0.0095 **

WT a3 a3N503Q 20.79 ± 2.21 10.87 ± 0.375 0.0141 *

WT a3 a3N483Q+N503Q 20.79 ± 2.21 5.59 ± 0.187 0.0067 **

WT a3 a3N483Q+N503Q + MG 20.79 ± 2.21 15.03 ± 1.72 0.0259 *

WT a3 a3N483Q+N503Q + Am 20.79 ± 2.21 5.33 ± 0.57 0.0045 **

WT a3 + MG a3N483Q+N503Q 29.59 ± 1.52 5.59 ± 0.187 0.0012 **

WT a3 + MG a3N483Q+N503Q + MG 29.59 ± 1.52 15.03 ± 1.72 0.0004 ***

WT a3 + Am a3N483Q+N503Q 23.48 ± 4.77 5.59 ± 0.187 0.0227 *

WT a3 + Am a3N483Q+N503Q + Am 23.48 ± 4.77 5.33 ± 0.57 0.0210 *

a3N483Q a3N503Q 7.79 ± 0.023 10.87 ± 0.375 0.0048 **

a3N483Q a3N483Q+N503Q 7.79 ± 0.023 5.59 ± 0.187 0.0021 ** 86

a3N503Q a3N483Q+N503Q 10.87 ± 0.375 5.59 ± 0.187 0.0002 ***

a3N483Q+N503Q a3N483Q+N503Q + MG 5.59 ± 0.187 15.03 ± 1.72 0.0102 *

a3N483Q+N503Q a3N483Q+N503Q + Am 5.59 ± 0.187 5.33 ± 0.57 0.5189 NS

6. Molecular Mechanisms of Cutis Laxa and Distal Renal Tubular Acidosis-Causing Mutations in V-ATPase a subunits, ATP6V0A2 and ATP6V0A4

Sally Esmail‡, Norbert Kartner‡, Yeqi Yao‡, Joo Wan Kim‡, Reinhart A. F.

Reithmeier§ and Morris F. Manolson‡§1

From the ‡Dental Research Institute, Faculty of Dentistry, University of Toronto,

Toronto, Ontario M5G 1G6, Canada, and §Department of Biochemistry, University

of Toronto, Toronto, Ontario M5S 1A8, Canada

Manuscript in preparation

6.1. Abstract

The a subunit is the largest of 15 different subunits that make up the V-ATPases complex, where it functions in proton translocation. In mammalian species it has four paralogous isoforms, a1–4 and evidence suggests that these isoforms may encode signals for targeting assembled V- ATPases to specific intracellular locations. Despite the functional importance of the a subunit, its structure remains controversial. By studying molecular mechanisms of human disease-causing missense mutations within a subunit isoforms, we hope to identify domains critical for V- ATPase targeting, activity and/or regulation. cDNA encoded FLAG-tagged human wild type a2 and a4 and mutants a2P405L (cutis laxa), a4R449H and a4G820R (renal tubular acidosis) were transiently expressed in HEK 293 cells. Glycosylation was assessed with endoglycosidases and data showed that a2P405L, a4R449H and a4G820R were N-glycosylated. Cycloheximide pulse-chase assay revealed that the a2P405L and a4R449H were unstable and degraded in the proteasomal pathway compared to their wild type counterparts. a2P405L was also degraded in the lysosomal pathway. Immunofluorescence data showed ER retention and defective cell-surface expression of a4R449H, and defective Golgi trafficking of a2P405L. Co-immunoprecipitation studies suggested R449H an increase in association of a4 with the V0 assembly factor VMA21 and lower association G820R with V1 B1. Molecular modeling of human a4 suggested that a4 causes distal renal tubular acidosis (dRTA) by interfering with proton translocation. This study provides new information that could be employed for drug interventions using protein rescue approaches to cure V- ATPases implicated dRTA and cutis laxa.

6.2. Introduction

Vacuolar H+-ATPases (V-ATPases) are evolutionarily conserved, multi-subunit rotary proton pumps that play crucial roles in regulating the pH of cells and intracellular compartments (151,204,205). Based on their subcellular localization, V-ATPases are categorized into either endomembrane or plasma membrane V-ATPases (86,151). Endomembrane V-ATPases are expressed in membranes of acidic organelles, such as lysosomes, endosomes and the Golgi apparatus, and are responsible for translocation of protons into the luminal compartment (206). Plasma membrane V-ATPases are found on the surfaces of some specialized cells such as 89

osteoclasts, kidney intercalated cells and metastatic cancer cells, where they extrude protons to the extracellular space (157,207,208).

V-ATPases form a complex that is composed of 15 subunits arranged into two major domains, the cytoplasmic V1 sector and the membrane-integrated V0 sector. V1 is responsible for ATP hydrolysis that provides the energy to rotate a central shaft that powers proton translocation

(161). V0 contains a coupled rotor that carries protons for extrusion through the largest V0 subunit, the c. 100-kDa a subunit, which forms the proton channel. In mammals there are four isoforms of the a subunit (a1–a4) that are thought also to encode signals for targeting assembled V-ATPase complexes to distinct cellular locations. Whereas a1 and a2 target the holocomplexes to endomembranes, a3 and a4 target V-ATPases to the plasma membranes of some specialized cells (156,209). Human missense mutations of the a subunits are implicated in diverse diseases (189,210). For example, a2 mutations results in cutis laxa (wrinkled skin syndrome) (211), a3 mutations result in osteopetrosis (dense, brittle bone) (162), and a4 mutations result in distal renal tubular acidosis (dRTA) (138). Despite the important implications for a subunit functions in disease, structures of human a subunit isoforms are still controversial due to the lack of high- resolution structural data. Recently, however, a relatively high resolution (6.4 Å) model of the yeast membrane-integrated domain of a subunit (Vph1p) was published, based on cryo-EM 3D reconstruction, evolutionary covariance mapping of key residues, and low resolution X-ray crystallography (190). This model confirms that the a subunit membrane domain consists of 8 transmembrane α-helices (TMs), as has been previously shown (161), with TM7 and TM8 highly tilted and forming an interface with the V0 rotor c-ring that powers proton translocation.

In spite of recent advances, knowledge of a subunit folding, targeting, and assembly into the V- ATPase holocomplex remains sparse, and considerably more investigation will be required to elucidate issues such as the mechanism of plasma membrane a subunit targeting, for example, whose resolution will be required before efforts at designing strategies for therapeutic interventions can realistically be made. To that end, we hypothesized that human disease-causing missense mutations within a subunits could be used to identify critical domains essential for V- ATPase targeting, activity and/or regulation. To test this hypothesis we have studied the molecular consequences of introducing the cutis laxa-causing mutation, a2P405L, and the dRTA- 90

causing mutations, a4R449H and a4G820R, with respect to subunit glycosylation, localization, stability and assembly.

6.3. Materials and Methods

ENZYMES AND REAGENTS: Restriction enzymes, endo-β-N-acetylglucosaminidase H (Endo H; catalog no. P0702S), and peptide N-glycosidase F (PNGase F; P0704S) were from New England Biolabs (Whitby, Canada). Octaethylene glycol mono-n-dodecyl ether (C12E8) was from NIKKO Chemicals (Barnet Products, Englewood Cliffs, NJ). Bradford protein assay reagent (500-0006) was from Bio-Rad (Mississauga, Canada), 4′,6-diamidine-2′-phenylindole dihydrochloride (DAPI; 10236276001) was from Roche Diagnostics (Mississauga, Canada) , and cycloheximide (CHX; CYC003) was from BioShop (Burlington, Canada). Phenylmethylsulfonyl fluoride (PMSF; P7626), Protease Inhibitor Cocktail (P8340), and the proteasome inhibitor N- Benzyloxycarbonyl-L-leucyl-L-leucyl-L-leucinal (MG132; C2211) were from Sigma-Aldrich (Oakville, Canada). Dulbecco’s modified Eagle’s medium (DMEM; 11965092), Dulbecco’s phosphate buffered saline (DPBS; 1404182), heat-inactivated fetal bovine serum (FBS; 16140071), penicillin/streptomycin mix (15140122), phosphate buffered saline (PBS; 10010023), 1X Trypsin/EDTA (25200056), and the Novex ECL horseradish peroxidase (HRP) chemiluminescent substrate reagent kit (WP20005) were obtained from Gibco (Fisher Scientific, Whitby, Canada). GenJet In Vitro DNA Transfection Reagent (SL100488) was purchased from SignaGen Laboratories (Rockville, MD).

ANTIBODIES: Mouse monoclonal IgG2b anti-calnexin (3H4A7; sc-130059), HRP-conjugated goat polyclonal IgG anti-rabbit IgG (sc-2004), mouse monoclonal IgG1 anti-glyceraldehyde-3- phosphate dehydrogenase (anti-GAPDH, 0411; sc-47724), and HRP-conjugated goat polyclonal IgG anti-mouse IgG (sc-2005) were purchased from Santa Cruz (Dallas, TX). Rabbit IgG anti- VMA21antibody (HPA010972) was from Sigma-Aldrich. Rabbit polyclonal IgG anti-FLAG (ab1162) and mouse monoclonal IgG1 anti-syntaxin 6 (ab56656) were from Abcam (Cedarlane; Burlington, Canada). Alexa Fluor 568-conjugated goat polyclonal IgG anti-rabbit IgG (A- 11011), Alexa Fluor 488-conjugated goat polyclonal IgG anti-rabbit IgG (A-11034), Alexa Fluor 91

568-conjugated goat polyclonal anti-mouse IgG (A-11004), and Alexa Fluor 488-conjugated goat polyclonal IgG anti-mouse IgG (A11001) were from Molecular Probes, Fisher Scientific.

cDNA CONSTRUCTS, PLASMIDS AND CELLS: The pCMV6-XL4 plasmid carrying a2- coding cDNA was purchased from Origene (SC115366). To prepare wild type (WT) a2-2FLAG (with tandem C-terminal FLAG epitope tags), the insert was transferred from pCMV6-XL4 to pBluescript SK+, then tagged with 2FLAG (2 × DYKDDDDK ) at its carboxy-terminus between XbaI/HindIII restriction sites. The 2FLAG-tagged construct was then transferred back to pCMV6-XL4 between APaI/HindIII sites. To prepare the a2P405L-2FLAG mutant construct, the pCMV6-XL4 carrying WT a2-2FLAG was modified by inserting a synthetic fragment bearing the a2 P405L mutation between BmgBI/pf1F1 sites (human a2 cDNA bp 1916–2487); the mutant synthetic cDNA was obtained from GeneArt in the PMA-T vector. WT a4-3HA-2FLAG was prepared as described previously (199). To prepare a4R449H-3HA-2FLAG and a4G820R-3HA- 2FLAG, GeneArt synthetic cDNA bearing the a4R449H mutation was inserted between EcoRI/SapI sites, and the a4 G820R fragment was inserted into the ApaI site in pcDNA3.1(+). The DNA sequences of all constructs were confirmed by commercial sequencing (ACGT; Toronto, Canada). Human embryonic kidney cells (HEK 293; CRL-1573TM) were from the American Type Culture Collection (ATCC; Manassas, VA).

CELL CULTURE AND TRANSFECTION: Liquid nitrogen-stored HEK 293 cells were rapidly thawed in a water bath at 37 °C followed by incubation in 75 cm2 tissue culture flasks containing 17 ml DMEM, supplemented with 10% FBS and 1% penicillin/streptomycin mix, in a humidified 5% CO2 incubator for 4 days at 37 °C. The cells, at 70–80% confluence, were then trypsinized with 1 ml of 1X Trypsin/EDTA and seeded into 6-well plates at a density of 4–7 × 105 cells/well and incubated for 24 h. Cells were subsequently transiently transfected with 1 μg/well of plasmid construct in a transfection complex containing GenJet reagent and plasmid DNA in a 3:1 ratio. The transfection complex was diluted to 200 µl final volume with serum-free DMEM and incubated for 10 min prior to transfection. Post-transfection cells were incubated for 24 h and then harvested for protein expression analysis. 92

PROTEIN EXPRESSION ANALYSIS AND DEGLYCOSYLATION ASSESSMENT: Whole- cell lysates were prepared as previously described (199). Briefly, cells were harvested in 0.2 ml/well lysis buffer (PBS containing 1% C12E8, 1 mM PMSF and 1:100 (v/v) Protease Inhibitor Cocktail) and incubated on ice for 30 min. Lysates were then centrifuged at 15,000g for 30 min at 4 °C, and supernatants were collected for further analysis. Protein concentrations of the supernatants were quantified using the Bradford protein assay.

Protein glycosylation was assessed by treatment of samples with either PNGase F or Endo H. Briefly, 30 µg of whole-cell lysate was denatured in 3 µl of 10X glycoprotein denaturation buffer (5% sodium dodecyl sulfate, 0.4 M dithiothreitol; New England Biolabs), the reaction mixture was adjusted to 20 µl and incubated at 65 °C for 10 min, then 2 µl 10X Glyco Buffer was added (for PNGase F, 0.5 M sodium phosphate, pH 7.4 at 25 °C, and for Endo H, 0.5 M sodium citrate, pH 7.5 at 25 °C). Subsequently, 2 μl 10% (w/v) NP-40 (New England Biolabs) and 2,000 u

PNGase F, or Endo H, were added. The final volume was adjusted to 40 μl with distilled H2O, incubated for 1 h at 37 °C, and then analyzed by immunoblotting.

IMMUNOBLOTTING: Immunoblotting was conducted as previously described (199). Briefly, 30 µg of whole-cell lysate was loaded per well and subjected to 7% SDS-PAGE. Proteins were then transferred to nitrocellulose membrane and incubated overnight at 4 °C with 1:2000–1:3000 diluted primary antibodies (anti-FLAG, anti-B1, or anti-VMA21). 1:5000 diluted GAPDH was used in some experiments to provide loading controls. The blots were then incubated for 1 h at room temperature with 1:5000 HRP-labeled secondary antibody and bands were developed with chemiluminescent substrate reagent.

PROTEIN STABILITY AND PROTEIN BAND QUANTIFICATION: Protein stability was evaluated using the cycloheximide (CHX) pulse-chase assay. Briefly, HEK 293 cells were transfected with WT and mutant cDNA constructs and, 24 h post-transfection, the cells were treated with 10 μg/ml CHX with or without proteasomal inhibitor (10 μM MG132), or lysosomal inhibitor (25 mM NH4Cl), for up to 12 h.The cells were subsequently harvested, and whole-cell lysates were prepared for immunoblotting with anti-FLAG, and anti-GAPDH as a loading control. 93

Protein band quantification of CHX immunoblots was performed using Bio-Rad Quantity One 4.6.9 software. Briefly, band intensities were quantified after subtracting background signals from band signals using the rolling-disc method. Relative protein levels were estimated after normalizing band intensities relative to GAPDH loading controls and zero-time controls. The Student’s t-test was used to determine the significance of differences in band comparisons, the use of the term ‘significant’ implying p<0.05. Protein-half lives (h) were calculated by GraphPad Prism 5 software. Briefely, curve fitting equation were modified either as one phase decay [Y= (Y0) *exp (-K*X)] or plateau followed by one phase decay [Y= IF ( X

CO-IMMUNOPRECIPITATION: HEK 293 were transfected with WT and mutants cDNA constructs and whole-cell extracts were prepared in IP buffer (150 mM NaCl, 25 mM Tris HCl, pH 7.2 at 25 °C, containing 1% C12E8, 1:100 (v/v) Protease Inhibitor Cocktail and 1 mM PMSF), as previously described (199). Co-immunoprecipitation of WT and mutants was conducted by treating 50 μg whole-cell lysate with 5 μg of anti-FLAG antibody and incubating overnight at 4 °C with agitation. Antigen-anti-FLAG antibody immunocomplexes were pulled down by incubation with 100 μl (50% packed volume) of protein A agarose beads for 2 h at room temperature with agitation. The antigen-beads were then incubated 5 min with SDS-PAGE sample buffer at 95 °C to elute antigens. Supernatants containing antigens were collected after centrifugation at 2,500g for 3 min, then were immunoblotted with anti-FLAG, anti-B1 and anti- VMA21 antibodies.

IMMUNOFLUORESCENCE AND COLOCALIZATION ANALYSIS: HEK 293 cells were grown on glass coverslips and transiently transfected with WT and mutant cDNA constructs. The cells were washed with DPBS and fixed with 3.7% (w/v) paraformaldehyde for 15 min at room temperature. Subsequently, cells were permeabilized with DPBS containing 0.2% Triton X-100 at room temperature for 15 min. Cells were then blocked with DPBS containing 5% bovine serum albumin for 1 h at room temperature, followed by immunostaining with anti-FLAG 94

(1:1,000), anti-calnexin (1:500), or anti-syntaxin 6 (1:500) antibodies in DPBS containing 5% bovine serum albumin for 45 min at room temperature. Cells then were washed 3 times with DPBS and immunostained with fluorescent second antibodies (1:500) for 45 min at room temperature. Nuclei were stained with 0.1 mg/ml DAPI in DPBS for 10 min and cells were mounted with ProLong Gold Antifade Reagent (Fisher Scientific). Photomicrography images were acquired using a Quorum Spinning Disk Confocal System equipped with a Hamamatsu C9100-13 EM-CCD, Yokogawa CSU X1 scan head, and Improvision Piezo focus drive (Imaging Facility, Hospital for Sick Children, Toronto, Canada).

Colocalization quantification of 20 images from three independent experiments was conducted using Volocity v6.3 3D image analysis software (PerkinElmer, Woodbridge, Canada). Colocalizations of two fluorescent signals (red and green) were quantified and expressed as Pearson’s correlation coefficients (r). Significance of differences between WT and mutants were estimated using two-tailed Student’s t tests.

CELL SURFACE BIOTINYLATION: Cell surface labeling was performed using EZ-Link NHSSS-Biotin reagent (Pierce 21328; Fisher Scientific), as descried previously (199). Briefly, HEK 293 cells were transiently transfected and, 24 h post-transfection, the cells were incubated with 1 mg/ml freshly prepared EZ-Link NHS-SS-Biotin for 1 h at 4°C with gentle agitation. The cells were then incubated with ice-cold quenching buffer (192 mM glycine, 25 mM Tris, pH 8.3 at 25 °C) to remove excess biotin. Cells were harvested in 0.4 ml ice-cold RIPA buffer (150 mM NaCl, 1% sodium deoxycholate, 0.1% SDS, 1% Triton X-100, 1 mM EDTA and 10 mM Tris- HCl, pH 7.5 at 25 °C ) containing Protease Inhibitor Cocktail (1:100 v/v) and 1 mM PMSF, and were incubated for 30 min on ice, then centrifuged at 15,000g for 30 min at 4 °C. Supernatant were collected and biotinylated cell-surface proteins were affinity purified by incubating

Scientific) for 2 h at 4 °C. The eluted, biotinylated cell-surface proteins and total lysate proteins were analyzed by 7% SDS-PAGE and immunoblotted, as previously described.

STRUCTURAL MODELING OF HUMAN a4 SUBUNIT: Homology modeling of the integral membrane domain of the human a4 subunit was generated by SWISS-MODEL, using the yeast a subunit ortholog, Vph1p, as a template (PDB: 5I1M and 5JT5) (190,212). Subsequently, the model was corrected and the 3D representation was generated using the 3D graphical YASARA 95

interface (213). A 3D representation of a4G820R was generated after substituting Gly 820 with Arg, using the YASARA FoldX plug-in.

6.4. Results

6.4.1. Amino acid residues a2 P405, a4 R449 AND a4 G820 are highly conserved

Alignment of a subunit segments affected by the human mutations causing cutis laxa and dRTA that are under study in the present work are shown in Fig. 6.1. The residues highlighted are identical in all four human a subunit isoforms, all four mouse a subunit isoforms, and also in the yeast a subunit isoform, Vph1p. Fig. 6.1A shows a segment of the integral membrane domain of the a subunit, where highlighted mutations in human a2 P405 (in TM1) and in human a4 R449 (in TM3) result in cutis laxa and dRTA, respectively. Fig. 6.1B shows a C-terminal segment of the a subunit comprising the CTD, where the highlighted mutation in human a4 G820 results in dRTA. Thus, the three mutations highlighted, a2P405L, a4R449H and a4G820R, all affect highly conserved amino acid residues.

6.4.2. Glycosylation and stability of cutis laxa mutant, a2P405L

We have previously shown that all human a subunit isoforms are N-glycosylated and that N- glycosylation is required for their stability (199) (see chapter five). We have also shown that in the case of the osteopetrosis mutation, R444L, the a3 subunit is misfolded, un-glycosylated, retained in the ER, and ultimately subjected to proteolytic degradation (162). It was of interest, therefore, to determine whether the cutis laxa and dRTA mutations have similar impacts on a2 and a4 subunit structures, respectively, using methods for assessing N-glycosylation and stability that were previously described (199). We used PNGase F to assess whether mutant proteins, a2P405L, a4R449H and a4G820R, are N-glycosylated, and Endo H to determine whether any bound glycans are of the high mannose or hybrid type (172,173). Stability was assessed using the CHX pulse-chase method (199). Fig. 6.1C shows immunoblots of wild type FLAG-tagged a2 protein (WT a2-2FLAG), and the cutis laxa mutant subunit a2P405L-2FLAG expressed transiently in HEK 293 cells, with and without PNGase F or Endo H treatment of the whole-cell lysates. WT a2-2FLAG was observed as a c. 110-kDa band, and upon Endo H treatment its relative mobility was reduced to c. 105 kDa, representing the deglycosylated a2-2FLAG. The mutant a2P405L- 97

2FLAG was also observed as a c. 110-kDa band, and upon Endo H treatment its relative mobility was also reduced to c. 105 kDa, representing the deglycosylated a2P405L-2FLAG.

In the same manner, protein stability of a2P405L was assessed by transient expression of the mutant protein or its wild type counterpart, then, 24 h post-transfection, the cells were treated with and without 10 µg/ml CHX, cells were harvested after the indicated time and whole-cell lysates were prepared (see Materials and Methods). Glycans were removed from all proteins, wild type and mutant, prior to immunoblotting, by treatment with PNGase F. Fig. 6.1D shows quantitative band analysis of the immunoblots used to assess stability of a2P405L-2FLAG transiently expressed in HEK 293 cells. All band intensities were normalized to GAPDH as a loading control, and to zero time controls. These data show that a2P405L-2FLAG was degraded at a faster rate than WT a2, the mutant protein having a half-life of 13.40±1.38 h compared to 20.95±1.55 h for WT a2-2FLAG (p<0.02).

6.4.3. Glycosylation and stability of dRTA mutants, a4R449H and a4G820R

Transient expression of FLAG-tagged human WT a4 and dRTA mutants was performed as for the a2 constructs. On immunoblotting, as shown in Fig. 6.1E, WT a4-2FLAG was observed as a c. 105-kDa band and, upon PNGase F and Endo H treatment, its relative mass was reduced to c. 98 kDa, representing the deglycosylated a4-2FLAG. Similarly, a4R449H-2FLAG and a4G820R- 2FLAG were observed as c. 105-kDa bands and, upon PNGase F or Endo H treatments, their relative mass were reduced to c. 98 kDa, representing deglycosylated a4R449H-2FLAG and deglycosylated a4G820R-2FLAG. Thus, both a4R449H and a4G820R appeared to be N-glycosylated with Endo H-sensitive glycans, consistent with what was observed for WT a4.

To determine stability, WT a4 and mutant proteins a4R449H and a4G820R were compared using transient expression, followed, 24 h post-transfection, by treatment of cells with and without 10 µg/ml CHX; whole-cell lysates were prepared at time intervals and analyzed. Glycans were removed from all proteins, wild type and mutants, by PNGase F treatment of cell lysates prior to immunoblotting; Fig. 6.1F shows quantitative band analysis of the immunoblots. All band intensities were normalized to GAPDH as a loading control, and to zero-time controls. Fig. 6.1F shows that the a4G820R-2FLAG have a half-life of 13.71±1.67 h compared to 17.02±1.20 h for 98

WT a4-2FLAG (p<0.05); however, a4R449H-2FLAG was degraded at a significantly faster rate than WT a4, having a half-life of 4.83± 0.39 h, compared with 17.02±1.20 h for WT a4-2FLAG (p=0.0015), (see table 6.1 for complete half-life time analysis).

6.4.4. Unstable a2P405L is degraded in proteasome and lysosome while a4R449H is degraded in proteasome

In order to determine if a2P405L and a4R449H are degraded in the proteasomal pathway or the lysosomal pathway, CHX-pulse chase experiment was repeated with and without either an inhibitor of proteasomes (MG132) or lysosomes (NH4Cl), as previously described (199). Fig. 6.2 and table 6.1 show quantitative band analyses for the immunoblots loaded with WT a2-2FLAG and a2P405L-2FLAG, or WT a4-2FLAG and a4R449H-2FLAG, before and after MG132 treatment. In Fig. 6.2A there appeared to be a small, but significant (p<0.009), decrease in the degradation rate of a2P405L-2FLAG upon proteasomal inhibition. In contrast, Fig. 6.2B shows that there was a highly significant decrease (p<0.003) in degradation rate for a4R449H-2FLAG upon proteasomal inhibition, with restoration to WT levels, upon proteasomal inhibition. Fig. 6.2C and D show that lysosomal inhibition had no significant effect on a4R449H-2FLAG degradation rate (p>0.05); However, the degradation rate of a2P405L-2FLAG was significantly decreased upon lysosomal inhibition (p=0.005).Taken together, the degradation of a2P405L appeared to occur in both proteasome and lysosome and the degradation of a4R449H definitively occurred in the proteasome, as there was significant abrogation of its degradation upon proteasomal pathway inhibition.

6.4.5. a2 P405 is required for Golgi trafficking, and a4 R449 for ER exit

Degradation of both a2P405L and a4R449H suggests that the proteins might fail to assemble into the V-ATPase complex and are degraded in the proteasomal pathway, which is activated in response to the presence of misfolded proteins in the ER (36). Thus, we conducted immunofluorescence localization experiments to establish whether there is colocalization of these mutant proteins with ER and/or Golgi compartments. 99

Fig. 6.3A and B show colocalization studies of WT a2-2FLAG and a2P405L-2FLAG with calnexin and syntaxin 6. Fig. 6.3A shows representative fluorescence photomicrography images of HEK 293 cells transfected with empty vector (left-most panel), WT a2-2FLAG (middle panel) and a2P405L-2FLAG (right-most panel), probed with anti-FLAG antibody (green) and antibodies to the ER marker, calnexin (red). Fig. 6.3A shows that a2P405L-2FLAG (green) colocalized with calnexin at a correlation coefficient similar to that seen for WT a2-2FLAG. In Fig. 6.3B, the a2P405L-2FLAG mutant protein appeared to colocalize with the Golgi marker protein, syntaxin 6, at a correlation coefficient lower than that was apparent for WT a2-2FLAG.

Fig. 6.3C shows representative fluorescence photomicrography images of control, empty vector- transfected cells (left-most panel), WT a4-2FLAG, a4R449H-2FLAG and a4G820R-2FLAG (right- most panel) probed with anti-FLAG (red) and anti-calnexin (green) antibodies. The data suggest that a4R449H-2FLAG colocalized with the ER marker, calnexin, at a higher correlation coefficient than the WT a4-2FLAG, whereas a4G820R-2FLAG was similar to WT a4-2FLAG in this respect. Fig. 6.3D shows representative micrographs of the same cell series probed with anti-FLAG and anti-syntaxin 6 (red). Mutant protein, a4R449H-2FLAG, was colocalized with syntaxin 6 at a lower correlation coefficient than the WT a4-2FLAG, whereas a4G820R-2FLAG was similar to the WT a4-2FLAG in this respect.

Fig. 6.3E shows colocalization analysis of images represented in Fig. 6.3A and B, revealing that a2P405L-2FLAG colocalized with calnexin in the ER, the same as WT a2-2FLAG. The localization of a2P405L-2FLAG to Golgi (syntaxin 6), however, was reduced with reference to the wild type (p < 0.001; r = 0.5–0.8). Similarly, Fig. 6.3F shows colocalization analysis of images represented in Fig. 6.3C and D, revealing significant retention of a4R449H-2FLAG in the ER, and significantly lower association with the Golgi marker, compared with WT a4 (p < 0.001; r = 0.5– 0.8). The a4G820R mutant, on the other hand, was indistinguishable from wild type in these respects.

6.4.6. Defective cell surface expression of a4R449H

As shown above, a4R449H was unstable relative to WT a4, was retained in the ER, and was ultimately degraded in the proteasome. As part of its characterization, it was of interest to 100

determine whether any of the mutant protein was able to traffic to its normal location at the cell surface. Fig. 6.4A–D show representative fluorescence micrographs of HEK 293 cells transfected with either WT a4-2FLAG, a4R449H-2FLAG or a4G820R-2FLAG, double-stained with anti-HA (red) on non-permeabilized cells followed by cell permeabilization and staining with anti-FLAG (green). Total protein expression is represented by anti-FLAG (green) staining, and cell-surface expression by anti-HA (red) staining. Fig. 6.4A shows empty vector-transfected (control) cells stained, Fig. 6.4B shows intracellular as well as cell surface expression for cells transfected with WT-a4-3HA-2FLAG, and Fig. 6.4C shows only intracellular expression in cells transfected with WT-a4R449H-3HA-2FLAG, with no cell surface expression detected. Fig. 6.7D shows intracellular, as well as cell-surface, expression for cells transfected with WT-a4G820R-3HA- 2FLAG, a mutant that has a half-life very similar to that of WT a4 (Fig. 6.2).

In order to confirm the above findings, cell-surface proteins of intact cells were biotinylated, and the biotinylated proteins were then affinity purified for further assessment. Fig. 6.4E shows an immunoblot of the whole-cell lysates and cell-surface fraction from cells that were transfected with WT a4-2FLAG, a4R449H-2FLAG, or a4G820R-2FLAG. WT a4-2FLAG and a4G820R-2FLAG were expressed on the surface, as expected, but there was no cell surface expression of a4R449H- 2FLAG. This result confirms the immunofluorescence findings in Fig. 6.4A–D.

6.4.7. a4R449H shows increased association with VMA21

As demonstrated above, a4R449H-2FLAG and a2P405L-2FLAG had shorter half-lives and defective ER and Golgi colocalization, as compared with WT a4. It was of interest to characterize the effect of these mutations on V-ATPase complex assembly. The assembly of human V-ATPase is not well characterized, but studies in yeast have revealed that the assembly of the V0 in the ER is dependent on three assembly factors, Vma12p, Vma21p, and Vma22p. Due to the high homology between mammalian a subunit and the yeast ortholog, Vph1p, a similar assembly mechanism was expected. VMA21, the human ortholog of yeast Vma21p, is the only characterized human V-ATPase assembly factor (202). A recent study proposed that TMEM 199 and CCDC 115 are the human homologues of the yeast Vma12p and Vma22p respectively. TMEM 199 had 24% sequence identity with yeast Vma12p, while CCDC 115 showed only 101

structural homology with yeast vma22p (214). Unlike Vma12p, an ER resident V0 assembly factor, TMEM 199 transits from the ER to intermediate Golgi and then recycles back to ER (87,215). All together, the function of TMEM 199 and CCDC 115 in V-ATPase assembly remain elusive and controversial and requires further investigation.

VMA21 is required for the assembly of the a subunit into the V0 sub-complex, but the dissociation of a subunit from V0 is required for further V1–V0 assembly. Thus, the prolonged association of VMA21 with V0 inhibits V1–V0 assembly (180,181). Fig. 6.5A–C show representative immunoblots loaded with immunoprecipitate fractions that were pulled down with anti-FLAG antibody from lysates of HEK 293 cells transfected with WT a2-2FLAG, a2P405L- 2FLAG, WT a4-2FLAG, a4R449H-2FLAG, or a4G820R-2FLAG and immunoblotted with either anti-VMA21 or anti-B1. Protein band quantification analysis (Fig. 6.5D and E) showed a difference in comparison with wild type for the mutant protein, a4R449H-2FLAG. This had a significantly higher association with VMA21 (representing a–V0 assembly) (p<0.02), and a lower association with B1 (representing V0–V1 assembly) compared with WT a4 (Fig. 6.5E). Interestingly, there was no significant (p>0.05) difference between the assembly of a4G820R- 2FLAG or a2P405L-2FLAG with either B1 or VMA21, compared with their wild type counterparts.

6.4.8. a4 G820 resides within the putative proton pathway

As shown above, the a4G820R-2FLAG have a half-life of 13.71±1.67 h, compared 17.02±1.20 h for WT a4 (p<0.05). a4G820R-2FLAG also showed no difference in term of localization in the secretory pathway (p<0.05), or cell surface expression relative to WT a4. Therefore, it remained of interest to determine the mechanism by which the a4G820R mutation causes dRTA. In an attempt to address this, we constructed a homology model for the CTD of the human a4 subunit, based on recent models for the CTD of yeast Vph1p(5,110). The latter were built based on low- resolution X-ray crystallography, high-resolution cryo-EM, mutagenesis studies, and analysis of evolutionary covariance (190). This model showed the locations of highly conserved, key functional residues within the proton translocation pathway, or proton channel. In a similar manner, we showed the same residues within our human a4 model, and found that the a4 G820 102

residue is located within the putative interface of the proton translocation pathway (Fig. 6.6A and B). We also created another homology model for the a4G820R mutant protein and showed that the positively-charged side chain of the mutant a4 R820 residue possibly interferes with the proton pathway by forming a salt bridge (3.2 Å) with the adjacent negatively-charged residue, E729, an important residue for proton translocation (5,142)(see fig. 6.6C).

6.5. Discussion

6.5.1. a2 P405, a4 R449 and a4 G820 are conserved and crucial for V- ATPase function

Mutation of the V-ATPase a2 subunit amino acid residue P405 results in cutis laxa, and a4 mutations in the residues R449 and G820 result in dRTA. In an effort to understand how these missense point mutations can lead to severe disease, we first conducted multiple amino acids sequence alignment, which revealed that the residues of interest were highly conserved. Interesting, Fig.6.1A also showed that the domains in which the a2 P405, a4 R449 and a4 G820 residues reside, i.e. TM1, TM3 and CTD, respectively, are also highly conserved from human to yeast. Characterization of highly conserved residues implicated in diseases has been successfully used before as a strategy to for determining protein domain function and has informed the discovery of novel targeted therapeutic (216-218). For example, deletion of the highly conserved residue F508 in the cystic fibrosis transmembrane-conductance regulator (CFTR) leads to cystic fibrosis (CF). The ΔF508 mutation leads to protein misfolding, misprocessing and aberrant trafficking (219). Characterization of the molecular mechanism of ΔF508 CFTR disease causation has lead to the discovery of molecular chaperone approaches to correct CFTR folding and trafficking to its functional destination that represents a promising approach for curing ∆F508 CF (218). In the present study, we attempt to elucidate the molecular disease mechanism of mutations in the highly conserved a subunit P405, R449 and G820 residues. By characterizing the role of these mutations on a subunit structural stability, trafficking and assembly we hope to shed light on functional/structural domains within human a subunit isoforms that could impact our understanding of human V-ATPases in general and provide a basis for rational drug design. 103

6.5.2. Human a2P405L and a4R449H are N-glycosylated, but are unstable and degraded in the proteasomal pathway

We have previously shown that human a1–a4 are N-glycosylated and that glycosylation is important for subunit stability (199); (chapter five). In the present study, results shown in Fig. 6.1C and D showed that mutant proteins, a2P405L, a4R449H and a4G820R, were all N-glycosylated; however, a2P405L and a4R449H, have shown a significantly higher rate of turnover relative to wild type (Fig. 6.2). These results also showed that the faster turnover rates of a2P405L and a4R449H, were significantly reduced, to the wild type levels, after treatment with the proteasomal inhibitor, P405L MG132. Upon the treatment with the lysosomal inhibitor, NH4Cl, the degradation rate of a2 , but not a4R449H, was significantly inhibited. This suggested that a2P405L was degraded in the lysosomal and proteasomal pathway, but a4R449H was only degraded in the proteasomal pathway (Fig. 6.2). Furthermore, degradation in the proteasomal pathway is evidence of ER-associated degradation (ERAD) due to the unfolded protein response (UPR) (185).

6.5.3. a4 R449 is required for ER exit and a2 P405 is required for Golgi trafficking

In spite of the somewhat higher turnover of a2P405L relative to the wild type (Fig. 6.2A and C), quantification of colocalization of a2P405L with calnexin showed that there was no significant difference in association of a2P405L with calnexin compared to WT a2. Interestingly, a2P405L showed significantly less association with the Golgi marker, syntaxin 6 (Fig. 6.3E), which suggests that the a2P405L was misprocessed leading to defective Golgi trafficking, but not ER retention. In contrast, quantification of colocalization analysis for a4R449H and a4G820R with the ER-resident marker, calnexin, revealed a significantly higher colocalization of a4R449H with calnexin, suggesting ER retention of a4R449H, but not a4G820R (Fig. 6.3C and F). However, the exact mechanism of a4R449H ER retention remained to be investigated. Taken together, as for the molecular implication of a2P405L, a4R449H, one can speculate that a2 P405 and a4-R449 residues, within TM1 and TM3, respectively, are essential for human a2 and a4 stability and trafficking in the secretory pathway.

104

6.5.4. a4 R449H is crucial for cell surface expression and a–V0 association

We have previously shown experimentally that the exogenously expressed WT a4 is able to traffic to the plasma membrane of HEK 293 cells (199). In the current study we have used the same strategy to determine the effect of the mutations in a4R449H and a4G820R on cell surface expression. Fig. 6.4B and D show that both WT a4 and a4G820R were able to traffic to the cell surface, while a4R449H showed defective cell surface expression (Fig. 6.4C). The same findings were subsequently confirmed by cell surface biotinylation (Fig. 6.4E).

It was of interest also to determine the effect of the mutations under investigation on V-ATPase complex assembly. To that end, we specifically characterized the association of a2P405L, a4R449H and a4G820R with the only characterized human V-ATPase assembly factor, VMA21. Protein band quantification of co-immunoprecipitates revealed that a4R449H had a significantly higher association with VMA21 (Fig. 6.5E). In yeast, Vma21p assembles with V0-associated a subunits, and dissociates only after V0 exits the ER (220); dissociation of Vma21p from V0 is required for

V1–V0 assembly, and prolonged Vma21p–V0 association reduces V1–V0 assembly. We propose that the significantly higher association observed between a4R449H and VMA21 indicates a R449H prolonged association of a4 -V0 and VMA21 that leads to failure of V1–V0 assembly, ER retention and, ultimately, proteasomal degradation and, hence, defective cell surface expression.

6.5.5. a4 G820 is a functional residues and resides in the putative proton pathway

As described before a4 G820 is highly conserved among species (Fig.6.1B). Thus, mechanism of a4G820R dRTA causing mutation has been studied before in yeast due to the lack of mammalian model (99,221). One of these studies reported that a4G820R mutation in the yeast homolog, Vph1p, did not affect pump assembly or targeting but decreased V-ATPase hydrolytic and proton pumping activities by 83–85%(221). Another study on yeast homolog showed that a4G820R were associated with sever loss in proton translocation (78%) and a moderate decrease in ATPase activity (36%). This study also showed that a4-G820 residue lies within the domain that interact with the glycolytic enzyme the phosphofructokinase-1 (PFK-1) and a4G820R inhibited this interaction(99). In our study we have used human kidney cells (HEK 293) to investigate the 105

role of this mutation in protein stability, glycosylation and trafficking in the secretory pathway and to the plasma membrane and our results showed that a4G820R neither altered protein stability (Fig.6.1F) nor trafficking to Golgi/ plasma membrane (Fig. 6.3D and Fig. 6.4D).

In an attempt to gather a preliminary evidence of how a4 G80R might impact V-ATPase function, we have used the recently published atomic model (5,110) to build human a4 C- terminal domain model (Fig.6.6A) and the results may suggest that a4 G820 resides in the interface of proton pathway, and swapping residues G with R (a4 R820) results in formation of salt bridge (3.2 Å) with the adjacent negatively charged residues a4 E729, which is highly conserved and thought to be important for proton translocation(5,142), (Fig.6.6B). Therefore, we have proposed that a4G820R mutation may cause dRTA by interfering with the proton translocation.

6.6. Conclusion and Future Perspective

V-ATPase a isoforms are valuable targets for therapeutics directed towards a number of diseases. Understanding of structural domains affecting the a subunit folding, trafficking, membrane targeting, function and regulation, will enhance our ability to target specialized V-ATPase isocomplexes. We have previously showed that N-glycosylation is required for a subunit stability, assembly and trafficking to the plasma membrane. Our current study has enhanced our knowledge of V-ATPase a subunit by pinpointing some key structural/functional domains. We showed that a2P405L and a4R449H resulted in cutis laxa and dRTA through interfering with protein stability, ER retention and intern subunits degradation in the proteasomal pathway. All together, we proposed a model for how we think the glycosylated a4 subunit assembled, trafficked in secretory pathway and to the plasma membrane (see Fig. 6.7). Our data also inform drug discovery and it is now possible to screen for chemical chaperons to get a subunits folded and out of the ER to cure cutis laxa and sever dRTA. As for the implication of a4 in cancer metastasis, the critical domains we have identified within a4 subunit could be specifically targeted to control cancer metastasis. One possibility we can think of is blocking plasma membrane targeting of a4 containing V-ATPase via inhibitors/peptides against a4 R449 resides domain TM3 or a4 G820 resides domain CTa. 106

Figure 6.1: Glycosylation and stability of a2P405L, a4R449H and a4G820R. Sequence alignments show a high degree of conservation for amino acids affected by cutis laxa and dRTA mutations in V-ATPase a subunit proteins. To determine if glycosylation is affected by mutations, HEK 293 cells were transiently transfected with FLAG-tagged WT and mutant a subunit constructs. Whole-cell lysates were prepared 24 h post-transfection, treated with PNGase F or Endo H, and immunoblotted (20 µg protein per lane) with anti-FLAG antibody. To determine stability, 24 h post-transfection cells were treated with CHX (10 μg/ml) for the indicated time, and then whole-cell lysates were prepared and immunoblotted with anti- FLAG antibody (GAPDH was used as a loading control). A: amino acid sequence alignments 107

encompassing the end of the N-terminal cytoplasmic domain to the end of the third (of eight) TM helices of human V-ATPase V0 a1–4 (H1–4), mouse a1–4 (M1–4) and yeast a subunit, Vph1p (YV). Domains are indicated below alignments in bold: CL, cytoplasmic loop; EL, extracellular (luminal) loop; NTa, N- terminal cytoplasmic domain; TM, transmembrane helix. Cyan highlights extrapolated from studies done in Vph1p indicate TM predictions (161). Red highlighted amino acid residues indicate human disease- causing mutations (noted above alignments). Yellow highlights indicate amino acids corresponding to the human mutations, within the subunit isoforms and species shown. B: alignments as in A, but of sequences encompassing the end of TM8 and the cytoplasmic C-terminal tail domain (CTD) to the C-terminus. C: WT a2-2FLAG (WT a2), or mutant a2P405L-2FLAG (a2P405L), treated with (+) or without (-) PNGase F or Endo H. D: Same constructs expressed as in C, but treated with CHX; plots show band intensities quantified from immunoblots of post-CHX chase. Data were normalized to GAPDH signal and zero time control. E: same as C, except HEK293 cells were transfected with either WT a4-2FLAG (WT a4), mutant a4R449H-2FLAG (a4R449H), or mutant a4G820R-2FLAG (a4G820R). F: same as D, except HEK293 cells were expressing either WT a4, mutant a4R449H, or mutant a4G820R. Data are representative of three independent experiments; error bars indicate ± SD.

108

Figure 6.2: a2P405L and a4R449H are degraded in the proteasomal pathway. A: plot of quantified bands from anti-FLAG antibody-probed immunoblots of whole-cell lysates from WT a2-2FLAG and a2P405L- 2FLAG-transfected HEK 293 cells. Cells were treated with CHX (10 μg/ml) for 24 hours and chased for the times indicated, with and without treatment with proteasome inhibitor (10 µM MG132, designated ‘MG’) as indicated. B: same as panel A, but cells were transfected with WT a4-2FLAG and a4R449HL-

2FLAG constructs. C: same as panel A, but cells were treated with lysosomal inhibitor (25 mM NH4Cl). D: same as panel B, but cells were treated with lysosomal inhibitor. Data were normalized to GAPDH and zero time control and are representative of three independent experiments; error bars indicate ± SD.

109

Figure 6.3: Localization of mutant a subunit proteins in the secretory pathway. A: representative confocal fluorescence images of empty vector-transfected HEK 293 cells (left panel), cells transiently transfected with WT a2-2FLAG (middle panel), or with a2P405L-2FLAG (right panel). All panels show cells stained with anti-calnexin (red) and anti-FLAG (green). Nuclei are counter-stained with DAPI (blue). B: same as A, except cells were stained with anti-syntaxin 6 (red) and anti-FLAG (green). C: fluorescence images of empty vector-transfected HEK 293 cells (left-most panel), cells transiently transfected with WT a4-2FLAG (second panel), with a4R449H-2FLAG (third panel), or with a4G820R- 110

2FLAG (right-most panel). All panels show cells stained with anti-calnexin (red) and anti-FLAG (green). D: same as C, except cells were stained with anti-syntaxin 6 (red) and anti-FLAG (green). E: quantitative colocalization analysis of data in panels A and B. Ordinate is Pearson’s correlation coefficient (r). Results show that a2P405L was localized with calnexin in a rate similar to the WT a2, but showed significantly less colocalization with syntaxin 6 than the WT a2, (p < 0.05). F: quantitative colocalization analysis of data in panels C and D. Results show that a4R449H is mostly retained in the ER (p < 0.05). Images are representative of 20 each from three independent experiments each.

Figure 6.4: Defective cell-surface expression of a4R449H. Fluorescence photomicrograph images show DAPI nuclear staining (blue) of HEK 293 cells transfected as indicated (left-most panels), fluorescent 111

double antibody staining, first with anti-HA on non-permeabilized cells (second-from-left panels), followed by anti-FLAG after cell permeabilization (second-from-right panels), and merged images (right- most panels). Staining of non-permeabilized cells with anti-HA antibody indicated cell-surface accessibility of the epitope tag. A: empty vector-transfected (control) cells stained with anti-FLAG (green) and anti-HA (red). B: same as A, but cells were transfected with WT a4-3HA-2FLAG. C: same as A, but cells were transfected with WT a4R449H-3HA-2FLAG. D: same as A, but cells were transfected with WT a4G820R-3HA-2FLAG. The scale bar in the bottom right panel (bottom right) is 5 µm; all panels are of the same magnification. Each panel is representative of 20 micrographs obtained from 3 independent experiments. E: surface proteins of intact transfected cells were subjected to biotinylation followed by streptavidin affinity purification (a.p.) and immunoblotting; blots were probed with anti-FLAG antibody (three independent experiments); lane 1 (from left), whole-cell lysate from WT a4-transfected HEK 293 cells (lysate); lane 2, surface protein biotinylation showing surface protein fraction from cells transfected with WT a4 (a.p.); lanes 3 and 4, same as lanes 1and 2, except cells were transfected with a4G820R; lanes 5 and 6, same as lanes 1and 2, except cells were transfected with a4R449H. Blot is representative of three independent experiments.

112

Figure 6.5: Association of a subunit with V-ATPase assembly chaperone, VMA21, and V marker, B 0 ATP6V0B1. HEK 293 cells were transfected with WT and mutant FLAG-tagged constructs. After 24 h expression, whole-cell lysates were immunoprecipitated with anti-FLAG antibody. A–C: immunoprecipitates were blotted and probed with anti-FLAG (A), anti-B1 (B) or anti-VMA21 (C) antibodies. D: quantification of WT and mutant a2 associations with VMA21 and B1 in blots shown in panels A–C. No significant differences were observed between WT and mutant a2 associations with either VMA21 or B1. E: quantification of WT and mutant a4 associations with VMA21 and B1 in blots shown in panels A–C. Results showed significantly higher association of a4R449H with VMA21, and reduced association with B1, compared with WT a4. No significant difference was seen for a4G820R. Results represent three independent experiments; error bars are ± SD, p<0.05.

113

Figure 6.6: a4 G820 resides within the putative proton translocation pathway. A: homology model for C-terminal integral membrane domain of the human a4 subunit. This model was constructed based on the recent high-resolution cryo-EM structure and evolutionary covariance analysis for the yeast a subunit, Vph1p (190). The indicated residues are highly conserved and essential for proton translocation. The red dashed line shows the hypothetical proton channel from the cytoplasmic side of the membrane to the luminal space. Cyan dashed box indicates the region where amino acid residue G820 is located. B: shows the close proximity of G820 and the highly conserved E729, a residue thought to be key in proton translocation. C: illustrates how the G820R mutation may results in a salt-bridge interaction (red asterisk) 114

between E729 and R820, possibly distorting or blocking the proton channel, resulting in inhibition of proton translocation.

115

Figure 6.7: Model for human a4 trafficking in the secretory pathway and to the plasma membrane. N489D A: un-glycosylated mutant a4 is unable to assemble into a V0 sub-complex. It is retained in the ER and is targeted to the ERAD pathway for proteolysis. In contrast, the glycosylated a4R449H mutant assembles within the V0 complex, but is ultimately degraded in the proteasome, and thus also fails to G820R reach the plasma membrane. B: the glycosylated a4 assembles within the V1V0 complex and 116

trafficked to the plasma membrane, but functional proton translocation appears to be inhibited by the mutation.

Table 6.1: Significance of half-life comparisons for data in figures 6.1 and 6.2

(P-values derived from unpaired 2-tailed t-tests; NS > 0.05, * ≤ 0.05, ** ≤ 0.01, *** ≤ 0.001)

Construct 1 vs. Construct 2 Half-life 1 Half-life 2 P-value

Fig. 6.1D

WT a2 a2P405L 23.81 ± 4.27 13.40 ± 1.02 0.0148 *

Fig. 6.1F

WT a4 a4R449H 17.02 ± 1.20 4.83 ± 0.39 0.0015 **

WT a4 a4G820R 17.02 ± 1.20 13.71 ± 1.67 0.0313 *

a4R449H a4G820R 4.83 ± 0.39 13.71 ± 1.67 0.0088 **

Fig. 6.2A

WT a2 WT a2 + MG 20.95 ± 1.55 17.77 ± 1.21 0.0488 *

WT a2 a2P405L 20.95 ± 1.55 13.40 ± 1.38 0.0032 **

WT a2 a2P405L + MG 20.95 ± 1.55 17.65 ± 0.69 0.0281 *

117

WT a2 + MG a2P405L 17.77 ± 1.21 13.40 ± 1.38 0.0146 *

WT a2 + MG a2P405L + MG 17.77 ± 1.21 17.65 ± 0.69 0.8886 NS

a2P405L a2P405L + MG 13.40 ± 1.38 17.65 ± 0.69 0.0088 **

Fig. 6.2B

WT a4 WT a4 + MG 21.31 ± 3.69 17.21 ± 0.98 0.1881 NS

WT a4 a4R449H 21.31 ± 3.69 5.59 ± 0.12 0.0178 *

WT a4 a4R449H + MG 21.31 ± 3.69 21.49 ± 1.46 0.9430 NS

WT a4 + MG a4R449H 17.21 ± 0.98 5.59 ± 0.12 0.0021 **

WT a4 + MG a4R449H + MG 17.21 ± 0.98 21.49 ± 1.46 0.0179 *

a4R449H a4R449H + MG 5.59 ± 0.12 21.49 ± 1.46 0.0026 **

Fig. 6.2C

WT a2 WT a2 + Am 18.91 ± 1.76 14.27 ± 0.38 0.0111 *

118

WT a2 a2P405L 18.91 ± 1.76 7.79 ± 0.51 0.0005 ***

WT a2 a2P405L + Am 18.91 ± 1.76 11.40 ± 1.00 0.0030 **

WT a2 + Am a2P405L 14.27 ± 0.38 7.79 ± 0.51 0.0001 ***

WT a2 + Am a2P405L + Am 14.27 ± 0.38 11.40 ± 1.00 0.0097 **

a2P405L a2P405L + Am 7.79 ± 0.51 11.40 ± 1.00 0.0051 **

Fig. 6.2D

WT a4 WT a4 + Am 19.72 ± 1.68 18.82 ± 2.56 0.6415 NS

WT a4 a4R449H 19.72 ± 1.68 5.07 ± 0.15 0.0041 **

WT a4 a4R449H + Am 19.72 ± 1.68 5.60 ± 0.34 0.0035 **

WT a4 + Am a4R449H 18.82 ± 2.56 5.07 ± 0.15 0.0112 *

WT a4 + Am a4R449H + Am 18.82 ± 2.56 5.60 ± 0.34 0.0112 *

a4R449H a4R449H + Am 5.07 ± 0.15 5.60 ± 0.34 0.0978 NS

119

7. General Discussion and Future Prespectives 7.1. General discussion

7.1.1 Role of a subunit N-glycosylation

Since the V-ATPase a subunit is implicated in various human diseases, it is essential to understand a subunit structure and function to enable therapeutic targeting. Generally, structures of integral membrane proteins have been hindered due to the low purification yields and the lack of well-ordered 3D crystals for diffraction studies (220). Amino acid sequence analysis, mapping functional domains using molecular approaches, such as site-directed mutagenesis, as well as the construction of experimentally-validated topology models have been the avenue in studying structural and functional features of hard to crystallize membrane proteins.

In this thesis, I have attempted to map the functional domains of human V-ATPase a subunits. Amino acid sequence analysis comparing a isoforms in different species showed that putative N- glycosylation sites within a subunit isoforms are conserved from amphibians to human (see figure (6.1). The high degree of conservation of these sites lead me to hypothesize that the a subunits are indeed N-glycosylated and that this glycosylation is important for subunit structure and function.

The results supported the hypothesis that EL2 of human a isoforms are N-glycosylated and deglycosylation lead to protein instability and consequent ER retention, impaired Golgi trafficking as well as proteasomal and lysosomal degradation. For a4, lack of glycosylation was associated with defective plasma membrane expression. Similarly, alteration of N-glycosylation led to proteasomal degradation of the human ATP-binding cassette transporter ABCA3 (222).

Additionally, altered glycosylation of a4 was associated with defective V0-V1 assembly.

The experimental characterization of N-glycosylation within human a isoforms have been in agreement with the recent topology model of the homologous yeast a subunit, Vph1p, in context of the extracellular orientation of EL2; since N-glycosylation can only occur in extracellular loops. There are no predicted glycosylation sites within the EL2 of Vph1p nor is there any experimental evidence to suggest that Vph1p is glycosylated. Nevertheless, considering the high degree of homology between Vph1p and human a subunits, one can argue that the EL2 loop 120

should also be luminal in Vph1p (139). Despite the fact that I found glycans on human a isoforms and these glycans are essential for stability, targeting and assembly, glycosylation of Vph1p is not required for yeast V-ATPases.

Interestingly, the data not only revealed that human a2 and a3 have two N-glycosylation sites (see figures 5.2 and 5.3), but also showed that these two glycosylation sites are not of equal importance for protein stability. This finding is in agreement with other studies that showed a differential effect on protein stability when multiple N-glycosylation sites were independently altered (222,223). This differential effect on stability might be due to presence of different glycan subtypes at different sites. Taken together, I concluded that N-glycosylation of a isoforms play a crucial role in protein stability, trafficking and plasma membrane targeting. In agreement, N-glycans regulate cell surface targeting of several neuronal channels such as acid sensing channel 1a (224), calcium channels Cav3.2 (225) and voltage-gated potassium channel Kv1.3 (226).

Additionally, N-glycosylation regulates functional activities of several plasma membrane proteins. For example, elimination of all four glycosylation sites on Oatp1, an organic anion transporter results in complete loss of taurocholate transport activity (220). N-glycans also control neurotransmitters release and cell excitability via the modulation of voltage gated Na+ and K+ channels (227). Considering how these studies show that N-glycosylation plays a crucial role in regulating cell surface protein structure and function, I further hypothesis that N- glycosylation regulates V-ATPase activity on the cell surface. To test this hypothesis, proton pumping as well as ATP hydrolysis must tested after eliminating all glycosylation sites on a subunits (see future perspectives).

N-glycans might serve as plasma membrane markers in drug targeted therapies/monoclonal antibodies directed towards a3 and/or a4 to treat cancer metastasis. Towards this goal, a precise understanding of the 3D structure of N-glycans and their effect on V-ATPase activity on the cell surface would be essential to the development of V-ATPase targeted therapeutics. 121

7.1.2. Disease causing mutations pinpoint functional domains within a isoforms

The a isoforms are characterized by highly conserved amino acid sequence among both isoforms and species (see figure 6.1) suggesting the potential structural and functional importance of these conserved residues. The fact that single point mutations within different isoforms led to diseases further illuminates the importance of specific residues. By determining how a single amino acid change alters a subunit structure and function, one can highlight structure/functional domains within human a subunit to characterize diseases mechanisms and inform drug discovery. To that end, my second hypothesis states that human disease-causing missense mutations within a subunit isoforms will identify critical domains essential for V-ATPase targeting, activity and/or regulation. To test this hypothesis, I have used mutations that affect highly conserved residues namely, Cutis laxa causing mutation; a2P405L and the renal tubular acidosis causing mutations; a4R449H and a4G820R. I characterized how these mutations affect protein glycosylation, stability, trafficking and assembly. The data shows that a2P405L and a4R449H affect highly conserved domains, TM1 and TM3, respectively. Further, these mutations cause diseases by altering protein stability, in turn leading to defective trafficking to their functional destination. A homologous mutation in the osteoclast-specific a3 isoform (R444L) causes autosomal recessive infantile malignant osteopetrosis (228). Our group studied this mutation using the mouse equivalent a3R445L and reported that this mutation affects protein folding and stability leading to misprocessing and ER retention of the protein (229).

Collectively, these findings enhance our knowledge of the structural/functional domains within a isoforms and showed that TM1 and TM3 could serve as potential targets to inhibit subunit activity in osteoporosis and cancer metastasis by using small molecule inhibitors to block plasma membrane targeting of a3 and a4. It could be also possible to screen for molecular or chemical chaperons to rescue a2P405L, a3R444L and a4R449H and aid protein folding and ER exit using similar strategy that have used to correct Δ F508 CFTR folding. In this study a large chemical library was used to screen for correctors/potentiators that were able to aid CFTR folding and enable plasma membrane targeting of functionally active mutant CFTR (230). However, I am expecting that rescuing V-ATPases a subunit structure may not necessary result in catalytically active V- ATPase complex on the cell surface. Thus, a good chaperon (s) would not only rescue a subunit 122

structure and stabilize it, but must also lead to functionally active V-ATPases complex (see future perspective section).

In contrast, a4G820R had similar stability, trafficking and cell surface expression to WT. My data suggests that a4G820R causes dRTA acidosis not by altering a4 structure, but may be by disturbing proton translocation through the formation of salt bridge between a4-R820 and a4-E729 (3.2 Å). This was speculated based on mapping the a4-G820 and a4-R820 residues in the human a4 atomic model (figure 6.6) that was built using the yeast Vph1p as a template (PDB:5I1M); this atomic model was generated using the low resolution structure of F-ATP synthase (231), high resolution cryo-EM mapping and evolutionary covariance of highly conserved residues that are likely to interact physically (5). Previous studies on Vph1p recreating the a4G820R in yeast resulted in decreased V-ATPase hydrolytic and proton pumping activities by 83–85% (221). Similarly, another study on yeast homolog showed that a4G820R were associated with dramatically decrease in proton translocation (78%) and a moderate decrease in ATPase activity (36%) (99). This data suggests that it may be possible to screen and or design for molecular inhibitors that are able to specifically bind to a4-G820 or to the domain containing this residue (CTa; C-terminal of the a subunit) and interfere with proton translocation in pathological conditions such as cancer metastasis and/or osteoporosis.

7.1.3. Refining the mechanism of Human V-ATPase assembly and membrane targeting

The assembly of mammalian V-ATPases is not well understood. In yeast it has been shown that Vma12p, Vma21p, and Vma22p are essential chaperones for the assembly of a subunit with the

V0 sub-complex in the ER. Subsequently, Vma21p must dissociate from a-V0 complex to

facilitate the V0-V1 assembly. Since the human homolog of Vma21p, VMA21, is the only well characterized V-ATPase assembly chaperon that has been characterized to date (126), it was of interest to study the effect of N-glycosylation and disease causing mutations on a-VMA21 and

V0-V1 assembly. The data show that deglycosylated a1–a4 subunits are unable to assemble with

VMA21 suggesting that glycans serve as recognition signal to enable the assembly of a-V0 complex and VMA21. 123

Studying the effects of a2P405L, a4R449H and a4G820R mutations revealed that a4R449H has a significantly higher association with VMA21suggesting a prolonged association as a result of protein instability. Yeast data shows that Vma21p dissociates from the a-V0 complex only after ER exit (220). In this study a4R449H was retained in the ER and had a lower association with the

V-ATPase B1 subunit suggesting a reduced V0-V1 assembly. Thus, by considering the high degree of similarity between human and yeast a subunits, we can infer that mammalian VMA21 only dissociates after ER exit and this dissociation is required for V0-V1 assembly in a mechanism similar to yeast. In contrast, I hypothesize that mammalian a isoforms require N- glycans to stabilize their structure and to serve as recognition signal for proper complex assembly.

I have shown that N-glycosylation of EL2 is required for structural stability, ER exit, assembly and plasma membrane trafficking of a isoforms; however, I still do not know if N-glycans could modulates a subunit function. In the future perspective section, I will discuss how plasma membrane a subunits could serve as valuable therapeutic target by addressing two major questions: First, could N-glycosylation regulate V-ATPase activity on the cell surface? Second, can human mutations identify good therapeutic targets?

7.2. Future perspectives

7.2.1. Advance the knowledge of mammalian V-ATPase assembly

This study is the first to experimentally reveal that human a isoforms are N-glycosylated and I was able to uncover some of the basic roles of N-glycan(s) in a subunit stability, assembly and trafficking within the secretory pathway and to the plasma membrane. Therefore, this study has generated many research questions that needed to be addressed. I have shown that N-glycans are important for VMA21 association and a subunit assembly. However, it is still unclear how N- glycan modulates V-ATPase complex assembly. I can think of two possibilities in which N- glycans regulate assembly. First, N-glycans may be important for subunit folding with un- glycosylated subunits unstable and targeted for degradation and hence unavailable for assembly. Alternatively, N-glycans can serve as assembly recognition signals facilitating V-ATPase 124

subunit assembly. The former is supported by the present thesis work. However, this work did not disprove the potential role of N-glycosylation as a recognition signal in protein folding and/or assembly in V-ATPases complex.

In order to test if glycosylating of a isoforms might serve as a recognition signal, we can use proteomics approaches to map other potential human V-ATPases assembly factors that bind directly to a subunit. For example, I can use specific anti-a subunit antibodies to immunoprecipitated ER fraction from different cells, then sort all assembly factors that interact with the a subunit. Consequently, I could test whether the identified assembly factors bind directly to the N-glycan of a isoforms or not, by comparing the difference in interaction strength when the assembly factor bind with glycosylated and N-glycosylated a subunit using in vivo methods such as fluorescent resonance energy transfer (FRET) or its bio-luminescent modification (BRET). These methods enable the accurate detection of strong and weak interactions between two fluorescently labeled exogenously proteins expressed in vivo (232).

The mechanism of human V-ATPases assembly is not yet characterized. In contrast, yeast V- ATPase assembly is better understood but still controversial. There is experimental evidence to support two alternative pathways. In the independent pathway, the fully assembled V0 bind to a fully assembled V1. This pathway was supported by mutational analysis that showed that a complete V0 complex could be independently assembled and targeted to the yeast vacuole even when the genes encoding the V1 subunits were eliminated (127). Visa versa, when genes encoding the V0 complex are eliminated, the V1 complex is assembled and waits patiently in the cytoplasm for a V0 to bind to (233,234). In contrast, in vivo pulse-chase experiments suggest early assembly between the a subunit in V0 and the A and B subunit in V1 , suggesting the subunits are added to the V-ATPase complex in a sequential manner (124) autonomous from V1 or V0 assembly. As mentioned before, VMA21 is the only characterized human V-ATPase assembly factor and human mutation in VMA21 results in XMEA, a disease condition associated with increase in lysosomal pH. However the exact role of VMA21 in V-ATPase assembly is not understood. Thus, it is necessary to understand the exact role of VMA21 in facilitating a subunit assembly. To this end, VMA21 expression should be knocked out in human cell cultures and the effect on a-V0 assembly and V0-V1 assembly would be assessed. That could be achieved by using

HEK 293 and use antibody against any of the endogenously expressed V0 proteolipid subunit and 125

immunoprecipitate it, followed by native PAGE to detect the co-immunoprecipitated subunit. This would provide evidence on the role of VMA21 in V-ATPase assembly, providing insight into the mechanism of human V-ATPase assembly. We may also understand why a4R449H showed a relatively higher association with the V0-VMA21 compared to WT-a4.

7.2.2. Role of N-glycosylation on V-ATPase activity

To assess if N-glycosylation has a role in modulating V-ATPase function on the cell surface, monoclonal antibodies should be raised against the EL2 glycosylation sites (anti-glycan) and tested whether the antibody binding affects ATPase activity. To design these monoclonal antibodies, a prior characterization of glycans 3D structure is required as recommended in (226) to target specific a isoform glycans on the cell surface. Consequently, mammalian cells expressing a3 and a4 containing V-ATPases may be used to test anti-glycan on V-ATPase activity. Proton pumping activity in live cells could be measured using Scanning Ion-Selective Electrode Technique (SIET), a highly sensitive non-invasive technique to specifically assess H+ efflux that could be tailored to test V-ATPase proton translocation activity upon anti-glycan treatment (227). It may be necessary to inhibit the activity of other plasma membrane proton pumps to measure V-ATPases specific proton extrusion. Alternatively, I can conduct the preliminary experiments by reconstituting purified V-ATPase in proteoliposomes and test specific proton efflux by V-ATPase with and without anti-glycan treatment.

It is also important to assess the role of targeting N-glycans on V-ATPase specific ATP hydrolysis. To that end, intact V-ATPases need to be purified under non-denaturing conditions and reconstituted in proteoliposomes followed by plate-reader colorimetric ATP hydrolysis assay before and after anti-glycan treatment. Briefly, ATPases remove the phosphate group from ATP and this phosphate binds to a phosphate-sensitive chromogen such as malachite green and a change in colour intensity represent ATPase activity. This assay is also well adapted for high- throughput screening assay (235). If the results indicate that N-glycans are important for V- ATPase activity, N-glycan could serve as a valuable therapeutic target and V-ATPases could be targeted on the cell surface of metastatic cells and/or osteoclast to cure cancer or lytic bone diseases. If the results show that N-glycans do not regulate V-ATPases activities on the cell 126

surface, glycans can still be used as a plasma membrane marker to target metastatic cancer cells by engineering chemotherapeutic/radioactive-conjugated-monoclonal antibodies against plasma membrane a3/a4. The use of conjugated antibodies to specifically target cancer cells is currently a powerful therapeutic treatment option (225).

7.2.3. High throughput screening for plasma membrane V-ATPases specific inhibitor(s)

As mentioned before, studying human disease mutation of a isoforms have identified critical structural and functional domains that are potential targets to cure cancer metastasis or osteoporosis. To test this possibility, we would need to run high throughput screens on libraries of inhibitors that bind/interfere with a4R449H, a3R444L or a4G820R with the highly conserved domains they reside in (TM3 and CTa).

It may be possible that a similar strategy used in our lab for the inhibition of a3-B2 and a3-d2 interaction could be employed to find inhibitors of the a4–B1 interaction or any putative important protein-protein interaction involving the a4 subunit. This would require using purified proteins, and implementation of a high-throughput ELISA screen to identify compounds that inhibits the interaction, as was previously described for a3 (8,9).

For further validation, we could test the potential inhibitors as anti-metastatic agent by using In vitro cell invasion assay. The MCF10CA1a cell line, could be used as a highly invasive breast cancer cells that over express a3 and a4 subunits. Then the anti-metastatic function of the potential a4/a3 interaction inhibitor(s) could be evaluated using in vitro metastasis model such as “BD BioCoat™ FluoroBlok™ Invasion System”. This is an in vitro model that mimics the in vivo environment for metastasizing cells and can be used in a 96-well plate to test multiple inhibitors simultaneously. This invasion system provides platforms to enable the study of the effect of inhibiting cell surface expression of a3/a4 not only one cancer cell migration, but also on cell invasion (degradation of extracellular matrix, which is mediated by V-ATPase activity).

To screen for anti-restorative agents that specifically act by either inhibiting surface expression of a3 and/or inhibit a3 function. I could use V-ATPase (ATP hydrolysis and proton 127

translocation) functional assays described above to assess V-ATPase activity in the presence of such inhibitors and then I could assess the anti-resorptive ability of these inhibitors in vitro. To test anti-resorptive activity, Murine macrophage RAW 264.7 cells could be differentiated into osteoclasts using RANKL and plated in 96-well plate coated with synthetic hydroxyapatite surfaces (mimicking bones), followed by addition of the potential a3 inhibitors and quantitative measurement of surface area of hydroxyapatite (from micrographs) in response to the selected inhibitors (8). In addition to anti-resorptive and anti-metastatic role of a3/a4 inhibitor, we may be further characterizing the mode of action of a3/a4 inhibitor(s), which might enhance our knowledge of plasma membrane V-ATPases potentially allowing the advancement of in silico approaches to drug discovery.

8. References

1. Casey, J. R., Grinstein, S., and Orlowski, J. (2010) Sensors and regulators of intracellular pH. Nature reviews Molecular cell biology 11, 50-61

2. Aoi, W., and Marunaka, Y. (2014) Importance of pH homeostasis in metabolic health and diseases: Crucial role of membrane proton transport. BioMed research international 2014

3. Cotter, K., Stransky, L., McGuire, C., and Forgac, M. (2015) Recent insights into the structure, regulation, and function of the V-ATPases. Trends in biochemical sciences 40, 611-622

4. Hinton, A., Sennoune, S. R., Bond, S., Fang, M., Reuveni, M., Sahagian, G. G., Jay, D., Martinez-Zaguilan, R., and Forgac, M. (2009) Function of a subunit isoforms of the V- ATPase in pH homeostasis and in vitro invasion of MDA-MB231 human breast cancer cells. Journal of Biological Chemistry 284, 16400-16408

5. Schep, D. G., Zhao, J., and Rubinstein, J. L. (2016) Models for the a subunits of the Thermus thermophilus V/A-ATPase and Saccharomyces cerevisiae V-ATPase enzymes by cryo-EM and evolutionary covariance. Proceedings of the National Academy of Sciences 113, 3245-3250

6. Pérez-Sayáns, M., Somoza-Martín, J. M., Barros-Angueira, F., Rey, J. M. G., and García- García, A. (2009) V-ATPase inhibitors and implication in cancer treatment. Cancer treatment reviews 35, 707-713

7. Capecci, J., and Forgac, M. (2013) The function of vacuolar ATPase (V-ATPase) a subunit isoforms in invasiveness of MCF10a and MCF10CA1a human breast cancer cells. Journal of Biological Chemistry 288, 32731-32741 128

8. Kartner, N., Yao, Y., Li, K., Crasto, G. J., Datti, A., and Manolson, M. F. (2010) Inhibition of osteoclast bone resorption by disrupting vacuolar H+-ATPase a3-B2 subunit interaction. Journal of Biological Chemistry 285, 37476-37490

9. Crasto, G. J., Kartner, N., Yao, Y., Li, K., Bullock, L., Datti, A., and Manolson, M. F. (2013) Luteolin inhibition of V‐ATPase a3–d2 interaction decreases osteoclast resorptive activity. Journal of cellular biochemistry 114, 929-941

10. Caboni, P., Tronci, L., Liori, B., Tocco, G., Sasanelli, N., and Diana, A. (2014) Tulipaline A: Structure–activity aspects as a nematicide and V-ATPase inhibitor. Pesticide biochemistry and physiology 112, 33-39

11. Landolt-Marticorena, C., and Reithmeier, R. A. (1994) Asparagine-linked oligosaccharides are localized to single extracytosolic segments in multi-span membrane glycoproteins. Biochemical Journal 302, 253-260

12. Aebi, M. (2013) N-linked protein glycosylation in the ER. Biochimica et Biophysica Acta (BBA)-Molecular Cell Research 1833, 2430-2437

13. Aebi, M., Bernasconi, R., Clerc, S., and Molinari, M. (2010) N-glycan structures: recognition and processing in the ER. Trends in biochemical sciences 35, 74-82

14. Schwarz, F., and Aebi, M. (2011) Mechanisms and principles of N-linked protein glycosylation. Current opinion in structural biology 21, 576-582

15. Supraha Goreta, S., Dabelic, S., and Dumic, J. (2012) Insights into complexity of congenital disorders of glycosylation. Biochemia medica 22, 156-170

16. Petrescu, A.-J., Milac, A.-L., Petrescu, S. M., Dwek, R. A., and Wormald, M. R. (2004) Statistical analysis of the protein environment of N-glycosylation sites: implications for occupancy, structure, and folding. Glycobiology 14, 103-114

17. Apweiler, R., Hermjakob, H., and Sharon, N. (1999) On the frequency of protein glycosylation, as deduced from analysis of the SWISS-PROT database. Biochimica et Biophysica Acta (BBA)-General Subjects 1473, 4-8

18. Schulz, B. L. (2012) Beyond the sequon: sites of N-glycosylation, INTECH Open Access Publisher

19. Nagae, M., and Yamaguchi, Y. (2012) Function and 3D structure of the N-glycans on glycoproteins. International journal of molecular sciences 13, 8398-8429

20. Shakin-Eshleman, S. H., Spitalnik, S. L., and Kasturi, L. (1996) The amino acid at the X position of an Asn-X-Ser sequon is an important determinant of N-linked core- glycosylation efficiency. Journal of Biological Chemistry 271, 6363-6366 129

21. Kasturi, L., Hegang, C., and Shakin-Eshleman, S. H. (1997) Regulation of N-linked core glycosylation: use of a site-directed mutagenesis approach to identify Asn-Xaa-Ser/Thr sequons that are poor oligosaccharide acceptors. Biochemical Journal 323, 415-419

22. Shrimal, S., and Gilmore, R. (2013) Glycosylation of closely spaced acceptor sites in human glycoproteins. J Cell Sci 126, 5513-5523

23. Rao, R. S. P., Buus, O. T., and Wollenweber, B. (2010) Evolutionary pattern of N- glycosylation sequon numbers in eukaryotic ABC protein superfamilies. Bioinformatics and biology insights 4, 9

24. Xu, C., and Ng, D. T. (2015) Glycosylation-directed quality control of protein folding. Nature Reviews Molecular Cell Biology 16, 742-752

25. Thaysen-Andersen, M., and Packer, N. H. (2012) Site-specific glycoproteomics confirms that protein structure dictates formation of N-glycan type, core fucosylation and branching. Glycobiology 22, 1440-1452

26. Vagin, O., Kraut, J. A., and Sachs, G. (2009) Role of N-glycosylation in trafficking of apical membrane proteins in epithelia. American Journal of Physiology-Renal Physiology 296, F459-F469

27. Esmail, S., Yao, Y., Kartner, N., Li, J., Reithmeier, R. A., and Manolson, M. F. (2016) N‐Linked Glycosylation Is Required for Vacuolar H+‐ATPase (V‐ATPase) a4 Subunit Stability, Assembly, and Cell Surface Expression. Journal of cellular biochemistry

28. Lee, H. S., Qi, Y., and Im, W. (2015) Effects of N-glycosylation on protein conformation and dynamics: Protein Data Bank analysis and molecular dynamics simulation study. Scientific reports 5, 8926

29. Crothers, J. M., Asano, S., Kimura, T., Yoshida, A., Wong, A., Kang, J. W., and Forte, J. G. (2004) Contribution of oligosaccharides to protection of the H, K‐ATPase β‐subunit against trypsinolysis. Electrophoresis 25, 2586-2592

30. Muthusamy, S., Malhotra, P., Hosameddin, M., Dudeja, A. K., Borthakur, S., Saksena, S., Gill, R. K., Dudeja, P. K., and Alrefai, W. A. (2015) N-glycosylation is essential for ileal ASBT function and protection against proteases. American Journal of Physiology- Cell Physiology 308, C964-C971

31. Hayashi, H., and Yamashita, Y. (2012) Role of N-glycosylation in cell surface expression and protection against proteolysis of the intestinal anion exchanger SLC26A3. American Journal of Physiology-Cell Physiology 302, C781-C795

32. Braakman, I., and Hebert, D. N. (2013) Protein folding in the endoplasmic reticulum. Cold Spring Harbor perspectives in biology 5, a013201 130

33. Tannous, A., Pisoni, G. B., Hebert, D. N., and Molinari, M. (2015) N-linked sugar- regulated protein folding and quality control in the ER. in Seminars in cell & developmental biology, Elsevier

34. Hagiwara, M., Ling, J., Koenig, P.-A., and Ploegh, H. L. (2016) Posttranscriptional Regulation of Glycoprotein Quality Control in the Endoplasmic Reticulum Is Controlled by the E2 Ub-Conjugating Enzyme UBC6e. Molecular Cell 63, 753-767

35. Ruggiano, A., Foresti, O., and Carvalho, P. (2014) ER-associated degradation: Protein quality control and beyond. The Journal of cell biology 204, 869-879

36. Lemus, L., and Goder, V. (2014) Regulation of endoplasmic reticulum-associated protein degradation (ERAD) by ubiquitin. Cells 3, 824-847

37. Stanley, P. (2011) Golgi glycosylation. Cold Spring Harbor perspectives in biology 3, a005199

38. Stanley, P., Schachter, H., and Taniguchi, N. (2009) N-glycans.

39. Fölsch, H. (2008) Regulation of membrane trafficking in polarized epithelial cells. Current opinion in cell biology 20, 208-213

40. Cantagrel, V., and Lefeber, D. J. (2011) From glycosylation disorders to dolichol biosynthesis defects: a new class of metabolic diseases. Journal of inherited metabolic disease 34, 859-867

41. Welti, M. (2013) Regulation of dolichol-linked glycosylation. Glycoconjugate journal 30, 51-56

42. Tuccillo, F. M., de Laurentiis, A., Palmieri, C., Fiume, G., Bonelli, P., Borrelli, A., Tassone, P., Scala, I., Buonaguro, F. M., and Quinto, I. (2014) Aberrant glycosylation as biomarker for cancer: focus on CD43. BioMed research international 2014

43. Vasconcelos-dos-Santos, A., Oliveira, I. A., Lucena, M. C., Mantuano, N. R., Whelan, S. A., Dias, W. B., and Todeschini, A. R. (2015) Biosynthetic machinery involved in aberrant glycosylation: promising targets for developing of drugs against cancer. Frontiers in oncology 5

44. Croset, A., Delafosse, L., Gaudry, J.-P., Arod, C., Glez, L., Losberger, C., Begue, D., Krstanovic, A., Robert, F., and Vilbois, F. (2012) Differences in the glycosylation of recombinant proteins expressed in HEK and CHO cells. Journal of biotechnology 161, 336-348

45. Butler, M., and Spearman, M. (2014) The choice of mammalian cell host and possibilities for glycosylation engineering. Current opinion in biotechnology 30, 107-112 131

46. Brodsky, A. N., Caldwell, M., and Harcum, S. W. (2012) Glycosylation and post- translational modification gene expression analysis by DNA microarrays for cultured mammalian cells. Methods 56, 408-417

47. Saravanan, C., Cao, Z., Head, S. R., and Panjwani, N. (2010) Analysis of differential expression of glycosyltransferases in healing corneas by glycogene microarrays. Glycobiology 20, 13-23

48. Agrawal, P., Kurcon, T., Pilobello, K. T., Rakus, J. F., Koppolu, S., Liu, Z., Batista, B. S., Eng, W. S., Hsu, K.-L., and Liang, Y. (2014) Mapping posttranscriptional regulation of the human glycome uncovers microRNA defining the glycocode. Proceedings of the National Academy of Sciences 111, 4338-4343

49. Ashkani, J., and Naidoo, K. J. (2016) Glycosyltransferase Gene Expression Profiles Classify Cancer Types and Propose Prognostic Subtypes. Scientific reports 6

50. Stowell, S. R., Ju, T., and Cummings, R. D. (2015) Protein glycosylation in cancer. Annual review of pathology 10, 473

51. Gu, J., Isaji, T., Xu, Q., Kariya, Y., Gu, W., Fukuda, T., and Du, Y. (2012) Potential roles of N-glycosylation in cell adhesion. Glycoconjugate journal 29, 599-607

52. Lyons, J. J., Milner, J. D., and Rosenzweig, S. D. (2015) Glycans instructing immunity: the emerging role of altered glycosylation in clinical immunology. Frontiers in pediatrics 3

53. Clark, G. F. (2015) Functional glycosylation in the human and mammalian uterus. Fertility Research and Practice 1, 1

54. Maverakis, E., Kim, K., Shimoda, M., Gershwin, M. E., Patel, F., Wilken, R., Raychaudhuri, S., Ruhaak, L. R., and Lebrilla, C. B. (2015) Glycans in the immune system and The Altered Glycan Theory of Autoimmunity: a critical review. Journal of autoimmunity 57, 1-13

55. Arey, B. J. (2012) The role of glycosylation in receptor signaling. InTech

56. Boscher, C., Dennis, J. W., and Nabi, I. R. (2011) Glycosylation, galectins and cellular signaling. Current opinion in cell biology 23, 383-392

57. Kawakami, S., and Hashida, M. (2014) Glycosylation-mediated targeting of carriers. Journal of Controlled Release 190, 542-555

58. Torreno-Pina, J. A., Castro, B. M., Manzo, C., Buschow, S. I., Cambi, A., and Garcia- Parajo, M. F. (2014) Enhanced receptor–clathrin interactions induced by N-glycan– mediated membrane micropatterning. Proceedings of the National Academy of Sciences 111, 11037-11042 132

59. Hatakeyama, H., Akita, H., and Harashima, H. (2013) The polyethyleneglycol dilemma: advantage and disadvantage of PEGylation of liposomes for systemic genes and nucleic acids delivery to tumors. Biological and Pharmaceutical Bulletin 36, 892-899

60. Cheng, M., Li, Q., Wan, T., Hong, X., Chen, H., He, B., Cheng, Z., Xu, H., Ye, T., and Zha, B. (2011) Synthesis and efficient hepatocyte targeting of galactosylated chitosan as a gene carrier in vitro and in vivo. Journal of Biomedical Materials Research Part B: Applied Biomaterials 99, 70-80

61. Freeze, H. H., Eklund, E. A., Ng, B. G., and Patterson, M. C. (2012) Neurology of inherited glycosylation disorders. The Lancet Neurology 11, 453-466

62. Scott, K., Gadomski, T., Kozicz, T., and Morava, E. (2014) Congenital disorders of glycosylation: new defects and still counting. Journal of inherited metabolic disease 37, 609-617

63. Timal, S., Hoischen, A., Lehle, L., Adamowicz, M., Huijben, K., Sykut-Cegielska, J., Paprocka, J., Jamroz, E., van Spronsen, F. J., and Körner, C. (2012) Gene identification in the congenital disorders of glycosylation type I by whole-exome sequencing. Human molecular genetics 21, 4151-4161

64. Pinho, S. S., and Reis, C. A. (2015) Glycosylation in cancer: mechanisms and clinical implications. Nature Reviews Cancer 15, 540-555

65. Munkley, J., Mills, I. G., and Elliott, D. J. (2016) The role of glycans in the development and progression of prostate cancer. Nature Reviews Urology

66. de Leoz, M. L. A., Young, L. J., An, H. J., Kronewitter, S. R., Kim, J., Miyamoto, S., Borowsky, A. D., Chew, H. K., and Lebrilla, C. B. (2011) High-Mannose Glycans are Elevated during Breast Cancer Progression*□ S.

67. Lau, K. S., Partridge, E. A., Grigorian, A., Silvescu, C. I., Reinhold, V. N., Demetriou, M., and Dennis, J. W. (2007) Complex N-glycan number and degree of branching cooperate to regulate cell proliferation and differentiation. Cell 129, 123-134

68. Andergassen, U., Liesche, F., Kölbl, A. C., Ilmer, M., Hutter, S., Friese, K., and Jeschke, U. (2015) Glycosyltransferases as markers for early tumorigenesis. BioMed research international 2015

69. Meany, D. L., and Chan, D. W. (2011) Aberrant glycosylation associated with enzymes as cancer biomarkers. Clinical proteomics 8, 1

70. Zhao, Y.-P., Xu, X.-Y., Fang, M., Wang, H., You, Q., Yi, C.-H., Ji, J., Gu, X., Zhou, P.- T., and Cheng, C. (2014) Decreased core-fucosylation contributes to malignancy in gastric cancer. PloS one 9, e94536 133

71. Miyoshi, E., Moriwaki, K., Terao, N., Tan, C.-C., Terao, M., Nakagawa, T., Matsumoto, H., Shinzaki, S., and Kamada, Y. (2012) Fucosylation is a promising target for cancer diagnosis and therapy. Biomolecules 2, 34-45

72. Gińdzieńska-Sieśkiewicz, E., Radziejewska, I., Domysławska, I., Klimiuk, P. A., Sulik, A., Rojewska, J., Gabryel-Porowska, H., and Sierakowski, S. (2016) Changes of glycosylation of IgG in rheumatoid arthritis patients treated with methotrexate. Advances in medical sciences 61, 193-197

73. Alavi, A., Pool, A. J., and Axford, J. S. (2005) New insights into rheumatoid arthritis associated glycosylation changes. in Glycobiology and Medicine, Springer. pp 129-138

74. Watson, M., Rudd, P., Dwek, R., and Axford, J. (1996) Sugar printing rheumatic diseases: A potential method for diagnosis and differentiation using immunoglobulin G oligosaccharides. in ARTHRITIS AND RHEUMATISM, LIPPINCOTT-RAVEN PUBL 227 EAST WASHINGTON SQ, PHILADELPHIA, PA 19106

75. Wojtowicz, K., Szaflarski, W., Januchowski, R., Zawierucha, P., Nowicki, M., and Zabel, M. (2012) Inhibitors of N-glycosylation as a potential tool for analysis of the mechanism of action and cellular localisation of glycoprotein P. Acta Biochim Pol 59, 44

76. Andreoli, F., Jorge Moura Barbosa, A., Daniele Parenti, M., and Del Rio, A. (2013) Modulation of epigenetic targets for anticancer therapy: clinicopathological relevance, structural data and drug discovery perspectives. Current pharmaceutical design 19, 578- 613

77. Sha, S., Agarabi, C., Brorson, K., Lee, D.-Y., and Yoon, S. (2016) N-Glycosylation Design and Control of Therapeutic Monoclonal Antibodies. Trends in biotechnology

78. Balzarini, J. (2007) Targeting the glycans of glycoproteins: a novel paradigm for antiviral therapy. Nature Reviews Microbiology 5, 583-597

79. Lindhout, T., Iqbal, U., Willis, L. M., Reid, A. N., Li, J., Liu, X., Moreno, M., and Wakarchuk, W. W. (2011) Site-specific enzymatic polysialylation of therapeutic proteins using bacterial enzymes. Proceedings of the National Academy of Sciences 108, 7397- 7402

80. Sun, X.-L. (2013) Glycosyltransferases as Potential Drug Targets. Medicinal chemistry 2013

81. Tiels, P., Baranova, E., Piens, K., De Visscher, C., Pynaert, G., Nerinckx, W., Stout, J., Fudalej, F., Hulpiau, P., and Tännler, S. (2012) A bacterial glycosidase enables mannose- 6-phosphate modification and improved cellular uptake of yeast-produced recombinant human lysosomal enzymes. Nature biotechnology 30, 1225-1231

82. Topin, J., Arnaud, J., Sarkar, A., Audfray, A., Gillon, E., Perez, S., Jamet, H., Varrot, A., Imberty, A., and Thomas, A. (2013) Deciphering the glycan preference of bacterial 134

lectins by glycan array and molecular docking with validation by microcalorimetry and crystallography. PloS one 8, e71149

83. Roth, Z., Yehezkel, G., and Khalaila, I. (2012) Identification and quantification of protein glycosylation. International Journal of Carbohydrate Chemistry 2012

84. Jo, S., and Im, W. (2013) Glycan fragment database: a database of PDB-based glycan 3D structures. Nucleic acids research 41, D470-D474

85. Jo, S., Lee, H. S., Skolnick, J., and Im, W. (2013) Restricted N-glycan conformational space in the PDB and its implication in glycan structure modeling. PLoS Comput Biol 9, e1002946

86. Breton, S., and Brown, D. (2013) Regulation of luminal acidification by the V-ATPase. Physiology 28, 318-329

87. Forgac, M. (2007) Vacuolar ATPases: rotary proton pumps in physiology and pathophysiology. Nature reviews Molecular cell biology 8, 917-929

88. Sun-Wada, G.-H., and Wada, Y. (2013) Vacuolar-type proton pump ATPases: acidification and pathological relationships. Histology and histopathology 28, 805-815

89. Bexiga, M. G., and Simpson, J. C. (2013) Human diseases associated with form and function of the Golgi complex. International journal of molecular sciences 14, 18670- 18681

90. Poëa-Guyon, S., Ammar, M. R., Erard, M., Amar, M., Moreau, A. W., Fossier, P., Gleize, V., Vitale, N., and Morel, N. (2013) The V-ATPase membrane domain is a sensor of granular pH that controls the exocytotic machinery. The Journal of cell biology 203, 283-298

91. Kozik, P., Hodson, N. A., Sahlender, D. A., Simecek, N., Soromani, C., Wu, J., Collinson, L. M., and Robinson, M. S. (2013) A human genome-wide screen for regulators of clathrin-coated vesicle formation reveals an unexpected role for the V- ATPase. Nature cell biology 15, 50-60

92. Huang, C., and Chang, A. (2011) pH-dependent cargo sorting from the Golgi. Journal of Biological Chemistry 286, 10058-10065

93. Zoncu, R., Bar-Peled, L., Efeyan, A., Wang, S., Sancak, Y., and Sabatini, D. M. (2011) mTORC1 senses lysosomal amino acids through an inside-out mechanism that requires the vacuolar H+-ATPase. Science 334, 678-683

94. Bar-Peled, L., and Sabatini, D. M. (2014) Regulation of mTORC1 by amino acids. Trends in cell biology 24, 400-406 135

95. Gao, Y., Liu, Y., Hong, L., Yang, Z., Cai, X., Chen, X., Fu, Y., Lin, Y., Wen, W., and Li, S. (2016) Golgi-associated LC3 lipidation requires V-ATPase in noncanonical autophagy. Cell Death & Disease 7, e2330

96. Mijaljica, D., Prescott, M., and Devenish, R. J. (2011) V-ATPase engagement in autophagic processes. Autophagy 7, 666-668

97. Lee, B. S., Gluck, S. L., and Holliday, L. S. (1999) Interaction between vacuolar H+- ATPase and microfilaments during osteoclast activation. Journal of Biological Chemistry 274, 29164-29171

98. Merkulova, M., Hurtado-Lorenzo, A., Hosokawa, H., Zhuang, Z., Brown, D., Ausiello, D. A., and Marshansky, V. (2011) Aldolase directly interacts with ARNO and modulates cell morphology and acidic vesicle distribution. American Journal of Physiology-Cell Physiology 300, C1442-C1455

99. Su, Y., Blake-Palmer, K. G., Sorrell, S., Javid, B., Bowers, K., Zhou, A., Chang, S. H., Qamar, S., and Karet, F. E. (2008) Human H+ ATPase a4 subunit mutations causing renal tubular acidosis reveal a role for interaction with phosphofructokinase-1. American Journal of Physiology-Renal Physiology 295, F950-F958

100. Brown, D., Paunescu, T. G., Breton, S., and Marshansky, V. (2009) Regulation of the V- ATPase in kidney epithelial cells: dual role in acid–base homeostasis and vesicle trafficking. Journal of Experimental Biology 212, 1762-1772

101. Xu, J., Cheng, T., Feng, H., Pavlos, N., and Zheng, M. H. (2007) Structure and function of V-ATPases in osteoclasts: potential therapeutic targets for the treatment of osteolysis.

102. Frattini, A., Orchard, P. J., Sobacchi, C., Giliani, S., Abinun, M., Mattsson, J. P., Keeling, D. J., Andersson, A.-K., Wallbrandt, P., and Zecca, L. (2000) Defects in TCIRG1 subunit of the vacuolar proton pump are responsible for a subset of human autosomal recessive osteopetrosis. Nature genetics 25, 343-346

103. Pietrement, C., Sun-Wada, G., Da Silva, N., McKee, M., Marshansky, V., Brown, D., Futai, M., and Breton, S. (2006) Distinct expression patterns of different subunit isoforms of the V-ATPase in the rat epididymis. Biology of reproduction 74, 185-194

104. Brisseau, G. F., Grinstein, S., Hackam, D. J., Nordström, T., Manolson, M. F., Khine, A. A., and Rotstein, O. D. (1996) Interleukin-1 increases vacuolar-type H-ATPase activity in murine peritoneal macrophages. Journal of Biological Chemistry 271, 2005-2011

105. Edwards, J. R., and Mundy, G. R. (2011) Advances in osteoclast biology: old findings and new insights from mouse models. Nature Reviews Rheumatology 7, 235-243

106. Lin, N.-Y., Beyer, C., Gießl, A., Kireva, T., Scholtysek, C., Uderhardt, S., Munoz, L. E., Dees, C., Distler, A., and Wirtz, S. (2013) Autophagy regulates TNFα-mediated joint destruction in experimental arthritis. Annals of the rheumatic diseases 72, 761-768 136

107. Van Liefferinge, J., Massie, A., Portelli, J., Di Giovanni, G., and Smolders, I. (2013) Are vesicular neurotransmitter transporters potential treatment targets for temporal lobe epilepsy?

108. Beyenbach, K. W., and Wieczorek, H. (2006) The V-type H+ ATPase: molecular structure and function, physiological roles and regulation. Journal of Experimental Biology 209, 577-589

109. Cotter, K., Capecci, J., Sennoune, S., Huss, M., Maier, M., Martinez-Zaguilan, R., and Forgac, M. (2015) Activity of plasma membrane V-ATPases is critical for the invasion of MDA-MB231 breast cancer cells. Journal of Biological Chemistry 290, 3680-3692

110. Mazhab-Jafari, M. T., Rohou, A., Schmidt, C., Bueler, S. A., Benlekbir, S., Robinson, C. V., and Rubinstein, J. L. (2016) Atomic model for the membrane-embedded VO motor of a eukaryotic V-ATPase. Nature 539, 118-122

111. Radermacher, M., Ruiz, T., Wieczorek, H., and Grüber, G. (2001) The structure of the V 1-ATPase determined by three-dimensional electron microscopy of single particles. Journal of structural biology 135, 26-37

112. Arai, S., Saijo, S., Suzuki, K., Mizutani, K., Kakinuma, Y., Ishizuka-Katsura, Y., Ohsawa, N., Terada, T., Shirouzu, M., and Yokoyama, S. (2013) Rotation mechanism of Enterococcus hirae V1-ATPase based on asymmetric crystal structures. Nature 493, 703- 707

113. Muench, S. P., Huss, M., Song, C. F., Phillips, C., Wieczorek, H., Trinick, J., and Harrison, M. A. (2009) Cryo-electron microscopy of the vacuolar ATPase motor reveals its mechanical and regulatory complexity. Journal of molecular biology 386, 989-999

114. Diepholz, M., Venzke, D., Prinz, S., Batisse, C., Flörchinger, B., Rössle, M., Svergun, D. I., Böttcher, B., and Féthière, J. (2008) A different conformation for EGC stator subcomplex in solution and in the assembled yeast V-ATPase: possible implications for regulatory disassembly. Structure 16, 1789-1798

115. Zhang, Z., Zheng, Y., Mazon, H., Milgrom, E., Kitagawa, N., Kish-Trier, E., Heck, A. J., Kane, P. M., and Wilkens, S. (2008) Structure of the yeast vacuolar ATPase. Journal of Biological Chemistry 283, 35983-35995

116. Benlekbir, S., Bueler, S. A., and Rubinstein, J. L. (2012) Structure of the vacuolar-type ATPase from Saccharomyces cerevisiae at 11-Å resolution. Nature structural & molecular biology 19, 1356-1362

117. Iwata, M., Imamura, H., Stambouli, E., Ikeda, C., Tamakoshi, M., Nagata, K., Makyio, H., Hankamer, B., Barber, J., and Yoshida, M. (2004) Crystal structure of a central stalk subunit C and reversible association/dissociation of vacuole-type ATPase. Proceedings of the National Academy of Sciences 101, 59-64 137

118. Powell, B., Graham, L. A., and Stevens, T. H. (2000) Molecular characterization of the yeast vacuolar H+-ATPase proton pore. Journal of Biological Chemistry 275, 23654- 23660

119. Couoh-Cardel, S., Hsueh, Y.-C., Wilkens, S., and Movileanu, L. (2016) Yeast V-ATPase proteolipid ring acts as a large-conductance transmembrane protein pore. Scientific reports 6

120. Wang, Y., Cipriano, D. J., and Forgac, M. (2007) Arrangement of subunits in the proteolipid ring of the V-ATPase. Journal of Biological Chemistry 282, 34058-34065

121. Wilkens, S., Zhang, Z., and Zheng, Y. (2005) A structural model of the vacuolar ATPase from transmission electron microscopy. Micron 36, 109-126

122. McGuire, C., Cotter, K., Stransky, L., and Forgac, M. (2016) Regulation of V-ATPase assembly and function of V-ATPases in tumor cell invasiveness. Biochimica et Biophysica Acta (BBA)-Bioenergetics 1857, 1213-1218

123. Marshansky, V., and Futai, M. (2008) The V-type H+-ATPase in vesicular trafficking: targeting, regulation and function. Current opinion in cell biology 20, 415-426

124. Kane, P. M., Tarsio, M., and Liu, J. (1999) Early steps in assembly of the yeast vacuolar H+-ATPase. Journal of Biological Chemistry 274, 17275-17283

125. Kane, P. M., and Smardon, A. M. (2003) Assembly and regulation of the yeast vacuolar H+-ATPase. Journal of bioenergetics and biomembranes 35, 313-321

126. Ramachandran, N., Munteanu, I., Wang, P., Ruggieri, A., Rilstone, J. J., Israelian, N., Naranian, T., Paroutis, P., Guo, R., and Ren, Z.-P. (2013) VMA21 deficiency prevents vacuolar ATPase assembly and causes autophagic vacuolar myopathy. Acta neuropathologica 125, 439-457

127. Kane, P. M. (1995) Disassembly and reassembly of the yeast vacuolar H+-ATPase in vivo. Journal of Biological Chemistry 270, 17025-17032

128. Diakov, T. T., and Kane, P. M. (2010) Regulation of vacuolar proton-translocating ATPase activity and assembly by extracellular pH. Journal of Biological Chemistry 285, 23771-23778

129. Smith, G. A., Howell, G. J., Phillips, C., Muench, S. P., Ponnambalam, S., and Harrison, M. A. (2016) Extracellular and luminal pH regulation by vacuolar H+-ATPase isoform expression and targeting to the plasma membrane and endosomes. Journal of Biological Chemistry 291, 8500-8515

130. Finnigan, G. C., Hanson-Smith, V., Houser, B. D., Park, H. J., and Stevens, T. H. (2011) The reconstructed ancestral subunit a functions as both V-ATPase isoforms Vph1p and Stv1p in Saccharomyces cerevisiae. Molecular biology of the cell 22, 3176-3191 138

131. Kawasaki-Nishi, S., Bowers, K., Nishi, T., Forgac, M., and Stevens, T. H. (2001) The amino-terminal domain of the vacuolar proton-translocating ATPase a subunit controls targeting and in vivo dissociation, and the carboxyl-terminal domain affects coupling of proton transport and ATP hydrolysis. Journal of Biological Chemistry 276, 47411-47420

132. Karet, F. E., Finberg, K. E., Nelson, R. D., Nayir, A., Mocan, H., Sanjad, S. A., Rodriguez-Soriano, J., Santos, F., Cremers, C. W., and Di Pietro, A. (1999) Mutations in the gene encoding B1 subunit of H+-ATPase cause renal tubular acidosis with sensorineural deafness. Nature genetics 21, 84-90

133. Hiesinger, P. R., Fayyazuddin, A., Mehta, S. Q., Rosenmund, T., Schulze, K. L., Zhai, R. G., Verstreken, P., Cao, Y., Zhou, Y., and Kunz, J. (2005) The v-ATPase V 0 subunit a1 is required for a late step in synaptic vesicle exocytosis in Drosophila. Cell 121, 607-620

134. Fischer, B., Dimopoulou, A., Egerer, J., Gardeitchik, T., Kidd, A., Jost, D., Kayserili, H., Alanay, Y., Tantcheva-Poor, I., and Mangold, E. (2012) Further characterization of ATP6V0A2-related autosomal recessive cutis laxa. Human genetics 131, 1761-1773

135. Guillard, M., Dimopoulou, A., Fischer, B., Morava, E., Lefeber, D. J., Kornak, U., and Wevers, R. A. (2009) Vacuolar H+-ATPase meets glycosylation in patients with cutis laxa. Biochimica et Biophysica Acta (BBA)-Molecular Basis of Disease 1792, 903-914

136. Pangrazio, A., Caldana, M., Iacono, N. L., Mantero, S., Vezzoni, P., Villa, A., and Sobacchi, C. (2012) Autosomal recessive osteopetrosis: report of 41 novel mutations in the TCIRG1 gene and diagnostic implications. Osteoporosis international 23, 2713-2718

137. Sobacchi, C., Schulz, A., Coxon, F. P., Villa, A., and Helfrich, M. H. (2013) Osteopetrosis: genetics, treatment and new insights into osteoclast function. Nature Reviews Endocrinology 9, 522-536

138. Batlle, D., and Haque, S. K. (2012) Genetic causes and mechanisms of distal renal tubular acidosis. Nephrology Dialysis Transplantation 27, 3691-3704

139. Kartner, N., Yao, Y., Bhargava, A., and Manolson, M. F. (2013) Topology, glycosylation and conformational changes in the membrane domain of the vacuolar H+‐ATPase a subunit. Journal of cellular biochemistry 114, 1474-1487

140. Srinivasan, S., Vyas, N. K., Baker, M. L., and Quiocho, F. A. (2011) Crystal structure of the cytoplasmic N-terminal domain of subunit I, a homolog of subunit a, of V-ATPase. Journal of molecular biology 412, 14-21

141. Marshansky, V., Rubinstein, J. L., and Grüber, G. (2014) Eukaryotic V-ATPase: novel structural findings and functional insights. Biochimica et Biophysica Acta (BBA)- Bioenergetics 1837, 857-879

142. Toei, M., Toei, S., and Forgac, M. (2011) Definition of membrane topology and identification of residues important for transport in subunit a of the vacuolar ATPase. Journal of Biological Chemistry 286, 35176-35186 139

143. Esmail, S., Kartner, N., Yao, Y., Kim, J. W., Reithmeier, R. A., and Manolson, M. F. (2017) N‐linked Glycosylation of a Subunit Isoforms is Critical for Vertebrate Vacuolar H+‐ATPase (V‐ATPase) Biosynthesis. Journal of Cellular Biochemistry

144. Bowman, E. J., Siebers, A., and Altendorf, K. (1988) Bafilomycins: a class of inhibitors of membrane ATPases from microorganisms, animal cells, and plant cells. Proceedings of the National Academy of Sciences 85, 7972-7976

145. Bowman, E. J., Graham, L. A., Stevens, T. H., and Bowman, B. J. (2004) The bafilomycin/concanamycin binding site in subunit c of the V-ATPases from Neurospora crassa and Saccharomyces cerevisiae. Journal of Biological Chemistry 279, 33131-33138

146. Huss, M., Ingenhorst, G., König, S., Gaßel, M., Dröse, S., Zeeck, A., Altendorf, K., and Wieczorek, H. (2002) Concanamycin A, the specific inhibitor of V-ATPases, binds to the Vo subunit c. Journal of Biological Chemistry 277, 40544-40548

147. Bockelmann, S., Menche, D., Rudolph, S., Bender, T., Grond, S., von Zezschwitz, P., Muench, S. P., Wieczorek, H., and Huss, M. (2010) Archazolid A binds to the equatorial region of the c-ring of the vacuolar H+-ATPase. Journal of Biological Chemistry 285, 38304-38314

148. Huss, M., Sasse, F., Kunze, B., Jansen, R., Steinmetz, H., Ingenhorst, G., Zeeck, A., and Wieczorek, H. (2005) Archazolid and apicularen: novel specific V-ATPase inhibitors. BMC biochemistry 6, 1

149. Boyd, M. R., Farina, C., Belfiore, P., Gagliardi, S., Kim, J. W., Hayakawa, Y., Beutler, J. A., McKee, T. C., Bowman, B. J., and Bowman, E. J. (2001) Discovery of a novel antitumor benzolactone enamide class that selectively inhibits mammalian vacuolar-type (H+)-ATPases. Journal of Pharmacology and Experimental Therapeutics 297, 114-120

150. Chouabe, C., Eyraud, V., Da Silva, P., Rahioui, I., Royer, C., Soulage, C., Bonvallet, R., Huss, M., and Gressent, F. (2011) New mode of action for a knottin protein bioinsecticide pea albumin 1 subunit b (PA1b) is the first peptidic inhibitor of V-ATPase. Journal of Biological Chemistry 286, 36291-36296

151. Forgac, M. (2007) Vacuolar ATPases: Rotary proton pumps in physiology and pathophysiology. Nat. Rev. Mol. Cell Biol. 8, 917–929

152. Futai, M., Nakanishi-Matsui, M., Okamoto, H., Sekiya, M., and Nakamoto, R. K. (2012) Rotational catalysis in proton pumping ATPases: from E. coli F-ATPase to mammalian V-ATPase. Biochim. Biophys. Acta 1817, 1711–1721

153. Saroussi, S., and Nelson, N. (2009) Vacuolar H+-ATPase—an enzyme for all seasons. Pfluegers Arch./Eur. J. Physiol. 457, 581–587

154. Kawasaki-Nishi, S., Bowers, K., Nishi, T., Forgac, M., and Stevens, T. H. (2001) The amino-terminal domain of the vacuolar proton-translocating ATPase a subunit controls 140

targeting and in vivo dissociation, and the carboxyl-terminal domain affects coupling of proton transport and ATP hydrolysis. J. Biol. Chem. 276, 47411–47420

155. Finnigan, G. C., Cronan, G. E., Park, H. J., Srinivasan, S., Quiocho, F. A., and Stevens, T. H. (2012) Sorting of the yeast vacuolar-type, proton-translocating ATPase enzyme complex (V-ATPase). J. Biol. Chem. 287, 19487–19500

156. Wagner, C. A., Finberg, K. E., Breton, S., Marshansky, V., Brown, D., and Geibel, J. P. (2004) Renal vacuolar H+-ATPase. Physiol. Rev. 84, 1263–1314

157. Hinton, A., Bond, S., and Forgac, M. (2009) V-ATPase functions in normal and disease processes. Pfluegers Arch./Eur. J. Physiol. 457, 589–598

158. Kartner, N., and Manolson, M. F. (2016) The vavuolar proton ATPase (V-ATPase): regulation and therapeutic targeting. in Regulation of Ca2+-ATPases, V-ATPases and F- ATPases (Chakraborti, S., and Dhalla, N. S. eds.), Springer International Publishing. pp 407–437

159. Lau, W. C. Y., and Rubinstein, J. L. (2010) Structure of intact Thermus thermophilus V- ATPase by cryo-EM reveals organization of the membrane-bound Vo motor. Proc. Natl. Acad. Sci. U. S. A. 107, 1367–1372

160. Kartner, N., Yao, Y., Li, K., Crasto, G. J., Datti, A., and Manolson, M. F. (2010) Inhibition of osteoclast bone resorption by disrupting vacuolar H+-ATPase a3–B2 subunit interaction. J. Biol. Chem. 285, 37476–37490

161. Kartner, N., Yao, Y., Bhargava, A., and Manolson, M. F. (2013) Topology, glycosylation and conformational changes in the membrane domain of the vacuolar H+-ATPase a subunit. J. Cell. Biochem. 114, 1474–1487

162. Bhargava, A., Voronov, I., Wang, Y., Glogauer, M., Kartner, N., and Manolson, M. F. (2012) Osteopetrosis mutation R444L causes ER retention and misprocessing of vacuolar H+-ATPase a3 subunit. J. Biol. Chem. 287, 26829–26839

163. Kaplan, H. A., Welply, J. K., and Lennarz, W. J. (1987) Oligosaccharyl transferase: the central enzyme in the pathway of glycoprotein assembly. Biochim. Biophys. Acta 906, 161–173

164. Mohorko, E., Glockshuber, R., and Aebi, M. (2011) Oligosaccharyltransferase: the central enzyme of N-linked protein glycosylation. J. Inherited Metab. Dis. 34, 869–878

165. Ferris, S. P., Kodali, V. K., and Kaufman, R. J. (2014) Glycoprotein folding and quality- control mechanisms in protein-folding diseases. Dis. Model. Mech. 7, 331–341

166. Moharir, A., Peck, S. H., Budden, T., and Lee, S. Y. (2013) The role of N-glycosylation in folding, trafficking, and functionality of lysosomal protein CLN5. PLoS One 8, e74299 141

167. Zhou, F., Xu, W., Hong, M., Pan, Z., Sinko, P. J., Ma, J., and You, G. (2005) The role of N-linked glycosylation in protein folding, membrane targeting, and substrate binding of human organic anion transporter hOAT4. Mol. Pharmacol. 63, 868–876

168. Landolt-Marticorena, C., and Reithmeier, R. A. F. (1994) Asparagine-linked oligosaccharides are localized to single extracytosolic segments in multi-span membrane glycoproteins. Biochem. J. 302, 253–260

169. Cheung, J. C., and Reithmeier, R. A. F. (2007) Scanning N-glycosylation mutagenesis of membrane proteins. Methods 41, 451–459

170. Li, J., Xia, F., and Reithmeier, R. A. F. (2014) N-glycosylation and topology of the human SLC26 family of anion transport membrane proteins. Am. J. Physiol. Cell Physiol. 306, C943–C960

171. Plummer Jr., T. H., and Tarantino, A. L. (19991) Purification of the oligosaccharide- cleaving enzymes of Flavobacterium meningosepticum. Glycobiology 1, 257–263

172. Maley, F., Trimble, R. B., Tarentino, A. L., and Plummer Jr., T. H. (1989) Characterization of glycoproteins and their associated oligosaccharides through the use of endoglycosidases. Anal. Biochem. 180, 195–204

173. Freeze, H. H., and Kranz, C. (2010) Endoglycosidase and glycoamidase release of N- linked glycans. Curr. Protoc. Mol. Biol.

174. Kisselev, A. F., and Goldberg, A. L. (2001) Proteasome inhibitors: from research tools to drug candidates. Chem. Biol. 8, 739–758

175. Koh, Y. H., von Arnim, C. A. F., Bradley, T. H., Tanzi, R. E., and Tesco, G. (2005) BACE is degraded via the lysosomal pathway. J. Biol. Chem. 280, 32499–32504

176. Lee, T.-T., Zhang, X.-D., Chuang, C.-C., Chen, J.-J., Chen, Y.-A., Chen, S.-C., Chen, T.- Y., and Tang, C.-Y. (2013) Myotonia congenita mutation enhances the degradation of human CLC-1 chloride channels. PLoS One 8, e55930

177. Okiyoneda, T., Barrière, H., Bagdány, M., Rabeh, W. M., Du, K., Höhfeld, J., Young, J. C., and Lukacs, G. L. (2010) Peripheral protein quality control removes unfolded CFTR from the plasma membrane. Science 329, 805–810

178. Lippincott-Schwartz, J., Roberts, T. H., and Hirschberg, K. (2000) Secretory protein trafficking and organelle dynamics in living cells. Annu. Rev. Cell. Dev. Biol. 16, 557– 589

179. Toei, M., Saum, R., and Forgac, M. (2010) Regulation and isoform function of the V- ATPases. Biochemistry 49, 4715–4723

180. Ryan, M., Graham, L. A., and Stevens, T. H. (2008) Voa1p functions in V-ATPase assembly in the yeast endoplasmic reticulum. Mol. Biol. Cell 19, 5131–5142 142

181. Hill, K., and Cooper, A. A. (2000) Degradation of unassembled Vph1p reveals novel aspects of the yeast ER quality control system. EMBO J. 19, 550–561

182. Solá, R. J., and Griebenow, K. (2009) Effects of glycosylation on the stability of protein pharmaceuticals. J. Pharm. Sci. 98, 1223–1245

183. Mitra, N., Sinha, S., Ramya, T., and Surolia, A. (2006) N-linked oligosaccharides as outfitters for glycoprotein folding, form and function. Trends Biochem. Sci. 31, 156–163

184. Watanabe, I., Zhu, J., Recio-Pinto, E., and Thornhill, W. B. (2004) Glycosylation affects the protein stability and cell surface expression of Kv1.4 but not Kv1.1 potassium channels. J. Biol. Chem. 279, 8879–8885

185. Lemus, L., and Goder, V. (2014) Regulation of endoplasmic reticulum-associated protein degradation (ERAD) by ubiquitin. Cells 3, 824–847

186. Fujita, E., Kouroku, Y., Isoai, A., Kumagai, H., Misutani, A., Matsuda, C., Hayashi, Y. K., and Momoi, T. (2007) Two endoplasmic reticulum-associated degradation (ERAD) systems for the novel variant of the mutant dysferlin: ubiquitin/proteasome ERAD(I) and autophagy/lysosome ERAD(II). Hum. Mol. Genet. 16, 618–629

187. Console, L., Scalise, M., Tarmakova, Z., Coe, I. R., and Indiveri, C. (2015) N-linked glycosylation of human SLC1A5 (ASCT2) transporter is critical for trafficking to membrane. Biochim. Biophys. Acta 1853, 1636–1645

188. Vagin, O., Kraut, J. A., and Sachs, G. (2009) Role of N-glycosylation in trafficking of apical membrane proteins in epithelia. Am. J. Physiol. Renal Physiol. 296, F459–F469

189. Kartner, N., and Manolson, M. F. (2016) The vacuolar proton ATPase (V-ATPase): regulation and therapeutic targeting. in Regulation of Ca2+-ATPases, V-ATPases and F- ATPases (Chakraborti, S., and Dhalla, N. S. eds.), Springer International Publishing, Switzerland. pp 407–437

190. Schep, D. G., Zhao, J., and Rubinstein, J. L. (2016) Models for the a subunits of the Thermus thermophilus V/A-ATPase and Saccharomyces cervisiae V-ATPase enzymes by cryo-EM and evolutionary covariance. Proc. Natl. Acad. Sci. U. S. A. 113, 3245–3250

191. Stam, N. J., and Wilkens, S. (2016) Structure of nanodisc reconstituted vacuolar ATPase proton channel: Definition of the interaction of rotor and stator and implications for enzyme regulation by reversible dissociation. J. Biol. Chem.

192. Nilsson, I., and von Heijne, G. (1993) Determination fo the distance between the oligosaccharyltransferase active site and the endoplasmic reticulum membrane. J. Biol. Chem. 268, 5797–5801

193. Malaby, H. L. H., and Kobertz, W. R. (2014) The middle X residue influences cotranslational N-glycosylation consensus site skipping. Biochemistry 53, 4884–4893 143

194. Shakin-Eshleman, S. H., Spitalnik, S. L., and Kasturi, L. (1996) The amino acid at the X position of an Asn-X-Ser sequon is an important determination of N-linked core- glycosylation efficiency. J. Biol. Chem. 271, 6363–6366

195. Bañó-Polo, M., Baldin, F., Tamborero, S., Marti-Renom, M. A., and Mingarro, I. (2011) N-glycosylation efficiency is determined by the distance to the C-terminus and the amino acid preceding an Asn-Ser-Thr sequon. Protein Sci. 20, 179–186

196. Mellquist, J. L., Ksturi, L., Spitalnik, S. L., and Shakin-Eshleman, S. H. (1998) The amino acid following an Asn-X-Ser/Thr sequon is an important determinant of N-linked core glycosylation efficiency. Biochemistry 37, 6833–6837

197. Apweiler, R., Hermjakob, H., and Sharon, N. (1999) On the frequency of protein glycosylation, as deduced from analysis of the SWISS-PROT database. Biochim. Biophys. Acta 1473, 4–8

198. Petrescu, A., Milac, A.-L., Petrescu, S. M., Dwek, R. A., and Wormald, M. R. (2004) Statistical analysis of the protein environment of N-glycosylation sites: implications for occupancy, structure, and folding. Glycobiology 14, 103–114

199. Esmail, S., Yao, Y., Kartner, N., Li, J., Reithmeier, R. A. F., and Manolson, M. F. (2016) N-linked glycosylation is required for vacuolar H+-ATPase (V-ATPase) a4 subunit stability, assembly, and cell surface expression. J. Cell. Biochem. 117, 2757

200. Plummer Jr., T. H., and Tarantino, A. L. (1991) Purification of the oligosaccharide- cleaving enzymes of Flavobacterium meningosepticum. Glycobiology 1, 257–263

201. Haga, Y., Ishii, K., and Suzuki, T. (2011) N-glycosylation is critical for stability and intracellular trafficking of glucose transporter GLUT4. J. Biol. Chem. 286, 31320–31327

202. Ramachandran, N., Munteanu, I., Wang, P., Ruggieri, A., Rilstone, J. J., Israelian, N., Naranian, T., Paroutis, P., Guo, R., Ren, Z.-P., Nishino, I., Charbrol, B., Pellissier, J.-F., Minetti, C., Udd, B., Fardeau, M., Tailor, C. S., Mahuran, D. J., Kissel, J. T., Kalimo, H., Levy, N., Manolson, M. F., Ackerley, C. A., and Mianssian, B. A. (2013) VMA21 deficiency prevents vacuolar ATPase assembly and causes autophagic vacuolar myopathy. Acta Neuropathol. (Berl). 125, 439–457

203. Branza-Nichitas, N., Negroiu, G., Petrescu, A. J., Garman, E. F., Platt, F. M., Wormald, M. R., Dwek, R. A., and Petrescu, S. M. (2000) Mutations at critical N-glycosylation sites reduce tyrosinase activity by altering folding and quality control. J. Biol. Chem. 275, 8169–8175

204. Casey, J. R., Grinstein, S., and Orlowski, J. (2009) Sensors and regulators of intracellular pH. Nat. Rev. Mol. Cell Biol. 11, 50–61

205. Beyenbach, K. W., and Wieczorek, H. (2006) The V-type H+ ATPase: molecular structure and function, physiological roles and regulation. J. Exp. Biol. 209, 577–589 144

206. Sun-Wada, G.-H., and Wada, Y. (2013) Vacuolar-type proton pump ATPases: acidification and pathological relationships. Histol. Histopathol. 28, 805–815

207. Ochotny, N., Van Vliet, A., Chan, N., Yao, Y., Morel, M., Kartner, N., von Schroeder, H. P., Heersche, J. N. M., and Manolson, M. F. (2006) Effects of human a3 and a4 mutations that result in osteopetrosis and distal renal tubular acidosis on yeast V-ATPase expression and activity. J. Biol. Chem. 281, 26102–26111

208. Kornak, U., Schultz, A., Friedrich, W., Uhlhaas, S., Kremens, B., Voit, T., Hasan, C., Bode, U., Jentsch, T. J., and Kubisch, C. (2000) Mutations in the a3 subunit of the vacuolar H+-ATPase cause infantile malignant osteopetrosis. Hum. Mol. Genet. 9, 2059– 2063

209. Toyomura, T., Oka, T., Yamaguchi, C., Wada, Y., and Futai, M. (2000) Three subunit a isoforms of mouse vacuolar H+-ATPase: Preferential expression of the a3 isoform during osteoclast differentiation. J. Biol. Chem. 275, 8760–8765

210. Bexiga, M. G., and Simpson, J. C. (2013) Human diseases associated with form and function of the Golgi complex. Int. J. Mol. Sci. 14, 18670–18681

211. Kornak, U., Reynders, E., Dimopoulou, A., Van Reeuwijk, J., Fischer, B., Rajab, A., Budde, B., Nürnberg, P., Foulquier, F., and Dobyns, W. B. (2008) Impaired glycosylation and cutis laxa caused by mutations in the vesicular H+-ATPase subunit ATP6V0A2. Nature genetics 40, 32-34

212. Biasini, M., Bienert, S., Waterhouse, A., Arnold, K., Studer, G., Schmidt, T., Kiefer, F., Cassarino, T. G., Bertoni, M., Bordoli, L., and Schwede, T. (2014) SWISS-MODEL: modelling protein tertiary and quaternary structure using evolutionary information. Nucleic Acids Res. 42, W252–W258

213. Van Durme, J., Delgado, J., Stricher, F., Serrano, L., Schymkowitz, J., and Rousseau, F. (2011) A graphical interface for the FoldX forcefield. Bioinformatics 27, 1711-1712

214. Miles, A. L., Burr, S. P., Grice, G. L., and Nathan, J. A. (2017) The vacuolar-ATPase complex and assembly factors, TMEM199 and CCDC115, control HIF1α prolyl hydroxylation by regulating cellular iron levels. eLife 6, e22693

215. Jansen, J. C., Timal, S., van Scherpenzeel, M., Michelakakis, H., Vicogne, D., Ashikov, A., Moraitou, M., Hoischen, A., Huijben, K., and Steenbergen, G. (2016) TMEM199 deficiency is a disorder of Golgi homeostasis characterized by elevated aminotransferases, alkaline phosphatase, and cholesterol and abnormal glycosylation. The American Journal of Human Genetics 98, 322-330

216. Guevara-Coto, J., Schwartz, C. E., and Wang, L. (2014) Protein sector analysis for the clustering of disease-associated mutations. BMC Genomics 15, 1–7

217. Valastyan, J. S., and Lindquist, S. (2014) Mechanisms of protein-folding diseases at a glance. Dis. Model. Mech. 7, 9–14 145

218. Okiyoneda, T., Veit, G., Dekkers, J. F., Bagdany, M., Soya, N., Xu, J., Verkman, A. S., Kurth, M., Simon, A., Hegedus, T., Beekman, J. M., and Lukacs, G. L. (2013) Mechanism-based corrector combination restores ∆F508-CFTR folding and function. Nat. Chem. Biol. 9, 1–27

219. Cheng, S. H., Gregory, R. J., Marshall, J., Paul, S., Souza, D., White, G. A., O'Riordan, C. R., and Smith, A. E. (1990) Defective intracellular transport and processing of CFTR is the molecular basis of most cystic fibrosis. Cell 63, 827–834

220. Malkus, P., Graham, L. A., Stevens, T. H., and Schekman, R. (2004) Role of Vma21p in assembly and transport of the yeast vacuolar ATPase. Molecular biology of the cell 15, 5075-5091

221. Ochotny, N., Van Vliet, A., Chan, N., Yao, Y., Morel, M., Kartner, N., von Schroeder, H. P., Heersche, J. N., and Manolson, M. F. (2006) Effects of human a3 and a4 mutations that result in osteopetrosis and distal renal tubular acidosis on yeast V-ATPase expression and activity. Journal of Biological Chemistry 281, 26102-26111

222. Beers, M. F., Zhao, M., Tomer, Y., Russo, S. J., Zhang, P., Gonzales, L. W., Guttentag, S. H., and Mulugeta, S. (2013) Disruption of N-linked glycosylation promotes proteasomal degradation of the human ATP-binding cassette transporter ABCA3. American Journal of Physiology-Lung Cellular and Molecular Physiology 305, L970- L980

223. Branza-Nichita, N., Negroiu, G., Petrescu, A. J., Garman, E. F., Platt, F. M., Wormald, M. R., Dwek, R. A., and Petrescu, S. M. (2000) Mutations at critical N-glycosylation sites reduce tyrosinase activity by altering folding and quality control. Journal of Biological Chemistry 275, 8169-8175

224. Kerr, K. M., Sauna, Z. E., and Ambudkar, S. V. (2001) Correlation between Steady-state ATP Hydrolysis and Vanadate-induced ADP Trapping in Human P-glycoprotein EVIDENCE FOR ADP RELEASE AS THE RATE-LIMITING STEP IN THE CATALYTIC CYCLE AND ITS MODULATION BY SUBSTRATES. Journal of Biological Chemistry 276, 8657-8664

225. Scott, A. M., Wolchok, J. D., and Old, L. J. (2012) Antibody therapy of cancer. Nature Reviews Cancer 12, 278-287

226. Higel, F., Seidl, A., Sörgel, F., and Friess, W. (2016) N-glycosylation heterogeneity and the influence on structure, function and pharmacokinetics of monoclonal antibodies and Fc fusion proteins. European Journal of Pharmaceutics and Biopharmaceutics 100, 94- 100

227. Monteiro, J., Aires, R., Becker, J. D., Jacinto, A., Certal, A. C., and Rodríguez-León, J. (2014) V-ATPase proton pumping activity is required for adult zebrafish appendage regeneration. PloS one 9, e92594 146

228. Del Fattore, A., Cappariello, A., and Teti, A. (2008) Genetics, pathogenesis and complications of osteopetrosis. Bone 42, 19-29

229. Bhargava, A., Voronov, I., Wang, Y., Glogauer, M., Kartner, N., and Manolson, M. F. (2012) Osteopetrosis mutation R444L causes endoplasmic reticulum retention and misprocessing of vacuolar H+-ATPase a3 subunit. Journal of Biological Chemistry 287, 26829-26839

230. Okiyoneda, T., Veit, G., Dekkers, J. F., Bagdany, M., Soya, N., Xu, H., Roldan, A., Verkman, A. S., Kurth, M., and Simon, A. (2013) Mechanism-based corrector combination restores ΔF508-CFTR folding and function. Nature chemical biology 9, 444-454

231. Allegretti, M., Klusch, N., Mills, D. J., Vonck, J., Kühlbrandt, W., and Davies, K. M. (2015) Horizontal membrane-intrinsic [agr]-helices in the stator a-subunit of an F-type ATP synthase. Nature 521, 237-240

232. Matoulková, E., and Vojtěšek, B. (2013) Detection of protein-protein interactions by FRET and BRET methods. Klinicka onkologie: casopis Ceske a Slovenske onkologicke spolecnosti 27, S82-86

233. Whyteside, G., Gibson, L., Scott, M., and Finbow, M. E. (2005) Assembly of the yeast vacuolar H+‐ATPase and ATP hydrolysis occurs in the absence of subunit c ″. FEBS letters 579, 2981-2985

234. Ho, M. N., Hirata, R., Umemoto, N., Ohya, Y., Takatsuki, A., Stevens, T., and Anraku, Y. (1993) VMA13 encodes a 54-kDa vacuolar H (+)-ATPase subunit required for activity but not assembly of the enzyme complex in Saccharomyces cerevisiae. Journal of Biological Chemistry 268, 18286-18292

235. Robinson, M. W., Alvarado, R., To, J., Hutchinson, A. T., Dowdell, S. N., Lund, M., Turnbull, L., Whitchurch, C. B., O'Brien, B. A., and Dalton, J. P. (2012) A helminth cathelicidin-like protein suppresses antigen processing and presentation in macrophages via inhibition of lysosomal vATPase. The FASEB Journal 26, 4614-4627