ABCDEFG

UNIVERS ITY OF OULU P.O.B . 7500 F I -90014 UNIVERS ITY OF OULU F INLAND

A C T A U N I V E R S I T A T I S O U L U E N S I S

S E R I E S E D I T O R S

SCIENTIAE RERUM NATURALIUM

HUMANIORA

TECHNICA

MEDICA

SCIENTIAE RERUM SOCIALIUM

SCRIPTA ACADEMICA

OECONOMICA

EDITOR IN CHIEF

PUBLICATIONS EDITOR

Professor Mikko Siponen

University Lecturer Elise Kärkkäinen

Professor Hannu Heusala

Professor Helvi Kyngäs

Senior Researcher Eila Estola

Information officer Tiina Pistokoski

University Lecturer Seppo Eriksson

University Lecturer Seppo Eriksson

Publications Editor Kirsti Nurkkala

ISBN 978-951-42-9268-2 (Paperback)ISBN 978-951-42-9269-9 (PDF)ISSN 0355-3191 (Print)ISSN 1796-220X (Online)

U N I V E R S I TAT I S O U L U E N S I SACTAA

SCIENTIAE RERUM NATURALIUM

U N I V E R S I TAT I S O U L U E N S I SACTAA

SCIENTIAE RERUM NATURALIUM

OULU 2009

A 535

Antti Rivinoja

GOLGI pH AND GLYCOSYLATION

FACULTY OF SCIENCE,DEPARTMENT OF BIOCHEMISTRY,UNIVERSITY OF OULU

A 535

ACTA

Antti R

ivinoja

A C T A U N I V E R S I T A T I S O U L U E N S I SA S c i e n t i a e R e r u m N a t u r a l i u m 5 3 5

ANTTI RIVINOJA

GOLGI pH AND GLYCOSYLATION

Academic dissertation to be presented with the assent ofthe Faculty of Science of the University of Oulu for publicdefence in Auditorium IT115, Linnanmaa, on 23 October2009, at 12 noon

OULUN YLIOPISTO, OULU 2009

Copyright © 2009Acta Univ. Oul. A 535, 2009

Supervised byDocent Sakari Kellokumpu

Reviewed byDocent Esa KuismanenDocent Vesa Olkkonen

ISBN 978-951-42-9268-2 (Paperback)ISBN 978-951-42-9269-9 (PDF)http://herkules.oulu.fi/isbn9789514292699/ISSN 0355-3191 (Printed)ISSN 1796-220X (Online)http://herkules.oulu.fi/issn03553191/

Cover designRaimo Ahonen

OULU UNIVERSITY PRESSOULU 2009

Rivinoja, Antti, Golgi pH and glycosylation. Faculty of Science, Department of Biochemistry, University of Oulu, P.O.Box 3000, FI-90014University of Oulu, Finland Acta Univ. Oul. A 535, 2009Oulu, Finland

AbstractGlycans, as a part of glycoproteins, glycolipids and other glycoconjugates, are involved in manyvital intra- and inter-cellular tasks, such as protein folding and sorting, protein quality control,vesicular trafficking, cell signalling, immunological defence, cell motility and adhesion.Therefore, their correct construction is crucial for the normal functioning of eukaryotic cells andorganisms they form. Most cellular glycans are constructed in the Golgi, and abnormalities in theirstructure may derive, for instance, from alkalinization of the Golgi lumen.

In this work we show that Golgi pH is generally higher and more variable in abnormallyglycosylating, i.e. strongly T-antigen (Gal-β1,3-GalNAc-ser/thr) expressing cancer cells, than innon-T-antigen expressing cells. We also confirmed that the Golgi pH alterations detected in cancercells have the potential to induce glycosylation changes. A mere 0.2 pH unit increase in Golgi pHis able to induce T-antigen expression and inhibit terminal N-glycosylation in normallyglycosylating cells. The mechanism of inhibition involves mislocalization of the correspondingglycosyltransferases.

We also studied potential factors that can promote Golgi pH misregulation in health anddisease, and found that cultured cancer cells, despite variation and elevation in Golgi pH, are fullycapable of acidifying the Golgi lumen under the normal Golgi pH. Moreover, we introduce a Golgilocalized Cl-/HCO3- exchanger, AE2a, that participates in Golgi pH regulation by alteringluminal bicarbonate concentration and thus also buffering capacity. Participation of AE2a in GolgipH regulation is especially intriguing, because it also provides a novel mechanism for expellingprotons from the Golgi lumen.

Keywords: anion exchanger, bicarbonate, cancer, glycosylation, Golgi apparatus, pHregulation

5

Acknowledgements

This work was carried out at the Department of Biochemistry, University of Oulu

under the Finnish Glycoscience Graduate School.

I would like to express my deepest gratitude to my supervisor Docent Sakari

Kellokumpu for guiding me in my research project and teaching me many

important things in cell biology. Special thanks also to the present and former SK-

group members, particularly Antti Hassinen, Annika Kauppila and Nina

Kokkonen. I also would like to express my gratitude to the head of the department,

Professor Kalervo Hiltunen, for running the department well during these years,

the technical staff, especially Kyösti Keränen, and other people, who have helped

me with my project and made the working atmosphere nice in our small

department. I also wish to thank Ari-Pekka Kvist for helping me with

mathematical and computer related issues, Professor Peter Neubauer for helping

me with analysis of adenosine nucleotides, Susanna Hamari for constructing

sCEA plasmid, Maija Risteli and Docent Vesa Olkkonen for providing me with

human fibroblasts and CaCO-2 cells for Golgi pH measurements, Maila Konttila

for recruiting me in the Sujakka-team, Tuula Koret, Pia Askonen and Anneli

Kaattari for helping me with the bureaucracy. I am also grateful to the scientists,

who let us to use their plasmids in our studies, and BioAcademia Inc., which

provided me with functional streptolysin O. I also wish to thank Docent Vesa

Olkkonen and Docent Esa Kuismanen for careful reviewing of this thesis and

Alun Parsons for the revision of the language.

Furthermore, I would like to thank my old friends from Santamäki and

Kempele high school, especially Mikko, Heikki, Jani and Ari, who have actively

stayed in my life during these years. I also express my appreciation to my newer

friends that I made at the university and the rock climbing, especially to Ilkka P

and a particular group of “boulder clowns” in OKS-91.

Finally, but not least, I would like to thank my family, especially my father

and my mother for supporting me emotionally and financially over the hard times,

my two sisters Anni and Mia, my brother Eero Jr. and other people, who have just

existed and enriched my family life during these years.

The study was funded by The Academy of Finland, University of Oulu,

Finnish Glycoscience Graduate School, Orion-Farmos Research Foundation, Emil

Aaltonen Foundation and Finnish cancer association.

Oulu, August 2009 Antti Rivinoja

6

7

Abbreviations

ADP adenosine diphosphate

AE anion exchanger

AEC adenylate energy charge

AMP adenosine monophosphate

ATP adenosine triphosphate

BafA1 bafilomycin A1

CaCO-2 human colorectal adenocarcinoma cells

CDG congenital disorders of glycosylation

CF cystic fibrosis

CGN cis-Golgi network

CMP cytidine monophosphate

COG conserved oligomeric complex

COS-7 African green monkey kidney cells

COP coatomer complex

CQ chloroquine

DNA deoxyribonucleic acid

DSA Dature stramonium agglutinin

DTT dithiotreitol

EDTA ethylenediaminetetraacetic acid

EGFP enhanced green fluorescent protein

EGTA glycol-bis(2-aminoethylether)-N,N,N',N'-tetraacetic acid

EM electron microscopy

Endo H endonuclease H

ER endoplasmic reticulum

ERGIC ER-Golgi-intermediate compartment

Fuc fucose

Gal galactose

GalNAc N-acetylgalactosamine

GalT I β1,4-galactosyltransferase

Glc glucose

GlcNAc N-acetylglucosamine

GM130 Golgi matrix protein 130

GPI glycosylphophatidylinositol

GT 80 aminoacids long signal sequence of GalT I

kDa kilodalton

8

HCO3- bicarbonate

H2DIDS 4,4'-diisothiocyanatodihydrostilbene-2,2'-disulfonic acid

HeLa human adenocarcinoma cells from cervix

HRP horseradish peroxidase

HT-29 human colorectal adenocarcinoma cells

LAMP-2 lysosomal-associated membrane protein 2

MAA Maackia amurensis agglutinin

Man mannose

MCF-7 human adenocarcinoma cells from mammary gland

mRNA messenger RNA

MTOC microtubulus organizing center

M6P-R mannose-6-phosphate receptor

NEM N-ethylmaleimide

NSF NEM sensitive factor

PAGE polyacrylamide gel electrophoresis

PNA peanut agglutinin

PNGase F peptide N-glycosidase F

rER rough ER

RNA ribonucleic acid

RNAi RNA interference

RT room temperature (20–25 ºC)

SDS sodium dodecyl sulphate

sER smooth ER

SLO streptolysin O

Sia sialic acid

siRNA small interfering RNA

SNA Sambucus nigra agglutinin

SNAREs soluble NSF attachment protein receptors

ST3Gal III N-acetyllactosaminide alpha2,3-sialyltransferase

ST6Gal I beta-galactoside alpha2,6-sialyltransferase

SW-48 human colorectal adenocarcinoma cells

T-antigen Gal-β1,3-GalNAc-ser/thr

tER transitional ER

TGN trans-Golgi network

UDP uridine diphosphate

V-ATPase vacuolar-H+ ATPase

VTCs vesicular tubular clusters

9

List of original articles

This thesis is based on two original articles and one manuscript, which are

referred to in the text by their roman numerals:

I Rivinoja A, Kokkonen N, Kellokumpu I & Kellokumpu S (2006) Elevated Golgi pH in breast and colorectal cancer cells correlates with the expression of oncofetal carbohydrate T-antigen. J Cell Physiol 208:167–74.

II Rivinoja A, Hassinen A, Kokkonen N, Kauppila A & Kellokumpu S (2009). Elevated Golgi pH impairs terminal N-glycosylation by inducing mislocalization of Golgi glycosyltransferases. J Cell Physiol. 220:144–54.

III Rivinoja A, Kauppila A, Kvist A-P & Kellokumpu S (2009) AE2 Cl-/HCO3-

exchanger is involved in the regulation of Golgi steady state pH. Manuscript.

10

11

Contents

Abstract Acknowledgements 5 Abbreviations 7 List of original articles 9 Contents 11 1 Introduction 15 2 Review of the literature 17

2.1 The Golgi apparatus ................................................................................ 17 2.2 Golgi membrane skeleton ....................................................................... 19

2.2.1 Microtubules ................................................................................. 20 2.2.2 Actin and spectrin ......................................................................... 21 2.2.3 Golgins and small GTPases .......................................................... 22 2.2.4 Golgi structure in mitosis ............................................................. 23

2.3 Membrane trafficking within the Golgi region ....................................... 24 2.3.1 ER-to-Golgi transport ................................................................... 24 2.3.2 Intra-Golgi transport ..................................................................... 25 2.3.3 Post-Golgi transport ..................................................................... 29 2.3.4 Fusion of membrane structures .................................................... 30

2.4 Protein sorting in the Golgi ..................................................................... 31 2.4.1 Golgi-to-ER sorting ...................................................................... 32 2.4.2 Sorting within the Golgi stack ...................................................... 32 2.4.3 Sorting to post-Golgi structures .................................................... 36

2.5 Golgi pH regulation ................................................................................ 37 2.5.1 Vacuolar H+ATPase mediated proton import ................................ 39 2.5.2 Proton export/elimination ............................................................. 41 2.5.3 Counterion conductance ............................................................... 44 2.5.4 Other factors that may affect Golgi pH ........................................ 46

2.6 Golgi functions ........................................................................................ 47 2.6.1 Glycosylation ................................................................................ 47

2.7 Golgi glycosyltransferases ...................................................................... 57 2.7.1 Protein structure and homology .................................................... 57 2.7.2 Localization .................................................................................. 58 2.7.3 β1,4-galactosyltransferase I (EC 2.4.1. 38) .................................. 59 2.7.4 α2,6-sialyltransferase I (EC2.4.99.1) and α2,3-

sialyltransferase III (EC 2.4.99.6) ................................................ 60

12

2.8 Biological role of glycoconjugates .......................................................... 61 2.8.1 Congenital disorders of glycosylation .......................................... 62 2.8.2 Cancer and glycosylation .............................................................. 63

3 Aims of the present work 67 4 Materials and methods 69

4.1 Antibodies, lectins and reagents .............................................................. 69 4.2 Plasmids .................................................................................................. 70 4.3 Cell culture .............................................................................................. 71 4.4 Indirect immunofluorescence and lectin stainings .................................. 72 4.5 Co-localization analysis .......................................................................... 72 4.6 Live cell microscopy ............................................................................... 73

4.6.1 Golgi pH measurements ............................................................... 73 4.6.2 Acidification measurements ......................................................... 73

4.7 Electron microscopy ................................................................................ 74 4.7.1 Preparation of cell pellets ............................................................. 74 4.7.2 Immunogold labeling of AE2 ....................................................... 75

4.8 Flow cytometry ....................................................................................... 75 4.9 Antibody internalization experiments ..................................................... 75 4.10 Preparation of cell lysates ....................................................................... 76

4.10.1 SDS based lysates ......................................................................... 76 4.10.2 Nonidet based lysates ................................................................... 76

4.11 Immunoprecipitation ............................................................................... 77 4.12 Glycosidase treatments ............................................................................ 77 4.13 AE2 gene silencing ................................................................................. 78 4.14 SDS-PAGE and Immuno- /lectin-blotting ............................................... 78

4.14.1 Immunodetection .......................................................................... 78 4.14.2 Lectin detection ............................................................................ 79

4.15 Adenosine nucleotide measurements ...................................................... 79 4.16 Statistical methods .................................................................................. 80

5 Results 81 5.1 Glycosylation and Golgi pH in cancer cells (I) ....................................... 81 5.2 The effects of drug treatments on Golgi pH (I, II) .................................. 82 5.3 Inhibition of glycosylation by Golgi pH dissipating drugs (I, II) ............ 83

5.3.1 Inhibition of terminal O-glycosylation (I) .................................... 83 5.3.2 Inhibition of terminal N-glycosylation (II) ................................... 84 5.3.3 Golgi pH dissipation is able to inhibit α2,3-sialylation

despite of its enhancement by ST3Gal III overexpression ........... 85

13

5.4 Perturbation of Golgi organization with pH gradient dissipating

drugs (II) ................................................................................................. 87 5.5 Chloroquine induced mislocalization of ST3Gal III (II) ......................... 88

5.5.1 Chloroquine induces mislocalization of ST3Gal III from

the Golgi to endosomes/lysosomes and the plasma

membrane ..................................................................................... 88 5.5.2 pH alterations in acidic organelles are associated with

ST3Gal III mislocalization (II) ..................................................... 89 5.5.3 ST3Gal III expression in CaCO-2 cells reveals the

dependency of correct localization and Golgi pH ........................ 89 5.6 Alterations in terminal glycosylation do not affect Golgi

localization (II) ........................................................................................ 90 5.7 The basis of the different pH sensitivity of the two

sialyltransferases, ST3Gal III and ST6Gal I (II) ..................................... 90 5.7.1 ST3Gal III and ST6Gal I localize to the trans-Golgi .................... 90 5.7.2 Covalent oligomerization of ST6Gal I does not affect

Golgi localization ......................................................................... 91 5.7.3 Signals in the luminal domain of ST6Gal I are responsible

for more stable Golgi localization in drug treated cells ................ 92 5.8 Live cell PNA stainings (I) ...................................................................... 92

5.8.1 Golgi pH in T-antigen expressing cells is higher than in

non-expressing cells (I) ................................................................ 92 5.8.2 Golgi acidification is not impaired in T-antigen expressing

MCF-7 cells .................................................................................. 93 5.9 Intracellular adenosine nucleotide (AXP) concentrations in

MCF-7 cells indicate a good energy status ............................................. 94 5.10 AE2a localization within the Golgi (III) ................................................. 95 5.11 Role of AE2a in Golgi acidification (III) ................................................ 96

5.11.1 AE2a overexpression and knock-down elevate Golgi pH

(III) ............................................................................................... 96 5.11.2 AE2 participates in Golgi acidification by altering luminal

buffering capacity (III) ................................................................. 97 6 Discussion 101

6.1 Increased Golgi pH in cancer correlates with increased T-antigen

expression ............................................................................................. 101 6.2 Terminal processing of O- and N-glycans is highly sensitive to

Golgi pH alterations .............................................................................. 101

14

6.3 Glycosylation processes are differently sensitive to Golgi pH

alterations .............................................................................................. 102 6.4 Inhibition of α2,3-sialylation involves mislocalization of the

corresponding sialyltransferase, ST3Gal III .......................................... 103 6.5 ST3Gal III mislocalization .................................................................... 103 6.6 Different pH sensitivities of ST3Gal III and ST6Gal I are mainly

dictated by their luminal domains ......................................................... 105 6.7 Dissipation of the acidic Golgi pH is able to inhibit α2,3-

sialylation despite ST3Gal III overexpression ...................................... 106 6.8 Cancer cell lines are often characterized by heterogeneous cell

surface glycan expression...................................................................... 107 6.9 Golgi pH regulation in health and disease ............................................. 107

6.9.1 V-ATPase is present and functional in the Golgi of MCF-7

cells ............................................................................................. 108 6.9.2 Energy status do not limit Golgi acidification in cultured

MCF-7 cells ................................................................................ 109 6.9.3 Role of AE2a in Golgi pH regulation ......................................... 110

7 Conclusions 113 References 115 Original articles 147

15

1 Introduction

Correct glycosylation of proteins and lipids is crucial for the normal functioning

of multicellular organisms. Glycoconjugates are involved in several central

cellular tasks, such as protein folding and sorting, protein quality control,

vesicular trafficking, cell signalling, immunological defence, cell motility and

adhesion. Altered glycosylation often leads to disease with difficult functional

disorders and malformations in cells and organs. Congenital disorders of

glycosylation (CDG), cystic fibrosis (CF) and cancer are perhaps the best known

examples of the glycosylation diseases. CDGs and CF are heritable, but cancer,

instead, can develop from accumulating mutations in the genome during an

individual´s life span.

Both heritable and non-heritable glycosylation diseases can originate from

single mutations in genes encoding glycosylation enzymes, misregulated

expression of glycosylation enzymes, or from physiological and functional

alterations in the organelles involved in glycosylation. In this work we focused on

the effects of pH alterations in the Golgi apparatus, which is the central organelle

in the secretory pathway and the site of most cellular glycosylation processes.

Our studies showed that Golgi pH is generally higher and more variable in

abnormally glycosylating, i.e. strongly T-antigen (Gal-β1,3-GalNAc-ser/thr)

expressing cancer cells, than in non-T-antigen expressing cells despite similarities

or differences in genetic background. We also confirmed that the Golgi pH

alterations detected in cancer cells have the potential to induce glycosylation

changes. A mere 0.2 pH unit increase in Golgi pH was able to induce T-antigen

expression and inhibit terminal N-glycosylation. The inhibition was found to

correlate with mislocalization of the corresponding glycosyltransferases.

In this work we also studied potential factors that can promote Golgi pH

misregulation in health and disease. We found that cultured cancer cells, despite

variation and elevation in Golgi pH, can be in a good energy status and fully

capable of acidifying Golgi lumen under the normal Golgi pH. Moreover, we

introduce here a novel candidate that can participate in Golgi pH regulation. The

candidate is a Golgi localized Cl-/HCO3- exchanger, anion exchanger 2a (AE2a),

which we suggest participates in Golgi pH regulation by altering luminal

buffering capacity. Participation of AE2a in Golgi pH regulation also provides a

logical explanation for the uncharacterised Golgi resident proton leakage pathway.

16

17

2 Review of the literature

2.1 The Golgi apparatus

The true nature of the Golgi apparatus has remained enigmatic, although the

earliest observations are already 110 years old. The Golgi apparatus was initially

described as an odd black staining in nerve cells by an Italian scientist, Camillo

Golgi, in 1898 (Golgi 1898). Since then the understanding of the Golgi apparatus

as a crucial component of eukaryotic cells has developed hand in hand with the

progress of biochemical techniques and microscopy (Farquhar & Palade 1981,

Farquhar & Palade 1998). Today it is known that the mildly acidic Golgi

apparatus, at the interface of the secretory pathway and the endosomal system, is

a site of several important post-synthetic modifications, such as glycosylation and

sulfation of proteins and lipids (Varki 1998, Brockhausen 1999, Wopereis et al. 2006, Maccioni 2007) and lipidation of proteins (Nadolski & Linder 2007). The

Golgi apparatus has also major roles in lipid biosynthesis (De Matteis &

D'Angelo 2007), protein and lipid sorting as well as cell polarization (Griffiths &

Simons 1986, Rodriguez-Boulan & Müsch 2005, De Matteis & D'Angelo 2007,

De Matteis & Luini 2008). However, many central mechanisms of the Golgi

function have remained unclear and sometimes it is still debated whether the

Golgi is a distinct organelle or just an extension of the endoplasmic reticulum

(ER).

The secretory pathway (Fig. 1), which is composed of the ER, the Golgi

apparatus and secretory vesicles, specializes in biosynthesis and transport of

proteins and lipids. Together with the endosomal system, which is mainly

specialized in endocytic processess and lysosomal degradation, the secretory

pathway forms a constantly fluctuating, polymorphic intracellular membrane

formation that has no static structures (Rodriguez-Boulan & Müsch 2005,

Bonifacino & Rojas 2006, De Matteis & Luini 2008). Therefore, it is often

dubious to define exact borders for the different organelles. According to the

prevailing opinion, however, the Golgi apparatus begins from or after the ER-

Golgi intermediate compartment (ERGIC) 1 (Saraste & Kuismanen 1984,

1 ERGIC, also known as a pre-Golgi intermediate compartment (IC), is a pleiomorphic membrane structure between the ER and the Golgi, where proteins are arrested at 15 ºC (Saraste & Kuismanen 1984)

18

Appenzeller-Herzog & Hauri 2006), and can be divided into the cis-Golgi

network (CGN), the cis-Golgi, the medial-Golgi, the trans-Golgi and the trans-

Golgi network (TGN), on the basis of different proteins and functions related to

them. Under a two dimensional electron microscope the Golgi of a higher

eukaryote cell looks like a stack of flattened concave sacs with plate-like centers

and dilated rims (Farquhar & Palade 1981, Farquhar & Palade 1998). These sacs

are named cisternae. Their number is cell type specific, but generally varies from

3 to 30 (Mollenhauer & Morre 1991). The convex side of the cisternae in the

stack is towards the cis-Golgi and the ER and the concave side is towards the

trans-Golgi and the endosomal system. Both ends of the stack structure, the CGN

and the TGN, are formed by tubular networks, whose lateral regions are occupied

by budding and fusing vesicles (Farquhar & Palade 1981, Farquhar & Palade

1998, Kepes et al. 2005). A three dimensional high resolution view on the Golgi

reveals additional details. The flattened sacs are actually pleiomorphic hollow

discs with tubular extrusions that sometimes connect neighbouring cisternae in

the same or adjacent stacks (Tanaka et al. 1986, Cooper et al. 1990, Rambourg &

Clermont 1990, Weidman 1995, Mironov et al. 1997, Griffiths 2000). A typical

higher eukaryote cell contains multiple Golgi stacks that generally organize into a

juxtanuclear ring around the microtubulus organizing center (MTOC) (Rogalski

& Singer 1984, Rios & Bornens 2003) or perinuclearly in association with the

nuclear envelope (Tassin et al. 1985). Despite the relatively high number of the

stacks, their summarized total volume is estimated to comprise only ~3–4% of the

total cellular volume in a normal mammalian cell (Sturgess & De la Iglesia 1972).

Due to the dynamic nature and the exposure to heavy transit of newly

synthesized and recycled proteins and lipids, the maintenance of the Golgi

architecture requires constant remodelling. The mechanisms of the remodelling

are not fully understood, but it is known that cytoskeleton and cytosolic Golgi

specific matrix proteins provide a scaffold for the cisternae, so that they can take

their unique concave shapes (Chapter 2.2). It is also known that in addition to the

scaffold, the presence and correct functioning of many luminal and

transmembrane proteins are required for the formation and the maintenance of the

normal organelle morphology. This in turn requires strictly controlled protein and

lipid trafficking within the Golgi (Chapters 2.3 & 2.4) and precise luminal pH

regulation to ensure optimal conditions for correct functioning of Golgi proteins

(Chapter 2.5). Once all these factors, the remodelling of the scaffold, the control

of the protein and lipid trafficking and the regulation of luminal pH, are in order,

the Golgi apparatus can perform its crucial functions without perturbation.

19

Fig. 1. The secretory pathway and the endosomal system form a complex fluctuating

entity. The Golgi apparatus, located in the heart of the secretory pathway and at the

interface of the endosomal system, is composed of the cis-Golgi network (CGN), the

Golgi stack and the trans-Golgi network (TGN). The Golgi stack can be further divided

to the cis-, medial- and trans-Golgi.

2.2 Golgi membrane skeleton

The Golgi membrane skeleton, which acts as a scaffold for the formation of the

Golgi apparatus, is composed of microtubules (Rios & Bornens 2003, Vaughan

20

2005), actin filaments (Fath & Burgess 1993, Valderrama et al. 2000), spectrins

(De Matteis & Morrow 2000, Beck 2005), golgins (Short et al. 2005) and their

associated proteins. These proteins and some of their interactions are briefly

introduced in the following chapters.

2.2.1 Microtubules

Microtubules and microtubule associated proteins determine the localization of

the Golgi stack (Rios & Bornens 2003, Vaughan 2005). The Golgi is typically

localized around the MTOC at the juxtanuclear region (Rogalski & Singer 1984,

Rios & Bornens 2003) or perinuclearly in association with the nuclear envelope

(Tassin et al. 1985) by a mechanism that is not fully understood, but is thought to

involve activities and regulation of several different microtubule binding motor

and non-motor proteins and regulation of microtubule elongation (Allan et al. 2002, Rios & Bornens 2003, Vaughan 2005). According to one simplified model

the correct positioning of the Golgi is defined in a continuous “tug-of-war”

between plus-end and minus-end directed motor proteins (Allan et al. 2002).

Microtubules have also crucial role in ER-to-Golgi transport (Saraste &

Svensson 1991, Presley et al. 1997, Rahkila et al. 1997, Murshid & Presley 2004).

Microtubules form tracks between the ER and the Golgi along which vesicles and

other cargo are transported by various motor proteins. Depolymerization of these

tracks, for instance, with nocodazole2, abolishes ER-to-Golgi transport and leads

to accumulation of Golgi proteins into the ER or to the proximity of the ER exit

sites (Presley et al. 1997, Pecot & Malhotra 2006). Perhaps the best known

ER/Golgi associated motor protein that moves along microtubules is a

temperature sensitive3, minus-end directed (i.e. ER-to-Golgi directed in this case)

and ATP powered cytosolic dynein (Paschal & Vallee 1987, Vaughan & Vallee

1995, Vaughan 2005). Recruitment of dynein to the microtubule bound ER/Golgi

membranes or other cargo is a highly regulated process that involves at least

functions of a large multi-subunit protein complex, dynactin (Gill et al. 1991,

Schroer & Sheetz 1991, Karki & Holzbaur 1995, Vaughan & Vallee 1995), and

cytoplasmic linker protein 170 (CLIP-170) (Rickard & Kreis 1990, Komarova et

2 an anti-tumoral synthetic drug (Atassi & Tagnon 1975) 3 Dynein mediated transport dercreases dramatically when temperture is decreased from +37 to +20 ºC (Vaughan 2005).

21

al. 2002). After succesful dynein recruitment, the tethered cargo is rapidly

transported to the microtubule minus-end (Vaughan et al. 1999, Vaughan 2005).

2.2.2 Actin and spectrin

Another part of the Golgi scaffold is formed by a network of actin and spectrin

and their associated proteins (De Matteis & Morrow 2000, Beck 2005, Egea et al. 2006). The exact structure of the Golgi associated actin and spectrin network is

unclear, but it is thought to be analogous to the scaffold that stabilizes the plasma

membrane in erythrocytes (De Matteis & Morrow 2000, Beck 2005). The

erythroid membrane scaffold is composed of heterodimeric αβ-spectrin bundles

that are linked to each other by short actin filaments and various linker proteins to

form a two dimensional elastic network (De Matteis & Morrow 2000). The

network is attached to the cytosolic side of the plasmamembrane by ankyrin

(Bennett & Stenbuck 1979, Bennett 1992), adducin (Gardner & Bennett 1987,

Mische et al. 1987), protein 4.1 (Pasternack et al. 1985, Correas et al. 1986) and

protein 4.2 (Korsgren & Cohen 1986) adaptor proteins that are able to interact

with a wide group of membrane spanning proteins (reviewed in De Matteis &

Morrow 2000). Spectrin can also interact directly with membranes by C-terminal

MAD2/PH domain that binds phosphoinositides (e.g. PtdIns(4,5)P2s) (Hyvönen et al. 1995), but the role of some central membrane spanning proteins in the

membrane anchoring is indispensable. For instance, a deletion of abundantly

expressed anion exchanger 1 (AE1) gene yields highly fragile erythrocytes (Peters et al. 1996, Southgate et al. 1996).

The analogy to plasmalemmal actin-spectrin network is based on the finding

that Golgi contains, or is associated with, several proteins that can participate in

the formation of a Golgi-oriented actin-spectrin network (De Matteis & Morrow

2000, Beck 2005). The Golgi has been shown to contain or associate with actin

(Fath & Burgess 1993, Valderrama et al. 2000), βIII spectrin (Beck et al. 1994,

Stankewich et al. 1998), two different ankyrins, AnkG119 (Devarajan et al. 1996)

and Ank195 (Beck et al. 1997), centractin and adducin (Holleran et al. 1996). A

potential Golgi localized counter part also exists for protein 4.1 (Kang et al. 2009).

The achoring of the actin-spectrin network to the Golgi membranes may also

be analogous to the anchoring of the erythroid actin-spectrin network. Related to

this, it has been shown that anion exchanger 2a (AE2a) localizes to the Golgi only

in the presence of Ank195 expression, indicating a potential interaction between

AE2a and Ank195 (Holappa & Kellokumpu 2003). Additionally, it has been

22

found that AE2a is important for the structural integrity of the Golgi (Holappa et al. 2004). Therefore, it is very likely that AE2a has a similar structural role in the

Golgi to AE1 in the erythroid plasma membrane as a structural stabilizer (Peters et al. 1996, Southgate et al. 1996).

The Golgi associated actin-spectrin network is not a separate entity, but is

linked to microtubules and other Golgi matrix components via several proteins

that are able to interact directly or indirectly with both the actin-spectrin network

and microtubules. One such protein recently identified WHAMM (WASP 4

homolog associated with actin, membranes and microtubules), whose N-terminal

domain binds to ERGIC or Golgi membranes and its central coiled coil region to

microtubules. The C-terminal WCA segment, in turn, has been shown to stimulate

actin polymerization (Campellone et al. 2008).

2.2.3 Golgins and small GTPases

The third part of the scaffold is formed by golgins and their interacting proteins

(reviewed in Short et al. 2005) that also participate in vesicle tethering (Chapter

2.3.4). The golgins are characterized by large coiled-coil regions that form an

extended rod-like structure (Burkhard et al. 2001, Short et al. 2005). Some of the

golgins, such as giantin (Linstedt & Hauri 1993) and golgin-84 (Diao et al. 2003),

are transmembrane proteins, but most of them are linked to membranes via

adaptor proteins, Golgi reassembly stacking proteins (GRASPs) (Barr et al. 1997,

Shorter et al. 1999) or small GTP activated GTPases from Rab, ARL or ARF

families (Short et al. 2005). The mechanisms by which the transmembrane

golgins, the GRAPSs or the Golgi localizing GTPases identify the correct

membrane domains are poorly understood, but the recognition is thought to

involve interactions with members of the p24 cargo receptor family and lipid

membranes (Barr et al. 2001, Pfeffer 2003). Once localized correctly the golgins

and their interacting proteins can support the Golgi structure independently by

interacting with each other and by forming intercisternal bridges and tethering

oligomers like p115, GM130 and giantin (Cluett & Brown 1992, Sönnichsen et al. 1998), or by interacting with the other components of the Golgi scaffold, similar

to GMAP-210, which has been shown to participate in the anchoring of the cis-

Golgi membranes to microtubules (Rios et al. 2004).

4 Wiskott-Aldrich syndrome protein

23

2.2.4 Golgi structure in mitosis

At the onset of mitosis in animal cells, the Golgi fragments into tubulo-vesicular

structures, which are divided evenly between daughter cells during cytokinesis

(Lucocq et al. 1987, Pypaert et al. 1993, Tang et al. 2008). The mitotic

fragmentation of the Golgi has been found to coincide with phosphorylation of

βIII-spectrin (Siddhanta et al. 2003), several golgins 5 and their interacting

proteins6 (Bailly et al. 1991, Lowe et al. 1998, Lowe et al. 2000, Diao et al. 2003,

Preisinger et al. 2005). The phosphorylations of the Golgi matrix proteins disrupt

their interactions with other Golgi matrix proteins and hence lead to the

dissassembly of the matrix and Golgi fragmentation. Conversely, post-mitotic

Golgi assembly is often associated with dephosphorylation of golgins and their

associated proteins (Lowe et al. 2000, Wang et al. 2003). However, there are also

some golgins that are phosphorylated and dephosphorylated differently.

Phosphorylation of p115, for instance, is required for post-mitotic Golgi

reassembly and not for mitotic disassembly (Dirac-Svejstrup et al. 2000).

Recent in vitro studies with isolated Golgi fractions revealed that the minimal

cytosolic machinery which is able to regulate Golgi fragmentation comprises only

two mitotic kinases, cdc2 (cell division cycle protein 2) and plk (polo-like kinase),

and components of COPI budding machinery, ADP ribosylation factor 1 (ARF1)

and the coatomer complex (Tang et al. 2008). Cdc2 and plk mediate Golgi

unstacking by phosphorylating various Golgi associated proteins (Lowe et al. 1998, Lowe et al. 2000, Wang et al. 2003), whereas ARF1 and components of the

COPI budding machinery mediate vesiculation (Xiang et al. 2007). Golgi

reassembly in the separated Golgi fractions, instead, requires addition of one of

the two soluble AAA ATPases, N-ethylmaleimide sensitive factor (NSF) or p96,

their associated adaptor proteins and protein phosphatase 2A (PP2A) (Tang et al. 2008). NSF with soluble adaptor proteins, α/γ-SNAPs and p115 (Shorter &

Warren 1999, Tang et al. 2008), and p96 with soluble adaptor proteins, p47 and

p115 (Kondo et al. 1997, Shorter & Warren 1999, Tang et al. 2008), are able to

induce fusion of Golgi membranes, whereas PP2A is required for the Golgi

restacking process, which it mediates by dephosphorylating central Golgi stacking

proteins (Tang et al. 2008). The upstream regulation of cytosolic factors

5 e.g. GM130 (Lowe et al. 1998, Lowe et al. 2000) and golgin-84 (Diao et al. 2003) 6 e.g. several Rabs (Bailly et al. 1991) and GRASP65 (Preisinger et al. 2005)

24

mediating Golgi assembly and disassembly during the cell cycle is mostly unclear.

Only the regulation of cdc2 is relatively well established.

2.3 Membrane trafficking within the Golgi region

Due to its central localization, the Golgi apparatus is constantly exchanging

material with the ER and endosomes. This exchange utilizes different kinds of

membrane formations and many aspects of this are still unclear. Some of the best

known mechanisms and hypotheses of membrane trafficking are introduced in

next chapters.

2.3.1 ER-to-Golgi transport

Anterograde transport of newly synthesized proteins starts from coat protein

complex II (COPII) coated exit domains in the ribosomal (rER) or ribosome-free

transitional ER (tER)(Orci et al. 1991, Sesso et al. 1994, Shaywitz et al. 1995,

Bannykh et al. 1996, Tang et al. 1997, Palmer & Stephens 2004, Kirk & Ward

2007). From these domains proteins are transported via vesicles (Farquhar &

Palade 1981, Orci et al. 1991, Rothman & Wieland 1996 and refs. therein) or

continuous tubules (Sesso et al. 1994, Bannykh et al. 1996, Mironov et al. 2003)

to the vesicular-tubular clusters (VTCs) that form the ER-Golgi intermediate

compartment (ERGIC). It is unclear whether COPII coat participates directly in

the biogenesis of vesicles and tubules (Lee & Miller 2007 and refs. therein) or the

generation and maintenance of the exit domains (Mironov et al. 2003, Palmer &

Stephens 2004). The crucial role of highly conserved COPII coating mechanism

for ER export, however, is widely agreed. Formation of COPII coat is initiated by a replacement of GDP in cytosolic

Sar1p-GDP with GTP by an ER bound type II membrane protein, Sec12p

(Barlowe & Schekman 1993). Sar1p-GTP then interacts with undefined ER

membrane components (Barlowe & Schekman 1993) and recruits cytosolic

heterodimer, Sec23/24p, to the proximity of the membrane (Barlowe et al. 1994).

The binding of Sec23/24p to Sar1p-GTP is instantly followed by a binding of

another cytosolic complex, heterotetrameric Sec13/31p (Antonny et al. 2001),

which forms the outer layer of the coat (Barlowe et al. 1994, Lederkremer et al. 2001, Matsuoka et al. 2001). The COPII coat is rather unstable and the

components undergo a constant disassembly-assembly cycle (Stephens et al. 2000,

Ward et al. 2001, Stephens 2003, Watson et al. 2005). The factors affecting the

25

coat stability are poorly known, but the disassembly of the coat is initiated with

Sar1p mediated hydrolysis of bound GTP, which is primarily stimulated by

Sec23p (Yoshihisa et al. 1993). In addition, the disassembly rate apparently has a

regulatory role in budding and cargo exit, disassembly of the coat is a common

prerequisite for the fusion of coated membrane structures to acceptor membranes

(Bonifacino & Glick 2004).

Before exiting the ER, proteins undergo selection and quality control in the

COPII coated tER, which is the last site of quality control from where improperly

folded proteins are still returned to the ER (Mezzacasa & Helenius 2002).

Therefore, the machinery participating in cargo selection must have a high fidelity.

Thus far the roles of Sec24p in selection of membrane spanning cargo proteins

(Kappeler et al. 1997, Nishimura & Balch 1997, Dominguez et al. 1998, Miller et al. 2002, Barlowe 2003) and p24 in selection of soluble cargo are recognized

(Barlowe 2003). When proteins have passed the quality control and selection,

they enter the VTCs (or the ERGIC). Dynein/dynacting complexes move the

VTCs along the microtubules to the proximity of the cis-Golgi network and the

cis-Golgi (Saraste & Svensson 1991, Mizuno & Singer 1994, Presley et al. 1997,

Murshid & Presley 2004), where the VTCs fuse with existing Golgi cisternae or

generate new ones (Chapter 2.3.2: Intra-Golgi transport). Transport of cargo from

VTCs to the Golgi may also occur via coat protein complex I (COPI) vesicles,

although their anterograde traffic may only be involved in recycling of the

machinery required for retrograde transport (Scales et al. 1997, Shima et al. 1999,

Stephens et al. 2000, Appenzeller-Herzog & Hauri 2006 and refs. therein), since

the main function of COPI vesicles is generally thought to be to mediate

retrograde transport from the Golgi and VTCs to the ER (Orci et al. 1997,

Bethune et al. 2006).

In addition to COPII and COPI mediated transport, also COP-free tubular

extensions between the ER and the Golgi may have role in ER-to-Golgi transport

(Mironov et al. 2003, Vivero-Salmeron et al. 2008).

2.3.2 Intra-Golgi transport

Vesicle transport model

The vesicle transport model, which is also known as the vesicular shuttle model

or the stable compartments model, postulates that maturing cargo in the Golgi is

26

transported via vesicles between stable compartments (Palade 1975, Rothman &

Wieland 1996, Orci et al. 1997). In this model anterograde traffic comprises

mainly of cargo transport and retrograde transport from recycling of the proteins

required for the anterograde cargo transport. Both anterograde and retrograde

transport are mediated by COPI vesicles (reviewed in Bethune et al. 2006). The

mechanism that distinguishes these is unclear, but at least three different kinds of

coat compositions (Wegmann et al. 2004) and a variety of potential “address

molecules”, such as golgins, COGs (conserved oligomeric complexes) and v-

SNAREs (soluble NSF attachment protein receptors), have been recognized thus

far (reviewed in Bethune et al. 2006, Pfeffer 2007). The vesicle transport between

cisternae is thought to occur via crosscurrent or countercurrent flow (Rothman &

Wieland 1996). In the crosscurrent flow, anterograde transport of vesicles

proceeds from previous cisternae to following ones, while retrograde transport to

the ER occurs directly from any stage of the Golgi. In the countercurrent flow

both antero- and retrograde transport occur between parallel cisternae.

Existence of vesicle concentrate at the proximity of the rims of the cisternae

(Palade 1975, Ladinsky et al. 1999) and evidence of cargo proteins (Orci et al. 1997, Nickel et al. 1998, Pepperkok et al. 2000, Lanoix et al. 2001) and Golgi

enzymes within the vesicles (Lanoix et al. 2001, Martinez-Menarguez et al. 2001,

Malsam et al. 2005) supports the vesicle transport model. However, a lot of

contradictory data exists. Several tested cargo proteins and enzymes are actually

absent from the separated vesicle structures (reviewed in Mironov et al. 2005).

The transport of large cargo e.g. procollagen whose aggregates have length of

~300 nm in vesicles having diameter of 50–80 nm is problematic (Bonfanti et al. 1998). The dilemma of large cargo has been, however, explained by competeting

intra-Golgi transport models that are introduced in the next chapters.

Formation of COPI coat is analogous to formation of COPII and clathrin

coats. First a small cytosolic GTPase, Arf1, bound to GDP is recruited to the

Golgi membrane by p23/p24 complex (Gommel et al. 2001, Majoul et al. 2001,

Contreras et al. 2004). At the vicinity of the membrane GBf1 (Zhao et al. 2002,

Niu et al. 2005), or another Sec7 domain containing guanine nucleotide exchange

factor (GEF), induces exchange of bound GDP to GTP (Chardin et al. 1996,

Bethune et al. 2006). Arf1-GTP dissociates from p23, and Arf1-GTP/p24 forms a

priming complex that recruits cytosolic coat components i.e. coatomers as a

heterotetramer (Hara-Kuge et al. 1994, Bethune et al. 2006). Nine coatomer

protein components: α, β, γ1, δ, β´, ε, ζ1 ,γ2 and ζ2 (revieved in Bethune et al. 2006) are known to compose at least three different types of heterotetramers

27

(Wegmann et al. 2004). The coat is stable until Arf1 hyrolyses the bound GTP

(Tanigawa et al. 1993). The hydrolysis is catalyzed by one of the ArfGAP

proteins: ArfGAP1 (Cukierman et al. 1995), ArfGAP2 (Randazzo 1997) or

ArfGAP3 (Liu et al. 2005). ArfGAPs are curvature sensitive proteins that achieve

maximal stimulation when the membrane they bound is curved in a way which is

correlative to COPI vesicles (Bigay et al. 2003, Bigay et al. 2005). However, it is

not clear whether disassembly occurs after or during the budding. The

disassembly of the COPI coat, similarly to other coat structures, is required for

the fusion of the vesicle with acceptor membranes.

Cisternal maturation model

According to the cisternal maturation model, cargo proteins are transported

forward and simultaneously processed within maturing cisternae. The maturation,

from cis- to medial- and then to trans-cisternae and the TGN, is driven by COPI

vesicles that retrieve Golgi enzymes from later and more “matured” cisternae to

earlier ones (Beams & Kessel 1968, Franke et al. 1971, Saraste & Kuismanen

1984, Bannykh & Balch 1997, Mironov et al. 1997, Glick & Malhotra 1998,

Pelham 1998, Martinez-Menarguez et al. 2001).

The observation that large cargo, such as algae cell surface scales, which are

~150 – 1600 nm in diameter (Becker & Melkonian 1996), and procollagen, which

forms ~300 nm long stable aggregates (Brodsky & Ramshaw 1997), are excluded

from COPI vesicle structures and remain within cisternea during their traverse

through the Golgi stack (Becker & Melkonian 1996, Bonfanti et al. 1998),

strongly supports the existence of the maturation model. In addition, the

maturation of the Golgi cisternae has recently been demonstrated in at least three

distinct studies with live microscopy of yeast cells (Wooding & Pelham 1998,

Losev et al. 2006, Matsuura-Tokita et al. 2006). These studies show that in the

distribution of the cisternae, resident membrane proteins do change over the time

from cis- to trans-Golgi proteins. The maturation model alone, however, is unable

to provide an explanation for different secretion rates of different cargo proteins

detected in some (Bonfanti et al. 1998) but not all experiments (Mironov et al. 2001).

28

Tubular model

According to the third and the newest model proteins and other cargo within the

Golgi are transported via tubular connections between adjacent cisternae

(Weidman 1995, Mironov et al. 1997, Griffiths 2000). The existence of the

connections between cisternae was described already in the 1980s (Tanaka et al. 1986, Cooper et al. 1990, Rambourg & Clermont 1990), but at that time they

were often regarded only as fixation artefacts. The tubular model provides a

potential explanation for anterograde protein transport within the Golgi, which

cannot be fully explained by COPI vesicles. Transport of large cargo (e.g.

procollagen > 300 nm) via vesicles is problematic, and according to various

reports enriched COPI vesicles also lack most of tested cargo proteins as well as

Golgi resident enzymes (reviewed in Mironov et al. 2005). Thus, the role of COPI

vesicles as cargo or enzyme transporters is rather questionable. The tubular model,

instead, allows a rapid flow of large, but also smaller cargo or enzymes between

cisternae (Mironov et al. 2003). According to the recent findings tubule mediated

retrograde transport may also be possible (Vivero-Salmeron et al. 2008).

The mechanism of tubular tranportation is unclear, but it is thought to occur

via transient tubules that undergo constant fusions and scissions at the pheriphery

of cisternae (Weidman 1995, Mironov et al. 1997). Another view suggests a “gut”

like organization (Griffiths 2000). According to this, the whole secretory pathway

from the ER to the TGN is organized as a single continuum with two gates, one

being between the ER and the cis-Golgi and another between the trans-Golgi and

the TGN. The empowering of the flow is unclear, but in both models it can be

driven, for instance, by a gradient change of lipids, salts or pH (Mironov et al. 2005) or by constrictions of tubules that could be driven by local lipid metabolism

(Shemesh et al. 2003) or by unknown matrix proteins.

Combined model

The three introduced intra-Golgi transport models need not be mutually exclusive.

Since a lot of supporting data for each of them has been collected, the model that

they may coexist or be cell type dependent (Mironov et al. 1997, Glick &

Malhotra 1998) is getting more and more popular. The model of three different

mechanisms acting in conjunction is often called the combined model.

29

2.3.3 Post-Golgi transport

The trans-Golgi network (TGN) functions as a traffic hub in the crossroads of the

endocytic and the exocytic pathways. Proteins and lipids received from

anterograde and retrograde direction are sorted there to the apical and basolateral

plasma membrane, early/sorting endosomes, recycling endosomes, late

endosomes, the Golgi stack, secretory granules, and to other specialized

organelles in some specialized cells (Griffiths & Simons 1986, Rodriguez-Boulan

& Müsch 2005, Bonifacino & Rojas 2006, De Matteis & Luini 2008).

Due to the multiple destinations on the hand the complexity of the post-Golgi

transport is stunning. Several pathways exist to both anterograde and retrograde

directions, but only some components of the transport machineries are shared

(Bonifacino & Rojas 2006). The diversity of the pathways involving the TGN

also applies to membrane carrier structures, which can exhibit vesicular, tubular,

or pleiomorphic shapes, and can be derived from the TGN, endosomal

membranes and the plasma membrane (Rodriguez-Boulan & Müsch 2005,

Bonifacino & Rojas 2006, De Matteis & Luini 2008). Additionally, some proteins

are able to follow multiple pathways, whereas some show more limited mobility

and are strictly dependent on one specific pathway (Bonifacino & Rojas 2006).

The best understood post-Golgi transport mechanism in mammalian cells is

that involved in the recycling of cation dependent and cation independent

mannose-6-phosphate 7 receptors (CD- and CI-MPRs) (Ghosh et al. 2003).

Enzymes tagged with mannose-6-phosphate, usually acid hydrolases, associate

with MPRs in the late-Golgi. The resulting complexes are packed to carrier

vesicles that fuse with early or late endosomes, where the enzyme-MPR

complexes dissociate due to acidic pH (Gonzalez-Noriega et al. 1980). After

dissociation MPRs are recycled back to the TGN and lysosomal enzymes are

destined for lysosomes (Bonifacino & Rojas 2006).

The transport cycle of MPRs between the TGN and endosomes is not fully

understood, but it has been shown to involve at least clathrin, which forms coat

structures both in the TGN and endosomes (Saint-Pol et al. 2004, Bonifacino &

Rojas 2006). The transport of MPRs from the TGN to endosomes involves also

Golgi associated – gamma adaptin ear containing – ARF binding proteins (GGAs)

(Doray et al. 2002) and heteromeric adaptor protein 1 (AP1) (Wan et al. 1998)

7 P-Man8GlcNAc2-Asn, where P denotes to inorganic phosphate, i.e. H2PO4

-.

30

that interact together and pack MPRs and the cargo bound to them into AP1 and

clathrin containing vesicles (Doray et al. 2002). The recycling of MPRs from

endosomes to the TGN requires the action of phosphofurin acidic-cluster-sorting

protein 1 (PACS1) (Wan et al. 1998), which recruits CI-MPRs and other cargo8

into clathrin coated recycling vesicles by interacting with AP1 (Crump et al. 2001). A multiprotein complex named the retromer (Bonifacino & Rojas 2006,

Collins 2008), a small GTPase named rab9 (Lombardi et al. 1993, Pfeffer 2003)

and a tail interacting protein of 47 kDa (TIP47)(Diaz & Pfeffer 1998) have been

shown to participate in the selection of retrograde cargo in endosomes, although,

it remains unclear whether they act cooperatively with PACS1 and AP1 or not.

The retrograde transport of MPRs is also dependent on epsin related clathrin

adaptor (epsinR) (Saint-Pol et al. 2004), which drives membrane curvature for

vesicle budding (Ford et al. 2002, Saint-Pol et al. 2004), and dynamin, a large

spiral-shaped GTPase oligomer, which generates a mechanical force for the

tubulation and vesicle scission on the surface of endosomes (Nicoziani et al. 2000,

Praefcke & McMahon 2004).

2.3.4 Fusion of membrane structures

Most of the mechanisms mediating membrane fusion are elucidated in studies

with vesicles or vesicular compartments. These mechanisms may, however, be

applicable also to fusion events of other kind of membrane structures, such as

tubular extrusions or VTCs (Chapters 2.3.1 & 2.3.2).

In the vesicle transport cycle, budding and scission from the donor membrane

are followed by uncoating, which is triggered by hydrolysis of GTPase bound

GTP within the vesicle coat structures (Tanigawa et al. 1993, Yoshihisa et al. 1993, Bonifacino & Glick 2004). Detached coat proteins are then recruited to

another cycle of vesicle formation and the uncoated vesicle is transported to the

proximity of an acceptor organelle, where the vesicle becomes tethered to the

target membrane by membrane associated Rab GTPases and long, generally

coiled coil shaped tethering factors9 (Bonifacino & Glick 2004, Sztul & Lupashin

2006).

8 e.g. furin 9 e.g. cis-Golgi resident tethering factor, giantin: MW ~400 kDa, length 250 nm (Linstedt et al. 1995)

31

The tethering to the target membrane is not obligatory for vesicle fusion

(Weber et al. 1998, Hu et al. 2003), but it has been shown to enhance vesicle

docking, which is mediated by a set of membrane specific SNAREs (Söllner et al. 1993, Rothman 1994). In the docking, one R-SNARE and three Q-SNAREs form

a stable four α-helix bundle, a trans-SNARE complex (Fasshauer et al. 1998,

Sutton et al. 1998) that initiates the fusion of the vesicle and the acceptor

membrane (Hanson et al. 1997, Weber et al. 1998, Chen & Scheller 2001). The

R-SNARE, which provides arginine to the highly conserved ionic 0 layer at the

center of the trans-SNARE complex, is usually provided for the complex by the

trafficking vesicle. A heterodimeric or trimeric group of Q-SNAREs that provide

glutamines to the ionic 0 layer, in turn, are often provided for the complex by the

acceptor membrane (Fasshauer et al. 1998). After fusion the SNARE complex

adopts a cis-conformation, which is required for dissociation. The dissociation of

the cis-SNARE complex is initiated by subsequent binding of soluble NSF

association protein (α-SNAP) (Clary et al. 1990) and NSF (Glick & Rothman

1987, Bonifacino & Glick 2004) that dissociates the cis-complex by hydrolysing

ATP (Mayer et al. 1996). The mechanism of dissociation is not clear, but ATP

hydrolysis is thought to generate the energy for NSF to untwist the four-helix

bundle (May et al. 1999, Yu et al. 1999). Recently, it has also been suggested that

the ionic 0 layer, which is composed of one arginine provided by an R-SNARE

and three glutamines provided by Q-SNAREs, may facilitate the dissociation

(Scales et al. 2001). After the dissociation of the cis-SNARE complex, SNAREs

associated with trafficking vesicles are kept inactive and recycled back to donor

membranes for another cycle (Bonifacino & Glick 2004).

2.4 Protein sorting in the Golgi

Sorting of proteins and lipids in the secretory pathway reaches a distinctively high

level of complexity and sophistication at the Golgi, where the newly synthesized

or recycled proteins and lipids are directed to multiple destinations, such as the

ER, different kind of endosomes, the apical and basolateral plasma membranes,

secretory granules and other cell type dependent specialized organelles (reviewed

in Griffiths & Simons 1986, Rodriguez-Boulan & Müsch 2005, Bonifacino &

Rojas 2006, De Matteis & Luini 2008). Additionally, some sorting occurs also

within the Golgi stack between cis-, medial- and trans-cisternae (Füllekrug &

Nilsson 1998, Paroutis et al. 2004, Pfeffer 2007). Succesful sorting is a

prerequisite for normal function of the cell, but also for many Golgi resident

32

processes. For instance, a minor localization change of a single glycosylation

enzyme from the medial-Golgi to the trans-Golgi can alter glycosylation output

significantly (Skrincosky et al. 1997). Some of the known sorting mechanisms are

introduced in following chapters.

2.4.1 Golgi-to-ER sorting

Despite of the high fidelity of the sorting in the ER some unfolded or normally

ER resident proteins can occasionally escape to the Golgi. Soluble ER proteins

are returned to the ER on the basis of their C-terminal KDEL (Lys-Asp-Glu-Leu)

signal peptide, which is recognized by two KDEL-receptors, KDELR1 and

KDELR2 (Pelham 1988, Lewis & Pelham 1992a, Lewis & Pelham 1992b, Pfeffer

2007). Many ER transmembrane proteins, including KDEL-receptors, contain a

dilysine retrieval signal KKXX (where X is any aminoacid) at their C-terminal

cytoplasmic tail (Nilsson et al. 1989, Jackson et al. 1990, Jackson et al. 1993).

Most unfolded proteins do not contain any specific signals, but are recognized by

luminal chaperones that carry a KDEL signal (Yamamoto et al. 2001).

The Golgi-to-ER retrieval mechanism for the KDEL-receptor is quite well

understood. Cargo binding triggers an aggregation of KDEL-receptors (Majoul et al. 1998), which may lead to recruitment of protein kinase A to phosphorylate

serine 209 in the cytosolic tail of the KDEL-receptor (Cabrera et al. 2003, Pfeffer

2007). The phosphorylation, in turn, is thought to facilitate the packing of KDEL-

receptors and KDEL-receptor bound cargo into COPI vesicles (Cabrera et al. 2003). The packing process involves functions of Arf1, ARFGAP and members of

p24 cargo receptor family (Majoul et al. 2001).

2.4.2 Sorting within the Golgi stack

Sorting within the Golgi stack between cis-, medial- and trans-cisternae is perhaps

the worst understood sorting process in the secretory pathway. This is partly due

to the uncertainty of the mechanisms involved in the intra-Golgi transport in

general (Chapter 2.3.2) and the fact that no common Golgi retention or targeting

signals have been identified. However, because the localization of many Golgi

proteins is sensitive to Golgi pH dissipation (Zhang et al. 1996, Axelsson et al. 2001, Bachert et al. 2001, Puri et al. 2002, Schaub et al. 2006, Schaub et al. 2008), it has been suggested that the descending pH gradient within the Golgi is

somehow coupled to formation of an overlapping distribution of Golgi enzymes

33

(Axelsson et al. 2001, Paroutis et al. 2004). This may involve, for instance, pH

sensitive oligomerization or kin recognition (Chapter: Kin recognition model).

Another factor that may have a role in intra-Golgi protein sorting is the Golgi

lipid membrane, whose composition changes in cis- to trans-direction. Cis-Golgi

membranes contain less cholesterol and sphingolipids than trans-Golgi

membranes and are thus generally thinner or occupied by fewer cholesterol rich

membrane thickenings, i.e. membrane microdomains10, than trans-Golgi cisternae

(Orci et al. 1981, Munro 2003). This, according to lipid bilayer sorting model

(Chapter: Lipid bilayer sorting model), leads to accumulation of proteins with

shorter transmembrane sequence to the cis-Golgi, where the lipid membrane is

thinner, and proteins with longer transmembrane sequences to the trans-Golgi or

the membrane microdomains, where the lipid membranes are thicker (Munro

2003).

Sorting within the Golgi can also be dependent on cytosolic components such

as Golgi matrix proteins (Barr et al. 1997, Shorter et al. 1999), or specialized

Golgi specific sorting elements such as conserved oligomeric Golgi (COG)

complexes (Chapter: Conserved oligomeric Golgi complex). Several members of

the p24 cargo receptor protein family have been shown to form complexes with

GRASPs that in turn interact directly with a set of various Golgi matrix associated

golgins (Barr et al. 1997, Shorter et al. 1999) and may thus anchor the cargo

receptors to their correct positions. COG complexes, instead, participate either in

retrograde COPI vesicle tethering or intra-Golgi trafficking (Smith & Lupashin

2008).

Oligomerization / Kin recognition model

Originally the oligomerization / kin recognition model postulated that Golgi

glycosyltransferases form homo- or hetero-oligomers that are too bulky to enter

transport vesicles and thus prevent their exit form the Golgi (Machamer 1991,

Nilsson et al. 1993b). To explain the model Machamer suggested that Golgi

membranes or other physiological conditions in the Golgi trigger an aggregation

that causes gathering of glycosyltransferases to bulky oligomers (Machamer

1991). Nilsson and colleagues, instead, suggested that the oligomerization is more

likely dependent on highly specific interactions between glycosylation enzymes

10 an alternative term is lipid raft

34

(Nilsson et al. 1993b). However, since the model of intra-Golgi transport has been

under the influence of new insights (Chapter 2.3.2), it is widely agreed nowadays

that mere bulkiness may not be enough to prevent export from the Golgi. Despite

that, the oligomerization model is not outdated. Many studies have shown that

oligomerization of glycosyltransferases and other Golgi resident proteins

improves their Golgi localization (Swift & Machamer 1991, Weisz et al. 1993,

Nilsson et al. 1994, Teasdale et al. 1994, Locker et al. 1995, Yamaguchi &

Fukuda 1995, Chen et al. 2000, Opat et al. 2001), although the explanations and

exact mechanisms for this are missing. The improved Golgi localization may

simply be related to higher number of available retention or retrieval signals in

complexes compared to single proteins (Opat et al. 2001).

The first direct evidence of oligomerization or kin recognition of

glycosylation enzymes was obtained with two medial-Golgi enzymes,

N-acetylglucosaminyltransferase I (GlcNAc-TI) and mannosidase II (ManII).

GlcNAc-TI tagged with ER retention motif retained also the non-tagged ManII to

the ER (Nilsson et al. 1994). Since then several glycosyltransferases, such as

β1,4-galactosyltransferase I (Teasdale et al. 1994, Yamaguchi & Fukuda 1995),

α2,6-sialyltransferase I (Colley 1997, Chen et al. 2000, Qian et al. 2001, Fenteany

& Colley 2005), β1,6-N-acetylglucosaminyl transferase V (Sasai et al. 2001) and

many others, have been shown to form non-covalent or disulfide bridge based

covalent dimers and oligomers.

Despite of the topological similarity of the Golgi glycosyltransferases the

oligomerization of different transferases is dependent on different domains.

Additionally, there seems to exist a common organizational difference between

oligomerization of medial- and trans-Golgi glycosyltransferases. Medial-Golgi

glycosyltransferases and glycosylation enzymes are generally in high molecular

weight hetero complexes and utilize the stem (Nilsson et al. 1996) or the catalytic

region at the luminal domain for oligomerization (Opat et al. 2000), whereas the

trans-Golgi glycosyltransferases exist mostly as homodimers and monomers

(Opat et al. 2000, Opat et al. 2001, Young 2004) and utilize both the

transmembrane and the luminal domains for dimerization (Colley 1997, Chen et al. 2000, Fenteany & Colley 2005).

Lipid bilayer sorting model

According to the lipid bilayer sorting model transmembrane proteins localize to

organelles, where the thickness of the membrane matches to the length of the

35

transmembrane spanning sequence (Bretscher & Munro 1993, Masibay et al. 1993). The model is based on the following findings. Firstly, the content of

cholesterol, which thickens lipid membranes (Levine & Wilkins 1971, Nezil &

Bloom 1992), increases from the ER to the Golgi and to the plasmamembrane

(Orci et al. 1981). Secondly, the transmembrane spanning sequences of the Golgi

resident proteins are on an average five aminoacids shorter than the

transmembrane spanning sequences of the plasma membrane resident proteins

(Bretscher & Munro 1993, Masibay et al. 1993, Munro 1998). Thirdly, increased

length of the membrane spanning sequences of Golgi resident proteins11 leads to

increased cell surface expression (Bretscher & Munro 1993) and, correspondingly,

decreased length of the membrane spanning sequences of the plasmamembrane

resident proteins12 reduces the cell surface expression of plasmalemmal proteins

(Sivasubramanian & Nayak 1987). Based on these findings the lipid bilayer

sorting model postulates that the transmembrane proteins travel freely on the

membranes until they arrive in a region where the thickness of the membrane

matches the length of their membrane spanning sequence(s). A residence at such a

region is thought to have the lowest potential energy and thus to be the most

favourable place for a transmembrane spanning protein to localize.

The model works at least in biologically passive, synthetic lipid membranes,

where it has been shown that addition of cholesterol is able to regulate insertion

of transmembrane domains (Ren et al. 1997). In live cell systems, however, the

lipid bilayer model has been proven to be insufficient to explain several aspects in

Golgi targeting. For instance, increased length of ST6Gal I membrane spanning

sequence is unable to induce increased surface expression in COS-1 cells (Dahdal

& Colley 1993). In another experiment the GlcNAcTI lacking the transmembrane

and cytoplasmic domains was shown to be retained in oligomers before secretion

(Opat et al. 2000). These kinds of results strongly indicate existence of other,

separately functioning, Golgi retention mechanisms such as the

“oligomerization/kin recognition” model introduced in the previous chapter.

11 β1,4-galactosyltransferase, EC 2.4.1.38; α2,6-sialyltransferase, EC 2.4.99.1 12 Influenza virus neuraminidase, EC 3.2.1.18

36

Conserved oligomeric Golgi complex

The conserved oligomeric Golgi (COG) complex is a peripheral multisubunit

Golgi protein complex that has been shown to participate in maintaining correct

Golgi localization of several type II transmembrane proteins including, for

instance, mannosidase II (ManII), α2,3-sialyltransferase 1 (ST3Gal1),

N-acetylglucosaminyltransferases 1 and 2 (GlcNAcT1 and 2) and β1,4-

galactosyltransferase I (GalT1) (Oka et al. 2004, Wu et al. 2004, Zolov &

Lupashin 2005, Shestakova et al. 2006, Steet & Kornfeld 2006). The exact

function of the complex is not known, but according to the latest data the complex

may act as a tether that connects COPI vesicles with Golgi membranes during

retrograde intra-Golgi or Golgi-to-ER transport (Vasile et al. 2006, Smith &

Lupashin 2008).

The COG complex is composed of eight subunits, which are distributed to A

and B lobes. In yeast the lobe A contains subunits Cog1p-Cog4p and the lobe B

contains subunits Cog5p-Cog8p (reviewed in Smith & Lupashin 2008). The lobes

are connected to each other by Cog1p and Cog8p (Ram et al. 2002, Fotso et al. 2005). Cog1p, additionally, connects the COG complex to Golgi membranes

(Smith & Lupashin 2008). The human COG complex is thought to be analogously

organized (Smith & Lupashin 2008). The crucial role of the COG complex for

Golgi resident glycosylation has been verified in several studies. Preventing COG

complex formation, for instance by a truncating mutation or a depletion of an

essential COG subunit, has been shown to lead to mislocalization of essential

glycosylation enzymes and hence to altered glycosylation and severe human

disease states (Oka et al. 2004, Wu et al. 2004, Spaapen et al. 2005, Foulquier et al. 2006, Shestakova et al. 2006, Steet & Kornfeld 2006).

2.4.3 Sorting to post-Golgi structures

The best characterized sorting mechanisms at the TGN are based on the

recognition of specific peptide motifs in the cytosolic tails of transmembrane

cargo or cargo receptor proteins. Proteins carrying tyrosine based (NPXY or

YXXØ) or dileucine motifs ([DE]XXXL[LI] or DXXLL), for instance, are

directed to endosomes or the basolateral plasma membrane (Bonifacino & Traub

2003, Rodriguez-Boulan & Müsch 2005, Rodriguez-Boulan et al. 2005). These

proteins are recruited to clathrin-coated vesicles in an interplay of different GGAs,

Arf1-GTP and AP1 (Wan et al. 1998, Doray et al. 2002, Rodriguez-Boulan &

37

Müsch 2005, De Matteis & Luini 2008). Soluble cargo can be recruited to

endosome directed vesicles, for instance, on the basis of mannose-6-phosphate

signal glycan, which is recognized by mannose-6-phophate receptor carrying a

DXXLL signal motif on its cytosolic tail (Ghosh et al. 2003).

Another sorting mechanism involves formation of protein complexes or

aggregates (Borgonovo et al. 2006), but the underlaying mechanisms, e.g. how

the oligomerized or aggregated cargo is packed to membrane carrier structures,

has remained unclear. The oligomerization, however, can be an intrinsic property

of the cargo protein (Borgonovo et al. 2006) or be induced by clustering of cargo

receptors (Hannan et al. 1993, Delacour et al. 2007).

Some of the TGN specific sorting signals arise also from post-translational

modifications. O- and N-glycans, for instance, are known to act as recessive

lumenal signals for apical targeting (Yeaman et al. 1997, Benting et al. 1999). The

sorting based on them is thought to be dependent either on specific lectin

receptors such as galectin-3 (Delacour et al. 2007) or proteinaceous interactions

that promote interactions causing oligomerization/aggregation (Potter et al. 2006).

Lipidation of proteins, which includes for instance palmitoylation or attachment

of glycosylphophatidylinositol (GPI) -anchor, enchances the affinity of cargo

proteins to cholesterol and sphingolipid containing membrane microdomains (or

lipid rafts), that might provide another means to direct cargo to the apical plasma

membrane (Simons & van Meer 1988, Schuck & Simons 2004, Greaves &

Chamberlain 2007). Additionally, the sorting of cargo proteins can be modulated

by phosphorylation and ubiquitinylation (Bonifacino & Traub 2003, Hinners &

Tooze 2003, Piper & Luzio 2007).

2.5 Golgi pH regulation

The ability of the Golgi apparatus to function correctly is dependent on a mildly

acidic luminal pH (Paroutis et al. 2004). Perturbation of organelle acidity with

different drugs has been found to impair at least membrane trafficking (Griffiths et al. 1983, Palokangas et al. 1998), protein sorting (Matlin 1986, Caplan et al. 1987, Ellis & Weisz 2006), and glycosylation of proteins and lipids (Kuismanen et al. 1985, Thorens & Vassalli 1986, Gawlitzek et al. 2000, Axelsson et al. 2001,

Campbell et al. 2001a, Kellokumpu et al. 2002).

The luminal pH of the secretory pathway decreases gradually from neutral in

the ER (pH ~7.0) to acidic in the TGN and secretory vesicles (pH 6.0–5.5). The

Golgi lumen pH is normally around 6.7–6.0, being near neutral in the cis-Golgi

38

and 6.3–6.0 in the trans-Golgi and the TGN (Anderson & Pathak 1985, Wu et al. 2000, Paroutis et al. 2004). The decrease of the luminal pH towards the end of the

secretory pathway correlates with simultaneous increase of proton import to the

luminal side and decrease of proton export to the cytosol. The membrane

potential, instead, shows only minor changes and remains modest all the way, due

to relatively high counter ion conductance (Kim et al. 1996, Demaurex et al. 1998,

Farinas & Verkman 1999, Schapiro & Grinstein 2000, Wu et al. 2001). The

absence of counter ion conductance, however, can impair organelle acidification

by preventing proton import (Glickman et al. 1983, Rybak et al. 1997). Therefore,

it has been suggested that the steady state pH in the Golgi and other organelles in

the secretory pathway is dictated by a balance of three components: proton import,

proton export/elimination and counter ion conductance. This balance in each

organelle is maintained by a specific set of regulatory proteins (Kim et al. 1996,

Demaurex et al. 1998, Farinas & Verkman 1999, Grabe & Oster 2001, Wu et al. 2001, Demaurex 2002, Paroutis et al. 2004). These proteins and some potential

canditates involved in the Golgi pH regulation (Fig. 2) are introduced in the

following chapters.

Fig. 2. Potential factors participating in Golgi pH regulation. The putative main factors

are proton import, proton export/elimination and counterion conductance, but also

luminal buffering, membrane trafficking and active cargo proteins are thought to

influence on the Golgi pH homeostasis. A- and C+ represent undefined anions and

cations, respectively.

39

2.5.1 Vacuolar H+ATPase mediated proton import

Vacuolar H+ATPase (V-ATPase, Fig. 3), an ATP powered proton pump, is the

main acidifyer in the Golgi and elsewhere in the secretory pathway (Forgac et al. 1983, Nishi & Forgac 2002, Sun-Wada et al. 2004, Jefferies et al. 2008).

V-ATPase comprises two multi-subunit domains, peripheral V1 and integral V0

(Nishi & Forgac 2002, Sun-Wada et al. 2004). The cytosolic V1 domain is

composed of eight subunits, labeled A-H, in a stoichiometry of A3B3CDE2FG2H1-2

(Arai et al. 1988, Xu et al. 1999, Ohira et al. 2006). In yeasts the integral V0

domain is composed of six subunits that are labeled a, d, e, c, c´ and c´´. The

stoichiometry of the V0 complex is adec3c´c´´ (Arai et al. 1988, Powell et al. 2000). Mammalian V-ATPase lacks c´ subunit, but contains an additional

glycoprotein, which is named Ac45 (Supek et al. 1994).

V-ATPase is structurally very similar to F0F1 ATP synthase (F-ATPase, Fig. 3),

which synthesizes ATP in mitochondria and chloroplasts (Sun-Wada et al. 2004,

Drory & Nelson 2006, Jefferies et al. 2008). F-ATPase mediated ATP synthesis is

powered by proton gradient between the mitochondrial intermembrane space and

the mitochondrial matrix. A flux of protons from the mitochondrial matrix to the

intermembrane space through F-ATPase complex is thought to cause a rotation of

the proteolipid c-ring in the F0 domain and the γ subunit in the F1 domain, whose

movement is required for attaching an inorganic phosphate to ADP (Boyer 1997,

Noji et al. 1997, Futai et al. 2000, Wada et al. 2000). V-ATPase, instead, utilizes

ATP to transport protons against the proton gradient and thus has a function

reverse to that of F-ATPase. Hydrolysis of ATP in a contact region of A and B

subunits rotates subunits D and F in the V1 domain and subunits d, c, c´ and c´´ in

the V0 domain. The rotary movement powered by the hydrolysis of a single ATP

molecule translocates two protons through the V0 domain to the luminal side (Xu et al. 1999, Nishi & Forgac 2002, Sun-Wada et al. 2004, Jefferies et al. 2008).

40

Fig. 3. Functional and structural comparison of V- and F-ATPases. V-ATPases utilize

ATP to transport protons against the proton gradient between the cytosol and the

Golgi lumen (or the endosomal/lysosomal/vacuolar lumen) whereas F-ATPases utilize

the proton gradient between the mitochondrial matrix and the intermembrane space to

produce ATP. Both ATPases are composed of multiple subunits that are organized

similarly to peripheral (V1/F1, light gray) and integral domains (V0/F0, dark gray).

Regulation of V-ATPase activity

The increase of proton import activity towards the end of the secretory pathway is

thought to reflect increased density of V-ATPases (Demaurex et al. 1998,

Schapiro & Grinstein 2000, Wu et al. 2001) or their active regulation (Paroutis et al. 2004, Sun-Wada et al. 2004, Jefferies et al. 2008). Most of the recent findings,

however, support the latter view. Currently, four different V-ATPase regulation

mechanisms are known. The first one involves glucose dependent reversible

dissociation of V0 domain from the V1 domain. Assembly of the functional

V-ATPase complex can occur only in the presence of glucose, whereas glucose

deprivation prevents assembly and also dissociates existing complexes (Kane &

Parra 2000). The glucose dependent regulation of the V-ATPase assembly has

been shown to involve aldolase, a glycolytic enzyme that interacts with a, B and E

subunits (Lu et al. 2004, Lu et al. 2007), and the RAVE complex (regulator of the

V-ATPase of the vacuolar and endosomal membranes), composed of Skp1, Rav1

41

and Rav2 proteins (Seol et al. 2001), that binds to the E and G subunits and

interacts with the C subunit (Smardon et al. 2002, Smardon & Kane 2007). The

disassembly, instead, is dependent on microtubules (Xu & Forgac 2001) and

independent on the RAVE complex, indicating that assembly and disassembly are

separately controlled processes. The glucose dependent regulation of V-ATPase

assembly to a functional complex has been detected in yeasts (Kane & Parra 2000)

as well as in mammalian kidney cells (Sautin et al. 2005).

The second regulation mechanism involves reversible disulfide bond

formation. For instance, a disulfide bond between cysteines 245 and 532 near the

ATP binding site in A subunit blocks ATP binding and leads to inactivation of the

V-ATPase (Feng & Forgac 1994). Consistent with this finding, V-ATPases in

different redox states have been purified from different cellular locations

(Rodman et al. 1994).

The third regulatory mechanism involves changes in coupling. The changes

in coupling may, for instance, be derived from different compositions of the

V-ATPase complexes. Most V-ATPase subunits have more than two isoforms,

which enables the existence of complexes that differ structurally and functionally

(Sun-Wada et al. 2004, Jefferies et al. 2008).

The fourth mechanism relies on changes in counter ion conductance. Since

the V-ATPase is electrogenic, it is dependent on the prevailing membrane

potential, which is dependent on the counter ion conductance (e.g. Cl- and K+, see

Chapter 2.5.3). The accumulation of the membrane potential in the absence of

proper counter ion conductance inhibits and finally prevents V-ATPase mediated

proton import (Glickman et al. 1983, Rybak et al. 1997, Paroutis et al. 2004).

2.5.2 Proton export/elimination

The mechanism of proton export/elimination, or proton leakage, from the Golgi

has not been fully characterized, but it is thought to involve exchange of lumenal

protons for cytosolic cations, free proton conductivity or elimination of protons

by base import (Paroutis et al. 2004). The most probable protein candidates that

are able to mediate these functions are introduced below.

Na+/H+ exchangers

Two ubiquitously expressed Na+/H+ exchanger (NHE) isoforms, NHE7 (Numata

& Orlowski 2001) and NHE8 (Nakamura et al. 2005), have been localized in the

42

Golgi. NHE7 resides predominantly in the TGN (Numata & Orlowski 2001),

whereas NHE8 can be found mainly at the medial- and trans-Golgi (Nakamura et al. 2005). NHEs are integral membrane proteins that mediate electroneutral

exchange of H+ for Na+ and K+ across the membrane (Bianchini & Poussegur

1994). Since no external energy source, such as ATP, is utilized, the exchange

occurs down the net gradient, which means that a high Na+ gradient can drive H+

transport against the gradient, if the Na+ gradient is higher than the H+ gradient.

NHEs are composed of an N-terminal multispanning transmembrane domain,

which mediates ion exchange, and a large cytosolic C-terminal regulatory domain

(Wakabayashi et al. 1992, Wakabayashi et al. 2000).

The evidence supporting the roles of NHE7 and NHE8 in Golgi pH

regulation and proton export has been obtained from overexpression studies. The

overexpression of NHE7HA or NHE8-ECF has been shown to increase Na+ and K+

influx to the Golgi (Numata & Orlowski 2001) and luminal Golgi pH (Nakamura et al. 2005), respectively. Both results indicate indirectly that overexpression of

the Golgi localizing NHEs leads to increased proton export.

Proton channels

The proton leakage from the Golgi may also occur via proton channels. The

channel protein itself has not been identified, but the characteristics of the H+

export from the Golgi are reminiscent of the conductance properties of plasma

membrane localized H+ channels (Schapiro & Grinstein 2000). Similar to the

conductance mediated by the channels on the plasma membrane, Golgi H+ export

can be inhibited with micromolar concentrations of Zn2+ (Cherny & DeCoursey

1999, Schapiro & Grinstein 2000). However, an interesting feature of the the

Golgi H+ export is that it is not fully blocked by Zn2+, not even with higher

concentrations (Schapiro & Grinstein 2000). The existence of this residual and

Zn2+ resistant H+ conductance may indicate presence of an additional, but

currently unknown, proton leakage pathway.

Anion exchanger 2a

An alternative way for the exchanger or channel mediated proton export to

eliminate protons from the Golgi lumen, is to import bases and buffering

compounds (Paroutis et al. 2004). One potential canditate to mediate this is the

electroneutral HCO3-/Cl- antiporter, anion exchanger 2a (AE2a or SLC4A2a),

43

which is ubiquitously expressed in all tissues and can be found on the cell

surfaces as well as on Golgi membranes (Kellokumpu et al. 1988, Alper et al. 1997, Stuart-Tilley et al. 1998, Holappa et al. 2001, Holappa et al. 2004, Romero et al. 2004, Alper 2006). The suggested role in Golgi pH regulation is based on an

assumption that AE2a has similar functions in the Golgi as it has at the plasma

membrane, where it participates in the regulation of cytosolic pH, chloride

concentration and cell volume by mediating bidirectional HCO3-/Cl- exchange

(Romero et al. 2004, Alper 2006).

AE2a belongs to SLC4 gene family, whose members share a common

structural pattern of three domains. They have a large ~400–700 amino acids long

cytoplasmic N-terminal domain, a polytopic ~500 amino acids long

transmembrane domain with 10–14 transmembrane spanning sequences and a

small ~30–100 amino acids long cytoplasmic C-terminal domain. The third

exofacial loop in the transmembrane domain is usually the largest and is often

N-glycosylated (Alper 2006). In humans AE2a is the most abundant of the four

currently identified AE2 isoforms. The other three isoforms: AE2b1, AE2b2 and

AE2c differ from AE2a by their N-terminal domains (Romero et al. 2004) and by

their more limited tissue- and cell-type specific expression in epithelia and

stomach (Wang et al. 1996, Medina et al. 2000). Currently, it is not fully clear

whether active AE2 proteins exist only as homodimers or also as homotetramers

(Zolotarev et al. 1999).

The role of plasmalemmal AE2a in pH regulation has been shown to involve

at least the N-terminal and transmembrane domains. Both of them have been

shown to be important in sensing and regulating intra- and extracellular pH

(Stewart et al. 2001, Stewart et al. 2002, Stewart et al. 2004, Stewart et al. 2007a,

Stewart et al. 2007b, Stewart et al. 2008). The mechanism by which the

N-terminal domain senses intracellular pH changes is unknown, but it may

involve, for instance, pH dependent blocking of regions in the transmembrane

domain that are mediating HCO3-/Cl- transport. However, the pH sensing

mechanism of the transmembrane domain is better understood. It is thought to

involve multiple transmembrane residues, often histidines, that are able to

undergo protonations and deprotonations in response to intra- and extracellular

pH alterations. These protonations and deprotonations are thought to cause

conformational changes that lead to reduced or increased transport of HCO3- and

Cl-, respectively (Stewart et al. 2007a, Stewart et al. 2007b). In a recent study it

was also shown that a putative re-entrant loop in the transmembrane domain may

44

have an additional role in pH sensing and regulation (Stewart et al. 2008). The

mechanism, however, is not fully understood.

In addition to the alterations in intra- and extra-cellular pH the activity of

AE2 may be regulated by phosphorylation. It has been reported that AE2a has one

highly conserved protein kinase C phosphorylation site and that a phorbol ester

which activates protein kinase C, stimulates also Cl-/HCO3- exchange mediated

by AE2a in rabbit mucous cells (Rossmann et al. 2001). Direct evidence for the

phosphorylation dependent regulation, however, is still missing.

The interactions of AE2 with carbonic anhydrases may also be important in

pH regulation. It has been shown that the transmembrane domain contains a

luminal/extracellular binding site for carbonic anhydrase IV (Sterling et al. 2002),

and that the cytosolic C-terminal domain contains a binding site for cytoplasmic

carbonic anhydrase II (Vince & Reithmeier 1998, Vince & Reithmeier 2000).

Together, carbonic anhydrase II, carbonic anhydrase IV and AE2 are thought to

form a bicarbonate metabolon (Sterling et al. 2002). Currently, however, it is not

known whether the Golgi lumen contains carbonic anhydrase IV.

2.5.3 Counterion conductance

Counterion conductance, which comprises influx of anions or efflux of cations,

maintains the organelle membrane potential low and favourable for normal

activity of electrogenic V-ATPase (Glickman et al. 1983, Rybak et al. 1997,

Paroutis et al. 2004). The absence of counterion conductance in the presence of

V-ATPase activity leads to accumulation of membrane potential, which in turn

leads to inhibition of V-ATPase activity and to limited organelle acidification

(Glickman et al. 1983). Therefore, it has been suggested that the counterion

conductance may have a central role in the pH regulation of some organelles

(Rybak et al. 1997). However, counterion conductance limited organelle pH

regulation can occur only in the presence of low conductivity, which is typical,

for instance, for endosomes (Rybak et al. 1997, Hara-Chikuma et al. 2005a, Hara-

Chikuma et al. 2005b). Estimations of the counterion conductivity of the Golgi,

conversely, have mostly indicated high conductivity (Demaurex et al. 1998,

Schapiro & Grinstein 2000, Wu et al. 2001, Demaurex 2002). In the presence of

high conductivity the counterion conductance cannot act as a primary factor that

limits acidification, but rather exist as an indispensable “background component”

that enables normal pH regulation by preventing formation of a too high

membrane potential (Rybak et al. 1997, Demaurex et al. 1998, Schapiro &

45

Grinstein 2000, Wu et al. 2001, Demaurex 2002). The indentified counterion

channels and some potential candidates that are able to mediate counterion

conductivity in the Golgi are introduced in next chapters.

GPHR/GOLACs

One potential protein to regulate counterion conductance is a recently identified

anion channel, Golgi pH regulator (GPHR)(Maeda et al. 2008). GPHR was found

in mutant cells in a screening study on the basis of a phenotype that showed

reduced secretion rate, alterations in N- and O-glycosylation and increased

luminal Golgi pH (Maeda et al. 2008). A significant role of GPHR for the Golgi

functions and the Golgi acidification was also confirmed by GPHR specific

siRNA treatment. The treatment caused a similar, but somewhat milder,

phenotype than that of the mutant cells (Maeda et al. 2008).

GPHR shares some physiological properties with two other earlier

characterized Golgi localized anion channels, GOLAC-1 (Nordeen et al. 2000)

and GOLAC-2 (Thompson et al. 2002), which have not been identified at the

molecular level. All three channels have 5–6 conductance states that are mostly

open, allowing high counterion conductance. Additionally, GPHR is, like

GOLAC-2, dependent on voltage, sensitive to DIDS and pH insensitive. Also

their relative permeabilities to different anions are very similar. The relative Cl- to

K+ selectivity13, however, is more like that of GOLAC-1 (Caldwell & Howell

2008, Maeda et al. 2008). Thus, it has been suggested that GOLAC-1, GOLAC-2

and GPHR may actually be the same protein, and that the earlier characterised

differences of GOLAC-1 and GOLAC-2 represent different regulation of the

same channel (Caldwell & Howell 2008).

Other potential counter ion channels and transporters

The group of other potential counterion channels contains several anion channels,

such as a cystic fibrosis transmembrane conductance regulator (CFTR) (Barasch et al. 1991, Barasch & al-Awqati 1993), a mid-1-related chloride channel (MClC)

(Nagasawa et al. 2001) and voltage gated chloride channels (ClCs) (Schwappach et al. 1998, Jentsch et al. 1999, Gentzsch et al. 2003, Jentsch 2007). CFTR is a

13 PCl/PK permeability ratio 2.8 ± 1.3 (mean ± S.D.) (Maeda et al. 2008)

46

cAMP (cyclic AMP) regulated chloride channel, which localizes to the trans-

Golgi, the TGN, endosomes and prelysosomes (Barasch et al. 1991). A mutation

in the CFTR gene in cystic fibrosis has been associated with defective

acidification of the lumens of these organelles (Barasch et al. 1991, Barasch & al-

Awqati 1993), but the result is controversial, since it has not been succesfully

repeated (Biwersi & Verkman 1994, Seksek et al. 1996, Machen et al. 2001).

MClC is an anion channel or an activator of anion channels that localizes to the

ER, the Golgi and the nucleus (Nagasawa et al. 2001). However, its role in Golgi

pH regulation awaits confirmation. ClCs, instead, comprise a large family of

voltage gated chloride channels, whose members are often required for proper

acidification of intra-cellular organelles (Jentsch et al. 1999, Jentsch 2007). Thus

far two ClCs, ClC-3B in mammalian cells (Gentzsch et al. 2003) and Gef-1 in

yeasts (Schwappach et al. 1998), have been found on Golgi membranes, but their

role in the Golgi pH regulation remains to be confirmed.

In addition to anion conductivity Golgi membranes are also highly permeable

to K+, although the channel itself remains to be identified at the molecular level

(Schapiro & Grinstein 2000). Some of the K+ conductivity may be

GPHR/GOLAC-1 mediated (Nordeen et al. 2000, Maeda et al. 2008), but due to

high permeability, it is also likely that other K+ conductive channels exist in the

Golgi.

2.5.4 Other factors that may affect Golgi pH

In addition to the proteins introduced above other factors such as cargo transiting

through the Golgi, vesicle transport and luminal buffering may also affect Golgi

pH (Paroutis et al. 2004). Since the Golgi locates in the center of the secretory

pathway, it is constantly exposed to the transit of multiple transporter and

channels proteins destined to endosomes, lysosomes and the plasmamembrane.

Some of the transiting proteins are fully active from their early biosynthesis

(Ruetz et al. 1993, Paroutis et al. 2004), and thus have been thought to cause

background noise to the Golgi pH.

The contribution of vesicle transport in Golgi pH regulation is also based on

the central role of the Golgi at the crossroads of the secretory and endocytic

pathways. The Golgi constantly receives material from anterograde and

retrograde directions that is carried by vesicles, tubules or pleimorphic membrane

formations (Chapter 2.3). The vesicles from the endosomal compartments bring

acid loads and the vesicles, or pleiomorphic membrane structures, from the ER

47

bring alkaline loads to the Golgi. If for some reason, one pathway overruns the

other, the Golgi pH is exposed to an increased acid or alkaline load, which may

alter Golgi pH (Paroutis et al. 2004).

The last factor listed, Golgi buffering, is a factor that is thought to reduce

Golgi pH changes and maintain Golgi homeostasis. Buffering in the Golgi is

thought to comprise lumenal matrix proteins that are rich of chemical groups with

suitable pKa-values i.e. groups that protonate and deprotonate around

physiological Golgi pH, and soluble buffering molecules, such as HCO3- (Paroutis

et al. 2004). Additionally, the presence of transporters and exchangers for

buffering molecules (e.g. AE2a that mediates HCO3-/Cl- antiport (Kellokumpu et

al. 1988, Alper et al. 1997, Stuart-Tilley et al. 1998, Holappa et al. 2001), may

enable active regulation of the buffering capacity and hence also active

maintenance of the correct Golgi pH.

2.6 Golgi functions

In addition to the role as a cellular sorting center (Chapter 2.4) the Golgi

apparatus is a site of several other important functions. For instance, conventional

glycosylation and sulfation of proteins and lipids (Varki 1998, Brockhausen 1999,

Wopereis et al. 2006, Maccioni 2007), protein lipidation (Nadolski & Linder 2007)

and some branches of lipid biosynthesis (De Matteis & D'Angelo 2007) are

performed in the Golgi. The Golgi resident glycosylation processes are introduced

more precisely in the next chapters.

2.6.1 Glycosylation

Glycosylation is the most common post-translational modification. It has been

estimated that more than half of all proteins expressed in eukaryotic cells are

glycosylated (Apweiler et al. 1999). Glycosylation of lipids is a rarer event. Only

5% or less of the cellular lipids are glycosylated in non-neural tissues and about

25% in neural tissues (Hougland et al. 1979, Maccioni 2007).

48

Reaction mechanism

Glycosyltransferases (EC 2.4.x.y, Chapter 2.7) catalyze transfer of a glycosyl

moiety from an activated donor to a specific acceptor in an inverting14 (Fig. 4) or

a retaining 15 (Fig. 5) reaction (Breton et al. 2006). The inverting reaction

mechanism is better understood and is classified as a bimolecular nucleophilic

substition or SN2-reaction, whereas the mechanism of the retaining reaction has

not been elucidated yet. The inverting reaction mechanism is believed to involve

an acidic amino acid that activates the acceptor hydroxyl group by deprotonation

and an N-terminal loop structure near the active domain that forms a closing lid.

The closure of the lid is thought to cause a conformational change in the donor

substrate and generate a binding site for the acceptor substrate. The mechanism

often, but not always, involves a divalent cation, generally Mn2+ or Mg2+, that acts

as a catalyst and participates also in substrate binding (Murray et al. 1996,

Lariviere et al. 2003, Breton et al. 2006, Qasba et al. 2008).

The potential donors in the glycosylation reactions mediated by

glycosyltransferases (EC 2.4.x.y) are limited to nucleoside diphosphosugars16,

nucleoside monophosphosugars17 , or lipid phosphosugars18 , and the potential

acceptors are usually different kinds of elongating or branching oligosaccharides,

but can also be monosaccharides as well as specific proteins or lipids (Breton et al. 2006, Varki et al. 2007). The concentrations of the donors and acceptors are

thought to be at a low millimolar range based on affinities of glycosyltransferases

measured in in vitro studies (Varki et al. 2007, Brenda enzyme databank,

http://www.brenda.uni-koeln.de/).

14 β-linkages 15 α-linkages 16 UDP-galactose, UDP-N-acetylgalactosamine, UDP-N-acetylglucosamine, UDP-glucose, UDP-xylose, UDP-glucuronic acid, GDP-fucose, GDP-mannose 17 CMP-Sia 18 dolichol-P-glucose, dolichol-P-mannose, dolichol-P-glucose3-mannose9-N-acetylglucosamine2

49

Fig. 4. Simplified reaction diagram for inverting glycosylation reaction. The reaction,

in this case β1,4-galactosylation, is mediated by β1,4-galactosyltransferase (GalT1, EC

2.4.1.38), which transfers galactose (Gal) from UDP-Gal to a terminal

N-acetylglucosamine (GlcNAc). The acceptor substrate in the figure is a synthetic

N-linked N-acetylglucosamine. The inverting reaction mechanism is classified as a

bimolecular nucleophilic substition or SN2-reaction (Qasba et al. 2008, Brenda

enzyme databank, http://www.brenda.uni-koeln.de/).

50

Fig. 5. Simplified reaction diagram for retaining glycosylation reaction. The reaction,

in this case α2,3-sialylation, is mediated by α2,3-sialyltransferase (ST3Gal III, EC

2.4.99.6), which transfers sialic acid (Sia) from CMP-Sia to a terminal galactose. The

acceptor substrate in the figure is a synthetic N-linked galactosyl-β1,4-N-

acetylglucosamine. Currently, the exact mechanism of retaining reaction is unclear

(Baum et al. 1996, Brenda enzyme databank, http://www.brenda.uni-koeln.de/).

N-glycosylation

N-linked glycosylation (Fig. 6) is initiated on the cytosolic side of the ER

membranes, where the first glycosyl moiety, N-acetylglucosamine phosphate

(GlcNAc-P), is transferred there from UDP-GlcNAc to a lipid glycosyl carrier,

dolichol phosphate (Dol-P), by GlcNAc-1-phosphotransferase (Schenk et al.

51

2001). The assembly continues on the cytosolic side with an addition of another

GlcNAc and five mannose residues in a sequential manner to generate

Man5GlcNAc2-PP-Dol (Helenius & Aebi 2002). The oligosaccharide intermediate

is then translocated across the ER membrane to the lumenal side via a mechanism

that is not completely understood (Abeijon & Hirschberg 1992, Helenius & Aebi

2002, Helenius et al. 2002). The translocation, however, might be assisted by a

concentrate of Dol-P and other polyisoprenols that have a polar phosphate head

and induce bilayer membrane instability (Vigo et al. 1984, Knudsen & Troy 1989,

Helenius & Aebi 2002).

On the luminal side of the ER the N-glycan precursor is processed to a

Glc3Man9GlcNAc2-PP-Dol form (Burda & Aebi 1999), and is then transferred as

“en bloc” to a potential glycosylation site on a nascent polypeptide chain by

oligosaccharyltransferase (OST) complex (Burda & Aebi 1999, Yan & Lennarz

1999). The potential N-glycosylation sites have an Asn-X-Ser/Thr-Y consensus

sequence, where X And Y are any amino acids except prolines (Kornfeld &

Kornfeld 1985, Gavel & von Heijne 1990). The transfer does not occur to every

potential site and the occupancy of N-glycosylation sites may vary even among

different molecules of the same protein19 (Jenkins et al. 1996). The exact reason

why such variation exists is poorly understood, but for instance a C-terminal

location of the concensus sequence in the peptide backbone (Gavel & von Heijne

1990) as well as the presence of certain aminoacids in the X or Y positions 20 has

been shown to inhibit the transfer (Bause 1983, Gavel & von Heijne 1990,

Shakin-Eshleman et al. 1996). After the N-glycan precursor is transferred to the

peptide, glycans are trimmed and processed while the peptide backbone

undergoes a folding process. Once the peptide backbone is correctly folded, the

N-glycan is modified to a Man8GlcNAc2-Asn form and the protein is exported to

the Golgi (Parodi 2000a, Ritter & Helenius 2000, Trombetta & Helenius 2000).

Further processing in the Golgi involves demannosylations as well as

N-acetylglucosaminylation, galactosylation, fucosylation and sialylation steps

(Roth 2002). Some of the enzymes mediating these steps recognize a common

oligosaccharide motif, and can therefore be shared with other Golgi resident

glycosylation pathways (Wopereis et al. 2006). For instance, β1,4-

19 The phenomenom is called macroheterogeneity. 20 X-position: tryptophan (Trp), aspartic acid (Asp), glutamic acid (Glu), leucine (Leu); Y-position: cysteine (Cys)

52

galactosyltransferase (GalT1) is able to galactosylate any terminal GlcNAc

residue and can thus participate in processing of N-glycans, O-glycans, as well as

glycolipids (Wopereis et al. 2006, Qasba et al. 2008). Mature N-glycans are a

diverse group and their structures may vary from biantennary to hexaantennary.

Some variation can also exist among glycans attached to the same glycosylation

site in different molecules of the same protein21 (Jenkins et al. 1996, Roth 2002).

The processing of N-glycans, however, is often finished with sialylation, but the

types of linkages vary depending on the tissue they are expressed in (Roth 2002).

For instance, terminal α2,3-linked sialic acids are typical of N-glycans expressed

in normal colon epithelium (Sata et al. 1991, Sata et al. 1992, Roth 2002),

whereas polymers of α2,8-linked sialic acids can be found in N-glycans present in

neural tissues (Zuber et al. 1992).

21 The phenomenom is known as a microheterogeneity.

53

Fig. 6. N-glycosylation. At the cytoplasmic side of the ER dolichol phosphate receives

first GlcNAc-1-P from UDP-GlcNAc and is then extended to Man5GlcNAc2-PP-Dol. After

that Man5GlcNAc2-PP-Dol is translocated across the ER membrane to the luminal side

and processed to a Glc3Man9GlcNAc2-PP-Dol form. Glc3Man9GlcNAc2 is then

transferred “en bloc” to a potential N-glycosylation site on a nascent polypeptide

chain. Once the polypetide is correctly folded, three Glc residues are removed and the

54

glycoproduct is transported to the Golgi for further processing. The processing in the

Golgi generally involves several demannosylation, N-acetylglucosaminylation,

galactosylation, fucosylation and sialylation steps. Some of the common branching

possibilities are introduced in the figure.

Mucin-type-O-Glycosylation

Mucins are the most common group of O-glycosylated proteins in humans

(Brockhausen 1999, Wopereis et al. 2006). Their synthesis is initiated in the ER

or the cis-Golgi with a transfer of an N-acetylgalactosamine (GalNAc) moiety

from UDP-GalNAc to a serine or threonine residue in a target protein. The

transfer is catalyzed by a polypeptide N-acetylgalactosaminyltransferase

(pp-GalNAc-T) (Brockhausen 1999, Wopereis et al. 2006). Mucin-type

O-glycosylation, unlike N-glycosylation, has no common consensus sequence.

This results from the coexistence of multiple pp-GalNAc-Ts with overlapping but

different substrate specificities (Ten Hagen et al. 2003, Wopereis et al. 2006).

Thus far 15 different tissue specific human pp-GalNAc-Ts have been discovered,

but at least 24 are thought to exist on the basis of sequence homology (Ten Hagen et al. 2003, Cheng et al. 2004).

Although no common consensus sequence exists, mucin-type

O-glycosylation follows some basic rules, which can be used to predict

glycosylation occupancy. Since mucin-type O-glycosylation is a postfolding event,

only the serine and threonine residues exposed on the protein surface are

glycosylated. Typically these sites reside in coils, turns and linker regions that are

rich in serine, threonine and proline (Julenius et al. 2005, Wopereis et al. 2006).

Additionally, the processing of mucin-type O-glycans shows some degree of

hierarchy and glycosylation of some specific sites can be a prerequisite for

glycosylation of other sites (Ten Hagen et al. 2003).

Once attached to a protein the O-linked GalNAc, also known as Tn-antigen

(GalNAcα1-Ser/Thr), can be used as a base for the processing of several different

core structures. It can be α2,6-sialylated (sialyl-Tn), β1,3-galactosylated (Core 1 /

Thomsen-Friedenreich i.e. T-antigen), β1,3-N-acetylglucosaminylated (Core 3),

α1,3-N-acetylgalactosaminylated (Core 5), β1,6-N-acetylglucosaminylated (Core

6), α1,6-N-acetylgalactosaminylated (Core 7) or α1,3-galactosylated (Core 8).

Sialyl-Tn can be further α2,6-sialylated to form branched disialyl-Tn, and Cores 1

and 3 can be β1,6-N-acetylglucosaminylated to branched Core 2 and Core 4

structures, respectively (Brockhausen 1999, Wopereis et al. 2006). Tn- and both

55

sialyl-Tn-antigens as well as cores 1-6 and 8 are found in humans, but Cores 1

and 2 are the most abundant (Brockhausen 1999).

The core structures can be further modified by adding a variety of different

sugar units such as GlcNAc, GalNAc, Gal, Fuc and Sia, which can be added with

several different linkages. The lengths and the branching of the resulting glycans

may also vary, and the attached sugar units can be acetylated (Varki 1992) and

sulfated (Carter et al. 1988, Capon et al. 1997, Brockhausen 1999). Together

these processing possibilities enable the existence of hundreds of different kind of

mucin-type O-glycan structures (Brockhausen 1999, Wopereis et al. 2006).

Other Golgi resident glycosylation reactions

In addition to N-glycosylation and mucin-type O-glycosylation, several other

glycosylation events take place in the Golgi apparatus. Synthesis of abundant

glycosaminoglycans (GAGs) is one of them. GAGs are long unbranched

disaccharide repeats that consist of glucuronic acid (GlcA) and GalNAc

(dermatan/chondroitin sulfate) (Kresse et al. 1994), GlcA and GlcNAc

(heparin/heparan sulfate) (Lin 2004), or Gal and GlcNAc sugars (keratan sulfate)

(Funderburgh 2000). Dermatan, chondroitin, heparin and heparan sulfates are

attached to a Ser residue in a protein via a linker tetrasaccharide, GlcAβ1-3Galβ1-

3Galβ1-4Xyl (Kresse et al. 1994, Lin 2004), whereas keratan sulfate is attached

to proteins via N- or Core 1 mucin-type O-glycans (Funderburgh 2000). GAGs

can also undergo some post-assembly modifications to obtain more heterogeneity.

GlcNAc residues in heparin and heparan sulfates can be acetylated and sulfated,

and GlcA in dermatan, heparin and heparan sulfates can be epimerized to

iduronate (Wopereis et al. 2006).

The Golgi is also a production site of complex glycolipids, such as

gangliosides (Maccioni 2007). Gangliosides are a large family of complex

glycolipids that are common in animal cells. Gangliosides are composed of a

ceramide and a negatively charged oligosaccharide with one or more sialic acids

(Hakomori 2003, Sonnino et al. 2007). Their synthesis is initiated in the ER or the

Golgi by ceramide glycosyltransferases and is finished in the Golgi by specific

glycolipid glycosyltransferases (Maccioni 2007). However, glycosyltransferases

involved in glycosylation of protein attached glycans may also participate in the

processing of gangliosides (Wopereis et al. 2006).

Other abundant groups of glycans that are produced in the Golgi are O-linked

galactoses and blood group and Lewis antigens. O-linked galactoses are attached

56

to special hydroxylysine aminoacids, which can be found only on collagens

(Pinnell et al. 1971, Kivirikko & Myllylä 1982). The blood group and Lewis

antigens, on the other hand, can be found on a variety of different proteins and

lipids. The synthesis of ABO blood group antigens begins with α1,2-fucosylation

of a terminal galactose in a precursor glycan moiety, Galα1-3GlcNAcβ1-3Gal-R22,

resulting in blood group O, and can continue with α1,3-N-

acetylgalactosaminylation or α1,3-galactosylation to result in blood groups A and

B, respectively (Brockhausen 1999, Lloyd 2000, Varki et al. 2007). The synthesis

of Lewis antigens is more complex, but is also initiated by fucosylation of certain

precursor glycans (Brockhausen 1999).

The rest of the known Golgi resident glycosylation processes are less

common or very rare. O-mannosylation (Endo 1999), O-glucosylation (Shao et al. 2002) and O-fucosylation (Shao & Haltiwanger 2003) occur only in certain

tissues and for special proteins or protein domains.

Synthesis and transport of nucleotide sugars

Nucleotide sugars for the glycosylation processes are synthesized in the cytosol,

except CMP-Sia, which is synthesized in the nucleus (Münster-Kühnel et al. 2004,

Wopereis et al. 2006, Varki et al. 2007). Monosaccharide moieties used in the

nucleotide sugars are derived from dietary sources or salvage pathways in

complex interconnected processes that involve a series of phosphorylation,

epimerization and acetylation steps. Glucose (Glc) and fructose (Fru) are the

major carbon sources, but other abundant monosaccharide substrates, such as

galactose (Gal), fucose (Fuc), mannose (Man), N-acetylgalactosamine (GalNAc)

and N-acetylglucosamine (GlcNAc), are also widely used (Wopereis et al. 2006,

Varki et al. 2007). Monosaccharides can be activated to nucleotide sugars in

several ways. The activation often involves monosaccharide monophosphate,

which can react with UTP/GTP or UDP-monosaccharide. Galactose, for instance,

is first phosphorylated at the 1-position. The formed Gal-1-P can then react with

UTP or UDP-Glc in reactions catalyzed by UTP-hexose-1-phosphate

uridylyltransferase (Isselbacher 1958) and UDP-glucose-hexose-1-phosphate

uridylyltrasferase (Frey et al. 1982, Varki et al. 2007), respectively. Conversely

sialic acid is not pre-phosphorylated for activation. A non-phoshorylated sialic

22 R group can be an O-glycan, an N-glycan or a glycan attached to a lipid (Brockhausen 1999)

57

acid reacts directly with CTP in a reaction catalysed by CMP-sialic acid

synthetase to form CMP-sialic acid (Kean 1969, Münster-Kühnel et al. 2004).

Since nucleotide sugars are membrane impermeable and are synthesized in

the cytosol or the nucleus, they or their sugar moities must be actively transported

to the luminal side of the ER or the Golgi, where the most abundant glycosylation

processes take place. Currently, two different transport mechanisms are known. A

mechanism, utilized by mannose and glucose, involves translocation of sugar

moieties through the ER membranes with dolichol phosphate (Dol-P). Sugar

moieties from GDP-Man and UDP-Glc are first transferred to Dol-P on the

cytosolic side of the membrane, after which the Dol-P-conjugates are translocated

to the luminal side via unknown mechanism (Schenk et al. 2001). Another

mechanism utilized by other sugar moieties involves transport of sugar

nucleotides. The transport of sugar nucleotides is mediated by a family of ER or

Golgi localized nucleotide sugar transporters (NSTs) that belong to solute carrier

family 35. NSTs are antiporters, whose sugar nucleotide importing activity is

coupled to simultaneous export of corresponding nucleoside monophosphates

(Hirschberg et al. 1998). Some NSTs are bi- or polyspecific and can transport

several substrates (Muraoka et al. 2001, Segawa et al. 2002), whereas some are

strictly monospecific (Eckhardt et al. 1996, Guillen et al. 1998, Luhn et al. 2001).

For instance, UPD-Gal transporter can transport both UDP-Gal and UDP-GalNAc

(Segawa et al. 2002), while CMP-Sia transporter transports only CMP-Sia

(Eckhardt et al. 1996).

2.7 Golgi glycosyltransferases

2.7.1 Protein structure and homology

Most mammalian glycosyltransferases are Golgi or ER residents that share

common topology and domain structure. They can be classified as type II

membrane proteins that have an amino-terminal cytoplasmic domain, a single

passing α-helical membrane domain, and a carboxy-terminal luminal domain,

which can be further divided into a stalk region and a catalytic domain (Fig. 7;

(Colley 1997, Opat et al. 2001, Young 2004, Breton et al. 2006). Although,

glycosyltransferases share structural characteristics, their sequence homology at

the aminoacid level is scarce. Recently listed sequences of human

glycosyltransferases fall into 42 distict families (Breton et al. 2006, http://

58

www.cazy.org). Members of the families, however, do not necessarily share

similar catalytic activity and glycosyltransferases with similar catalytic activities

do not necessarily belong to the same family (Campbell et al. 1997, Campbell et al. 1998, Breton et al. 2006). Crystallographic structures exist only for a limited

set of glycosyltransferases, but the folds of the catalytic domains seem to consist

primarily of α/β/α sandwiches (Breton et al. 2006). The structures solved so far

are classified in three groups; GT-A, GT-B and Cst II fold; according to the

orientation of α helices to the central β sheet (Breton et al. 2006).

Fig. 7. Most Golgi glycosyltransferases are type II membrane proteins. Type II

membrane proteins consist of amino-terminal cytoplasmic domain, transmembrane

domain and luminal or extracellular domain depending on the residence at the cell

membranes.

2.7.2 Localization

Glycosyltransferases and other glycosylation enzymes are primarily localized at

the ER or the Golgi membranes, but in some specialized cells and tissues they can

also be found in mucin droplets, the plasma membrane, and body fluids (reviewed

in Colley 1997). Like all other proteins within the secretory pathway or

trafficking through it, glycosyltransferases are also synthesized with a

59

hydrophobic α-helical signal peptide or anchor 23 that directs them co-

translationally to the ER (von Heijne 1985). After translation, co- and post-

translational modifications (e.g. glycosylation, disulfide bond formation), and

folding, Golgi glycosyltransferases and residents of post-Golgi organelles exit

from the ER by a mechanism that utilizes COPII vesicles, tubules, or maturating

membrane compartments that eventually fuse to or form the Golgi cisternae

(Chapter 2.3.2). Regardless of the mechanism of exit from the ER, this process is,

for type II proteins, triggered by a dibasic cytosolic motif near the membrane

spanning sequence of the exported protein (Giraudo & Maccioni 2003).

Within the Golgi, glycosyltransferases organize into overlapping gradients

along the descending, mildly acidic Golgi pH gradient, so that their localizations

mainly follow their subsequent order in glycosylation procesesses they participate

in (Nilsson et al. 1993a, Rabouille et al. 1995, Rottger et al. 1998, Paroutis et al. 2004). Most Golgi resident glycosyltransferases utilize the transmembrane

domain and/or the luminal stem region for correct Golgi targeting and retention,

but sometimes the cytosolic tail and the catalytic region are also utilized

(reviewed in Colley 1997, Opat et al. 2001, Young 2004). The exact Golgi

localization mechanisms are unclear, but according to current knowledge they

may involve oligomerization, i.e. “kin recognition” (Machamer 1991, Nilsson et al. 1993b), a sorting based on the variable thicknesses of lipid membranes within

the secretory pathway (Bretscher & Munro 1993, Masibay et al. 1993), cytosolic

factors, such as conserved oligomeric Golgi (COG) complexes (Oka et al. 2004,

Shestakova et al. 2006), or all of them together (Colley 1997, Opat et al. 2001,

Young 2004). Sometimes glycosyltransferases are also found in post-Golgi

structures, but the exact mechanisms of glycosyltransferases direction to these

structures remain unknown. The post-Golgi localization, however, is often

associated with proteolytic cleavage (reviewed in Young 2004) or dissipation of

the acidic Golgi pH (Axelsson et al. 2001, Schaub et al. 2006, Schaub et al. 2008).

2.7.3 β1,4-galactosyltransferase I (EC 2.4.1. 38)

Galactosyltransferases transfer galactose from a donor susbstrate, UDP-Gal, to a

specific acceptor substrate, which is usually a growing glycan chain attached to a

23 ~20-30 aminoacids with a short positively charged N-terminal region, a central α-helical hydrophobic core and a polar C-terminal region (von Heijne 1985)

60

peptide or a lipid backbone (Chapter 2.4.1). It has been estimated that there are at

least seven galactosyltransferases that synhesize β1,4-linkages and multiple of

those that synthesize other kind of linkages between galactose and the acceptor

substrate (reviewed in Qasba et al. 2008). The β1,4-galactosyltransferase studied

in this work, named GalT1 (EC 2.4.1.38), participates in the synthesis of

glycoconjugates by transferring galactose from UDP-Gal to a terminal

N-acetylglucosamine in a glycan chain with a β1,4-linkage (Fig. 4, (Qasba et al. 2008, Brenda enzyme databank, http://www.brenda.uni-koeln.de/).

GalT1 localizes into medial-/trans-Golgi (Berger et al. 1981, Roth & Berger

1982, Rabouille et al. 1995) mainly by utilizing its transmembrane domain

(Nilsson et al. 1991, Aoki et al. 1992, Teasdale et al. 1992, Yamaguchi & Fukuda

1995). However, the presence of a cytosolic tail (Nilsson et al. 1991, Evans et al. 1993) or stem region has been shown to enhance Golgi localization in some cell

types (Teasdale et al. 1992) indicating that these regions may also contain

additional localization signals. The pH probe, GT-EGFP (Llopis et al. 1998),

which was used in this study in most of the Golgi pH measurements, is composed

of a 81 amino acids long N-terminal region of GalT1 and a full length EGFP. The

81 amino acids long N-terminal region contains the cytosolic tail, the

transmembrane sequence and parts of the stem region and is thus fully capable of

localizing correctly in the Golgi (Llopis et al. 1998).

2.7.4 α2,6-sialyltransferase I (EC2.4.99.1) and α2,3-sialyltransferase III (EC 2.4.99.6)

Sialyltransferases transfer sialic acid from a donor substrate, CMP-Sia, to a

specific acceptor substrate. Mammalian sialyltransferases are known to synthesize

α2,3-, α2,6- or α2,8-linkages. The acceptor substrate is often galactose, but can

also be N-acetylgalactosamine for α2,6-sialylation. α2,8-sialylation, in turn, can

only occur on terminal sialic acid (reviewed in Takashima 2008). The two

sialyltransferases studied in this work, β-galactoside-α2,6-sialyltransferase I

(ST6Gal I, EC 2.4.99.1) and β-galactoside-α2,3-sialyltransferase III (ST3Gal III,

EC 2.4.99.6), participate in synthesis of N-glycans by sialylating terminal

galactose. They show different but overlapping specifities on the substrate and

may in some circumstances act competitively (Weinstein et al. 1982, Yeh &

Cummings 1997). ST6Gal I synthesizes α2,6-linkage and ST3Gal III α2,3-linkage

between the sialic acid and the terminal galactose (Fig. 5, Baum et al. 1996,

Brenda enzyme databank, http://www.brenda.uni-koeln.de/).

61

Both ST6Gal I and ST3Gal III localize dominantly to the trans-Golgi, but

minor amounts of them can be found also from the medial-Golgi and the TGN

(Berger et al. 1993, Burger et al. 1998, Grabenhorst & Conradt 1999). ST6Gal I

can be found in monomeric and homodimeric forms (Colley 1997, Chen et al. 2000, Qian et al. 2001, Fenteany & Colley 2005) and its localization is mediated

by signals in the transmembrane and the luminal domains (Colley et al. 1989,

Dahdal & Colley 1993, Fenteany & Colley 2005). The oligomerization state of

ST3Gal III and the domains containing localization information are currently

unknown.

2.8 Biological role of glycoconjugates

Glycogonjugates participate in several cellular functions and recognition events in

health and disease. They are involved, for instance, in protein folding and sorting,

vesicular trafficking, cell signalling, cell adhesion and motility, organogenesis and

immunological defence (Paulson 1989, Varki 1993, Scheiffele et al. 1995, Kim &

Varki 1997, Parodi 2000b, Yu 2007, Marth & Grewal 2008, van Kooyk &

Rabinovich 2008, Zhao et al. 2008). Altered glycosylation can impair all these

functions and is involved in the pathogenesis of various diseases such as

congenital disorders of glycosylation (CDG) (Freeze & Aebi 2005), cystic fibrosis

(CF) (Rhim et al. 2001), various muscular dystrophies (Martin-Rendon & Blake

2003, Freeze 2006) and cancer (Kim & Varki 1997, Campbell et al. 2001b, Yu

2007, Zhao et al. 2008). The mechanism of interference often involves inhibition

of normal or promotion of disease typical inter- or intramolecular interactions

(Fig. 8). In some special cases when glycogonjugates contribute majority of the

mass and the charge of the glycoprotein, the altered glycosylation can change the

general appearance or the chemical properties of the glycoprotein and thus

prevent its correct function (Fig. 8). Experiences from various diseases and

genetically engineered disease models have shown that glycosylation defects

become more obvious in multicellular organisms than in cell cultures and that the

earlier the defect is in the glycosylation pathway, the more severe are the

symptoms arising from it (Ohtsubo & Marth 2006). Total inhibition of N-

glycosylation, for example, is lethal and leads to death in early embryogenesis

(Marek et al. 1999).

62

Fig. 8. Glycoconjugates can modulate molecular interactions. Alterations in

glycosylation can inhibit existing and induce abrnomal inter- and intra-molecular

interactions or change the general appearance and chemical propertiens of the highly

glycosylated molecule. Figure is based on a review article authored by Ohtsubo and

Marth (Ohtsubo & Marth 2006).

2.8.1 Congenital disorders of glycosylation

Congenital disorders of glycosylation (CDG) are one family of rare heterogenous

glycosylation diseases that have greatly enhanced understanding of the

importance of glycoconjugates in multicellular organisms. CDG are classified as

CDG types I and II (CDG-I and CDG-II), depending on the nature and location of

63

the biochemical defect in the glycosylation pathway (Freeze & Aebi 2005, Freeze

2006). Lower case letters in the abbreviation of the disease indicate different

subtypes in chronological order of identification (e.g. CDG-Ia is the first

identified CDG-I). Thus far 19 different CDGs have been identified (Freeze 2006).

Twelve of them belong to CDG-I, which result from deficiencies in biosynthesis

of oligosaccharide precursors. Seven of them belong to CDG-II, which result

from defects in assembly of protein linked glycans. Additionally, there are some

CDG-I and CDG-II-like diseases that have not been characterized (Freeze 2006,

Zeevaert et al. 2008). CDG have a wide range of clinical features and are thus

difficult to recognize. Their features, however, are often associated with severe or

mild malformations of various organs and organ systems (Freeze & Aebi 2005,

Freeze 2006).

2.8.2 Cancer and glycosylation

Glycosylation alterations are typical for all kinds of cancers (Kim & Varki 1997,

Campbell et al. 2001b, Ono & Hakomori 2004, Yu 2007, Zhao et al. 2008). Even

if the detected alterations are always cell- and tissue-type dependent and often

unique, they usually have general characteristics that can include, for instance,

shortening and increased sialylation of O-linked oligosaccharide side chains

(Boland & Deshmukh 1990, Campbell et al. 1995, Itzkowitz et al. 1995, Karlen et al. 1998), reduced sulphation (Kuhns et al. 1995), reduced O-acetylation of

sialic acid (Jass et al. 1994, Ogata et al. 1995), elevated fucosylation (Parodi et al. 1982), altered sialylation and branching of N-glycans (Kobata et al. 1995, Roth

2002) and increased expression of blood group structures (Kim et al. 1986). Some

of these glycosylation alterations, such as increased α2,8-sialylation of N-glycans

(Roth et al. 1988) and increased expression of truncated O-glycans e.g. sialyl-Tn

(Neu5Acα2-6GalNAcα1-Ser/Thr) or T-antigen (Galβ1-3GalNAcα1-Ser/Thr)

(Moriyama et al. 1987, Wolf et al. 1988, Itzkowitz et al. 1995), are associated

with increased proliferation, invasion and metastasis, and are thus useful

prognostic markers of malignant disease states. The underlaying mechanisms

behind the aggressive growth associated with these glycosylation epitopes,

however, are not understood. The mechanisms may be related to altered cell-cell

interactions and recognition events in which glycoproteins have central roles (Yu

2007, van Kooyk & Rabinovich 2008, Zhao et al. 2008). Consistent with this idea

it has been hypothesized that high expression of structures with terminal Galβ1-

3/4GalNAc and/or sialic acid may induce abnormal interactions with special

64

lectins named galectins 24 and siglecs 25 that have roles in several cellular

recognition and signalling events and are able to modulate, for example, innate

and adaptive immune responses (Yu 2007, van Kooyk & Rabinovich 2008). Such

interactions are, at least in theory, capable of inducing and perpetuating malignant

behaviour of cancer cells.

Mechanisms of altered glycosylation

In many cases it is unclear how various cancer associated glycosylation

alterations develop. Sometimes expression of altered oligosaccharide structures

correlates with overexpression and/or depletion of corresponding

glycosyltransferases (Whitehouse et al. 1997, Brockhausen 1999, Burchell et al. 1999, Dalziel et al. 2001), but often such correlation is inadequate or lacking

(Yang et al. 1994, Brockhausen et al. 1995, Hanisch et al. 1996, Lloyd et al. 1996,

Taylor-Papadimitriou et al. 1999, Müller & Hanisch 2002). This indicates that

other factors than variations in the expression levels of glycosyltransferases can

also induce cancerous glycosylation. Consistent with this, some correlation has

been shown for the increased presence of substrates and altered glycosylation.

Increased expression of UDP-Gal transporter, for instance, is thought to induce

increased T-antigen expression by enhancing availability of UDP-Gal for core 1

β1,3-galactosyl transferase (EC 2.4.1.122) (Kumamoto et al. 2001). Another

mechanism that can control glycosylation involves glycosyltransferase specific

chaperones. One such is Core 1 β1,3-galactosyl transferase-specific molecular

chaperone (Cosmc), whose deficiency has been associated with increased Tn-

antigen expression, albeit in a non-cancerous specimen (Schietinger et al. 2006).

In the absence of Cosmc Core 1 β1,3-galactosyl transferase folds incorrectly and

is targeted to proteosomes for degradation. As a consequence expression of Core

1 structures are remarkably reduced. In some instances it has also been reported

that changes in the splicing of a glycoprotein can lead to increased expression of

certain oncofetal glycan epitopes. In the case of CD44 it has been shown that only

the splicing variant CD44v6 is associated with T-antigen expression in cancer and

normal tissue (Singh et al. 2001). Cancer associated glycosylation alterations may

24 Galectins are soluble oligosaccharide binding proteins that recognize N-acetyl lactosamine-like (Galβ1-3/4GalNAcα1-Ser/Thr) structures (Yang et al. 2008). 25 Siglecs are type I membrane spannign proteins that bind to sialic acid containing glycans (Crocker et al. 2007).

65

also arise from general defects in Golgi function and homeostasis. In line with

this view, perturbation of Golgi pH with different drug treatments has been found

to induce T-antigen expression and cancer typical alterations in Golgi morphology

(Campbell et al. 2001a, Kellokumpu et al. 2002).

66

67

3 Aims of the present work

Glycosylation alterations are linked with several severe diseases. One of these

diseases is cancer. Nearly every cancer patient has a tumor or tumors with

recognisable glycosylation alterations. Some glycosylation alterations can be seen

on benign cysts and in long term inflammation states. Sometimes these alterations

are explained by increased or decreased expression of glycosylation enzymes, but

often no clear correlation exists. As the Golgi resident glycosylation is known to

be sensitive to Golgi lumenal pH disturbances, we decided to examine whether

glycosylation abnormalities in cancer are associated with altered Golgi pH. The

second aim was to solve how Golgi pH increase inhibits glycosylation. The third

aim was to understand how Golgi pH is regulated in health and disease. Due to

the complexity of the processes involved in Golgi pH regulation, however, we

focused on the functions of a Golgi localized proton pump, vacuolar-H+ATPase

(V-ATPase), and a Cl-/HCO3- exchanger, anion exchanger 2a (AE2a).

68

69

4 Materials and methods

4.1 Antibodies, lectins and reagents

The antibodies used in the work were purchased from commercial suppliers or

were generous gifts from individual researchers. These included antibodies

against the following antigens: β1,4-galactosyltransferase (anti-GalT, Dr. E.

Berger, Zürich, Switzerland), GM130 and GFP (monoclonal BD Biosciences,

Palo Alto, CA, USA), mannosidase II (anti-MAN II, Chemicon International,

Temecula, CA, USA), LAMP-2 (Developmental Studies Hybridoma Bank, The

University of Iowa, Department of Biological Sciences, Iowa City, USA), M6P-

receptor (M6P-R), early endosomal antigen-1 (EEA1), Rab9, β-COP (Affinity

Bioreagents, Inc., Golden, CO, USA), TGN46 (Serotec, Oxford, UK); α-tubulin

(Sigma, Saint Louis, MO, USA), KDEL-receptor (StressGen Biotechnologies

Corp., Victoria, BC, Canada), Carcinoembryonic Antigen (polyclonal anti-CEA,

Dako, Glostrup, Denmark; monoclonal COL-1, Zymed Laboratories, Inc., San

Francisco, CA, USA) and myc-epitope (polyclonal, Abcam, Cambridge, UK;

monoclonal, Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA). A

polyclonal antibody against AE2 C-terminal 12 aminoacids long peptide (anti-

AE2ct, against aminoacids 1229–1241 or GVDEYNEMPMPV) was specially

ordered from Davids Biotechnologies (Regensburg, Germany). Two separate

stocks were ordered, because the first one used mainly in earlier works ran out.

As secondary antibodies we used horseradish peroxidase (HRP)-conjugated

goat anti-mouse IgG (P.A.R.I.S., Compiègne, France) and donkey anti-goat IgG

(Santa Cruz Biotechnology, Inc., Santa Cruz, California), Alexa Fluor® 488-

conjugated donkey anti-mouse IgG, Alexa Fluor® 596-conjugated anti-rabbit and

Alexa Fluor® 596-conjugated anti-sheep (Molecular probes, Eugene, OR, USA)

antibodies and peroxidase-conjugated anti-digoxigenin Fab fragments (Roche

Diagnostics GmbH, Mannheim, Germany).

Lectin stainings were performed using the DIG Glycan Differentation Kit

(Roche Applied Science, Basel, Switzerland) or HRP-conjugated WGA and ConA

lectins (Sigma-Aldrich, St. Louis, MO, USA). HRP-conjugated MAA lectin was

purchased from Ey Laboratories, Inc. (San Mateo, CA, USA).

Chloroquine (CQ), ammonium chloride, Hoechst dye (33258), monensin,

nigericin, brefeldin A and DNA-oligos were purchased from Sigma-Aldrich (St.

Louis, MO, USA); endoglycosidase H (Endo H) and peptide N-glycosidase F

70

(PNGase F) from New England Biolabs (Ipswich, MA, USA) and Sialidase

STM/NANase I from Glyko®/ProZyme (San Leandro, CA, USA). Proteins on the

western blots were visualized using the ECL system (GE Healthcare,

Buckinhamshire, UK). Streptolysin O (SLO) was purchased from Sigma Aldrich

or BioAcademia, Inc. (Osaka, Japan).

4.2 Plasmids

The plasmid encoding the medial/trans-Golgi targeted GT-EGFP fusion protein

used for Golgi pH measurements (Llopis et al. 1998) was kindly donated to us by

Dr. R. Tsien (San Diego, CA, USA).

The plasmid encoding the full length (3 kb) carcinoembryonic antigen (CEA)

was kindly provided to us by Dr. C. Stanners (Montreal, QB, Canada). The

plasmid that encodes a secretory form of CEA (sCEA) was constructed from that

as follows. At first the 3 kb full length CEA cDNA insert was subcloned into the

EcoRI site of the pCDNA3 vector (Invitrogen, CA, USA) from which a 2.1 kb

fragment (30-2130 bp) was amplified by PCR using forward (5’-CACAGCAA-

GCTTGACAAAAC-3’) and reverse primers (5’-CAGGAgGcAATTCAGATG-

CAGAG-3’). The forward primer contained two mutated nucleotides (GC>AG,

bold) to create a new HindIII site (underlined). Two deletions (g2129, c2131)

were included in the reverse primer to create a new EcoRI site (underlined). After

digestion with the appropriate enzymes, the gel purified PCR product was

inserted into the HindIII-EcoRI sites of a previously modified pCDNA 3 vector,

which also contains an in frame c-myc tag sequence and a 3’ stop codon. The

translation of the cDNA gives a secreted C-terminally truncated 76 kDa CEA-

polypeptide which lacks the GPI-anchor found in the authentic membrane-bound

CEA protein.

To construct a plasmid that encodes EGFP tagged α2,3-sialyltransferase III

(ST3Gal III-EGFP), its full length ORF (cDNA clone, Unigene RZPD, Berlin,

Germany) was amplified by PCR using forward (5´-AAAAAGGTACCATGGG-

ACTCTTGGTATTTGTG-3´) and reverse (5´-AAAAAGGATCCGATGCCA-

CTGCTTAGATCAG-3´) primers. After digestion with KpnI and BamHI enzymes

and gel purification, the PCR-product was cloned in frame with the EGFP within

the cloning site of the pEGFP-N3 plasmid (Clontech Laboratories, Inc., Mountain

View, CA, USA). A plasmid that encodes EGFP tagged α2,6-sialyltransferase I

(ST6Gal I-EGFP) was constructed in a similar way. Briefly, ST6Gal I cDNA

(Unigene RZPD, Berlin, Germany) was amplified by PCR using forward

71

(5´-AAAAAGGTACCATGATTCACACCAACCTG-3´) and reverse (5´-AAA-

AAGGATCCGCAGTGAATGGTCCG-3´) primers. After partial digestion with

KpnI and HindIII enzymes and gel purification, the PCR-product was cloned in

the pEGFP-N3 plasmid.

Mutations of nine cysteines in human ST6Gal I -cDNA were performed with

Quikchange II site-directed mutagenesis kit (Stratagene, La Jolla, CA, USA)

using primers designed with Stratagene´s web-based primer design program.

Domain swapping of transmembrane domains from ST3Gal III to ST6Gal I and

vice versa were performed with same kit according to Kirsch and Joly (Kirsch &

Joly 1998). Briefly, primers 5´-GCTAGCGCTACCGGACTCAG-3´, 5´-GCAAT-

TTAAAGGAATCATAGTACTGGAGTAAGTGTAGCTTCCACGC-3´ and 5´-CTG-

AATTGGAGTCCTCCTCCCAACTCCCTTTCTTCTTTTCCTT-CCAC-3´ were first

used to generate megaprimers from ST3Gal III-EGFP and ST6Gal I-EGFP

plasmids. The megaprimers were then used to change transmembrane domains in

ST3Gal III and ST6Gal I-EGFP plasmids to create chimeric EGFP-conjugated

sialyltransferases.

AE2-myc construct used in overexpression assays was prepared by cloning

full length AE2 into pcDNATM3.1/myc-His(-) A vector (Invitrogen, Carlsbad, CA,

USA) from GFP-AE2 vector (Holappa et al. 2001) with forward

(5´-AAAAGAATTCCGGCCATGAAGCAGCG-3´) and reverse primers

(5´-TACAAAGCTTCACAGGCATGGCCATCTC-3`). AE2-DsRed vector was

constructed by detaching AE2 cDNA from AE2-myc vector using NheI and

HindIII restriction enzymes and ligating the detached cDNA to linearized DsRed-

Monomer vector (Clontech Laboratories, Inc., Mountain View, CA, USA).

4.3 Cell culture

African green monkey kidney cells (COS-7), human mammary adenocarcinoma

cells (MCF-7), human cervical adenocarcinoma cells (HeLa) and colorectal

adenocarcinoma cells (HT-29, SW-48) were purchased from American Type

Culture Collection (ATCC, Manassas, VA, USA). Human primary fibroblasts

(F XII) and another colorectal adenocarcinoma cell line (CaCO-2) were received

as generous gifts from Dr. Maija Risteli and Dr. Vesa Olkkonen, respectively.

All cells were cultivated on plastic dishes or glass coverslips in Dulbecco´s

modified essential medium (DMEM) supplemented with GlutamaxTM, 10% fetal

calf serum (FCS; HyClone, Cramlington, UK) and 100 units of penicillin –

100 μg/ml streptomycin (Gibco BRL, Grand Island, NY, USA) at 37 ºC, 5.0%

72

CO2. Cell transfections were performed with Fugene 6 or Fugene HD reagents

(Roche Applied Sciences, Basel, Switzerland). Stable cell lines were selected

from transiently transfected cells with 500 μM geneticin (Sigma-Aldrich, St.

Louis, MO, USA).

4.4 Indirect immunofluorescence and lectin stainings

Cells grown on coverslips were washed with PBS and fixed with 4%

paraformaldehyde in PBS for 15 min at room temperature (RT). After subsequent

washes with PBS the cells were blocked with 1% BSA in PBS for 1 hour. If

permeabilization of the cells was required, the blocking was performed in the

presence of 0.05% saponin. Antibodies and lectins were diluted in the blocking

buffer and incubations with them were carried out at RT, and in dark if they were

conjugated with fluorochromes. Between primary and secondary antibody

incubations the cells were washed with the blocking buffer and after secondary

antibody and lectin incubations the cells were washed with PBS and water to

remove excess staining and salts.

When HRP-conjugated lectins were used the initial procedure was similar to

that described above, but PBS was replaced with 0.1 M Na-phosphate (Na-P).

After washes and fixation with 4% paraformaldehyde, the samples were post-

fixed with 2.5% glutaraldehyde in 0.1 M Na-P for 30 min at RT. Visualization of

HRP conjugates was performed by adding diaminobenzidine (0.5 mg/ml) and

H2O2 in 0.1 M Na-P. The peroxidase reaction was allowed to proceed for ~15–30

min on ice and the excess reaction product was gently washed away with 0.1 M

Na-P and water.

The stained specimens were mounted on microscope slides with Immumount

(Shandon, Pittsburgh, PA, USA) and examined under an epifluorescence or a

confocal microscope.

4.5 Co-localization analysis

Co-localization analyses were performed with ImageJ 1.38w software (National

institutes of health, USA) and Mander´s/ Pearson´s coefficient -plugin created by

Tony Collins (Wright Cell Imaging Facility, Toronto, Canada) and Wayne

Rasband (National institutes of health, USA) from 0.5µm thick representative

image layers captured using a confocal microscope.

73

4.6 Live cell microscopy

4.6.1 Golgi pH measurements

Golgi pH measurements were carried out according to Llopis et al. (1998). Briefly,

~20–24 h post-transfection with vectors encoding EGFP or an EGFP tagged

protein the cells were bathed in PBS and examined under a microscope using a

60 × water immersion objective and appropriate filter set for EGFP. An image of

intact cells was captured and the cells were calibrated in situ with different pH

buffers (125 mM KCl/20 mM NaCl/1.0 mM CaCl2/1.0 mM MgCl2) adjusted to

pH 8.0, pH 7.0 or pH 6.0 with 20 mM HEPES, MOPS or MES, respectively. The

pH buffers used in calibration contain 5 μM nigericin and 5 μM monensin to

abolish gradients of monovalent cations (Na+, K+, H+). The images of intact cells

and the images of cells in the pH buffers were analysed with AnalySIS Pro 3.2

software (Soft Imaging System Inc., Lakewood, CO, USA). Fluorescence

intensity was measured from the regions of interest (ROI) in each cell against the

background and converted to pH values using a sigmoidial calibration curve.

Numerical data were processed using Sigmablot 8.0 (SPSS Inc., Chicago, IL,

USA) or Microsoft Excel solver (Redmond, WA, USA).

When appropriate, fresh drug solutions (30 nM BafA1; 10–60 μM CQ, 10

mM NH4Cl) were prepared in PBS or in complete medium and incubated with

live cells for an appropriate length of time, dependending on the drug type. In

some experiments, live MCF-7 cells were stained with Alexa Fluor® 594-

conjugated PNA (10 μg/ml in PBS) for 5 min at 37 ºC to enable discrimination

between T-antigen expressing and non-expressing cells prior to Golgi pH

measurements.

4.6.2 Acidification measurements

Active proton pumping into the Golgi lumen (i.e. Golgi acidification) was

measured in GT-EGFP expressing cells as described by Demareux et al.

(Demaurex et al. 1998). Cells were permeabilized using 1–5 μg/ml Streptolysin O

(SLO) in high potassium buffer (120 mM KCl, 30 mM NaCl, 10 mM EGTA, 10

mM MgCl2, 10 mM Hepes, pH 7.2) ~18–20 h post-transfection. SLO perforates

only the plasma membrane and other cholestrol rich membranes when used at

proper concentrations (Bhakdi et al. 1993), and therefore maintains Golgi

membranes intact. After 3 to 5 minutes of equilibration excess ATP, to a final

74

concentration of 4 mM, was added to the cells. Images of examined cells were

captured at appropriate intervals for 5 to 7 minutes. At the end of each experiment

the pH was calibrated in situ as described above.

In some experiments chloride was replaced with gluconate or sulphate (e.g.

NaCl Na-gluconate, MgCl2 MgSO4) and some experiments were performed

in presence of 20–40 mM bicarbonate (HCO3-). Live cell staining with Alexa

Fluor® 594-conjugated PNA was used to distinguish T-antigen expressing and

non-expressing MCF-7 cells prior to Golgi acidification measurements. Staining

was performed as described above.

4.7 Electron microscopy

4.7.1 Preparation of cell pellets

To study Golgi cisternal dilatation in CQ treatments, cells grown to 90–95%

confluency were washed with PBS (or Hank´s buffer) and fixed with 1%

glutaraldehyde, 4% paraformaldehyde (or 2.5% glutaraldehyde) in 0.1 M Na-P for

10 min at RT. Cells were then gently detached from the plate using a cell scraper

and fixation was allowed to further proceed for at least 1 hour in the fixative.

Fixed cells were harvested by centrifugation (500g, 5 min) and washed three

times with 0.1 M Na-P and twice with water for 10 min. Washed pellet was casted

into 2% agarose in water. After agarose was cooled and had set, the pellet was

post-fixed with 1% osmiumtetroxide, 2.2% CaCl2 at +4 ºC for 30 min and then

stained with 1–2% uranyl acetate for 30 min and washed three times in water for

10 min between and after incubations. The pellet was then washed with ascending

acetone series 30%, 50%, 70% for 10 min each and twice with 100% acetone for

10 min at +4 °C to remove all the water before casting into epon. In the epon

casting procedure, the pellet was first incubated in epon acetone 1:1 for 45 min

and then in 100% epon o/n. Embedding was performed with 100% Epon at 40 ºC

for 1 hour and polymerization at 60°C for ~24–48 h. Thin sections were cut and

the specimens were examined using a Philips CM100 transmission electron

microscope.

75

4.7.2 Immunogold labeling of AE2

Immunogold labeling of the AE2 epitope was performed in a similar way as

described earlier (Holappa et al. 2004, Sipilä et al. 2007). Briefly, cells grown to

~90–95% confluency were fixed with 4% paraformaldehyde in PBS, detached

from the plate with cell scraper and embebbed in 4% gelatin. The specimens were

then snap-frozen in liquid nitrogen in the presence of 2.3 M saccharose and cutted

into 80–100 nm thin sections. The sections were immunostained with anti-AE2

rabbit antibody (1st stock) and secondary rabbit anti-mouse antibody and finally

with protein A-coated colloidal gold.

4.8 Flow cytometry

Cells grown to ~70–80% confluency were washed with PBS, detached from the

plates using trypsin and gently harvested by centrifugation (300 × g, 5 min, +

4 ºC). Loosely pelleted cells were resuspended in blocking buffer (PBS, 1% BSA,

20 mM EDTA) for Alexa 488 conjugated PNA staining (1 μg/ml, 5 min, on ice).

Filtered samples were then analysed with flow cytometry (CyFlow, Partec,

Münster, Germany) using appropriate filter sets.

4.9 Antibody internalization experiments

Stably ST3Gal III-EGFP expressing cells were labeled with anti-GFP antibody as

described by Chapman & Munro (1994). Briefly, cells grown in the presence or

absence of 20µM CQ were incubated with 1 µg/ml anti-GFP antibody in the

growth medium for 3 h at +37˚C, 5% CO2 before fixation with 4%

paraformaldehyde and immunostaining with Alexa-594 conjugated anti-mouse

IgG.

76

4.10 Preparation of cell lysates

4.10.1 SDS based lysates

General procedure

Cells were washed with PBS and dissolved in 2 × SDS-PAGE loading buffer (4%

SDS, 20% glycerol, bromphenol blue, 125 mM Tris, pH 6.8) containing 100 mM

DTT when reduction of samples was required. Samples were heated for 5 min at

+95°C before loading into a SDS-PAGE gel.

Procedure for EGFP tagged glycosyltransferases

The total lysates of cells stably expressing ST6Gal I-EGFP or ST3Gal III-EGFP

were pretreated as described earlier (Molinari & Helenius 1999) to prevent a

formation of nonspecific intermolecular cysteine bridges during lysis. Briefly, the

stably expressing cells were preincubated in 25mM N-ethylmaleimide (NEM) in

PBS, pH 6.8 for 5 min. The cells were then lysed in loading buffer (1% SDS, 20%

glycerol, 125 mM Tris, pH 6.8) and analysed with SDS-PAGE without heating the

samples. The EGFP fusion proteins were visualized directly from the gel with

Typhoon 9400 variable mode imager (GE Healthcare, Little Chalfont, UK) using

appropriate laser and filter sets.

4.10.2 Nonidet based lysates

Lysates for AE2 immunoblotting were prepapared in Nonidet P-40, because the

carboxyterminal AE2 epitope is SDS-sensitive. Trypzinised cells were suspended

in native lysis-buffer (50 mM Tris-HCl, 150 mM NaCl, 1% Nonidet P-40, pH 8.0)

supplemented with protease inhibitors and incubated in shaker for 30 min, at

+4 ºC. Insoluble material was removed from the suspension by centrifugation

with tabletop centrifuge (16100 × g, 4 min, +4 ºC). Samples were heated for 5

min at +50 ºC before loading into native-PAGE gel.

77

4.11 Immunoprecipitation

Cells expressing the secretory form of CEA (sCEA) were grown for 48 h in the

absence or presence of Golgi pH dissipating drugs after which the media were

collected and suplemented with protease inhibitors (CompleteTM, Roche Applied

Science, Basel, Switzerland). Cell debris was removed form the media by

centrifugation (16100 × g, 10 min, +4 ºC). The cells from the same plates were

washed and lysed in 2 × SDS-PAGE loading buffer. sCEA protein was

immunoprecipitated from the cleared media samples using 2 µg of polyclonal

anti-CEA antibody and protein A sepharose (GE Healthcare Bio-Sciences, Little

Chalfont, UK) by incubating o/n at +4 ºC. The beads were then washed 5 times

with 1.5 ml of RIPA-buffer (1% Triton X-100, 0.5% deoxycholate, 0.1% SDS,

0.15 M NaCl, 0.05 M Tris, pH 7.5) and once with 10 mM Tris-HCl, pH 7.5.

Immunoprecipitated sCEA was detached from the beads with 50 µl of 2 × SDS

loading buffer. Both immunoprecipitated and cell-associated sCEA protein were

analysed using SDS-PAGE and visualized on the blots using COL-1 antibody

(1:500), MAA and DSA lectins included within the DIG Glycan Differentation

Kit (Roche Applied Sciences, Basel, Schwitzerland) or HRP-conjugated MAA

lectin (Ey Laboratories, Inc., San Mateo, CA, USA). α-tubulin in lysates, detected

with a specific antibody (1:5000), was used for quantification of the secretion and

also as a loading control. Quantification of protein bands was performed using the

AnalySIS Pro 3.2 software (Soft Imaging System Inc., Lakewood, CO, USA).

4.12 Glycosidase treatments

Immunoprecipitated sCEA samples were mixed with glycoprotein denaturing

buffer (supplied with the Endo H or PNGase F kit). Denatured samples were

incubated with the Endo H or PNGase F enzyme according to manufacturer’s

protocol (New England Biolabs, Ipswich, MA, USA) prior to analysis with SDS-

PAGE and immunoblotting. The inhibition of α2,3-sialylation of sCEA by CQ

was confirmed with a sialidase specific for α2,3-linked sialic acids (Sialidase

STM/NANase I, Glyko®/ProZyme). The immunoprecipitated sCEA samples were

mixed with 0.5M Na-acetate, pH 5.5 and incubated with the enzyme o/n at +37 ºC

prior to analysis with SDS-PAGE and immunoblotting.

78

4.13 AE2 gene silencing

Cells were grown overnight to ~70–80% confluency and transfected with AE2

gene specific siRNA SMARTpool® (Dharmacon, Lafayette, CO, USA) using

DharmaFECT™ 2 (Dharmacon, Lafayette, CO, USA) as a transfection reagent.

Transfection was performed according to the manufacturers instructions. The

siRNA SMARTpool® was removed after 24h or 48h incubation and the medium

was changed before GT-EGFP transfection that was required for Golgi pH

measurements.

The efficiency of gene silencing was tested using native-PAGE and

immunoblotting. Briefly, harvested cells were dissolved in native lysis-buffer

(50mM Tris-HCl, 150 mM NaCl, 1% Nonidet P-40, pH 8.0) as described above.

After sample preparation, total protein concentration was measured using the

Bradford assay (BioRad Laboratories, Hercules, CA, USA) and the sample

volumes were adjusted in a way that each sample in the gel contained an equal

amount of protein. The amount of expressed AE2 on the native gel was detected

using antibody against AE2 C-terminal peptide (Davids Biotechnologies,

Regensburg, Germany).

4.14 SDS-PAGE and Immuno- /lectin-blotting

SDS-PAGE and Western blotting were performed according to Laemmli (1970).

Native-PAGE was performed similarly but in the absence of SDS.

4.14.1 Immunodetection

The blot was blocked in 5% non-fat dry milk in TBST (50mM Tris, 150 mM

NaCl, 0.02% Tween 20, 0.1% BSA, pH 7.4) for at least 2 hours at +4 ºC before

the blot was rinsed with TBST and incubated for 2 hours separately with primary

and HRP conjugated secondary antibodies diluted in TBST. Excess and unbound

antibodies were washed away between and after antibody incubations. Bound

antibody complexes were visualized using ECL detection reagent (GE Healthcare,

Buckinhamshire, UK).

79

4.14.2 Lectin detection

Procedure for HRP-conjugated lectins

The blot was blocked with 2% Tween in TBS (0.05 M Tris-HCl, 0.15 M NaCl, pH

7.5) for 2 min at RT. The blot was then rinsed twice with TBS and incubated for

16 hour at RT with 1μg/ml lectin (MAA-HRP: Ey-Laboratories, Inc., San Mateo,

CA, USA) diluted in TBS containing 0.05% Tween, 1 mM CaCl2, 1 mM MgCl2

and 1mM MnCl2. The blot was rinsed twice and bound lectins were visualized

with ECL detection reagent.

Procedure for digoxigenin conjugated lectins

The blot was blocked with diluted blocking reagent for 30 min at RT or for o/n at

+4 ºC. After blocking, the filter was washed twice with TBS (0.05 M Tris-HCl,

0.15 M NaCl, pH 7.5) and once with buffer 1 (TBS; 1 mM CaCl2, 1 mM MgCl2

and 1 mM MnCl2, pH 7.5). Lectins (MAA, DSA or SNA, DIG Glycan

Differentation Kit: Roche Applied Sciences, Basel, Schwitzerland) were diluted

in buffer 1 at appropriate concentration and incubated for 1 hour at RT. The blot

was then washed three times with TBS and incubated for 1 h with 1:10000 anti-

digoxigenin-HRP. Finally, the blot was washed three times with TBS to remove

unbound antibodies. The lectin-antibody complexes were visualized using ECL

detection reagent.

4.15 Adenosine nucleotide measurements

Cells grown to ~70–80% confluency were chilled, rinsed briefly with ice-cold

PBS and harvested in 1 ml of PBS using a cell scraper. After centrifugation

(300 × g, 2 min, +4˚C) cells were resuspended in ice-cold 7% perchloric acid

(HClO4) and vortexed twice for 20 s, before collecting cell debris and precipitated

proteins by centrifugation (13000 × g, 10 min, +4˚C). The resulting supernatant

and the protein precipitate were stored at –70˚C for analyses. Adenosine

nucleotide levels (ATP, ADP and AMP) in the samples were determined in

triplicate according to (Folley et al. 1983) using Supelcosil LC-18-T (Ø 3 μm)

reverse phase column (Supelco, Bellefonte, PA, USA). Elution rates and spectra

of standard nucleotides were used to identify the different nucleotide peaks in the

samples. Proteins in the precipitates were dissolved in 1% SDS and the protein

80

concentrations were measured using the Bradford assay (Biorad, Hercules, CA,

USA). For successful protein measurements SDS content of the samples was

diluted to ≤0.025%.

4.16 Statistical methods

Statistical analyses were performed with Mann-Whitney U-test

(http://elegans.swmed.edu/~leon/stats/utest.html) or when the sample number was

limited (n = 2–3) with Student´s t-test. Other statistical quantities were calculated

with Microsoft® Excel (Redmond, WA, USA).

81

5 Results

5.1 Glycosylation and Golgi pH in cancer cells (I)

Perturbation of Golgi pH with different drug treatments has been shown to induce

T-antigen expression (Campbell et al. 2001a, Kellokumpu et al. 2002). To study

whether T-antigen expression is associated with increased Golgi pH in cancer, we

selected two non-malignant cell lines as controls (human fibroblasts and COS-7

cells) and four different cancer cell lines from breast (MCF-7) or colon cancers

(CaCO-2, HT-29, SW-48) for Golgi pH measurements. T-antigen expression was

tested by staining non-permeabilized cells with HRP conjugated PNA. Non-

malignant control cells and CaCO-2 cells expressed no or very low amounts of T-

antigen on their surfaces, whereas the other cancer cells (MCF-7, HT-29, SW-48)

expressed it more intensively. The expression of T-antigen was, however,

relatively heterogenous (Fig. 9) and closely resembled a staining of a tumorigenic

tissue in vivo (I, Fig. 4C). The proportion of highly T-antigen expressing cells in

HT-29, MCF-7 and SW-48 cells varied from 18 to 52% (I, Fig. 4B).

Golgi pH was measured using a medial-/trans-Golgi localizing GT-EGFP as

described in the materials and methods section. The Golgi pH in two non-

malignant cell lines, human fibroblasts and COS-7, was 6.15 ± 0.07 (median ±

S.E.) and 6.32 ± 0.03, respectively. In low T-antigen expressing CaCO-2 cells the

Golgi pH was 6.29 ± 0.07, while in the rest of the cancer cells (MCF-7, HT-29

and SW-48) it was 6.77 ± 0.05, 6.73 ± 0.10 and 6.77 ± 0.07, respectively. The

results show that increased Golgi pH correlates with high T-antigen expression

(r = 0.845, p < 0.05; I, Fig. 4B).

82

Fig. 9. T-antigen expression in cancer and control cells. Non-permeabilized cells were

stained with HRP-conjugated PNA-lectin. The dark precipitate identifies to T-antigen

expressing cells, while non-expressing cancer cells and control cells are unstained.

Scale bar 10 µm.

5.2 The effects of drug treatments on Golgi pH (I, II)

Different drugs such as ammonium chloride (NH4Cl), chloroquine (CQ),

bafilomycin A1 (BafA1) and monensin have been widely used for Golgi

alkalinization and inhibition of glycosylation (e.g. Thorens & Vassalli 1986,

Gawlitzek et al. 2000, Axelsson et al. 2001, Campbell et al. 2001a, Kellokumpu et al. 2002), but their relative potencies has never been estimated. To make the

effects of the different drug treatments comparable, we treated GT-EGFP

expressing COS-7 cells with 30 nM BafA1, 10 mM NH4Cl, 40 µM and 20 µM

CQ for 24h and measured Golgi pH (I, Figure 3A). Incubation with 30 nM BafA1

83

and 10 mM NH4Cl increased the Golgi pH above neutral i.e. to 7.8 ± 0.27 (median ± S.E.) and 7.05 ± 0.10, respectively. With 40 µM and 20 µM CQ the

Golgi pH increase was lower being 6.74 ± 0.08 and 6.48 ± 0.03, respectively. 10

µM CQ has no statistically significant effect on Golgi pH.

5.3 Inhibition of glycosylation by Golgi pH dissipating drugs (I, II)

5.3.1 Inhibition of terminal O-glycosylation (I)

To inhibit terminal O-glycosylation we treated COS-7 cells with 30 nM BafA1,

10 mM NH4Cl and 40 µM, 20 µM or 10 µM CQ and stained them after

permeabilization with Alexa-488 conjugated PNA. We found that accumulation of

T-antigen into the Golgi was initiated at 20 µM CQ (I, Figure 3B), which

corresponds to increase of ~0.2 pH units. When Golgi pH was further increased

using higher concentrations of CQ or more potent drugs, the presence of T-

antigen in the Golgi became more evident. Some of the expressed T-antigen,

however, was additionally distributed to vesicular structures indicating potential

secretion or cell surface expression. To test the existence of cell surface

expression we used flow cytometry.

Cells for the flow cytometry were treated with 0, 10, 20 or 40 µM CQ. After

gentle detachment the cells were stained with Alexa-488 conjugated PNA as

described in the materials and methods section. The PNA staining of detached

live cells showed that the cell surface expression of T-antigen also increases with

20 µM CQ (I, Figure 3C). The cell surface staining, however, did not get more

intense when higher CQ concentrations were used, but instead seemed to decrease.

This might be due to inhibition of more proximal glycosylation steps or reduced

transport of the T-antigen to the cell surface.

84

5.3.2 Inhibition of terminal N-glycosylation (II)

To study the pH sensitivity of N-glycosylation, we expressed a plasmid encoding

a secretory form of carcinoembryonic antigen (sCEA26) for 48 h in the presence

or absence of Golgi pH dissipating drugs. The secreted sCEA was then collected

from the growth media and analysed with SDS-PAGE as described in the

materials and methods section. Cells on the plates were lysed for α-tubulin

quantification and normalization of sCEA secretion.

Treatments with 0-60 µM CQ showed that the molecular size of the sCEA

can be decreased gradually by increasing CQ concentration (II, Fig. 2A & B). The

highest concentrations used in the test i.e. 40 and 60 µM CQ reduced the

molecular size from ~200 to ~160 kDa. At the same time secretion of sCEA

remained unaltered (II, Fig. 2A left panel). To study whether the decrease of the

molecular size of sCEA involves proteolytic degradation or inhibition of

glycosylation or both, we treated each sample with PNGase F to remove N-linked

carbohydrates. The size of the resulting peptide backbone was unaffected (~77

kDa) in every sample (data not shown) indicating that the decrease of the

molecular size represents only inhibition of N-glycosylation.

N-glycosylation of sCEA was also studied with Endo H, which specifically

removes all high-mannose N-glycans but leaves Golgi modified N-glycans intact.

Molecular weight decreased by ~20 kDa in every sample i.e. to ~180–140 kDa

dependending on the sample treated (II, Fig. 2A). The decrement of the molecular

size and the absence ~77 kDa band in Endo H treated samples revealed that sCEA,

also secreted by non-treated control cells, contains Golgi processed and high-

mannose N-glycans, consistent with earlier data (Kobata et al. 1995). Secondly,

and more importantly, the finding also indicates that the drug concentrations used

in the tests do not prevent initial processing of the high-mannose N-glycans to

hybrid and complex type N-glycans in the cis-Golgi.

To study the inhibited glycosylation in more detail, we utilised an ascending

series of CQ, from 0 to 40 µM, with 10 µM increments and stained sCEA blots

with MAA- (Maackia amurensis agglutinin), DSA- (Dature Stramonium

agglutinin) and SNA- (Sambucus nigra agglutinin) lectins (II, Fig. 2B). Since

26 Secretory form of CEA (sCEA) is genetically modified from wild type CEA by removing GPI-anchor attacment site (Chapter 4.2). Highly N-glycosylated wild type CEA contains 28 potential N-glycosylation sites and has approximately 120-130 kDa of glycans in a 77kDa peptide backbone being thus ideal marker for lectin blotting assays (Oikawa et al. 1987, Yamashita et al. 1987).

85

SNA-lectin, detecting α2,6-sialylation, failed to stain the immunoprecipitated

sCEA, we studied α2,6-sialylation using an unknown 55kDa protein present in the

cell lysate (II, Fig. 2B left panel). Our studies showed that α2,3-sialylation,

detected with MAA-lectin, was the most sensitive to drug treatments and thus to

Golgi pH increase. With 20 µM CQ, which corresponds to Golgi pH increase of

~0.2 pH units, α2,3-sialylation was reduced to ~40% of the initial level

(*p < 0.005), while β1,4-galactosylation, detected with DSA-lectin, and α2,6-

sialylation, detected with SNA-lectin were not significantly altered (II, Fig. 2B &

C). The inhibition of α2,3-sialylation was more evident when higher CQ

concentrations or more potent drugs were used (II, Fig. 2B & 2E). 40 µM CQ, for

instance, reduced α2,3-sialylation by over 80% (**p < 0.01). The inhibition of

α2,3-sialylation was also confirmed with sialidase S (II, Fig. 2D), which

specifically cleaves only α2,3-linked sialic acids. Sialidase induced decrement of

the molecular size of sCEA was lower in samples collected from the medium of

cells treated with 40 µM CQ compared to samples collected from the medium of

non-treated cells.

5.3.3 Golgi pH dissipation is able to inhibit α2,3-sialylation despite of

its enhancement by ST3Gal III overexpression

Glycosylation alterations in cancer often correlate with altered expression of

glycosylation enzymes (Whitehouse et al. 1997, Brockhausen 1999, Burchell et al. 1999, Dalziel et al. 2001), in many cases however, this correlation is absent (Yang et al. 1994, Brockhausen et al. 1995, Hanisch et al. 1996, Lloyd et al. 1996,

Taylor-Papadimitriou et al. 1999, Müller & Hanisch 2002). The lack of

correlation can derive form several different factors (Chapter 2.8.2: Mechanisms

of altered glycosylation). One of the suggested factors is a general disorder in

Golgi pH homeostasis. Consistent with this suggestion perturbation of Golgi pH

has been shown to induce glycosylation alterations that are typical for cancers

(Campbell et al. 2001a, Kellokumpu et al. 2002). To study whether the inhibition

of glycosylation achieved by elevating Golgi pH can be overrun by increased

expression of glycosylation enzymes, we overexpressed ST3Gal III in stably

sCEA expressing COS-7 cells prior to Golgi pH elevation with CQ.

The overexpression of ST3Gal III was found to increase α2,3-sialylation of

sCEA by ~3.6 fold as evidenced by the more intensive MAA-lectin staining (Fig.

10A, middle panel) and increased molecular size of the sCEA band from ~200 to

~220 kDa (Fig. 10A, the bottom panel). In drug treatments the sCEA expressed in

86

ST3Gal III overexpressing cells got smaller and also the intesity of the MAA-

staining decreased, indicating reduced α2,3-sialylation (Fig. 10B). α2,3-

sialylation was reduced by ~30 or ~80% from the control when the cells were

treated with 20 or 40 µM CQ, respectively. Simultaneously, the average molecular

size was decreasing from ~220 to ~200 or ~180 kDa, respectively. The result

indicates that overexpression of ST3Gal III cannot overrun the inhibitory effect of

Golgi pH elevation on glycosylation.

Fig. 10. A. Overexpression of ST3Gal III increases α2,3-glycosylation. B. The effect of

ST3Gal III overexpression can be overrun with dissipation of the acidic Golgi pH with

CQ. ST3Gal III and sCEA coexpressing cells were treated with different concentrations

of CQ prior to collection of the secreted sCEA from the growth medium and analysis

with SDS-PAGE and MAA-lectin.

87

5.4 Perturbation of Golgi organization with pH gradient dissipating drugs (II)

Correct Golgi glycosylation is known to be dependent on intact Golgi structure

and pH regulation (Axelsson et al. 2001, Campbell et al. 2001a, Kellokumpu et al. 2002, Paroutis et al. 2004), but the exact mechanisms behind the dependency are

not fully understood. It has been suggested that glycosylation alterations induced

by Golgi pH increase may be due to nonoptimal luminal pH for glycosylation

enzyme activity (Gawlitzek et al. 2000), reduced availability of the sugar

nucleotides (Waldman & Rudnick 1990) or mislocalization of the corresponding

glycosylation enzymes (Axelsson et al. 2001). However, the tests on which these

suggestions are based, were performed with relatively strong drug treatments that,

according to our experience, lead to extensive alkalinization and complete

fragmentation of the Golgi. Therefore, we tested how the Golgi structure is

altered when cells are treated with milder drugs which raise Golgi pH moderately

and inhibit some, but not all, glycosylation processes. For that purpose we

incubated COS-7 cells in the presence of 0-60 µM CQ for 24 h prior to fixation

and immunostainings. In the absence of available antibodies that are suitable for

immunostainings we used stably expressed EGFP-transferase fusions.

The studies showed that Golgi proteins can be divided roughly into three

groups according to their sensitivity to Golgi pH dissipation (II, Fig. 1A). GM130,

KDEL-R, β-COP and MANII belong to the most insensitive group expressing

more than 95% of the antibody staining intensity at the juxtanuclear Golgi when

the cells are treated with 40 µM CQ. GalT and ST6Gal I belong to moderately

sensitive group and are gradually distributed to vesicular structures when CQ

concentration is increased above 20–40 µM. With 40 µM CQ, for instance, ~80%

of the stably expressed ST6Gal I-EGFP remains at the juxta-nuclear Golgi region

(II, Fig. 1B). ST3Gal III, TGN46 and M6P-R belong to the most sensitive group

and are extensively mislocalized from the juxta-nuclear Golgi region when the

cells are treated with 20 µM CQ. For instance, in the cells treated with 20 µM CQ

only ~20% of the stably expressed ST3Gal III-EGFP remains at the juxta-nuclear

Golgi region (II, Fig. 1B). The use of 60 µM CQ was found to lead to a total

fragmentation of the Golgi, similarly to NH4Cl and BafA1 treatments, and

therefore higher CQ concentrations were not tested.

Disruption of Golgi organization was also studied by using electron

microscopy (EM). Cells for the samples were treated with 20 and 40 µM CQ. The

treatment with 20 µM CQ induced a slight cisternal swelling and formation of

88

large vesicles (II, Fig. 1D). In cells treated with 40 µM, instead, the cisternae

were already extensively dilated, so that the identification of Golgi structures was

difficult.

5.5 Chloroquine induced mislocalization of ST3Gal III (II)

5.5.1 Chloroquine induces mislocalization of ST3Gal III from the

Golgi to endosomes/lysosomes and the plasma membrane

Several TGN proteins are known to constantly cycle between the TGN,

endosomes and the plasma membrane (Reaves et al. 1993, Stanley & Howell

1993, Chapman & Munro 1994, Bonifacino & Rojas 2006). Since, ST3Gal III

behaves similarly to TGN proteins in the CQ treatments, we tested whether

ST3Gal III is recycled in the drug treated cells in a way that is typical for many

TGN proteins.

Stably ST3Gal III-EGFP expressing cells treated with 20 µM CQ were first

fixed and then stained with antibodies against common endosomal markers (II,

Fig. 3). The examination of the stained specimens with confocal microscope

showed that in the drug treated cells ST3Gal III localizes dominantly into large

vesicles with the late-endosomal/lysosomal marker, LAMP-2 (Pearson´s

coefficient 0.888, II Fig. 3A), whereas ~25% of the protein remains at the Golgi

membranes (II, Fig. 1B). Colocalization with other endosomal markers, a marker

for recycling endosomes (Rab9) and a marker for early endosomes (EEA1), was

less remarkable (Pearson´s coefficients 0.590 and 0.439, respectively, II Fig. 3B

& 3C).

To study whether the detected mislocalization of ST3Gal III into the LAMP-2

containing vesicles is direct or involves transport via the plasma membrane, we

cultivated the cells in medium containing anti-EGFP antibody. After that the cells

were fixed and stained with a fluorescence-conjugated secondary antibody against

anti-EGFP. Antibody labeling was detected only in the CQ treated cells (II, Fig. 5),

indicating that the mislocalization to the LAMP-2 containing vesicles involves

ST3Gal III-EGFP transport via the plasma membrane. The absence of labeling in

the control cells, in turn, indicates that ST3Gal III is not normally recycled via the

plasma membrane.

89

5.5.2 pH alterations in acidic organelles are associated with ST3Gal III mislocalization (II)

To better understand the mechanisms involved in the mislocalization of ST3Gal

III, we measured luminal pH in ST3Gal III containing compartments in the drug

treated cells (II, Fig. 4). The measurements with ST3Gal III-EGFP showed that

after a 24 h incubation with 20 µM CQ, juxtanuclear Golgi pH measured using

residual Golgi localized sensor (~25% of the total ST3Gal III-EGFP) was 6.53 ±

0.23 (median ± S.D.) being remarkably higher than in non-treated control cells

(6.27 ± 0.20, ***p < 0.001). At the same time pH in the vesicular structures,

where most of the ST3Gal III in cells treated with 20 µM CQ is localized, was

5.96 ± 0.23 being within the expected pH range of late endosomes/lysosomes (pH

5.5–6.0, e.g. (Paroutis et al. 2004). Endosomal acidity was further examined with

Lysosensor Yellow/Blue stainings. The stainings showed that endosomal/

lysosomal acidity is partly recovered already after two hours of initiation of CQ

treatments, the luminal pH being, however, somewhat higher than in non-treated

cells (data not shown). Together these data show that both Golgi and endosomal

pH have been elevated upon the drug treatments indicating that both of them, in

theory, have the potential to cause the mislocalization, if ST3Gal III is in normal

conditions recycled through the post-Golgi structures. The antibody

internalization assay (II, Fig. 5), however, strongly denotes that ST3Gal III is not

recycled through the post-Golgi structures similarly to TGN proteins.

5.5.3 ST3Gal III expression in CaCO-2 cells reveals the dependency of correct localization and Golgi pH

To further test the potential effect of impaired endosomal acidification on the

correct localization of ST3Gal III and the possibility that ST3Gal III is normally

recycled through the post-Golgi structures similarly to some TGN resident

proteins, we used CaCO-2 cells that have normal Golgi pH (I, Fig. 1B), but which

are unable to properly acidify endosomes and lysosomes (Kokkonen et al. 2004;

II, Fig. 6A). Thus, if the CQ induced mislocalization is dependent only on a pH

change in endosomes, the ST3Gal III expressed in CaCO-2 cells should

accumulate into endosomes. This, however, was not the case. Instead, in CaCO-2

cells the expressed ST3Gal III remained strictly Golgi associated and showed

more significant co-localization with GM130 than with LAMP-2 (Pearson´s

coefficients 0.751 and 0.109, respectively; II, Fig. 6B & C). The treatment of

90

CaCO-2 cells with CQ, however, caused similar mislocalization of ST3Gal III to

COS-7 cells. These results therefore strongly suggest that the drug induced

ST3Gal III mislocalization is dependent on luminal pH increase in the trans-Golgi

and that ST3Gal III is not normally recycled through the post-Golgi structures.

5.6 Alterations in terminal glycosylation do not affect Golgi

localization (II)

It has been previously shown that dissipation of Golgi pH inhibits terminal

glycosylation and simultaneously induces mislocalization of Golgi resident

proteins (Axelsson et al. 2001, Campbell et al. 2001a, Kellokumpu et al. 2002).

To study whether impaired terminal glycosylation itself is a factor capable of

causing altered localization of Golgi proteins, we treated ST3Gal III-EGFP

expressing cells with two glycosylation inhibitors that do not disturb Golgi pH

homeostasis. ST3Gal III was selected as a marker protein, because it contains one

or two N-glycosylation sites and is mislocalized with the same concentrations of

Golgi pH dissipating drugs that cause inhibition of terminal N-glycosylation.

The selected inhibitors were benzyl-N-acetyl-α-galactosamine (BGN) and

swainsonine. BGN inhibits glycosylation by competing with natural substrates

and thus prevents elongation of hybrid and complex type N-glycans (Kuan et al. 1989). Swainsonine, in turn, inhibits mannosidase and thus prevents conversion

of hybrid type N-glycans to complex type glycans (Elbein et al. 1981). Neither of

the inhibitors, however, was able to cause mislocalization of ST3Gal III. The

same test was also performed with ST6Gal I, which is also N-glycosylated, but

the result was similar. These results therefore indicate that terminal glycosylation

is not an important determinant for the correct localization of ST3Gal III in the

Golgi.

5.7 The basis of the different pH sensitivity of the two sialyltransferases, ST3Gal III and ST6Gal I (II)

5.7.1 ST3Gal III and ST6Gal I localize to the trans-Golgi

Next, we studied whether the different pH sensitivity of two sialyltransferases is

dependent on their differential subcompartmentalization. Potential Golgi/TGN

localization difference was tested by using brefeldin A1 (BFA), which induces

91

accumulation of Golgi stack proteins, but not of TGN residents, in the ER. The

BFA treatment caused similar and relatively rapid accumulation of both ST3Gal

III and ST6Gal I in the ER (II, Fig. 7A) indicating that both them are Golgi stack

residents. In contrast, a TGN marker protein, TGN46, remained at the Golgi

region despite the drug treatment. We also noticed that the luminal pH values

measured with ST3Gal III-EGFP (pHG ST3 6.27 ± 0.20) and ST6Gal I-EGFP (pHG

ST6 6.29 ± 0.14, *p > 0.05) were nearly identical and typical for trans-Golgi lumen

(Paroutis et al. 2004), indicating a similar intra-Golgi localization, consistent with

earlier data (Berger et al. 1993, Burger et al. 1998, Grabenhorst & Conradt 1999).

Therefore, the observed differences in the pH sensitivity of the two

sialyltransferses are apparently not dependent on their differential localization

within the Golgi subcompartments.

5.7.2 Covalent oligomerization of ST6Gal I does not affect Golgi localization

Previous studies have shown that rat ST6Gal I is able to form disulphide-

mediated dimers (Ma & Colley 1996, Qian et al. 2001). The exact function of the

dimers is not known, but they may enchance Golgi localization, for instance,

according to the oligomerization / kin recognition hypothesis (Machamer 1991,

Nilsson et al. 1993b). To study whether the differential pH sensitivity of ST3Gal

III and ST6Gal I is dependent on different tendency to form covalent oligomers,

we analysed overexpressed sialyltransferase-EGFPs with SDS-PAGE and

Typhoon 9400 variable mode imager. Analyses revealed that ST3Gal III-EGFP

exist only as a 72 kDa monomer (45 kDa + 27 kDa; II, Fig. 7B left panel),

whereas ST6Gal I-EGFP was found to exist, similarly to its rat ortholog, as a 80

kDa monomer (52 kDa + 27 kDa) and a DTT sensitive 160 kDa dimer. The

proportion of the dimer, however, was only 15–20% of the total ST6Gal I protein.

To examine whether this disulphide mediated dimerization is responsible for the

apparent pH insensitivity of ST6Gal I compared to ST3Gal III, we mutated all

cysteines in ST6Gal I to prevent covalent dimerization. Consistent with earlier

studies (Qian et al. 2001), we found that only the cysteine 24 mediates covalent

dimerization (II, Fig. 7B middle panel). The mutation, however, had no effect on

Golgi localization in the presence or absence of Golgi pH elevating drugs (II, Fig.

7C).

92

5.7.3 Signals in the luminal domain of ST6Gal I are responsible for more stable Golgi localization in drug treated cells

According to earlier data the correct localization of ST6Gal I in the Golgi is

dependent on multiple signals that are distributed between the transmembrane and

the luminal domain (Colley 1997, Fenteany & Colley 2005). To study the effect

of the different domains on apparent pH insensitive localization, we generated

two chimeric proteins, one containing ST6Gal I transmembrane domain

(aminoacids 1-33) and ST3Gal III luminal domain (aminoacids 34–375) and the

other containing ST3Gal III transmembrane domain (aminoacids 1–34) and

ST6Gal I luminal domain (aminoacids 33–406). Consistent with earlier findings

from cysteine mutation studies (II, Fig. 7B middle panel), we found that the first

chimera (ST6tmST3) was able to form covalent dimers, thus providing additional

evidence that the ST6Gal I transmembrane domain is required for covalent

dimerization. Both chimeras, ST6tmST3 and the second chimera (ST3tmST6),

localized correctly in the Golgi when expressed in non-treated cells. In the

presence of 20µM CQ about 65% of the ST3tmST6 and 35% of the ST6tmST3

expressing cells showed predominant juxtanuclear localization of the chimeras (II,

Fig. 7C). This result therefore indicates that the luminal domain of ST6Gal I

contains dominant signals or interaction sites that are responsible for its more pH

insensitive localization in the Golgi.

5.8 Live cell PNA stainings (I)

5.8.1 Golgi pH in T-antigen expressing cells is higher than in non-

expressing cells (I)

T-antigen expressing cells were identified from GT-EGFP transfected MCF-7

cells with live cell staining using Alexa-594 conjugated PNA as described in the

materials and methods section. The median Golgi pH was about 0.3 units higher

in T-antigen expressing cells than in non-T-antigen expressing cells. The method,

however, decreased the steady state Golgi pH approximately by 0.3 units

compared to cells that were not stained with PNA. Therefore, the measured Golgi

pH was 6.42 ± 0.08 (median ± S.E.) in T-antigen expressing cells, 6.10 ± 0.06 in

non-T-antigen expressing cells and 6.05 ± 0.05 in COS-7 cells that were used as a

second control (I, Fig. 4A). The pH difference between the T-antigen expressing

93

and non-expressing cells corresponds roughly to a two-fold difference in lumenal

H+ concentration.

5.8.2 Golgi acidification is not impaired in T-antigen expressing MCF-7 cells

Golgi acidification is mainly dependent on the function of vacuolar-H+ATPase

(V-ATPase), which transports H+ from cytosol to the Golgi lumen (Grabe & Oster

2001, Paroutis et al. 2004). Because a reduced activity of V-ATPase would be

logical explanation for the increased Golgi pH in cancer cells, we tested the

V-ATPase activity in T-antigen expressing and non-expressing cells. The

measurements were performed according to Demaurex et al. (1998), and the pH

changes were monitored using transiently expressed GT-EGFP. T-antigen

expressing cells were identified with Alexa-594 conjugated PNA as described in

the materials and methods section. The measured acidification rates were nearly

identical (k -PNA≈ -1.91, k +PNA≈ -1.94) and both PNA-stained and unstained

MCF-7 cells were capable of acidifying the Golgi lumen to lower pH than

required for normal Golgi function (Fig. 11).

94

Fig. 11. ATP induced Golgi acidification in MCF-7 cells. Both T-antigen and non-

expressing cells are capable of acidifying the Golgi lumen below physiological pH at

almost identical rates. T-antigen expressing cells were identified with live-cell staining

using Alexa-594 conjugated PNA. After SLO permeabilization excess ATP was added

to the cells in high potassium medium and Golgi acidification was measured by the

intensity change in GT-EGFP fluorescence. T-antigen expressing cells are denoted

with grey boxes and non-expressing cells with black circles. Error bars indicate

standard deviation (S.D.). The straight lines show the initial acidification rates.

5.9 Intracellular adenosine nucleotide (AXP) concentrations in

MCF-7 cells indicate a good energy status

Adenosine triphosphate (ATP) powers V-ATPase mediated H+ transport from the

cytosol to the Golgi lumen (Grabe & Oster 2001, Paroutis et al. 2004), while

adenosine diphosphate (ADP) has been shown to inhibit it (Kettner et al. 2003).

Therefore, a poor cellular energy status with an increased amount of intracellular

ADP and reduced amount of intracellular ATP can diminish Golgi acidification.

To study this possibility more closely, we measured adenosine nucleotide levels in

COS-7 and MCF-7 cells with reverse phase chromatography according to Folley

et al. (1983). Adenosine nucleotides were extracted from the cells during their

exponential growth phase, i.e. before reaching 80% confluency. The measured

ATP, ADP and adenosine monophosphate (AMP) levels in MCF-7 cells (ATP:

10596 ± 760 nM; ADP: 1869 ± 160 nM and AMP: 794 ± 81 nM / 1000000 cells)

were comparable to those measured in COS-7 cells (ATP: 8890 ± 623 nM; ADP:

95

1346 ± 186 nM and AMP: 425 ± 39 nM / 1000000 cells) (Fig. 12). The ATP/ADP

ratios were 5.67 and 6.60 for MCF-7 and COS-7 cells, and adenylate energy

charges (AECs27) (Wiebe & Bancroft 1975), which are used to estimate overal

cellular energy status, were 0.896 and 0.870. AEC-value around 0.80–0.95

indicates a good energy status. The results indicate that both cell lines are in a

good energy status and that energy deficiency is an unlikely explanation for the

reduced Golgi acidification in cultured MCF-7 cells.

Fig. 12. Intracellular AXP concentrations in COS-7 and MCF-7 cells. Proteins from the

exponentially growing cells were precipitated with ice cold HClO4 and the supernatant

was analysed with reverse phase column. Both cell lines are viable and in a good

energy status. The AEC-values for COS-7 and MCF-7 cells were 0.896 and 0.870,

respectively.

5.10 AE2a localization within the Golgi (III)

AE2a has been previously shown to localize in the Golgi in some cells lines

(Holappa et al. 2001), but its exact distribution there has remained unclear. To

define that, we stained COS-7 cryo-sections prepared for electron microscopy

first with anti-AE2ct antibody (1st stock), then with secondary anti-rabbit

antibody and finally with gold-labeled protein A as described earlier (Holappa et

27 AEC = (ATP+0.5*ADP)/(ATP+ADP+AMP) (Wiebe & Bancroft 1975)

96

al. 2004, Sipilä et al. 2007). Quantification of the gold particles in the different

Golgi structures was performed according to Rabouille et al. (1995). Briefly, the

first and the last continuos cisternae were defined as the cis- and trans-Golgi and

the cisternae between them as the medial-Golgi. The TGN consists of mildly

swollen tubulo-reticular network next to the trans-Golgi cisternae. The relative

distribution of AE2a was calculated from counted numbers of gold particles in

each cisternae, the TGN, and cisternal rims, as well as in separate vesicles close

to the Golgi (III, Fig. 1). Based on the above quantification method approximately

80% of the endogenous AE2a is localized in the medial-/trans-Golgi and the TGN.

5.11 Role of AE2a in Golgi acidification (III)

5.11.1 AE2a overexpression and knock-down elevate Golgi pH (III)

AE anion exchangers localized at the plasma membrane participate in regulation

of cytosolic pH, Cl--concentration and cell volume (Romero et al. 2004, Alper

2006). To study whether Golgi localized AE2a (Kellokumpu et al. 1988, Alper et al. 1997, Stuart-Tilley et al. 1998, Holappa et al. 2001, Holappa et al. 2004) has a

similar role in the Golgi, we used overexpression and siRNA based protein

depletion. The Golgi pH in AE2a overexpressing and depleted COS-7 cells was

measured with GT-EGFP, which localizes to the medial-/trans-Golgi nearly

identically with AE2a (III, Fig. 1 & 2A). Five fold overexpression of myc-tagged

AE2a increased Golgi pH by ~0.3 pH-units, to 6.73 ± 0.11 (median ± S.E.; III,

Fig. 2C) and induced intra-Golgi T-antigen expression (III, Fig. 2E). In non- or

myc vector-transfected cells the measured Golgi pH was 6.35 ± 0.05 or 6.37 ±

0.06, respectively (*p < 0.05, Mann-Whitney U-test).

Depletion of AE2a to ~40% of its normal level (III, Fig. 3B) with a mixture

of four specific siRNAs increased Golgi pH to 6.51 ± 0.10 (III, Fig. 3C). Golgi

pH in cells treated with non-specific siRNA was 6.42 ± 0.06 (*p < 0.05; Mann-

Whitney U-test), indicating that the true Golgi pH increase achieved with AE2a

depletion was ~0.1 pH units. Longer exposure of the cells to siRNA treatments

induced more efficient AE2a depletion, but it was also more harmful to the cells

making reliable Golgi pH measurements unfeasible due to increased cell death

and detachment. A similar fall in cell viability has also been detected in AE2

knock-down studies performed with hepatocellular carcinoma cells (Hwang et al. 2009). Therefore, a too efficient and extended AE2 knock-down may harm the

97

cells so that the performance of Golgi pH measurements becomes impossible. In

conclusion, however, both overexpression and depletion of AE2a increase the

steady state Golgi pH, which provides evidence that AE2a has a role in Golgi pH

regulation.

5.11.2 AE2 participates in Golgi acidification by altering luminal

buffering capacity (III)

To understand how AE2a participates in Golgi pH regulation we performed a

series of Golgi acidification studies in different buffers. The acidification studies

were performed according to Demaurex et al. (1998) using GT-EGFP as a pH

sensor and SLO to permeabilize the plasma membrane.

The acidification was found to be ATP dependent (III, Fig. 4A) and to require

Cl- to proceed below physiological Golgi pH, i.e. below pHG ~6.0–6.3 (III, Fig.

4B). Chloride apparently functions here as a counter anion for H+ transport, and

helps to keep membrane potential low within the Golgi to enable undisturbed

V-ATPase function. In the absence or presence of limited counter ion conductance

the membrane potential increases and leads to inhibition of V-ATPase mediated

proton transport (Glickman et al. 1983).

The presence of 20 mM bicarbonate (HCO3-) was found to reduce Golgi

acidification, but interestingly only below the physiological Golgi pH (III, Fig.

4C). With 40mM HCO3- the reduction of Golgi acidification was more efficient,

but again was initiated after the lumenal pH was decreased below the

physiological Golgi pH. Together these results indicate that AE2a is capable of

importing HCO3- to the Golgi lumen and thus participates in Golgi pH regulation

by increasing luminal buffering capacity. The import, however, is not constantly

active, but is stimulated only when the Golgi lumen is acidified below the

physiological Golgi pH.

The active role of AE2a in HCO3- import was further supported by

mathematical estimations (Fig. 13). Calculations based on the Henderson-

Hasselbalch equation 28 and pKa ~6.3 of HCO3- showed that if no HCO3

- is

imported to the Golgi during acidification and the contribution of V-ATPase for

acidification is the same in all conditions (and equal to the actual measurements

performed in the absence of HCO3-, III, Fig. 4A), the acidification curve would be

28 Henderson-Hasselbalch equation: pH = pKa - log([HA]/[A-]

98

nearly linear instead of curve shaped, which is seen in the actual measurements

performed in the presence of HCO3- (III, Fig. 4C). The curved shape of the

acidification curve indicates that the HCO3- import increases in response to

luminal acidification. At the beginning of the acidification and just after ATP

addition the HCO3- import is thus lower than, for instance, after three minutes.

Fig. 13. Calculated HCO3- buffering in the absence of active HCO3

- import. Dashed

lines denote the actual representative acidification measurements performed in the

absence (black line) or presence of HCO3- (gray line). White dots, diamonds and boxes

represent calculated buffering in conditions where luminal HCO3- concentration is

constant and no HCO3- transport occurs. The time point zero represents the time of

ATP addition. Note the different shape of the curves as compared to the actual

measurements performed in the presence of HCO3-.

Acidification studies performed in the presence of 20 mM HCO3- using AE2a-

DsRed expressing cells showed that the initial acidification, as well as recovery

from ATP dependent acidification, occurs more rapidly in AE2a overexpressing

cells than in control cells (III, Fig. 5). This further supports the role of AE2a as a

HCO3- importer, but also indicates that AE2a may have an additional role in the

acidification above physiological Golgi pH.

Further support for the function of AE2a in Golgi acidification was obtained

from a chemical knock-down, performed using H2DIDS (4,4'-

diisothiocyanatodihydrostilbene-2,2'-disulfonic acid). We found that the presence

of 200µM H2DIDS reduces Golgi acidification in a similar fashion to Cl-

depletion, supporting a potential contribution of AE2a to Cl- influx and HCO3-

99

efflux (III, Fig. 6). Together with the result that the initial acidification occurs

more rapidly in AE2a overexpressing cells, this result suggests that AE2a may

have a dual role in the Golgi. Therefore, in addition to the capacity to reduce

excess acidification of the Golgi lumen by importing HCO3- and exporting Cl-,

AE2a may also participate in Golgi acidification above the physiological Golgi

pH by importing Cl- and exporting HCO3-. To be able to do that AE2a should be

able to change the direction of the HCO3-/Cl- exchange. By comparing Golgi

acidification curves that were converted to a H+ concentration scale instead of pH

scale, we estimated that this point would be at pHG ~6.0. At that point the slope of

the tangent function describing Golgi acidification in the presence of 20mM

HCO3- starts to decrease compared to the tangent function describing acidification

in the absence of HCO3- (data not shown). Before that point the slopes of the

tangent functions are equal or the slope of the tangent function describing Golgi

acidification in the presence of 20mM HCO3- is higher.

100

101

6 Discussion

6.1 Increased Golgi pH in cancer correlates with increased

T-antigen expression

Consistent with earlier hypothesis (Campbell et al. 2001a, Kellokumpu et al. 2002) we have shown in this work that Golgi pH is elevated in several cancer

cells lines and that the elevation of Golgi pH correlates with increased T-antigen

expression inspite of a similar or different genetic background (I, Fig. 4). Highly

T-antigen expressing MCF-7, SW-48 and HT-29 cancer cells were found to have a

generally higher Golgi pH than non-expressing cancer or control cells. A similar

trend was also noted within a single cell line i.e. cells having the same genetic

origin. Highly T-antigen expressing MCF-7 cells were found to have generally

higher Golgi pH than non-expressing cells.

The Golgi pH elevation, however, is not the only potential factor that can

induce T-antigen expression, but also, for example, alterations in the folding

(Schietinger et al. 2006), substrate availability (Kumamoto et al. 2001) or the

expression levels of the corresponding glycosylation enzymes (Whitehouse et al. 1997, Brockhausen 1999, Burchell et al. 1999, Dalziel et al. 2001) have been

associated with increased T-antigen expression. Our data do not exclude a

contribution of these or other potential factors to the glycosylation in the cell lines

used in our study, but rather shows that despite the evident correlation between

elevated Golgi pH and increased expression of T-antigen, there exist some cell

type dependent factors that cause different T-antigen expression in the cell lines

with almost identical Golgi pH profiles (compare e.g. MCF-7, HT-29 and SW-48

cells; I, Fig. 1B & 4B). The contribution of these additional factors affecting

glycosylation, however, was not assessed in this study.

6.2 Terminal processing of O- and N-glycans is highly sensitive to

Golgi pH alterations

Neutralization/alkalinization of mildly acidic Golgi pH with various drugs is

known to cause alterations in glycosylation (Thorens & Vassalli 1986, Gawlitzek et al. 2000, Axelsson et al. 2001, Campbell et al. 2001a, Kellokumpu et al. 2002).

In this work, however, we showed that Golgi resident glycosylation processes are

much more sensitive to Golgi pH changes than previously thought. We found that

102

only a 0.2 unit increase in Golgi pH in COS-7 cells is sufficient to inhibit terminal

O- and N-glycosylation. The perturbation of terminal O-glycosylation was

detected as an increase of intra-Golgi and cell surface T-antigen expression (I, Fig.

3) and the perturbation of the terminal N-glycosylation was detected as a

decreased level of α2,3-sialylation (II, Fig. 2B & C). These data further support

the finding that the Golgi pH elevation detected in cancer cells (~0.3-0.4 pH units;

I, Fig. 1B) is sufficient to impair terminal O- and N-glycosylation processes.

Analogous and hence supporting results were also obtained from the AE2

overexpression studies. The overexpression of AE2 was found to increase Golgi

pH by ~0.3-0.4 pH units (III, Fig. 2C) and concurrently induce T-antigen

expression (III, Fig. 2E).

Recently, others have also noted the high pH sensitivity of the Golgi resident

glycosylation processes. Maeda and coworkers (Maeda et al. 2008) reported that

depletion of functional GPHR protein leads to a similar Golgi pH elevation to the

observations from our Golgi pH dissipation experiments (i.e. ~0.3–0.4 pH units).

This is consistent with our findings and is sufficient to inhibit both O- and

N-glycosylation.

6.3 Glycosylation processes are differently sensitive to Golgi pH alterations

In addition to high pH sensitivity of terminal glycosylation processes, we note

that different glycosylation processes are differently sensitive to Golgi pH

alterations. Previously, distinction of the pH sensitivities has been noted only

between cis-Golgi resident proximal and trans-Golgi resident terminal

glycosylation processes (Gawlitzek et al. 2000, Axelsson et al. 2001), but our data

show that terminal glycosylation can also be differently sensitive to Golgi pH

alterations. We found that α2,3-sialylation can be inhibited with lower Golgi pH

increase than, for example, α2,6-sialylation or β1,4-galactosylation (II, Fig. 2B &

C).

Especially the finding that the pH sensitivities of the two terminal sialylation

reactions are different, is very intriguing, because it may explain one rather

common colon cancer associated glycosylation alteration. CEA extracted from

colon cancer patients is often characterized by a reduced level of α2,3-sialylation

and an increased level of α2,6-sialylation (Kobata et al. 1995). Therefore, it

would be interesting to study whether a mere Golgi pH elevation is able to induce

this colon cancer associated sialylation change in cells originating from colon

103

tissue. In COS-7 cells, however, Golgi pH elevation reduced α2,3-sialylation of

sCEA, but did not induce its α2,6-sialylation (II, Fig. 2B & C).

6.4 Inhibition of α2,3-sialylation involves mislocalization of the corresponding sialyltransferase, ST3Gal III

Glycosylation alterations induced by Golgi pH elevation are thought to derive

from a nonoptimal luminal pH for enzyme activities (Gawlitzek et al. 2000),

reduced availability of the sugar nucleotides (Waldman & Rudnick 1990) or

mislocalization of the corresponding glycosylation enzymes (Axelsson et al. 2001). Currently, however, it is not known which one of these is the primary

affecting factor that takes place upon the lowest Golgi pH elevation and initially

causes inhibition of glycosylation. This uncertainty is partly due to the fact that

most of the studies in which the effects of Golgi pH elevation are studied, are

performed using different drugs and drug concentrations without measuring the

exact Golgi pH. For that reason in this work we measured the drug induced

elevation of Golgi pH and tested, by elevating Golgi pH in a step-wise manner,

what kind of alterations in the Golgi are associated with initial inhibition of

glycosylation. We found that the first effect detected simultaneously with Golgi

pH elevation induced inhibition of α2,3-sialylation is the mislocalization of the

corresponding α2,3-sialyltransferase, ST3Gal III (II, Fig. 1A & 2B). At the same

time α2,6-sialylation was practically unaltered (II, Fig. 2B), which excludes the

possibility that availability of the sugar nucleotide, CMP-Sia, is reduced. The

effect of the pH shift to potentially less optimal Golgi pH for ST3Gal III activity

was not studied in this work, but according to Brenda enzyme database

(http://www.brenda.uni-koeln.de/) the pH optimum for human ST3Gal III peaks

at pH 6.7. Therefore, the CQ induced Golgi pH elevation from 6.3 to 6.5–6.7

should rather increase than inhibit α2,3-sialylation. However, our data does not

exclude the possibility that other factors than the mislocalization can participate

in inhibition of glycosylation when more potent drugs or drug concentrations are

used.

6.5 ST3Gal III mislocalization

By utilizing Golgi pH measurements and CaCO-2 cells the mislocalization of

ST3Gal III was confirmed to be dependent on Golgi pH elevation and not, for

instance, on elevated endosomal pH. The Golgi pH measurements showed that

104

mislocalization of ST3Gal III is tightly associated with luminal pH elevation in

the juxta-nuclear region (II, Fig. 4) and the tests with CaCO-2 cells showed that

ST3Gal III remains mainly in the Golgi despite the endosomal acidification defect

in these cells (II, Fig. 6).

ST3Gal III, however, is not the only glycosyltransferase whose

mislocalization has been associated with Golgi pH elevation. Previously,

GalNAcT2, NAGT1 and GalT have also been shown to mislocalize to the plasma

membrane or to undefined vesicular structures when the mildly acidic Golgi pH is

neutralised with BafA1 or NH4Cl (Axelsson et al. 2001). Additionally, monensin,

an ionophore that allows free exchange of Na+ and H+ through membranes, has

been shown to cause signal mediated redistribution of GalT1, but not of ST6Gal I,

to the TGN (Schaub et al. 2006). Together these data show that Golgi pH

elevation induced mislocalization of Golgi glycosyltransferases is a rather

common event. However, because the underlying mechanisms, transport

pathways as well as destinations involved in the pH dependent mislocalization are

poorly known, we studied the mislocalization more closely using ST3Gal III as a

model glycosyltransferase.

We found that during CQ induced Golgi pH elevation ST3Gal III

mislocalizes predominantly to LAMP-2 containing vesicles (II, Fig. 3). The

luminal pH in these vesicles is around ~6.0 (II, Fig. 4), which together with the

presence of LAMP-2 strongly indicates that the vesicles are late

endosomes/lysosomes or at least of endosomal/lysosomal origin. We also found

that during the drug treatments ST3Gal III behaves similarly to a TGN marker,

TGN46 (II, Fig. 1A), which is known to recycle between the TGN, the plasma

membrane and endosomes (Chapman & Munro 1994, Bonifacino & Rojas 2006).

Further studies revealed that the similar behaviour of ST3Gal III and TGN46 in

drug treated cells was not limited to the accumulation of both proteins to the

LAMP-2 containing vesicles, but also ST3Gal III, similarly to TGN46, is recycled

via the plasma membrane in drug treated cells (II, Fig. 5). Normally, i.e. in cells

that have normal Golgi pH, ST3Gal III is not recycled via the plasma membrane

(II, Fig. 5). Due to methodological limitations, however, it remained unclear

whether ST3Gal III is first transported to the plasma membrane and then to the

LAMP-2 containing vesicles or vice versa.

Nevertheless, the mislocalization of ST3Gal III is a phenomenon that

involves disturbance of two distinct processes. The disturbance of the first one is

caused by Golgi pH elevation, which either inhibits active retrieval or releases

some constraints that normally retain ST3Gal III in the Golgi. The disturbance of

105

the second process, in turn, is caused by an elevation of endosomal pH, which

inhibits retrieval of mislocalized ST3Gal III from the plasma membrane and

endosomes back to the Golgi.

6.6 Different pH sensitivities of ST3Gal III and ST6Gal I are mainly dictated by their luminal domains

Domain swapping mutagenesis revealed that the difference in the pH sensitivity

of the two terminal glycosyltransferases, ST3Gal III and ST6Gal I, is mainly

dictated by their luminal domains (II, Fig. 7C). At this point, however, it is

unclear what features in the luminal domain of ST6Gal I make ST6Gal I or a

chimera carrying its luminal domain more insensitive to Golgi pH elevation

compared to ST3Gal III. One possibility is that the luminal domain of ST6Gal I

contains some signals or interaction sites that, for instance, interact with similar

kind of molecules according to the oligomerization / kin recognition hypothesis

(Machamer 1991, Nilsson et al. 1993b). These interactions, however, must be

relatively weak and non-covalent, because SDS-PAGE analysis in non-reducing

conditions did not reveal any interactions associated with the luminal domain of

ST6Gal I (II, Fig. 7B). Another completely opposite possibility is that, instead of

a stabilizing effect of the luminal domain of ST6Gal I, the luminal domain of

ST3Gal III could somehow destabilize Golgi localization of the protein upon

Golgi pH elevation. Luminal pH elevation may, for instance, induce specific or

unspecific association of ST3Gal III with Golgi/TGN associated aggregates,

which have been suggested to participate in protein sorting (Borgonovo et al. 2006). Non-reducing SDS-PAGE of samples prepared from CQ treated cells

having elevated Golgi pH, however, did not reveal any signs of ST3Gal III

aggregation (data not shown). This either indicates that the aggregates do not

form or that they are, similar to the suggested interactions that may stabilize the

localization of ST6Gal I luminal domain, too weak to be detected using SDS-

PAGE analysis. A better understanding on this issue, however, awaits recognition

of the luminal proteins that interact with ST6Gal I or ST3Gal III in normal and/or

elevated Golgi pH. This could be possible, for instance, with methods such as

BiFC (Bimolecular fluorescence complementation) or FRET (Förster resonance

energy transfer) that allow detection of molecular interactions in live cells.

Despite the central roles of the luminal domains in pH sensitive Golgi

localization, it is important to note that also the transmembrane domain of

ST6Gal I is able to stabilize Golgi localization (II, Fig. 7C). Its ability to enhance

106

Golgi localization during luminal Golgi pH elevation is currently unclear but may

be related to the different lengths of the membrane spanning sequences of ST6Gal

I and ST3Gal III. ST6Gal I has a three amino acids shorter membrane spanning

sequence and may thus be directed, according to the lipid bilayer sorting model

(Bretscher & Munro 1993, Masibay et al. 1993), to thinner membrane domains

than ST3Gal III. The thinner membrane domains are thought to be less active as

Golgi export sites than thicker membrane domains29 and thus to provide more

stable Golgi residence. Confirmation of the involvement of lipid bilayer sorting,

however, would require genetic manipulation of the lengths of the membrane

spanning sequences.

Another potential mechanism that could contribute to Golgi localization by

utilizing the transmembrane domain was, in turn, excluded. Cysteine mediated

covalent dimerization, which has been previously associated with the

transmembrane domain (cysteine 24) and improved Golgi retention (Ma & Colley

1996, Qian et al. 2001), was found to be unimportant for Golgi localization of

ST6Gal I in non- or drug-treated COS-7 cells. The mutant lacking cysteine 24 and

being thus incapable of forming covalent dimers, was found to be equally

insensitive to Golgi pH elevation as the wild type ST6Gal I (II, Fig. 7C).

6.7 Dissipation of the acidic Golgi pH is able to inhibit α2,3-sialylation despite ST3Gal III overexpression

Glycosylation alterations in cancer correlate sometimes with increased or reduced

expression of corresponding glycosylation enzymes (Whitehouse et al. 1997,

Brockhausen 1999, Burchell et al. 1999, Dalziel et al. 2001), but often such

correlation is inconsistent or totally lacking (Yang et al. 1994, Brockhausen et al. 1995, Hanisch et al. 1996, Lloyd et al. 1996, Taylor-Papadimitriou et al. 1999,

Müller & Hanisch 2002). Our results show that one potential factor that can cause

this phenomenon is elevated Golgi pH. We found that dissipation of the mildly

acidic Golgi pH is able the overrun the effect of ST3Gal III overexpression. In

cells that have a normal Golgi pH, transient ST3Gal III overexpression increases

α2,3-sialylation roughly 3.6 fold (Fig. 10A), but when the ST3Gal III

overexpressing cells are treated with a sufficient load of CQ (i.e. > 40µM), α2,3-

sialylation can be diminished to a lower level than in non-ST3Gal III

29 i.e. lipid rafts

107

overexpressing cells (Fig. 10B). This result further emphasizes the role of

properly regulated Golgi pH for succesful glycosylation, and also explains why

glycosyltransferase expression levels, even in the case of overexpression, do not

always correlate with detected glycosylation alterations.

Some glycosylation enzymes, however, can be practically insensitive to Golgi

pH alterations. Obviously many cis-Golgi resident proximal glycosylation

enzymes belong to this group, since, for instance, acquiring Endo H sensitivity of

N-glycans is not remarkably affected by the drug treatments that elevate Golgi pH

(II, Fig. 2A).

6.8 Cancer cell lines are often characterized by heterogeneous cell surface glycan expression

In addition to the elevated and highly variable Golgi pH, many cancer cell lines

are characterized by heterogeneous cell surface T-antigen expression (Fig. 9). The

PNA staining of such cells clearly shows that even sister cells, growing side by

side in the same colony, can express T-antigen at different levels. The finding is

an intriguing demonstration how differently the cancer cells can behave and how

different they can appear, even if they share the same genetic origin and the same

growth environment and stimuli. The same trend can be seen also in the stainings

of histological cancer sections (e.g. colorectal carcinoma in article I, Fig. 4C),

although the actual tumors may also contain non-malignant cells. At this point,

however, it is unclear how the heterogeneity is primarily induced and whether it

represents a constantly changing mosaic or a static state. This heterogeneity

apparently reflects a general cellular regulation defect, which somehow also

impairs Golgi pH regulation and processing of glycans either directly or indirectly.

A general cellular regulation defect can, for instance, cause alterations in the

expression levels of Golgi pH regulators, glycosylation enzymes and glycosylated

proteins that can all affect glycosylation. Another issue that remains to be solved

in the future is the physiological significance of the heterogeneous cell surface

glycan expression. Does it, for instance, enhance metastasis or protect cancer cells

from the immunodefence?

6.9 Golgi pH regulation in health and disease

Mechanisms involved in Golgi pH regulation (Chapter 2.5) are not fully

understood, but according to current knowledge the machinery mediating Golgi

108

pH regulation is composed of three main components: proton import to the Golgi

lumen, proton export/elimination from the Golgi lumen and counterion

conductance (Kim et al. 1996, Demaurex et al. 1998, Farinas & Verkman 1999,

Grabe & Oster 2001, Wu et al. 2001, Demaurex 2002, Paroutis et al. 2004).

Additionally, some other components such as buffering, membrane traffic and

activities of transiting cargo proteins can affect Golgi pH homeostasis (Paroutis et al. 2004). Each of these components can be mediated by a single protein or

protein complex, but there may also be several mediators with partially

overlapping functions. Therefore, to solve why Golgi pH is elevated and more

variable in cancer cells than in non-malignant control cells is an extremely

difficult task. One thing that makes the task even more complicated is the fact that

also the mechanisms that impair Golgi pH regulation can be different in different

cancer types. In this work, however, we focused the study on the function of

V-ATPase in differentially T-antigen expressing MCF-7 cells. We also studied the

function of another potential Golgi localizing pH regulator, AE2a, which is

variably expressed in many cancer cell lines (data not shown).

6.9.1 V-ATPase is present and functional in the Golgi of MCF-7 cells

Previously, it has been shown that the Golgi acidification is mainly mediated by

V-ATPase (Grabe & Oster 2001, Paroutis et al. 2004). Additionally, alterations in

the localization and the expression of V-ATPase subunits are quite commonly

associated with cancer (Sennoune et al. 2004a, Sennoune et al. 2004b). Therefore,

one potential reason for the cancer associated Golgi pH elevation is decreased

presence of functional V-ATPase in the Golgi. The number of Golgi resident

V- H+ATPases, for instance, can be lowered due to a defective sorting or

trafficking or the Golgi resident V-ATPases may be inactivated by some

regulatory mechanism. Our studies, however, show that the Golgi lumen in SLO-

permeabilized MCF-7 cells can be rapidly acidified below normal physiological

value (i.e. pHG < 6.0–6.3) after addition of ATP into the medium (Fig. 11).

Additionally, there were no remarkable differences in the rates of ATP induced

Golgi acidification between T-antigen expressing and non-expressing MCF-7

cells, although the initial steady state Golgi pH of the T-antigen expressing cells is

generally higher. These results thus demonstrate that despite the level of T-antigen

expression and Golgi pH elevation, both MCF-7 cell types have sufficient amount

of functional V-ATPase on their Golgi membranes to maintain normal Golgi

acidity.

109

However, the question whether V-ATPase activity on the Golgi membranes is

somehow down-regulated in intact T-antigen expressing cells, is a more difficult

issue to assess because most of the regulatory mechanisms (Chapter 2.5.1:

Regulation of V-ATPase activity) are reversible and may be highly sensitive to

cytosolic changes. The procedure used in the acidification measurements requires

replacement of the cytosol with an artificial buffer and can thus easily affect, for

instance, RAVE and disulfide bond dependent regulation. The possibility of

different coupling i.e. that V-ATPases are composed of different subunits in

T-antigen expressing cells, can be excluded with quite a high confidence, because

alterations in the coupling are generally relatively slow and at least partly

irreversible. Assembled V-ATPase domains, V0 and V1, are more likely

proteolytically degraded than disassembled for assembly of a new complex.

Results from the acidification studies additionally show indirectly that the

counterion conductance or the proton leakage in the Golgi of T-antigen expressing

MCF-7 cells are not remarkably altered. Otherwise, there would be detectable

differences in the shapes of acidification curves of differently T-antigen

expressing cells. Increased proton leakage and reduced counterion conductance,

for instance, would be capable of reducing the rate of acidification.

6.9.2 Energy status do not limit Golgi acidification in cultured MCF-7 cells

Another factor that can prevent proper Golgi acidification in cancer cells is a poor

cellular energy status, because the capability of V-ATPase to transfer protons from

the cytosol to the Golgi lumen is dependent the on presence of ATP and inhibited

by ADP (Kettner et al. 2003). However, we found that cultured MCF-7 cells are

not energy deficient. Their adenylate energy charge (AEC) was 0.870 (Fig. 12),

which indicates a very good cellular energy status. The Golgi pH elevation is thus

not caused by a catabolic or metabolic defect that impairs the production of ATP,

and hence also the activity of V-ATPase. Our studies performed with cultured

cells, however, do not exclude the possibility that energy deficiency reduces

Golgi acidification in tumors. Even though the cultured cancer cells are capable

of using available energy efficiently, as our studies indicate, tumors may suffer

from energy deficiency due to a too high growth rate of the cancer cells, and

insufficiently developed blood vessels that are unable to supply rapidly

multiplying cells with appropriate amounts of nutrients and oxygen.

110

Due to the limitations of the available methods, we were unable to test here

the potential differences in the energy status between highly T-antigen expressing

and non-expressing cells. However, the percentage of cells with increased Golgi

pH (i.e. pHG > 6.75) is so high (52%) that one should observe a notable

decrement in the AEC-value of a cultured MCF-7 population, if the poor energy

status is the main factor that elevates Golgi pH and increases T-antigen expression.

Since such decrement could not be noted in the AEC, the energy deficiency in

culture conditions appears to be very unlikely reason for the elevated Golgi pH

and increased T-antigen expression.

6.9.3 Role of AE2a in Golgi pH regulation

Since the plasmalemmal AE2a participates in the regulation of cytosolic pH,

chloride concentration and cell volume (Romero et al. 2004, Alper 2006), it has

been suggested that also the Golgi resident AE2a may have a similar role in the

Golgi. Consistent with this suggestion, our results show that the Golgi resident

AE2a participates in Golgi pH regulation by altering luminal buffering capacity.

At this point, however, the mechanism is not fully understood, but it involves at

least HCO3--import/Cl--export in response to luminal acidification (III, Fig. 4).

One very interesting characteristic of the HCO3- import is that it can be detected

only at or below the physiological Golgi pH (i.e. pHG < 6.0–6.3) regardless of the

HCO3- concentration in the extra-Golgi buffer (III, Fig. 4C) or the expression

level of AE2a on the Golgi membranes (III, Fig. 5).

Additionally, the Golgi resident AE2a may also participate in acidification

above physiological Golgi pH by mediating HCO3--export/Cl--import. This

suggestion is based on findings that overexpression of AE2a enchances initial

acidification (III, Fig. 5) and that Golgi acidification can be inhibited by using

H2DIDS, a highly specific inhibitor against AE2a (III, Fig. 6). Recently, however,

the specificity of stilbenesulphonates to anion exchangers has been challenged by

showing that other proteins can also be inhibited by them. One such protein is a

recently characterized GPHR, which may also participate in Golgi pH regulation

(Maeda et al. 2008). We also understand that some caution is necessary in the

interpretation of these results, since the numbers of successful pH recordings are

for technical reasons relatively low.

Despite these uncertainties related to specificity of H2DIDS, and the limited

number of successful pH recordings, our findings indicate that Golgi resident

AE2a is not constantly active, and may be capable of mediating bidirectional

111

HCO3-/Cl- exchange. Therefore, the Golgi resident AE2a must have a specific

mechanism for sensing luminal acidification, and activating HCO3- import, or

switch from HCO3- export to import. Consistent with this possibility, the

plasmalemmal AE2a has been shown to be able to sense pH in the extracellular

space, which corresponds topologically to the Golgi lumen, with specific regions

in its transmembrane domain (Stewart et al. 2007a, Stewart et al. 2007b). The

plasmalemmal AE2a, however, has a pHo(50) value ~6.9 (Stewart et al. 2007b),

which indicates that the plasmalemmal AE2a is active mainly when extra-cellular

pH is near neutral or alkaline. Our results, however, show that the HCO3- import

mediated by Golgi resident AE2a is in normal conditions activated at or below

pHG 6.1. Therefore, if the same regulatory mechanism is used in both cases, it

must be somehow modulated to be able to sense different extracellular/luminal

pH values. Such modulation could, for instance, involve phosphorylation or non-

covalent molecular interactions with other Golgi or plasma membrane resident

proteins.

The involvement of AE2a activity in Golgi pH regulation also provides an

explanation for the uncharacterized proton export. Luminal HCO3- reacts

efficiently with H+ in acidic conditions and forms carbonic acid. Since carbonic

acid is highly unstable in aqueous conditions, it is rapidly dissociated into water

and carbon dioxide that can pass lipid membranes via aquaporin water channels

or by diffusion, respectively. Therefore, the proton export from the Golgi, or the

“proton leakage pathway”, does not necessarily require any proton specific

channels or exchangers.

In the future it would be interesting to study the contribution of AE2a to

Golgi pH regulation in cancer. According to the preliminary data, AE2a

expression is variable between different cancer cell lines (data not shown), and

both its overexpression and depletion can cause Golgi pH elevation. Confirmation

of the involvement of AE2 in Golgi pH regulation in cancer, however, would

require authentic controls (i.e. non-malignant cells from the same tissue and the

same patient). Otherwise the normal localization and tissue specific levels of

AE2a expression cannot be reliably defined.

112

113

7 Conclusions

In this work we have shown that cancer cells have often more variable and higher

Golgi pH than cells of non-cancerous origin. The Golgi pH elevation was also

found to correlate with increased expression of a typical cancer associated glycan

epitope, Thomsen-Friedenreich antigen (T-antigen / Core-1 O-glycan structure,

Gal-β1,3-GalNAc-Ser/Thr). The importance of Golgi pH regulation for successful

glycosylation was further evidenced with a variety of Golgi pH gradient

dissipating drug treatments. Even a relatively small Golgi pH increase, ~0.2 pH

units, is enough to impair terminal processing of both O- and N-linked glycans.

Additionally, we show in this work that terminal glycosylation processes can

be differentially sensitive to Golgi pH elevation. Terminal α2,3-sialylation was

found to be more sensitive to drug treatments than α2,6-sialylation due to the high

tendency of the corresponding α2,3-sialyltransferase, ST3Gal III, to mislocalize

after minor Golgi pH elevation. Higher Golgi pH elevation was required for

mislocalization of α2,6-sialyltransferase, ST6Gal I, and other tested Golgi

resident glycosyltransferases. The different sensitivities of the two

sialyltransferases to Golgi pH elevation were found to be mainly dictated by

currently unknown properties within the luminal domains.

In this work we also show that, despite the more variable and elevated Golgi

pH, the cellular energy status and the ability of V-ATPase to acidify the Golgi

lumen can both be normal in cancer cells. Therefore, the cancer associated Golgi

pH alterations may derive from a general cellular regulation defect, which may

lead, for instance, to an abnormal expression of a protein or proteins participating

in Golgi pH regulation. Related to this, we found that a Golgi localized HCO3-/Cl-

exchanger, AE2a, can induce Golgi pH alterations when over- or underexpressed.

Our data shows that AE2a participates in Golgi pH regulation by altering luminal

buffering capacity. The import of HCO3- increases luminal buffering and thus

reduces Golgi acidification, whereas the HCO3- export reduces buffering capacity

and thus helps acidification. Involvement of AE2 mediated HCO3- transport in

Golgi pH regulation also provides a novel mechanism to expel protons from the

Golgi lumen. The more acidic the luminal pH is, the more efficiently HCO3-

reacts with excess H+ to form water and CO2. Therefore, the elimination of excess

H+ from the Golgi lumen does not require any specific H+ channels or exporters.

The role of AE2a in the misregulation of Golgi pH in cancer cells, however,

remains to be elucidated.

114

115

References

Abeijon C & Hirschberg CB (1992) Topography of glycosylation reactions in the endoplasmic reticulum. Trends Biochem Sci 17(1): 32–36.

Allan VJ, Thompson HM & McNiven MA (2002) Motoring around the Golgi. Nat Cell Biol 4(10): E236–242.

Alper SL, Stuart-Tilley AK, Biemesderfer D, Shmukler BE & Brown D (1997) Immunolocalization of AE2 anion exchanger in rat kidney. Am J Physiol 273(4 Pt 2): F601–14.

Alper SL (2006) Molecular physiology of SLC4 anion exchangers. Exp Physiol 91(1): 153–161.

Anderson RG & Pathak RK (1985) Vesicles and cisternae in the trans Golgi apparatus of human fibroblasts are acidic compartments. Cell 40(3): 635–643.

Antonny B, Madden D, Hamamoto S, Orci L & Schekman R (2001) Dynamics of the COPII coat with GTP and stable analogues. Nat Cell Biol 3(6): 531–537.

Aoki D, Lee N, Yamaguchi N, Dubois C & Fukuda MN (1992) Golgi retention of a trans-Golgi membrane protein, galactosyltransferase, requires cysteine and histidine residues within the membrane-anchoring domain. Proc Natl Acad Sci U S A 89(10): 4319–4323.

Appenzeller-Herzog C & Hauri HP (2006) The ER-Golgi intermediate compartment (ERGIC): in search of its identity and function. J Cell Sci 119(Pt 11): 2173–2183.

Apweiler R, Hermjakob H & Sharon N (1999) On the frequency of protein glycosylation, as deduced from analysis of the SWISS-PROT database. Biochim Biophys Acta 1473(1): 4–8.

Arai H, Terres G, Pink S & Forgac M (1988) Topography and subunit stoichiometry of the coated vesicle proton pump. J Biol Chem 263(18): 8796–8802.

Atassi G & Tagnon HJ (1975) R17934-NSC 238159: a new antitumor drug--I. Effect on experimental tumors and factors influencing effectiveness. Eur J Cancer 11(9): 599–607.

Axelsson MA, Karlsson NG, Steel DM, Ouwendijk J, Nilsson T & Hansson GC (2001) Neutralization of pH in the Golgi apparatus causes redistribution of glycosyltransferases and changes in the O-glycosylation of mucins. Glycobiology 11(8): 633–644.

Bachert C, Lee TH & Linstedt AD (2001) Lumenal endosomal and Golgi-retrieval determinants involved in pH-sensitive targeting of an early Golgi protein. Mol Biol Cell 12(10): 3152–3160.

Bailly E, McCaffrey M, Touchot N, Zahraoui A, Goud B & Bornens M (1991) Phosphorylation of two small GTP-binding proteins of the Rab family by p34cdc2. Nature 350(6320): 715–718.

Bannykh SI, Rowe T & Balch WE (1996) The organization of endoplasmic reticulum export complexes. J Cell Biol 135(1): 19–35.

Bannykh SI & Balch WE (1997) Membrane dynamics at the endoplasmic reticulum-Golgi interface. J Cell Biol 138(1): 1–4.

116

Barasch J, Kiss B, Prince A, Saiman L, Gruenert D & al-Awqati Q (1991) Defective acidification of intracellular organelles in cystic fibrosis. Nature 352(6330): 70–73.

Barasch J & al-Awqati Q (1993) Defective acidification of the biosynthetic pathway in cystic fibrosis. J Cell Sci Suppl 17: 229–233.

Barlowe C & Schekman R (1993) SEC12 encodes a guanine-nucleotide-exchange factor essential for transport vesicle budding from the ER. Nature 365(6444): 347–349.

Barlowe C, Orci L, Yeung T, Hosobuchi M, Hamamoto S, Salama N, Rexach MF, Ravazzola M, Amherdt M & Schekman R (1994) COPII: a membrane coat formed by Sec proteins that drive vesicle budding from the endoplasmic reticulum. Cell 77(6): 895–907.

Barlowe C (2003) Signals for COPII-dependent export from the ER: what's the ticket out? Trends Cell Biol 13(6): 295–300.

Barr FA, Puype M, Vandekerckhove J & Warren G (1997) GRASP65, a protein involved in the stacking of Golgi cisternae. Cell 91(2): 253–262.

Barr FA, Preisinger C, Kopajtich R & Korner R (2001) Golgi matrix proteins interact with p24 cargo receptors and aid their efficient retention in the Golgi apparatus. J Cell Biol 155(6): 885–891.

Baum LG, Derbin K, Perillo NL, Wu T, Pang M & Uittenbogaart C (1996) Characterization of terminal sialic acid linkages on human thymocytes. Correlation between lectin-binding phenotype and sialyltransferase expression. J Biol Chem 271(18): 10793–10799.

Bause E (1983) Structural requirements of N-glycosylation of proteins. Studies with proline peptides as conformational probes. Biochem J 209(2): 331–336.

Beams HW & Kessel RG (1968) The Golgi apparatus: structure and function. Int Rev Cytol 23: 209–276.

Beck KA, Buchanan JA, Malhotra V & Nelson WJ (1994) Golgi spectrin: identification of an erythroid beta-spectrin homolog associated with the Golgi complex. J Cell Biol 127(3): 707–723.

Beck KA, Buchanan JA & Nelson WJ (1997) Golgi membrane skeleton: identification, localization and oligomerization of a 195 kDa ankyrin isoform associated with the Golgi complex. J Cell Sci 110 ( Pt 10)(Pt 10): 1239–1249.

Beck KA (2005) Spectrins and the Golgi. Biochim Biophys Acta 1744(3): 374–382. Becker B & Melkonian M (1996) The secretory pathway of protists: spatial and functional

organization and evolution. Microbiol Rev 60(4): 697–721. Bennett V & Stenbuck PJ (1979) Identification and partial purification of ankyrin, the high

affinity membrane attachment site for human erythrocyte spectrin. J Biol Chem 254(7): 2533–2541.

Bennett V (1992) Ankyrins. Adaptors between diverse plasma membrane proteins and the cytoplasm. J Biol Chem 267(13): 8703–8706.

Benting JH, Rietveld AG & Simons K (1999) N-Glycans mediate the apical sorting of a GPI-anchored, raft-associated protein in Madin-Darby canine kidney cells. J Cell Biol 146(2): 313–320.

117

Berger EG, Mandel T & Schilt U (1981) Immunohistochemical localization of galactosyltransferase in human fibroblasts and HeLa cells. J Histochem Cytochem 29(3): 364–370.

Berger EG, Grimm K, Bachi T, Bosshart H, Kleene R & Watzele M (1993) Double immunofluorescent staining of alpha 2,6 sialyltransferase and beta 1,4 galactosyltransferase in monensin-treated cells: evidence for different Golgi compartments? J Cell Biochem 52(3): 275–288.

Bethune J, Wieland F & Moelleken J (2006) COPI-mediated transport. J Membr Biol 211(2): 65–79.

Bhakdi S, Weller U, Walev I, Martin E, Jonas D & Palmer M (1993) A guide to the use of pore-forming toxins for controlled permeabilization of cell membranes. Med Microbiol Immunol 182(4): 167–175.

Bianchini L & Poussegur J (1994) Molecular structure and regulation of vertebrate Na+/H+ exchangers. J Exp Biol 196: 337–345.

Bigay J, Gounon P, Robineau S & Antonny B (2003) Lipid packing sensed by ArfGAP1 couples COPI coat disassembly to membrane bilayer curvature. Nature 426(6966): 563–566.

Bigay J, Casella JF, Drin G, Mesmin B & Antonny B (2005) ArfGAP1 responds to membrane curvature through the folding of a lipid packing sensor motif. EMBO J 24(13): 2244–2253.

Biwersi J & Verkman AS (1994) Functional CFTR in endosomal compartment of CFTR-expressing fibroblasts and T84 cells. Am J Physiol 266(1 Pt 1): C149–156.

Boland CR & Deshmukh GD (1990) The carbohydrate composition of mucin in colonic cancer. Gastroenterology 98(5 Pt 1): 1170–1177.

Bonfanti L, Mironov AA,Jr, Martinez-Menarguez JA, Martella O, Fusella A, Baldassarre M, Buccione R, Geuze HJ, Mironov AA & Luini A (1998) Procollagen traverses the Golgi stack without leaving the lumen of cisternae: evidence for cisternal maturation. Cell 95(7): 993–1003.

Bonifacino JS & Traub LM (2003) Signals for sorting of transmembrane proteins to endosomes and lysosomes. Annu Rev Biochem 72: 395–447.

Bonifacino JS & Glick BS (2004) The mechanisms of vesicle budding and fusion. Cell 116(2): 153–166.

Bonifacino JS & Rojas R (2006) Retrograde transport from endosomes to the trans-Golgi network. Nat Rev Mol Cell Biol 7(8): 568–579.

Borgonovo B, Ouwendijk J & Solimena M (2006) Biogenesis of secretory granules. Curr Opin Cell Biol 18(4): 365–370.

Boyer PD (1997) The ATP synthase–a splendid molecular machine. Annu Rev Biochem 66: 717–749.

Breton C, Snajdrova L, Jeanneau C, Koca J & Imberty A (2006) Structures and mechanisms of glycosyltransferases. Glycobiology 16(2): 29R–37R.

Bretscher MS & Munro S (1993) Cholesterol and the Golgi apparatus. Science 261(5126): 1280–1281.

118

Brockhausen I, Yang JM, Burchell J, Whitehouse C & Taylor-Papadimitriou J (1995) Mechanisms underlying aberrant glycosylation of MUC1 mucin in breast cancer cells. Eur J Biochem 233(2): 607–617.

Brockhausen I (1999) Pathways of O-glycan biosynthesis in cancer cells. Biochim Biophys Acta 1473(1): 67–95.

Brodsky B & Ramshaw JA (1997) The collagen triple-helix structure. Matrix Biol 15(8-9): 545–554.

Burchell J, Poulsom R, Hanby A, Whitehouse C, Cooper L, Clausen H, Miles D & Taylor-Papadimitriou J (1999) An alpha2,3 sialyltransferase (ST3Gal I) is elevated in primary breast carcinomas. Glycobiology 9(12): 1307–1311.

Burda P & Aebi M (1999) The dolichol pathway of N-linked glycosylation. Biochim Biophys Acta 1426(2): 239–257.

Burger PC, Lötscher M, Streiff M, Kleene R, Kaissling B & Berger EG (1998) Immunocytochemical localization of alpha2,3(N)-sialyltransferase (ST3Gal III) in cell lines and rat kidney tissue sections: evidence for golgi and post-golgi localization. Glycobiology 8(3): 245–257.

Burkhard P, Stetefeld J & Strelkov SV (2001) Coiled coils: a highly versatile protein folding motif. Trends Cell Biol 11(2): 82–88.

Cabrera M, Muniz M, Hidalgo J, Vega L, Martin ME & Velasco A (2003) The retrieval function of the KDEL receptor requires PKA phosphorylation of its C-terminus. Mol Biol Cell 14(10): 4114–4125.

Caldwell JH & Howell KE (2008) Pores galore for the Golgi. Nat Cell Biol 10(10): 1125–1126.

Campbell BJ, Finnie IA, Hounsell EF & Rhodes JM (1995) Direct demonstration of increased expression of Thomsen-Friedenreich (TF) antigen in colonic adenocarcinoma and ulcerative colitis mucin and its concealment in normal mucin. J Clin Invest 95(2): 571–576.

Campbell BJ, Rowe GE, Leiper K & Rhodes JM (2001a) Increasing the intra-Golgi pH of cultured LS174T goblet-differentiated cells mimics the decreased mucin sulfation and increased Thomsen-Friedenreich antigen (Gal beta1-3GalNac alpha-) expression seen in colon cancer. Glycobiology 11(5): 385–393.

Campbell BJ, Yu LG & Rhodes JM (2001b) Altered glycosylation in inflammatory bowel disease: a possible role in cancer development. Glycoconj J 18(11-12): 851–858.

Campbell JA, Davies GJ, Bulone V & Henrissat B (1997) A classification of nucleotide-diphospho-sugar glycosyltransferases based on amino acid sequence similarities. Biochem J 326(Pt 3): 929–939.

Campbell JA, Davies GJ, Bulone VV & Henrissat B (1998) A classification of nucleotide-diphospho-sugar glycosyltransferases based on amino acid sequence similarities. Biochem J 329(Pt 3): 719.

Campellone KG, Webb NJ, Znameroski EA & Welch MD (2008) WHAMM is an Arp2/3 complex activator that binds microtubules and functions in ER to Golgi transport. Cell 134(1): 148–161.

119

Caplan MJ, Stow JL, Newman AP, Madri J, Anderson HC, Farquhar MG, Palade GE & Jamieson JD (1987) Dependence on pH of polarized sorting of secreted proteins. Nature 329(6140): 632–635.

Capon C, Wieruszeski JM, Lemoine J, Byrd JC, Leffler H & Kim YS (1997) Sulfated lewis X determinants as a major structural motif in glycans from LS174T-HM7 human colon carcinoma mucin. J Biol Chem 272(51): 31957–31968.

Carter SR, Slomiany A, Gwozdzinski K, Liau YH & Slomiany BL (1988) Enzymatic sulfation of mucus glycoprotein in gastric mucosa. Effect of ethanol. J Biol Chem 263(24): 11977–11984.

Chapman RE & Munro S (1994) Retrieval of TGN proteins from the cell surface requires endosomal acidification. EMBO J 13(10): 2305–2312.

Chardin P, Paris S, Antonny B, Robineau S, Beraud-Dufour S, Jackson CL & Chabre M (1996) A human exchange factor for ARF contains Sec7- and pleckstrin-homology domains. Nature 384(6608): 481–484.

Chen C, Ma J, Lazic A, Backovic M & Colley KJ (2000) Formation of insoluble oligomers correlates with ST6Gal I stable localization in the golgi. J Biol Chem 275(18): 13819–13826.

Chen YA & Scheller RH (2001) SNARE-mediated membrane fusion. Nat Rev Mol Cell Biol 2(2): 98–106.

Cheng L, Tachibana K, Iwasaki H, Kameyama A, Zhang Y, Kubota T, Hiruma T, Tachibana K, Kudo T, Guo JM & Narimatsu H (2004) Characterization of a novel human UDP-GalNAc transferase, pp-GalNAc-T15. FEBS Lett 566(1-3): 17–24.

Cherny VV & DeCoursey TE (1999) pH-dependent inhibition of voltage-gated H(+) currents in rat alveolar epithelial cells by Zn(2+) and other divalent cations. J Gen Physiol 114(6): 819–838.

Clary DO, Griff IC & Rothman JE (1990) SNAPs, a family of NSF attachment proteins involved in intracellular membrane fusion in animals and yeast. Cell 61(4): 709–721.

Cluett EB & Brown WJ (1992) Adhesion of Golgi cisternae by proteinaceous interactions: intercisternal bridges as putative adhesive structures. J Cell Sci 103(Pt 3): 773–784.

Colley KJ, Lee EU, Adler B, Browne JK & Paulson JC (1989) Conversion of a Golgi apparatus sialyltransferase to a secretory protein by replacement of the NH2-terminal signal anchor with a signal peptide. J Biol Chem 264(30): 17619–17622.

Colley KJ (1997) Golgi localization of glycosyltransferases: more questions than answers. Glycobiology 7(1): 1–13.

Collins BM (2008) The Structure and Function of the Retromer Protein Complex. Traffic 9(11): 1811–1822.

Contreras I, Ortiz-Zapater E & Aniento F (2004) Sorting signals in the cytosolic tail of membrane proteins involved in the interaction with plant ARF1 and coatomer. Plant J 38(4): 685–698.

Cooper MS, Cornell-Bell AH, Chernjavsky A, Dani JW & Smith SJ (1990) Tubulovesicular processes emerge from trans-Golgi cisternae, extend along microtubules, and interlink adjacent trans-golgi elements into a reticulum. Cell 61(1): 135–145.

120

Correas I, Leto TL, Speicher DW & Marchesi VT (1986) Identification of the functional site of erythrocyte protein 4.1 involved in spectrin-actin associations. J Biol Chem 261(7): 3310–3315.

Crocker PR, Paulson JC & Varki A (2007) Siglecs and their roles in the immune system. Nat Rev Immunol 7(4): 255–266.

Crump CM, Xiang Y, Thomas L, Gu F, Austin C, Tooze SA & Thomas G (2001) PACS-1 binding to adaptors is required for acidic cluster motif-mediated protein traffic. EMBO J 20(9): 2191–2201.

Cukierman E, Huber I, Rotman M & Cassel D (1995) The ARF1 GTPase-activating protein: zinc finger motif and Golgi complex localization. Science 270(5244): 1999–2002.

Dahdal RY & Colley KJ (1993) Specific sequences in the signal anchor of the beta-galactoside alpha-2,6-sialyltransferase are not essential for Golgi localization. Membrane flanking sequences may specify Golgi retention. J Biol Chem 268(35): 26310–26319.

Dalziel M, Whitehouse C, McFarlane I, Brockhausen I, Gschmeissner S, Schwientek T, Clausen H, Burchell JM & Taylor-Papadimitriou J (2001) The relative activities of the C2GnT1 and ST3Gal-I glycosyltransferases determine O-glycan structure and expression of a tumor-associated epitope on MUC1. J Biol Chem 276(14): 11007–11015.

De Matteis MA & Morrow JS (2000) Spectrin tethers and mesh in the biosynthetic pathway. J Cell Sci 113(Pt 13): 2331–2343.

De Matteis MA & D'Angelo G (2007) The role of the phosphoinositides at the Golgi complex. Biochem Soc Symp (74): 107–116.

De Matteis MA & Luini A (2008) Exiting the Golgi complex. Nat Rev Mol Cell Biol 9(4): 273–284.

Delacour D, Greb C, Koch A, Salomonsson E, Leffler H, Le Bivic A & Jacob R (2007) Apical sorting by galectin-3-dependent glycoprotein clustering. Traffic 8(4): 379–388.

Demaurex N, Furuya W, D'Souza S, Bonifacino JS & Grinstein S (1998) Mechanism of acidification of the trans-Golgi network (TGN). In situ measurements of pH using retrieval of TGN38 and furin from the cell surface. J Biol Chem 273(4): 2044–2051.

Demaurex N (2002) pH Homeostasis of cellular organelles. News Physiol Sci 17: 1–5. Devarajan P, Stabach PR, Mann AS, Ardito T, Kashgarian M & Morrow JS (1996)

Identification of a small cytoplasmic ankyrin (AnkG119) in the kidney and muscle that binds beta I sigma spectrin and associates with the Golgi apparatus. J Cell Biol 133(4): 819–830.

Diao A, Rahman D, Pappin DJ, Lucocq J & Lowe M (2003) The coiled-coil membrane protein golgin-84 is a novel rab effector required for Golgi ribbon formation. J Cell Biol 160(2): 201–212.

Diaz E & Pfeffer SR (1998) TIP47: a cargo selection device for mannose 6-phosphate receptor trafficking. Cell 93(3): 433–443.

121

Dirac-Svejstrup AB, Shorter J, Waters MG & Warren G (2000) Phosphorylation of the vesicle-tethering protein p115 by a casein kinase II-like enzyme is required for Golgi reassembly from isolated mitotic fragments. J Cell Biol 150(3): 475–488.

Dominguez M, Dejgaard K, Füllekrug J, Dahan S, Fazel A, Paccaud JP, Thomas DY, Bergeron JJ & Nilsson T (1998) gp25L/emp24/p24 protein family members of the cis-Golgi network bind both COP I and II coatomer. J Cell Biol 140(4): 751–765.

Doray B, Ghosh P, Griffith J, Geuze HJ & Kornfeld S (2002) Cooperation of GGAs and AP-1 in packaging MPRs at the trans-Golgi network. Science 297(5587): 1700–1703.

Drory O & Nelson N (2006) The emerging structure of vacuolar ATPases. Physiology (Bethesda) 21: 317–325.

Eckhardt M, Muhlenhoff M, Bethe A & Gerardy-Schahn R (1996) Expression cloning of the Golgi CMP-sialic acid transporter. Proc Natl Acad Sci U S A 93(15): 7572–7576.

Egea G, Lazaro-Dieguez F & Vilella M (2006) Actin dynamics at the Golgi complex in mammalian cells. Curr Opin Cell Biol 18(2): 168–178.

Elbein AD, Solf R, Dorling PR & Vosbeck K (1981) Swainsonine: an inhibitor of glycoprotein processing. Proc Natl Acad Sci U S A 78(12): 7393–7397.

Ellis MA & Weisz OA (2006) In vitro assays differentially recapitulate protein export from the trans-Golgi network. Anal Biochem 354(2): 314–316.

Endo T (1999) O-mannosyl glycans in mammals. Biochim Biophys Acta 1473(1): 237–246.

Evans SC, Lopez LC & Shur BD (1993) Dominant negative mutation in cell surface beta 1,4-galactosyltransferase inhibits cell-cell and cell-matrix interactions. J Cell Biol 120(4): 1045–1057.

Farinas J & Verkman AS (1999) Receptor-mediated targeting of fluorescent probes in living cells. J Biol Chem 274(12): 7603–7606.

Farquhar MG & Palade GE (1981) The Golgi apparatus (complex)-(1954–1981)-from artifact to center stage. J Cell Biol 91(3 Pt 2): 77s–103s.

Farquhar MG & Palade GE (1998) The Golgi apparatus: 100 years of progress and controversy. Trends Cell Biol 8(1): 2–10.

Fasshauer D, Sutton RB, Brünger AT & Jahn R (1998) Conserved structural features of the synaptic fusion complex: SNARE proteins reclassified as Q- and R-SNAREs. Proc Natl Acad Sci U S A 95(26):15781–15786.

Fath KR & Burgess DR (1993) Golgi-derived vesicles from developing epithelial cells bind actin filaments and possess myosin-I as a cytoplasmically oriented peripheral membrane protein. J Cell Biol 120(1): 117–127.

Feng Y & Forgac M (1994) Inhibition of vacuolar H(+)-ATPase by disulfide bond formation between cysteine 254 and cysteine 532 in subunit A J Biol Chem 269(18): 13224–13230.

Fenteany FH & Colley KJ (2005) Multiple signals are required for alpha2,6-sialyltransferase (ST6Gal I) oligomerization and Golgi localization. J Biol Chem 280(7): 5423–5429.

122

Folley LS, Power SD & Poyton RO (1983) Separation of nucleotides by ion-pair, reversed-phase high-performance liquid chromatography Use of Mg(II) and triethylamine as competing hetaerons in the separation of adenine and guanine nucleotides. Journal of Chromatography A 281: 199–207.

Ford MG, Mills IG, Peter BJ, Vallis Y, Praefcke GJ, Evans PR & McMahon HT (2002) Curvature of clathrin-coated pits driven by epsin. Nature 419(6905): 361–366.

Forgac M, Cantley L, Wiedenmann B, Altstiel L & Branton D (1983) Clathrin-coated vesicles contain an ATP-dependent proton pump. Proc Natl Acad Sci U S A 80(5): 1300–1303.

Fotso P, Koryakina Y, Pavliv O, Tsiomenko AB & Lupashin VV (2005) Cog1p plays a central role in the organization of the yeast conserved oligomeric Golgi complex. J Biol Chem 280(30): 27613–27623.

Foulquier F, Vasile E, Schollen E, Callewaert N, Raemaekers T, Quelhas D, Jaeken J, Mills P, Winchester B, Krieger M, Annaert W & Matthijs G (2006) Conserved oligomeric Golgi complex subunit 1 deficiency reveals a previously uncharacterized congenital disorder of glycosylation type II. Proc Natl Acad Sci U S A 103(10): 3764–3769.

Franke WW, Morre DJ, Deumling B, Cheetham RD, Kartenbeck J, Jarasch ED & Zentgraf HW (1971) Synthesis and turnover of membrane proteins in rat liver: an examination of the membrane flow hypothesis. Z Naturforsch B 26(10): 1031–1039.

Freeze HH & Aebi M (2005) Altered glycan structures: the molecular basis of congenital disorders of glycosylation. Curr Opin Struct Biol 15(5): 490–498.

Freeze HH (2006) Genetic defects in the human glycome. Nat Rev Genet 7(7): 537–551. Frey PA, Wong LJ, Sheu KF & Yang SL (1982) Galactose-1-phosphate uridylyltransferase:

detection, isolation, and characterization of the uridylyl enzyme. Methods Enzymol 87: 20–36.

Füllekrug J & Nilsson T (1998) Protein sorting in the Golgi complex. Biochim Biophys Acta 1404(1-2): 77–84.

Funderburgh JL (2000) Keratan sulfate: structure, biosynthesis, and function. Glycobiology 10(10): 951–958.

Futai M, Omote H, Sambongi Y & Wada Y (2000) Synthase (H(+) ATPase): coupling between catalysis, mechanical work, and proton translocation. Biochim Biophys Acta 1458(2-3): 276–288.

Gardner K & Bennett V (1987) Modulation of spectrin-actin assembly by erythrocyte adducin. Nature 328(6128): 359–362.

Gavel Y & von Heijne G (1990) Sequence differences between glycosylated and non-glycosylated Asn-X-Thr/Ser acceptor sites: implications for protein engineering. Protein Eng 3(5): 433–442.

Gawlitzek M, Ryll T, Lofgren J & Sliwkowski MB (2000) Ammonium alters N-glycan structures of recombinant TNFR-IgG: degradative versus biosynthetic mechanisms. Biotechnol Bioeng 68(6): 637–646.

123

Gentzsch M, Cui L, Mengos A, Chang XB, Chen JH & Riordan JR (2003) The PDZ-binding chloride channel ClC-3B localizes to the Golgi and associates with cystic fibrosis transmembrane conductance regulator-interacting PDZ proteins. J Biol Chem 278(8): 6440–6449.

Ghosh P, Dahms NM & Kornfeld S (2003) Mannose 6-phosphate receptors: new twists in the tale. Nat Rev Mol Cell Biol 4(3): 202–212.

Gill SR, Schroer TA, Szilak I, Steuer ER, Sheetz MP & Cleveland DW (1991) Dynactin, a conserved, ubiquitously expressed component of an activator of vesicle motility mediated by cytoplasmic dynein. J Cell Biol 115(6): 1639–1650.

Giraudo CG & Maccioni HJ (2003) Endoplasmic reticulum export of glycosyltransferases depends on interaction of a cytoplasmic dibasic motif with Sar1. Mol Biol Cell 14(9): 3753–3766.

Glick BS & Rothman JE (1987) Possible role for fatty acyl-coenzyme A in intracellular protein transport. Nature 326(6110): 309–312.

Glick BS & Malhotra V (1998) The curious status of the Golgi apparatus. Cell 95(7): 883–889.

Glickman J, Croen K, Kelly S & Al-Awqati Q (1983) Golgi membranes contain an electrogenic H+ pump in parallel to a chloride conductance. J Cell Biol 97(4): 1303–1308.

Golgi C (1898) Sur la structure des cellules nerveuses des ganglions spinaux. Arch Ital Biol 30: 278–286.

Gommel DU, Memon AR, Heiss A, Lottspeich F, Pfannstiel J, Lechner J, Reinhard C, Helms JB, Nickel W & Wieland FT (2001) Recruitment to Golgi membranes of ADP-ribosylation factor 1 is mediated by the cytoplasmic domain of p23. EMBO J 20(23): 6751–6760.

Gonzalez-Noriega A, Grubb JH, Talkad V & Sly WS (1980) Chloroquine inhibits lysosomal enzyme pinocytosis and enhances lysosomal enzyme secretion by impairing receptor recycling. J Cell Biol 85(3): 839–852.

Grabe M & Oster G (2001) Regulation of organelle acidity. J Gen Physiol 117(4): 329–344. Grabenhorst E & Conradt HS (1999) The cytoplasmic, transmembrane, and stem regions

of glycosyltransferases specify their in vivo functional sublocalization and stability in the Golgi. J Biol Chem 274(51): 36107–36116.

Greaves J & Chamberlain LH (2007) Palmitoylation-dependent protein sorting. J Cell Biol 176(3): 249–254.

Griffiths G, Quinn P & Warren G (1983) Dissection of the Golgi complex. I. Monensin inhibits the transport of viral membrane proteins from medial to trans Golgi cisternae in baby hamster kidney cells infected with Semliki Forest virus. J Cell Biol 96(3): 835–850.

Griffiths G & Simons K (1986) The trans Golgi network: sorting at the exit site of the Golgi complex. Science 234(4775): 438–443.

Griffiths G (2000) Gut thoughts on the Golgi complex. Traffic 1(9): 738–745.

124

Guillen E, Abeijon C & Hirschberg CB (1998) Mammalian Golgi apparatus UDP-N-acetylglucosamine transporter: molecular cloning by phenotypic correction of a yeast mutant. Proc Natl Acad Sci U S A 95(14): 7888–7892.

Hakomori S (2003) Structure, organization, and function of glycosphingolipids in membrane. Curr Opin Hematol 10(1): 16–24.

Hanisch FG, Stadie TR, Deutzmann F & Peter-Katalinic J (1996) MUC1 glycoforms in breast cancer–cell line T47D as a model for carcinoma-associated alterations of 0-glycosylation. Eur J Biochem 236(1): 318–327.

Hannan LA, Lisanti MP, Rodriguez-Boulan E & Edidin M (1993) Correctly sorted molecules of a GPI-anchored protein are clustered and immobile when they arrive at the apical surface of MDCK cells. J Cell Biol 120(2): 353–358.

Hanson PI, Roth R, Morisaki H, Jahn R & Heuser JE (1997) Structure and conformational changes in NSF and its membrane receptor complexes visualized by quick-freeze/deep-etch electron microscopy. Cell 90(3): 523–535.

Hara-Chikuma M, Wang Y, Guggino SE, Guggino WB & Verkman AS (2005a) Impaired acidification in early endosomes of ClC-5 deficient proximal tubule. Biochem Biophys Res Commun 329(3): 941–946.

Hara-Chikuma M, Yang B, Sonawane ND, Sasaki S, Uchida S & Verkman AS (2005b) ClC-3 chloride channels facilitate endosomal acidification and chloride accumulation. J Biol Chem 280(2): 1241–1247.

Hara-Kuge S, Kuge O, Orci L, Amherdt M, Ravazzola M, Wieland FT & Rothman JE (1994) En bloc incorporation of coatomer subunits during the assembly of COP-coated vesicles. J Cell Biol 124(6): 883–892.

Helenius J & Aebi M (2002) Transmembrane movement of dolichol linked carbohydrates during N-glycoprotein biosynthesis in the endoplasmic reticulum. Semin Cell Dev Biol 13(3): 171–178.

Helenius J, Ng DT, Marolda CL, Walter P, Valvano MA & Aebi M (2002) Translocation of lipid-linked oligosaccharides across the ER membrane requires Rft1 protein. Nature 415(6870): 447–450.

Hinners I & Tooze SA (2003) Changing directions: clathrin-mediated transport between the Golgi and endosomes. J Cell Sci 116(Pt 5): 763–771.

Hirschberg CB, Robbins PW & Abeijon C (1998) Transporters of nucleotide sugars, ATP, and nucleotide sulfate in the endoplasmic reticulum and Golgi apparatus. Annu Rev Biochem 67: 49–69.

Holappa K, Suokas M, Soininen P & Kellokumpu S (2001) Identification of the full-length AE2 (AE2a) isoform as the Golgi-associated anion exchanger in fibroblasts. J Histochem Cytochem 49(2): 259–269.

Holappa K & Kellokumpu S (2003) Targeting of the AE2 anion exchanger to the Golgi apparatus is cell type-dependent and correlates with the expression of Ank(195), a Golgi membrane skeletal protein. FEBS Lett 546(2-3): 257–264.

Holappa K, Munoz MT, Egea G & Kellokumpu S (2004) The AE2 anion exchanger is necessary for the structural integrity of the Golgi apparatus in mammalian cells. FEBS Lett 564(1-2): 97–103.

125

Holleran EA, Tokito MK, Karki S & Holzbaur EL (1996) Centractin (ARP1) associates with spectrin revealing a potential mechanism to link dynactin to intracellular organelles. J Cell Biol 135(6 Pt 2): 1815–1829.

Hougland AE, Gaush CR & Mayberry WR (1979) Glycolipids and fatty acids of two dog kidney cell lines. Biochem J 177(1): 311–317.

Hu C, Ahmed M, Melia TJ, Söllner TH, Mayer T & Rothman JE (2003) Fusion of cells by flipped SNAREs. Science 300(5626): 1745–1749.

Hwang JM, Kao SH, Hsieh YH, Li KL, Wang PH, Hsu LS & Liu JY (2009) Reduction of anion exchanger 2 expression induces apoptosis of human hepatocellular carcinoma cells. Mol Cell Biochem 327(1-2):135–144.

Hyvönen M, Macias MJ, Nilges M, Oschkinat H, Saraste M & Wilmanns M (1995) Structure of the binding site for inositol phosphates in a PH domain. EMBO J 14(19): 4676–4685.

Isselbacher KJ (1958) A mammalian uridinediphosphate galactose pyrophosphorylase. J Biol Chem 232(1): 429–444.

Itzkowitz SH, Marshall A, Kornbluth A, Harpaz N, McHugh JB, Ahnen D & Sachar DB (1995) Sialosyl-Tn antigen: initial report of a new marker of malignant progression in long-standing ulcerative colitis. Gastroenterology 109(2): 490–497.

Jackson MR, Nilsson T & Peterson PA (1990) Identification of a consensus motif for retention of transmembrane proteins in the endoplasmic reticulum. EMBO J 9(10): 3153–3162.

Jackson MR, Nilsson T & Peterson PA (1993) Retrieval of transmembrane proteins to the endoplasmic reticulum. J Cell Biol 121(2): 317–333.

Jass JR, Allison LM & Edgar S (1994) Monoclonal antibody TKH2 to the cancer-associated epitope sialosyl Tn shows cross-reactivity with variants of normal colorectal goblet cell mucin. Pathology 26(4): 418–422.

Jefferies KC, Cipriano DJ & Forgac M (2008) Function, structure and regulation of the vacuolar (H+)-ATPases. Arch Biochem Biophys 476(1): 33–42.

Jenkins N, Parekh RB & James DC (1996) Getting the glycosylation right: implications for the biotechnology industry. Nat Biotechnol 14(8): 975–981.

Jentsch TJ, Friedrich T, Schriever A & Yamada H (1999) The CLC chloride channel family. Pflugers Arch 437(6): 783–795.

Jentsch TJ (2007) Chloride and the endosomal-lysosomal pathway: emerging roles of CLC chloride transporters. J Physiol 578(Pt 3): 633–640.

Julenius K, Mølgaard A, Gupta R & Brunak S (2005) Prediction, conservation analysis, and structural characterization of mammalian mucin-type O-glycosylation sites. Glycobiology 15(2): 153–164.

Kane PM & Parra KJ (2000) Assembly and regulation of the yeast vacuolar H(+)-ATPase. J Exp Biol 203(Pt 1): 81–87.

Kang Q, Wang T, Zhang H, Mohandas N & An X (2009) A Golgi-associated protein 4.1B variant is required for assimilation of proteins in the membrane. J Cell Sci

126

Kappeler F, Klopfenstein DR, Foguet M, Paccaud JP & Hauri HP (1997) The recycling of ERGIC-53 in the early secretory pathway. ERGIC-53 carries a cytosolic endoplasmic reticulum-exit determinant interacting with COPII. J Biol Chem 272(50): 31801–31808.

Karki S & Holzbaur EL (1995) Affinity chromatography demonstrates a direct binding between cytoplasmic dynein and the dynactin complex. J Biol Chem 270(48): 28806–28811.

Karlen P, Young E, Broström O, Löfberg R, Tribukait B, Ost K, Bodian C & Itzkowitz S (1998) Sialyl-Tn antigen as a marker of colon cancer risk in ulcerative colitis: relation to dysplasia and DNA aneuploidy. Gastroenterology 115(6): 1395–1404.

Kean EL (1969) Sialic acid activating enzyme in ocular tissue. Exp Eye Res 8(1): 44–54. Kellokumpu S, Neff L, Jämsä-Kellokumpu S, Kopito R & Baron R (1988) A 115-kD

polypeptide immunologically related to erythrocyte band 3 is present in Golgi membranes. Science 242(4883): 1308–1311.

Kellokumpu S, Sormunen R & Kellokumpu I (2002) Abnormal glycosylation and altered Golgi structure in colorectal cancer: dependence on intra-Golgi pH. FEBS Lett 516(1-3): 217–224.

Kepes F, Rambourg A & Satiat-Jeunemaitre B (2005) Morphodynamics of the secretory pathway. Int Rev Cytol 242: 55–120.

Kettner C, Obermeyer G & Bertl A (2003) Inhibition of the yeast V-type ATPase by cytosolic ADP. FEBS Lett 535(1-3): 119–124.

Kim JH, Lingwood CA, Williams DB, Furuya W, Manolson MF & Grinstein S (1996) Dynamic measurement of the pH of the Golgi complex in living cells using retrograde transport of the verotoxin receptor. J Cell Biol 134(6): 1387–1399.

Kim YJ & Varki A (1997) Perspectives on the significance of altered glycosylation of glycoproteins in cancer. Glycoconj J 14(5): 569–576.

Kim YS, Yuan M, Itzkowitz SH, Sun QB, Kaizu T, Palekar A, Trump BF & Hakomori S (1986) Expression of LeY and extended LeY blood group-related antigens in human malignant, premalignant, and nonmalignant colonic tissues. Cancer Res 46(11): 5985–5992.

Kirk SJ & Ward TH (2007) COPII under the microscope. Semin Cell Dev Biol 18(4): 435–447.

Kirsch RD & Joly E (1998) An improved PCR-mutagenesis strategy for two-site mutagenesis or sequence swapping between related genes. Nucleic Acids Res. 26(7): 1848–1850.

Kivirikko KI & Myllylä R (1982) Posttranslational enzymes in the biosynthesis of collagen: intracellular enzymes. Methods Enzymol 82(Pt A): 245–304.

Knudsen MJ & Troy FA (1989) Nuclear magnetic resonance studies of polyisoprenols in model membranes. Chem Phys Lipids 51(3-4): 205–212.

Kobata A, Matsuoka Y, Kuroku M & Yamashita K (1995) The sugar chain structures of carsinoembryonic antigens and related normal antigens. Pure & Appl Chem 67(10): 1689–1698.

127

Kokkonen N, Rivinoja A, Kauppila A, Suokas M, Kellokumpu I & Kellokumpu S (2004) Defective acidification of intracellular organelles results in aberrant secretion of cathepsin D in cancer cells. J Biol Chem 279(38): 39982–39988.

Komarova YA, Akhmanova AS, Kojima S, Galjart N & Borisy GG (2002) Cytoplasmic linker proteins promote microtubule rescue in vivo. J Cell Biol 159(4): 589–599.

Kondo H, Rabouille C, Newman R, Levine TP, Pappin D, Freemont P & Warren G (1997) P47 is a Cofactor for P97-Mediated Membrane Fusion. Nature 388(6637): 75–78.

Kornfeld R & Kornfeld S (1985) Assembly of asparagine-linked oligosaccharides. Annu Rev Biochem 54: 631–664.

Korsgren C & Cohen CM (1986) Purification and properties of human erythrocyte band 4.2. Association with the cytoplasmic domain of band 3. J Biol Chem 261(12): 5536–5543.

Kresse H, Hausser H, Schonherr E & Bittner K (1994) Biosynthesis and interactions of small chondroitin/dermatan sulphate proteoglycans. Eur.J.Clin ChemClin.Biochem. 32(4): 259–264.

Kuan SF, Byrd JC, Basbaum C & Kim YS (1989) Inhibition of mucin glycosylation by aryl-N-acetyl-alpha-galactosaminides in human colon cancer cells. J Biol Chem 264(32): 19271–19277.

Kuhns W, Jain RK, Matta KL, Paulsen H, Baker MA, Geyer R & Brockhausen I (1995) Characterization of a novel mucin sulphotransferase activity synthesizing sulphated O-glycan core 1,3-sulphate-Gal beta 1-3GalNAc alpha-R. Glycobiology 5(7): 689–697.

Kuismanen E, Saraste J & Pettersson RF (1985) Effect of monensin on the assembly of Uukuniemi virus in the Golgi complex. J Virol 55(3): 813–822.

Kumamoto K, Goto Y, Sekikawa K, Takenoshita S, Ishida N, Kawakita M & Kannagi R (2001) Increased expression of UDP-galactose transporter messenger RNA in human colon cancer tissues and its implication in synthesis of Thomsen-Friedenreich antigen and sialyl Lewis A/X determinants. Cancer Res 61(11): 4620–4627.

Ladinsky MS, Mastronarde DN, McIntosh JR, Howell KE & Staehelin LA (1999) Golgi structure in three dimensions: functional insights from the normal rat kidney cell. J Cell Biol 144(6): 1135–1149.

Laemmli UK (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227(5259): 680–685.

Lanoix J, Ouwendijk J, Stark A, Szafer E, Cassel D, Dejgaard K, Weiss M & Nilsson T (2001) Sorting of Golgi resident proteins into different subpopulations of COPI vesicles: a role for ArfGAP1. J Cell Biol 155(7): 1199–1212.

Lariviere L, Gueguen-Chaignon V & Morera S (2003) Crystal structures of the T4 phage beta-glucosyltransferase and the D100A mutant in complex with UDP-glucose: glucose binding and identification of the catalytic base for a direct displacement mechanism. J Mol Biol 330(5): 1077–1086.

Lederkremer GZ, Cheng Y, Petre BM, Vogan E, Springer S, Schekman R, Walz T & Kirchhausen T (2001) Structure of the Sec23p/24p and Sec13p/31p complexes of COPII. Proc Natl Acad Sci U S A 98(19): 10704–10709.

128

Lee MC & Miller EA (2007) Molecular mechanisms of COPII vesicle formation. Semin Cell Dev Biol 18(4): 424–434.

Levine YK & Wilkins MH (1971) Structure of oriented lipid bilayers. Nat New Biol 230(11): 69–72.

Lewis MJ & Pelham HR (1992a) Ligand-induced redistribution of a human KDEL receptor from the Golgi complex to the endoplasmic reticulum. Cell 68(2): 353–364.

Lewis MJ & Pelham HR (1992b) Sequence of a second human KDEL receptor. J Mol Biol 226(4): 913–916.

Lin X (2004) Functions of heparan sulfate proteoglycans in cell signaling during development. Development 131(24): 6009–6021.

Linstedt AD & Hauri HP (1993) Giantin, a novel conserved Golgi membrane protein containing a cytoplasmic domain of at least 350 kDa. Mol Biol Cell 4(7): 679–693.

Linstedt AD, Foguet M, Renz M, Seelig HP, Glick BS & Hauri HP (1995) A C-terminally-anchored Golgi protein is inserted into the endoplasmic reticulum and then transported to the Golgi apparatus. Proc Natl Acad Sci U S A 92(11): 5102–5105.

Liu W, Duden R, Phair RD & Lippincott-Schwartz J (2005) ArfGAP1 dynamics and its role in COPI coat assembly on Golgi membranes of living cells. J Cell Biol 168(7): 1053–1063.

Llopis J, McCaffery JM, Miyawaki A, Farquhar MG & Tsien RY (1998) Measurement of cytosolic, mitochondrial, and Golgi pH in single living cells with green fluorescent proteins. Proc Natl Acad Sci U S A 95(12): 6803–6808.

Lloyd KO, Burchell J, Kudryashov V, Yin BW & Taylor-Papadimitriou J (1996) Comparison of O-linked carbohydrate chains in MUC-1 mucin from normal breast epithelial cell lines and breast carcinoma cell lines. Demonstration of simpler and fewer glycan chains in tumor cells. J Biol Chem 271(52): 33325–33334.

Lloyd KO (2000) The chemistry and immunochemistry of blood group A, B, H, and Lewis antigens: past, present and future. Glycoconj J 17(7-9): 531–541.

Locker JK, Opstelten DJ, Ericsson M, Horzinek MC & Rottier PJ (1995) Oligomerization of a trans-Golgi/trans-Golgi network retained protein occurs in the Golgi complex and may be part of its retention. J Biol Chem 270(15): 8815–8821.

Lombardi D, Soldati T, Riederer MA, Goda Y, Zerial M & Pfeffer SR (1993) Rab9 functions in transport between late endosomes and the trans Golgi network. EMBO J 12(2): 677–682.

Losev E, Reinke CA, Jellen J, Strongin DE, Bevis BJ & Glick BS (2006) Golgi maturation visualized in living yeast. Nature 441(7096): 1002–1006.

Lowe M, Rabouille C, Nakamura N, Watson R, Jackman M, Jämsä E, Rahman D, Pappin DJ & Warren G (1998) Cdc2 kinase directly phosphorylates the cis-Golgi matrix protein GM130 and is required for Golgi fragmentation in mitosis. Cell 94(6): 783–793.

Lowe M, Gonatas NK & Warren G (2000) The mitotic phosphorylation cycle of the cis-Golgi matrix protein GM130. J Cell Biol 149(2): 341–356.

129

Lu M, Sautin YY, Holliday LS & Gluck SL (2004) The glycolytic enzyme aldolase mediates assembly, expression, and activity of vacuolar H+-ATPase. J Biol Chem 279(10): 8732–8739.

Lu M, Ammar D, Ives H, Albrecht F & Gluck SL (2007) Physical interaction between aldolase and vacuolar H+-ATPase is essential for the assembly and activity of the proton pump. J Biol Chem 282(34): 24495–24503.

Lucocq JM, Pryde JG, Berger EG & Warren G (1987) A mitotic form of the Golgi apparatus in HeLa cells. J Cell Biol 104(4): 865–874.

Luhn K, Wild MK, Eckhardt M, Gerardy-Schahn R & Vestweber D (2001) The gene defective in leukocyte adhesion deficiency II encodes a putative GDP-fucose transporter. Nat Genet 28(1): 69–72.

Ma J & Colley KJ (1996) A disulfide-bonded dimer of the Golgi beta-galactoside alpha2,6-sialyltransferase is catalytically inactive yet still retains the ability to bind galactose. J Biol Chem 271(13): 7758–7766.

Maccioni HJ (2007) Glycosylation of glycolipids in the Golgi complex. J Neurochem 103(1): 81–90.

Machamer CE (1991) Golgi retention signals: do membranes hold the key? Trends Cell Biol 1(6): 141–144.

Machen TE, Chandy G, Wu M, Grabe M & Moore HP (2001) Cystic fibrosis transmembrane conductance regulator and H+ permeability in regulation of Golgi pH. JOP 2(4 Suppl): 229–236.

Maeda Y, Ide T, Koike M, Uchiyama Y & Kinoshita T (2008) GPHR is a novel anion channel critical for acidification and functions of the Golgi apparatus. Nat Cell Biol 10(10): 1135–1145.

Majoul I, Sohn K, Wieland FT, Pepperkok R, Pizza M, Hillemann J & Söling HD (1998) KDEL receptor (Erd2p)-mediated retrograde transport of the cholera toxin A subunit from the Golgi involves COPI, p23, and the COOH terminus of Erd2p. J Cell Biol 143(3): 601–612.

Majoul I, Straub M, Hell SW, Duden R & Söling HD (2001) KDEL-cargo regulates interactions between proteins involved in COPI vesicle traffic: measurements in living cells using FRET. Dev Cell 1(1): 139–153.

Malsam J, Satoh A, Pelletier L & Warren G (2005) Golgin tethers define subpopulations of COPI vesicles. Science 307(5712): 1095–1098.

Marek KW, Vijay IK & Marth JD (1999) A recessive deletion in the GlcNAc-1-phosphotransferase gene results in peri-implantation embryonic lethality. Glycobiology 9(11): 1263–1271.

Marth JD & Grewal PK (2008) Mammalian glycosylation in immunity. Nat Rev Immunol 8(11): 874–887.

Martinez-Menarguez JA, Prekeris R, Oorschot VM, Scheller R, Slot JW, Geuze HJ & Klumperman J (2001) Peri-Golgi vesicles contain retrograde but not anterograde proteins consistent with the cisternal progression model of intra-Golgi transport. J Cell Biol 155(7): 1213–1224.

130

Martin-Rendon E & Blake DJ (2003) Protein glycosylation in disease: new insights into the congenital muscular dystrophies. Trends Pharmacol Sci 24(4): 178–183.

Masibay AS, Balaji PV, Boeggeman EE & Qasba PK (1993) Mutational analysis of the Golgi retention signal of bovine beta-1,4-galactosyltransferase. J Biol Chem 268(13): 9908–9916.

Matlin KS (1986) Ammonium chloride slows transport of the influenza virus hemagglutinin but does not cause mis-sorting in a polarized epithelial cell line. J Biol Chem 261(32): 15172–15178.

Matsuoka K, Schekman R, Orci L & Heuser JE (2001) Surface structure of the COPII-coated vesicle. Proc Natl Acad Sci U S A 98(24): 13705–13709.

Matsuura-Tokita K, Takeuchi M, Ichihara A, Mikuriya K & Nakano A (2006) Live imaging of yeast Golgi cisternal maturation. Nature 441(7096): 1007–1010.

May AP, Misura KM, Whiteheart SW & Weis WI (1999) Crystal structure of the amino-terminal domain of N-ethylmaleimide-sensitive fusion protein. Nat Cell Biol 1(3): 175–182.

Mayer A, Wickner W & Haas A (1996) Sec18p (NSF)-driven release of Sec17p (alpha-SNAP) can precede docking and fusion of yeast vacuoles. Cell 85(1): 83–94.

Medina JF, Lecanda J, Acin A, Ciesielczyk P & Prieto J (2000) Tissue-specific N-terminal isoforms from overlapping alternate promoters of the human AE2 anion exchanger gene. Biochem Biophys Res Commun 267(1): 228–235.

Mezzacasa A & Helenius A (2002) The transitional ER defines a boundary for quality control in the secretion of tsO45 VSV glycoprotein. Traffic 3(11): 833–849.

Miller E, Antonny B, Hamamoto S & Schekman R (2002) Cargo selection into COPII vesicles is driven by the Sec24p subunit. EMBO J 21(22): 6105–6113.

Mironov AA, Weidman P & Luini A (1997) Variations on the intracellular transport theme: maturing cisternae and trafficking tubules. J Cell Biol 138(3): 481–484.

Mironov AA, Beznoussenko GV, Nicoziani P, Martella O, Trucco A, Kweon HS, Di Giandomenico D, Polishchuk RS, Fusella A, Lupetti P, Berger EG, Geerts WJ, Koster AJ, Burger KN & Luini A (2001) Small cargo proteins and large aggregates can traverse the Golgi by a common mechanism without leaving the lumen of cisternae. J Cell Biol 155(7): 1225–1238.

Mironov AA, Mironov AA,Jr, Beznoussenko GV, Trucco A, Lupetti P, Smith JD, Geerts WJ, Koster AJ, Burger KN, Martone ME, Deerinck TJ, Ellisman MH & Luini A (2003) ER-to-Golgi carriers arise through direct en bloc protrusion and multistage maturation of specialized ER exit domains. Dev Cell 5(4): 583–594.

Mironov AA, Beznoussenko GV, Polishchuk RS & Trucco A (2005) Intra-Golgi transport: a way to a new paradigm? Biochim Biophys Acta 1744(3): 340–350.

Mische SM, Mooseker MS & Morrow JS (1987) Erythrocyte adducin: a calmodulin-regulated actin-bundling protein that stimulates spectrin-actin binding. J Cell Biol 105(6 Pt 1): 2837–2845.

Mizuno M & Singer SJ (1994) A possible role for stable microtubules in intracellular transport from the endoplasmic reticulum to the Golgi apparatus. J Cell Sci 107(Pt 5): 1321–1331.

131

Molinari M & Helenius A (1999) Glycoproteins form mixed disulphides with oxidoreductases during folding in living cells. Nature 402(6757): 90–93.

Mollenhauer HH & Morre DJ (1991) Perspectives on Golgi apparatus form and function. J Electron Microsc Tech 17(1): 2–14.

Moriyama H, Nakano H, Igawa M & Nihira H (1987) T antigen expression in benign hyperplasia and adenocarcinoma of the prostate. Urol.Int. 42(2): 120–123.

Müller S & Hanisch FG (2002) Recombinant MUC1 probe authentically reflects cell-specific O-glycosylation profiles of endogenous breast cancer mucin. High density and prevalent core 2-based glycosylation. J Biol Chem 277(29): 26103–26112.

Munro S (1998) Localization of proteins to the Golgi apparatus. Trends Cell Biol 8(1): 11–15.

Munro S (2003) Lipid rafts: elusive or illusive? Cell 115(4): 377–388. Münster-Kühnel AK, Tiralongo J, Krapp S, Weinhold B, Ritz-Sedlacek V, Jacob U &

Gerardy-Schahn R (2004) Structure and function of vertebrate CMP-sialic acid synthetases. Glycobiology 14(10): 43R–51R.

Muraoka M, Kawakita M & Ishida N (2001) Molecular characterization of human UDP-glucuronic acid/UDP-N-acetylgalactosamine transporter, a novel nucleotide sugar transporter with dual substrate specificity. FEBS Lett 495(1-2): 87–93.

Murray BW, Takayama S, Schultz J & Wong CH (1996) Mechanism and specificity of human alpha-1,3-fucosyltransferase V. Biochemistry 35(34): 11183–11195.

Murshid A & Presley JF (2004) ER-to-Golgi transport and cytoskeletal interactions in animal cells. Cell Mol Life Sci 61(2): 133–145.

Nadolski MJ & Linder ME (2007) Protein lipidation. FEBS J. 274(20): 5202–5210. Nagasawa M, Kanzaki M, Iino Y, Morishita Y & Kojima I (2001) Identification of a novel

chloride channel expressed in the endoplasmic reticulum, golgi apparatus, and nucleus. J Biol Chem 276(23): 20413–20418.

Nakamura N, Tanaka S, Teko Y, Mitsui K & Kanazawa H (2005) Four Na+/H+ exchanger isoforms are distributed to Golgi and post-Golgi compartments and are involved in organelle pH regulation. J Biol Chem 280(2): 1561–1572.

Nezil FA & Bloom M (1992) Combined influence of cholesterol and synthetic amphiphillic peptides upon bilayer thickness in model membranes. Biophys J 61(5): 1176–1183.

Nickel W, Malsam J, Gorgas K, Ravazzola M, Jenne N, Helms JB & Wieland FT (1998) Uptake by COPI-coated vesicles of both anterograde and retrograde cargo is inhibited by GTPgammaS in vitro. J Cell Sci 111(Pt 20): 3081–3090.

Nicoziani P, Vilhardt F, Llorente A, Hilout L, Courtoy PJ, Sandvig K & van Deurs B (2000) Role for dynamin in late endosome dynamics and trafficking of the cation-independent mannose 6-phosphate receptor. Mol Biol Cell 11(2): 481–495.

Nilsson T, Jackson M & Peterson PA (1989) Short cytoplasmic sequences serve as retention signals for transmembrane proteins in the endoplasmic reticulum. Cell 58(4): 707–718.

132

Nilsson T, Lucocq JM, Mackay D & Warren G (1991) The membrane spanning domain of beta-1,4-galactosyltransferase specifies trans Golgi localization. EMBO J 10(12): 3567–3575.

Nilsson T, Pypaert M, Hoe MH, Slusarewicz P, Berger EG & Warren G (1993a) Overlapping distribution of two glycosyltransferases in the Golgi apparatus of HeLa cells. J Cell Biol 120(1): 5–13.

Nilsson T, Slusarewicz P, Hoe MH & Warren G (1993b) Kin recognition. A model for the retention of Golgi enzymes. FEBS Lett 330(1): 1–4.

Nilsson T, Hoe MH, Slusarewicz P, Rabouille C, Watson R, Hunte F, Watzele G, Berger EG & Warren G (1994) Kin recognition between medial Golgi enzymes in HeLa cells. EMBO J 13(3): 562–574.

Nilsson T, Rabouille C, Hui N, Watson R & Warren G (1996) The role of the membrane-spanning domain and stalk region of N-acetylglucosaminyltransferase I in retention, kin recognition and structural maintenance of the Golgi apparatus in HeLa cells. J Cell Sci 109(Pt 7): 1975–1989.

Nishi T & Forgac M (2002) The vacuolar (H+)-ATPases–nature's most versatile proton pumps. Nat Rev Mol Cell Biol 3(2): 94–103.

Nishimura N & Balch WE (1997) A di-acidic signal required for selective export from the endoplasmic reticulum. Science 277(5325): 556–558.

Niu TK, Pfeifer AC, Lippincott-Schwartz J & Jackson CL (2005) Dynamics of GBF1, a Brefeldin A-sensitive Arf1 exchange factor at the Golgi. Mol Biol Cell 16(3): 1213–1222.

Noji H, Yasuda R, Yoshida M & Kinosita K,Jr (1997) Direct observation of the rotation of F1-ATPase. Nature 386(6622): 299–302.

Nordeen MH, Jones SM, Howell KE & Caldwell JH (2000) GOLAC: an endogenous anion channel of the Golgi complex. Biophys J 78(6): 2918–2928.

Numata M & Orlowski J (2001) Molecular cloning and characterization of a novel (Na+,K+)/H+ exchanger localized to the trans-Golgi network. J Biol Chem 276(20): 17387–17394.

Ogata S, Ho I, Chen A, Dubois D, Maklansky J, Singhal A, Hakomori S & Itzkowitz SH (1995) Tumor-associated sialylated antigens are constitutively expressed in normal human colonic mucosa. Cancer Res 55(9): 1869–1874.

Ohira M, Smardon AM, Charsky CM, Liu J, Tarsio M & Kane PM (2006) The E and G subunits of the yeast V-ATPase interact tightly and are both present at more than one copy per V1 complex. J Biol Chem 281(32): 22752–22760.

Ohtsubo K & Marth JD (2006) Glycosylation in cellular mechanisms of health and disease. Cell 126(5): 855–867.

Oikawa S, Kosaki G & Nakazato H (1987) Molecular cloning of a gene for a member of carcinoembryonic antigen (CEA) gene family; signal peptide and N-terminal domain sequences of nonspecific crossreacting antigen (NCA). Biochem Biophys Res Commun 146(2): 464–469.

133

Oka T, Ungar D, Hughson FM & Krieger M (2004) The COG and COPI complexes interact to control the abundance of GEARs, a subset of Golgi integral membrane proteins. Mol Biol Cell 15(5): 2423–2435.

Ono M & Hakomori S (2004) Glycosylation defining cancer cell motility and invasiveness. Glycoconj J 20(1): 71–78.

Opat AS, Houghton F & Gleeson PA (2000) Medial Golgi but not late Golgi glycosyltransferases exist as high molecular weight complexes. Role of luminal domain in complex formation and localization. J Biol Chem 275(16): 11836–11845.

Opat AS, van Vliet C & Gleeson PA (2001) Trafficking and localisation of resident Golgi glycosylation enzymes. Biochimie 83(8): 763–773.

Orci L, Montesano R, Meda P, Malaisse-Lagae F, Brown D, Perrelet A & Vassalli P (1981) Heterogeneous distribution of filipin–cholesterol complexes across the cisternae of the Golgi apparatus. Proc Natl Acad Sci U S A 78(1): 293–297.

Orci L, Ravazzola M, Meda P, Holcomb C, Moore HP, Hicke L & Schekman R (1991) Mammalian Sec23p homologue is restricted to the endoplasmic reticulum transitional cytoplasm. Proc Natl Acad Sci U S A 88(19): 8611–8615.

Orci L, Stamnes M, Ravazzola M, Amherdt M, Perrelet A, Söllner TH & Rothman JE (1997) Bidirectional transport by distinct populations of COPI-coated vesicles. Cell 90(2): 335–349.

Palade G (1975) Intracellular aspects of the process of protein synthesis. Science 189(4200): 347–358.

Palmer KJ & Stephens DJ (2004) Biogenesis of ER-to-Golgi transport carriers: complex roles of COPII in ER export. Trends Cell Biol 14(2): 57–61.

Palokangas H, Ying M, Väänänen K & Saraste J (1998) Retrograde transport from the pre-Golgi intermediate compartment and the Golgi complex is affected by the vacuolar H+-ATPase inhibitor bafilomycin A1. Mol Biol Cell 9(12): 3561–3578.

Parodi AJ, Blank EW, Peterson JA & Ceriani RL (1982) Dolichol-bound oligosaccharides and the transfer of distal monosaccharides in the synthesis of glycoproteins by normal and tumor mammary epithelial cells. Breast Cancer ResTreat 2(3): 227–237.

Parodi AJ (2000a) Role of N-oligosaccharide endoplasmic reticulum processing reactions in glycoprotein folding and degradation. Biochem J 348 Pt 1: 1–13.

Parodi AJ (2000b) Protein glucosylation and its role in protein folding. Annu Rev Biochem 69: 69–93.

Paroutis P, Touret N & Grinstein S (2004) The pH of the secretory pathway: measurement, determinants, and regulation. Physiology (Bethesda) 19: 207–215.

Paschal BM & Vallee RB (1987) Retrograde transport by the microtubule-associated protein MAP 1C. Nature 330(6144): 181–183.

Pasternack GR, Anderson RA, Leto TL & Marchesi VT (1985) Interactions between protein 4.1 and band 3. An alternative binding site for an element of the membrane skeleton. J Biol Chem 260(6): 3676–3683.

Paulson JC (1989) Glycoproteins: what are the sugar chains for? Trends Biochem Sci 14(7): 272–276.

134

Pecot MY & Malhotra V (2006) The Golgi apparatus maintains its organization independent of the endoplasmic reticulum. Mol Biol Cell 17(12): 5372–5380.

Pelham HR (1988) Evidence that luminal ER proteins are sorted from secreted proteins in a post-ER compartment. EMBO J 7(4): 913–918.

Pelham HR (1998) Getting through the Golgi complex. Trends Cell Biol 8(1): 45–49. Pepperkok R, Whitney JA, Gomez M & Kreis TE (2000) COPI vesicles accumulating in

the presence of a GTP restricted arf1 mutant are depleted of anterograde and retrograde cargo. J Cell Sci 113(Pt 1): 135–144.

Peters LL, Shivdasani RA, Liu SC, Hanspal M, John KM, Gonzalez JM, Brugnara C, Gwynn B, Mohandas N, Alper SL, Orkin SH & Lux SE (1996) Anion exchanger 1 (band 3) is required to prevent erythrocyte membrane surface loss but not to form the membrane skeleton. Cell 86(6): 917–927.

Pfeffer S (2003) Membrane domains in the secretory and endocytic pathways. Cell 112(4): 507–517.

Pfeffer SR (2007) Unsolved mysteries in membrane traffic. Annu Rev Biochem 76: 629–645.

Pinnell SR, Fox R & Krane SM (1971) Human collagens: differences in glycosylated hydroxylysines in skin and bone. Biochim Biophys Acta 229(1): 119–122.

Piper RC & Luzio JP (2007) Ubiquitin-dependent sorting of integral membrane proteins for degradation in lysosomes. Curr Opin Cell Biol 19(4): 459–465.

Potter BA, Hughey RP & Weisz OA (2006) Role of N- and O-glycans in polarized biosynthetic sorting. Am J PhysiolCell Physiol 290(1): C1–C10.

Powell B, Graham LA & Stevens TH (2000) Molecular characterization of the yeast vacuolar H+-ATPase proton pore. J Biol Chem 275(31): 23654–23660.

Praefcke GJ & McMahon HT (2004) The dynamin superfamily: universal membrane tubulation and fission molecules? Nat Rev Mol Cell Biol 5(2): 133–147.

Preisinger C, Korner R, Wind M, Lehmann WD, Kopajtich R & Barr FA (2005) Plk1 docking to GRASP65 phosphorylated by Cdk1 suggests a mechanism for Golgi checkpoint signalling. EMBO J 24(4): 753–765.

Presley JF, Cole NB, Schroer TA, Hirschberg K, Zaal KJ & Lippincott-Schwartz J (1997) ER-to-Golgi transport visualized in living cells. Nature 389(6646): 81–85.

Puri S, Bachert C, Fimmel CJ & Linstedt AD (2002) Cycling of early Golgi proteins via the cell surface and endosomes upon lumenal pH disruption. Traffic 3(9): 641–653.

Pypaert M, Nilsson T, Berger EG & Warren G (1993) Mitotic Golgi clusters are not tubular endosomes. J Cell Sci 104(Pt 3): 811–818.

Qasba PK, Ramakrishnan B & Boeggeman E (2008) Structure and function of beta -1,4-galactosyltransferase. Curr.Drug Targets 9(4): 292–309.

Qian R, Chen C & Colley KJ (2001) Location and mechanism of alpha 2,6-sialyltransferase dimer formation. Role of cysteine residues in enzyme dimerization, localization, activity, and processing. J Biol Chem 276(31): 28641–28649.

Rabouille C, Hui N, Hunte F, Kieckbusch R, Berger EG, Warren G & Nilsson T (1995) Mapping the distribution of Golgi enzymes involved in the construction of complex oligosaccharides. J Cell Sci 108(Pt 4): 1617–1627.

135

Rahkila P, Väänänen K, Saraste J & Metsikkö K (1997) Endoplasmic reticulum to Golgi trafficking in multinucleated skeletal muscle fibers. Exp Cell Res 234(2): 452–464.

Ram RJ, Li B & Kaiser CA (2002) Identification of Sec36p, Sec37p, and Sec38p: components of yeast complex that contains Sec34p and Sec35p. Mol Biol Cell 13(5): 1484–1500.

Rambourg A & Clermont Y (1990) Three-dimensional electron microscopy: structure of the Golgi apparatus. Eur J Cell Biol 51(2): 189–200.

Randazzo PA (1997) Resolution of two ADP-ribosylation factor 1 GTPase-activating proteins from rat liver. Biochem J 324(Pt 2): 413–419.

Reaves B, Horn M & Banting G (1993) TGN38/41 recycles between the cell surface and the TGN: brefeldin A affects its rate of return to the TGN. Mol Biol Cell 4(1): 93–105.

Ren J, Lew S, Wang Z & London E (1997) Transmembrane orientation of hydrophobic alpha-helices is regulated both by the relationship of helix length to bilayer thickness and by the cholesterol concentration. Biochemistry 36(33): 10213–10220.

Rhim AD, Stoykova L, Glick MC & Scanlin TF (2001) Terminal glycosylation in cystic fibrosis (CF): a review emphasizing the airway epithelial cell. Glycoconj J 18(9): 649–659.

Rickard JE & Kreis TE (1990) Identification of a novel nucleotide-sensitive microtubule-binding protein in HeLa cells. J Cell Biol 110(5): 1623–1633.

Rios RM & Bornens M (2003) The Golgi apparatus at the cell centre. Curr Opin Cell Biol 15(1): 60–66.

Rios RM, Sanchis A, Tassin AM, Fedriani C & Bornens M (2004) GMAP-210 recruits gamma-tubulin complexes to cis-Golgi membranes and is required for Golgi ribbon formation. Cell 118(3): 323–335.

Ritter C & Helenius A (2000) Recognition of local glycoprotein misfolding by the ER folding sensor UDP-glucose:glycoprotein glucosyltransferase. Nat Struct Biol 7(4): 278–280.

Rodman J, Feng Y, Myers M, Zhang J, Magner R & Forgac M (1994) Comparison of the coated-vesicle and synaptic-vesicle vacuolar (H+)-ATPases. Ann N Y Acad Sci 733: 203–211.

Rodriguez-Boulan E, Kreitzer G & Müsch A (2005) Organization of vesicular trafficking in epithelia. Nat Rev Mol Cell Biol 6(3): 233–247.

Rodriguez-Boulan E & Müsch A (2005) Protein sorting in the Golgi complex: shifting paradigms. Biochim Biophys Acta 1744(3): 455–464.

Rogalski AA & Singer SJ (1984) Associations of elements of the Golgi apparatus with microtubules. J Cell Biol 99(3): 1092–1100.

Romero MF, Fulton CM & Boron WF (2004) The SLC4 family of HCO 3 - transporters. Pflugers Arch 447(5): 495–509.

Rossmann H, Bachmann O, Wang Z, Shull GE, Obermaier B, Stuart-Tilley A, Alper SL & Seidler U (2001) Differential expression and regulation of AE2 anion exchanger subtypes in rabbit parietal and mucous cells. J Physiol 534(Pt 3): 837–848.

136

Roth J & Berger EG (1982) Immunocytochemical localization of galactosyltransferase in HeLa cells: codistribution with thiamine pyrophosphatase in trans-Golgi cisternae. J Cell Biol 93(1): 223–229.

Roth J, Zuber C, Wagner P, Taatjes DJ, Weisgerber C, Heitz PU, Goridis C & Bitter-Suermann D (1988) Reexpression of poly(sialic acid) units of the neural cell adhesion molecule in Wilms tumor. Proc Natl Acad Sci U S A 85(9): 2999–3003.

Roth J (2002) Protein N-glycosylation along the secretory pathway: relationship to organelle topography and function, protein quality control, and cell interactions. Chem Rev 102(2): 285–303.

Rothman JE (1994) Mechanisms of intracellular protein transport. Nature 372(6501): 55–63.

Rothman JE & Wieland FT (1996) Protein sorting by transport vesicles. Science 272(5259): 227–234.

Rottger S, White J, Wandall HH, Olivo JC, Stark A, Bennett EP, Whitehouse C, Berger EG, Clausen H & Nilsson T (1998) Localization of three human polypeptide GalNAc-transferases in HeLa cells suggests initiation of O-linked glycosylation throughout the Golgi apparatus. J Cell Sci 111(Pt 1): 45–60.

Ruetz S, Lindsey AE, Ward CL & Kopito RR (1993) Functional activation of plasma membrane anion exchangers occurs in a pre-Golgi compartment. J Cell Biol 121(1): 37–48.

Rybak SL, Lanni F & Murphy RF (1997) Theoretical considerations on the role of membrane potential in the regulation of endosomal pH. Biophys J 73(2): 674–687.

Saint-Pol A, Yelamos B, Amessou M, Mills IG, Dugast M, Tenza D, Schu P, Antony C, McMahon HT, Lamaze C & Johannes L (2004) Clathrin adaptor epsinR is required for retrograde sorting on early endosomal membranes. Dev Cell 6(4): 525–538.

Saraste J & Kuismanen E (1984) Pre- and post-Golgi vacuoles operate in the transport of Semliki Forest virus membrane glycoproteins to the cell surface. Cell 38(2): 535–549.

Saraste J & Svensson K (1991) Distribution of the intermediate elements operating in ER to Golgi transport. J Cell Sci 100(Pt 3): 415–430.

Sasai K, Ikeda Y, Tsuda T, Ihara H, Korekane H, Shiota K & Taniguchi N (2001) The critical role of the stem region as a functional domain responsible for the oligomerization and Golgi localization of N-acetylglucosaminyltransferase V. The involvement of a domain homophilic interaction. J Biol Chem 276(1): 759–765.

Sata T, Roth J, Zuber C, Stamm B & Heitz PU (1991) Expression of alpha 2,6-linked sialic acid residues in neoplastic but not in normal human colonic mucosa. A lectin-gold cytochemical study with Sambucus nigra and Maackia amurensis lectins. Am J Pathol 139(6): 1435–1448.

Sata T, Roth J, Zuber C, Stamm B, Rinderle SJ, Goldstein IJ & Heitz PU (1992) Studies on the Thomsen-Friedenreich antigen in human colon with the lectin Amaranthin. Normal and neoplastic epithelium express only cryptic T antigen. Lab Invest 66(2): 175–186.

137

Sautin YY, Lu M, Gaugler A, Zhang L & Gluck SL (2005) Phosphatidylinositol 3-kinase-mediated effects of glucose on vacuolar H+-ATPase assembly, translocation, and acidification of intracellular compartments in renal epithelial cells. Mol Cell Biol 25(2): 575–589.

Scales SJ, Pepperkok R & Kreis TE (1997) Visualization of ER-to-Golgi transport in living cells reveals a sequential mode of action for COPII and COPI. Cell 90(6): 1137–1148.

Scales SJ, Yoo BY & Scheller RH (2001) The ionic layer is required for efficient dissociation of the SNARE complex by alpha-SNAP and NSF. Proc Natl Acad Sci U S A 98(25):14262–14267.

Schapiro FB & Grinstein S (2000) Determinants of the pH of the Golgi complex. J Biol Chem 275(28): 21025–21032.

Schaub BE, Berger B, Berger EG & Rohrer J (2006) Transition of galactosyltransferase 1 from trans-Golgi cisterna to the trans-Golgi network is signal mediated. Mol Biol Cell 17(12): 5153–5162.

Schaub BE, Berger EG & Rohrer J (2008) Dissection of a novel molecular determinant mediating Golgi to trans-Golgi network transition. Cell Mol Life Sci 65(22): 3677–3687.

Scheiffele P, Peränen J & Simons K (1995) N-glycans as apical sorting signals in epithelial cells. Nature 378(6552): 96–98.

Schenk B, Fernandez F & Waechter CJ (2001) The ins(ide) and out(side) of dolichyl phosphate biosynthesis and recycling in the endoplasmic reticulum. Glycobiology 11(5): 61R–70R.

Schietinger A, Philip M, Yoshida BA, Azadi P, Liu H, Meredith SC & Schreiber H (2006) A mutant chaperone converts a wild-type protein into a tumor-specific antigen. Science 314(5797): 304–308.

Schroer TA & Sheetz MP (1991) Two activators of microtubule-based vesicle transport. J Cell Biol 115(5): 1309–1318.

Schuck S & Simons K (2004) Polarized sorting in epithelial cells: raft clustering and the biogenesis of the apical membrane. J Cell Sci 117(Pt 25): 5955–5964.

Schwappach B, Stobrawa S, Hechenberger M, Steinmeyer K & Jentsch TJ (1998) Golgi localization and functionally important domains in the NH2 and COOH terminus of the yeast CLC putative chloride channel Gef1p. J Biol Chem 273(24): 15110–15118.

Segawa H, Kawakita M & Ishida N (2002) Human and Drosophila UDP-galactose transporters transport UDP-N-acetylgalactosamine in addition to UDP-galactose. Eur J Biochem 269(1): 128–138.

Seksek O, Biwersi J & Verkman AS (1996) Evidence against defective trans-Golgi acidification in cystic fibrosis. J Biol Chem 271(26): 15542–15548.

Sennoune SR, Bakunts K, Martinez GM, Chua-Tuan JL, Kebir Y, Attaya MN & Martinez-Zaguilan R (2004a) Vacuolar H+-ATPase in human breast cancer cells with distinct metastatic potential: distribution and functional activity. Am J PhysiolCell Physiol 286(6): C1443–1452.

138

Sennoune SR, Luo D & Martinez-Zaguilan R (2004b) Plasmalemmal vacuolar-type H+-ATPase in cancer biology. Cell Biochem Biophys 40(2): 185–206.

Seol JH, Shevchenko A, Shevchenko A & Deshaies RJ (2001) Skp1 forms multiple protein complexes, including RAVE, a regulator of V-ATPase assembly. Nat Cell Biol 3(4): 384–391.

Sesso A, de Faria FP, Iwamura ES & Correa H (1994) A three-dimensional reconstruction study of the rough ER-Golgi interface in serial thin sections of the pancreatic acinar cell of the rat. J Cell Sci 107(Pt 3): 517–528.

Shakin-Eshleman SH, Spitalnik SL & 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. J Biol Chem 271(11): 6363–6366.

Shao L, Luo Y, Moloney DJ & Haltiwanger R (2002) O-glycosylation of EGF repeats: identification and initial characterization of a UDP-glucose: protein O-glucosyltransferase. Glycobiology 12(11): 763–770.

Shao L & Haltiwanger RS (2003) O-fucose modifications of epidermal growth factor-like repeats and thrombospondin type 1 repeats: unusual modifications in unusual places. Cell Mol Life Sci 60(2): 241–250.

Shaywitz DA, Orci L, Ravazzola M, Swaroop A & Kaiser CA (1995) Human SEC13Rp functions in yeast and is located on transport vesicles budding from the endoplasmic reticulum. J Cell Biol 128(5): 769–777.

Shemesh T, Luini A, Malhotra V, Burger KN & Kozlov MM (2003) Prefission constriction of Golgi tubular carriers driven by local lipid metabolism: a theoretical model. Biophys J 85(6): 3813–3827.

Shestakova A, Zolov S & Lupashin V (2006) COG complex-mediated recycling of Golgi glycosyltransferases is essential for normal protein glycosylation. Traffic 7(2): 191–204.

Shima DT, Scales SJ, Kreis TE & Pepperkok R (1999) Segregation of COPI-rich and anterograde-cargo-rich domains in endoplasmic-reticulum-to-Golgi transport complexes. Curr Biol 9(15): 821–824.

Short B, Haas A & Barr FA (2005) Golgins and GTPases, giving identity and structure to the Golgi apparatus. Biochim Biophys Acta 1744(3): 383–395.

Shorter J & Warren G (1999) A role for the vesicle tethering protein, p115, in the post-mitotic stacking of reassembling Golgi cisternae in a cell-free system. J Cell Biol 146(1): 57–70.

Shorter J, Watson R, Giannakou ME, Clarke M, Warren G & Barr FA (1999) GRASP55, a second mammalian GRASP protein involved in the stacking of Golgi cisternae in a cell-free system. EMBO J 18(18): 4949–4960.

Siddhanta A, Radulescu A, Stankewich MC, Morrow JS & Shields D (2003) Fragmentation of the Golgi apparatus. A role for beta III spectrin and synthesis of phosphatidylinositol 4,5-bisphosphate. J Biol Chem 278(3): 1957–1965.

Simons K & van Meer G (1988) Lipid sorting in epithelial cells. Biochemistry 27(17): 6197–6202.

139

Singh R, Campbell BJ, Yu LG, Fernig DG, Milton JD, Goodlad RA, FitzGerald AJ & Rhodes JM (2001) Cell surface-expressed Thomsen-Friedenreich antigen in colon cancer is predominantly carried on high molecular weight splice variants of CD44. Glycobiology 11(7): 587–592.

Sipilä L, Ruotsalainen H, Sormunen R, Baker NL, Lamande SR, Vapola M, Wang C, Sado Y, Aszodi A & Myllylä R (2007) Secretion and assembly of type IV and VI collagens depend on glycosylation of hydroxylysines. J Biol Chem 282(46): 33381–33388.

Sivasubramanian N & Nayak DP (1987) Mutational analysis of the signal-anchor domain of influenza virus neuraminidase. Proc Natl Acad Sci U S A 84(1): 1–5.

Skrincosky D, Kain R, El-Battari A, Exner M, Kerjaschki D & Fukuda M (1997) Altered Golgi localization of core 2 beta-1,6-N-acetylglucosaminyltransferase leads to decreased synthesis of branched O-glycans. J Biol Chem 272(36): 22695–22702.

Smardon AM, Tarsio M & Kane PM (2002) The RAVE complex is essential for stable assembly of the yeast V-ATPase. J Biol Chem 277(16): 13831–13839.

Smardon AM & Kane PM (2007) RAVE is essential for the efficient assembly of the C subunit with the vacuolar H(+)-ATPase. J Biol Chem 282(36): 26185–26194.

Smith RD & Lupashin VV (2008) Role of the conserved oligomeric Golgi (COG) complex in protein glycosylation. Carbohydr.Res. 343(12): 2024–2031.

Söllner T, Whiteheart SW, Brunner M, Erdjument-Bromage H, Geromanos S, Tempst P & Rothman JE (1993) SNAP receptors implicated in vesicle targeting and fusion. Nature 362(6418): 318–324.

Sönnichsen B, Lowe M, Levine T, Jämsä E, Dirac-Svejstrup B & Warren G (1998) A role for giantin in docking COPI vesicles to Golgi membranes. J Cell Biol 140(5): 1013–1021.

Sonnino S, Mauri L, Chigorno V & Prinetti A (2007) Gangliosides as components of lipid membrane domains. Glycobiology 17(1): 1R-13R.

Southgate CD, Chishti AH, Mitchell B, Yi SJ & Palek J (1996) Targeted disruption of the murine erythroid band 3 gene results in spherocytosis and severe haemolytic anaemia despite a normal membrane skeleton. Nat Genet 14(2): 227–230.

Spaapen LJ, Bakker JA, van der Meer SB, Sijstermans HJ, Steet RA, Wevers RA & Jaeken J (2005) Clinical and biochemical presentation of siblings with COG-7 deficiency, a lethal multiple O- and N-glycosylation disorder. J Inherit Metab Dis 28(5): 707–714.

Stankewich MC, Tse WT, Peters LL, Ch'ng Y, John KM, Stabach PR, Devarajan P, Morrow JS & Lux SE (1998) A widely expressed betaIII spectrin associated with Golgi and cytoplasmic vesicles. Proc Natl Acad Sci U S A 95(24): 14158–14163.

Stanley KK & Howell KE (1993) TGN38/41: a molecule on the move. Trends Cell Biol 3(8): 252–255.

Steet R & Kornfeld S (2006) COG-7-deficient Human Fibroblasts Exhibit Altered Recycling of Golgi Proteins. Mol Biol Cell 17(5): 2312–2321.

Stephens DJ, Lin-Marq N, Pagano A, Pepperkok R & Paccaud JP (2000) COPI-coated ER-to-Golgi transport complexes segregate from COPII in close proximity to ER exit sites. J Cell Sci 113(Pt 12): 2177–2185.

140

Stephens DJ (2003) De novo formation, fusion and fission of mammalian COPII-coated endoplasmic reticulum exit sites. EMBO Rep 4(2): 210–217.

Sterling D, Alvarez BV & Casey JR (2002) The extracellular component of a transport metabolon. Extracellular loop 4 of the human AE1 Cl-/HCO3- exchanger binds carbonic anhydrase IV. J Biol Chem 277(28): 25239–25246.

Stewart AK, Chernova MN, Kunes YZ & Alper SL (2001) Regulation of AE2 anion exchanger by intracellular pH: critical regions of the NH(2)-terminal cytoplasmic domain. Am J PhysiolCell Physiol 281(4): C1344–1354.

Stewart AK, Chernova MN, Shmukler BE, Wilhelm S & Alper SL (2002) Regulation of AE2-mediated Cl- transport by intracellular or by extracellular pH requires highly conserved amino acid residues of the AE2 NH2-terminal cytoplasmic domain. J Gen Physiol 120(5): 707–722.

Stewart AK, Kerr N, Chernova MN, Alper SL & Vaughan-Jones RD (2004) Acute pH-dependent regulation of AE2-mediated anion exchange involves discrete local surfaces of the NH2-terminal cytoplasmic domain. J Biol Chem 279(50): 52664–52676.

Stewart AK, Kurschat CE, Burns D, Banger N, Vaughan-Jones RD & Alper SL (2007a) Transmembrane domain histidines contribute to regulation of AE2-mediated anion exchange by pH. Am J Physiol Cell Physiol 292(2): C909–918.

Stewart AK, Kurschat CE, Vaughan-Jones RD, Shmukler BE & Alper SL (2007b) Acute regulation of mouse AE2 anion exchanger requires isoform-specific amino acid residues from most of the transmembrane domain. J Physiol 584(Pt 1): 59–73.

Stewart AK, Kurschat CK, Vaughan-Jones RD & Alper SL (2008) Putative re-entrant loop 1 of Ae2 transmembrane domain has a major role in acute regulation of anion exchange by pH. J Biol Chem 284(10): 6126–39.

Stuart-Tilley AK, Shmukler BE, Brown D & Alper SL (1998) Immunolocalization and tissue-specific splicing of AE2 anion exchanger in mouse kidney. J Am Soc Nephrol 9(6): 946–959.

Sturgess JM & De la Iglesia FA (1972) Morphometry of the golgi apparatus in developing liver. J Cell Biol 55(2): 524–530.

Sun-Wada GH, Wada Y & Futai M (2004) Diverse and essential roles of mammalian vacuolar-type proton pump ATPase: toward the physiological understanding of inside acidic compartments. Biochim Biophys Acta 1658(1-2): 106–114.

Supek F, Supekova L, Mandiyan S, Pan YC, Nelson H & Nelson N (1994) A novel accessory subunit for vacuolar H(+)-ATPase from chromaffin granules. J Biol Chem 269(39): 24102–24106.

Sutton RB, Fasshauer D, Jahn R & AT (1998) Crystal structure of a SNARE complex involved in synaptic exocytosis at 2.4 A resolution. Nature 395(6700): 347–353.

Swift AM & Machamer CE (1991) A Golgi retention signal in a membrane-spanning domain of coronavirus E1 protein. J Cell Biol 115(1): 19–30.

Sztul E & Lupashin V (2006) Role of tethering factors in secretory membrane traffic. Am J PhysiolCell Physiol 290(1): C11–26.

141

Takashima S (2008) Characterization of mouse sialyltransferase genes: their evolution and diversity. Biosci Biotechnol Biochem 72(5): 1155–1167.

Tanaka K, Mitsushima A, Fukudome H & Kashima Y (1986) Three-dimensional architecture of the Golgi complex observed by high resolution scanning electron microscopy. J Submicrosc Cytol 18(1): 1–9.

Tang BL, Peter F, Krijnse-Locker J, Low SH, Griffiths G & Hong W (1997) The mammalian homolog of yeast Sec13p is enriched in the intermediate compartment and is essential for protein transport from the endoplasmic reticulum to the Golgi apparatus. Mol Cell Biol 17(1): 256–266.

Tang D, Mar K, Warren G & Wang Y (2008) Molecular mechanism of mitotic Golgi disassembly and reassembly revealed by a defined reconstitution assay. J Biol Chem 283(10): 6085–6094.

Tanigawa G, Orci L, Amherdt M, Ravazzola M, Helms JB & Rothman JE (1993) Hydrolysis of bound GTP by ARF protein triggers uncoating of Golgi-derived COP-coated vesicles. J Cell Biol 123(6 Pt 1): 1365–1371.

Tassin AM, Paintrand M, Berger EG & Bornens M (1985) The Golgi apparatus remains associated with microtubule organizing centers during myogenesis. J Cell Biol 101(2): 630–638.

Taylor-Papadimitriou J, Burchell J, Miles DW & Dalziel M (1999) MUC1 and cancer. Biochim Biophys Acta 1455(2-3): 301–313.

Teasdale RD, D'Agostaro G & Gleeson PA (1992) The signal for Golgi retention of bovine beta 1,4-galactosyltransferase is in the transmembrane domain. J Biol Chem 267(18): 13113.

Teasdale RD, Matheson F & Gleeson PA (1994) Post-translational modifications distinguish cell surface from Golgi-retained beta 1,4 galactosyltransferase molecules. Golgi localization involves active retention. Glycobiology 4(6): 917–928.

Ten Hagen KG, Fritz TA & Tabak LA (2003) All in the family: the UDP-GalNAc:polypeptide N-acetylgalactosaminyltransferases. Glycobiology 13(1): 1R-16R.

Thompson RJ, Nordeen MH, Howell KE & Caldwell JH (2002) A large-conductance anion channel of the Golgi complex. Biophys J 83(1): 278–289.

Thorens B & Vassalli P (1986) Chloroquine and ammonium chloride prevent terminal glycosylation of immunoglobulins in plasma cells without affecting secretion. Nature 321(6070): 618–620.

Trombetta ES & Helenius A (2000) Conformational requirements for glycoprotein reglucosylation in the endoplasmic reticulum. J Cell Biol 148(6): 1123–1129.

Valderrama F, Luna A, Babia T, Martinez-Menarguez JA, Ballesta J, Barth H, Chaponnier C, Renau-Piqueras J & Egea G (2000) The golgi-associated COPI-coated buds and vesicles contain beta/gamma -actin. Proc Natl Acad Sci U S A 97(4): 1560–1565.

van Kooyk Y & Rabinovich GA (2008) Protein-glycan interactions in the control of innate and adaptive immune responses. Nat Immunol 9(6): 593–601.

Varki A (1992) Diversity in the sialic acids. Glycobiology 2(1): 25–40.

142

Varki A (1993) Biological roles of oligosaccharides: all of the theories are correct. Glycobiology 3(2): 97–130.

Varki A (1998) Factors controlling the glycosylation potential of the Golgi apparatus. Trends Cell Biol 8(1): 34–40.

Varki A, Cummings RD, Esko JD, Freeze HH, Stanley P, Bertozzi CR, Hart GW & Etzler ME (eds) (2007) Essential of glycobiology. 2 La Jolla, California: The Consortium of Glycobiology Editors.

Vasile E, Oka T, Ericsson M, Nakamura N & Krieger M (2006) IntraGolgi distribution of the Conserved Oligomeric Golgi (COG) complex. Exp Cell Res 312(16): 3132–3141.

Vaughan KT & Vallee RB (1995) Cytoplasmic dynein binds dynactin through a direct interaction between the intermediate chains and p150Glued. J Cell Biol 131(6 Pt 1): 1507–1516.

Vaughan KT, Tynan SH, Faulkner NE, Echeverri CJ & Vallee RB (1999) Colocalization of cytoplasmic dynein with dynactin and CLIP-170 at microtubule distal ends. J Cell Sci 112(Pt 10): 1437–1447.

Vaughan KT (2005) Microtubule plus ends, motors, and traffic of Golgi membranes. Biochim Biophys Acta 1744(3): 316–324.

Vigo C, Grossman SH & Drost-Hansen W (1984) Interaction of dolichol and dolichyl phosphate with phospholipid bilayers. Biochim Biophys Acta 774(2): 221–226.

Vince JW & Reithmeier RA (1998) Carbonic anhydrase II binds to the carboxyl terminus of human band 3, the erythrocyte C1-/HCO3- exchanger. J Biol Chem 273(43): 28430–28437.

Vince JW & Reithmeier RA (2000) Identification of the carbonic anhydrase II binding site in the Cl(-)/HCO(3)(-) anion exchanger AE1. Biochemistry 39(18): 5527–5533.

Vivero-Salmeron G, Ballesta J & Martinez-Menarguez JA (2008) Heterotypic tubular connections at the endoplasmic reticulum-Golgi complex interface. Histochem Cell Biol 130(4):709–17.

von Heijne G (1985) Ribosome-SRP-signal sequence interactions. The relay helix hypothesis. FEBS Lett 190(1): 1–5.

Wada Y, Sambongi Y & Futai M (2000) Biological nano motor, ATP synthase F(o)F(1): from catalysis to gammaepsilonc(10-12) subunit assembly rotation. Biochim Biophys Acta 1459(2-3): 499–505.

Wakabayashi S, Fafournoux P, Sardet C & Pouyssegur J (1992) The Na+/H+ antiporter cytoplasmic domain mediates growth factor signals and controls "H(+)-sensing". Proc Natl Acad Sci U S A 89(6): 2424–2428.

Wakabayashi S, Pang T, Su X & Shigekawa M (2000) Second mutations rescue point mutant of the Na(+)/H(+) exchanger NHE1 showing defective surface expression. FEBS Lett 487(2): 257–261.

Waldman BC & Rudnick G (1990) UDP-GlcNAc transport across the Golgi membrane: electroneutral exchange for dianionic UMP Biochemistry 29(1): 44–52.

Wan L, Molloy SS, Thomas L, Liu G, Xiang Y, Rybak SL & Thomas G (1998) PACS-1 defines a novel gene family of cytosolic sorting proteins required for trans-Golgi network localization. Cell 94(2): 205–216.

143

Wang Y, Seemann J, Pypaert M, Shorter J & Warren G (2003) A direct role for GRASP65 as a mitotically regulated Golgi stacking factor. EMBO J 22(13): 3279–3290.

Wang Z, Schultheis PJ & Shull GE (1996) Three N-terminal variants of the AE2 Cl-/HCO3- exchanger are encoded by mRNAs transcribed from alternative promoters. J Biol Chem 271(13): 7835–7843.

Ward TH, Polishchuk RS, Caplan S, Hirschberg K & Lippincott-Schwartz J (2001) Maintenance of Golgi structure and function depends on the integrity of ER export. J Cell Biol 155(4): 557–570.

Watson P, Forster R, Palmer KJ, Pepperkok R & Stephens DJ (2005) Coupling of ER exit to microtubules through direct interaction of COPII with dynactin. Nat Cell Biol 7(1): 48–55.

Weber T, Zemelman BV, McNew JA, Westermann B, Gmachl M, Parlati F, Söllner TH & Rothman JE (1998) SNAREpins: minimal machinery for membrane fusion. Cell 92(6): 759–772.

Wegmann D, Hess P, Baier C, Wieland FT & Reinhard C (2004) Novel isotypic gamma/zeta subunits reveal three coatomer complexes in mammals. Mol Cell Biol 24(3): 1070–1080.

Weidman PJ (1995) Anterograde transport through the Golgi complex: do Golgi tubules hold the key? Trends Cell Biol 5(8): 302–305.

Weinstein J, de Souza-e-Silva U & Paulson JC (1982) Sialylation of glycoprotein oligosaccharides N-linked to asparagine. Enzymatic characterization of a Gal beta 1 to 3(4)GlcNAc alpha 2 to 3 sialyltransferase and a Gal beta 1 to 4GlcNAc alpha 2 to 6 sialyltransferase from rat liver. J Biol Chem 257(22): 13845–13853.

Weisz OA, Swift AM & Machamer CE (1993) Oligomerization of a membrane protein correlates with its retention in the Golgi complex. J Cell Biol 122(6): 1185–1196.

Whitehouse C, Burchell J, Gschmeissner S, Brockhausen I, Lloyd KO & Taylor-Papadimitriou J (1997) A transfected sialyltransferase that is elevated in breast cancer and localizes to the medial/trans-Golgi apparatus inhibits the development of core-2-based O-glycans. J Cell Biol 137(6): 1229–1241.

Wiebe WJ & Bancroft K (1975) Use of the adenylate energy charge ratio to measure growth state of natural microbial communities. Proc Natl Acad Sci U S A 72(6): 2112–2115.

Wolf MF, Ludwig A, Fritz P & Schumacher K (1988) Increased expression of Thomsen-Friedenreich antigens during tumor progression in breast cancer patients. Tumour Biol 9(4): 190–194.

Wooding S & Pelham HR (1998) The dynamics of golgi protein traffic visualized in living yeast cells. Mol Biol Cell 9(9): 2667–2680.

Wopereis S, Lefeber DJ, Morava E & Wevers RA (2006) Mechanisms in protein O-glycan biosynthesis and clinical and molecular aspects of protein O-glycan biosynthesis defects: a review. Clin Chem 52(4): 574–600.

Wu MM, Llopis J, Adams S, McCaffery JM, Kulomaa MS, Machen TE, Moore HP & Tsien RY (2000) Organelle pH studies using targeted avidin and fluorescein-biotin. Chem Biol 7(3): 197–209.

144

Wu MM, Grabe M, Adams S, Tsien RY, Moore HP & Machen TE (2001) Mechanisms of pH regulation in the regulated secretory pathway. J Biol Chem 276(35): 33027–33035.

Wu X, Steet RA, Bohorov O, Bakker J, Newell J, Krieger M, Spaapen L, Kornfeld S & Freeze HH (2004) Mutation of the COG complex subunit gene COG7 causes a lethal congenital disorder. Nat Med 10(5): 518–523.

Xiang Y, Seemann J, Bisel B, Punthambaker S & Wang Y (2007) Active ADP-ribosylation factor-1 (ARF1) is required for mitotic Golgi fragmentation. J Biol Chem 282(30): 21829–21837.

Xu T, Vasilyeva E & Forgac M (1999) Subunit interactions in the clathrin-coated vesicle vacuolar (H(+))-ATPase complex. J Biol Chem 274(41): 28909–28915.

Xu T & Forgac M (2001) Microtubules are involved in glucose-dependent dissociation of the yeast vacuolar [H+]-ATPase in vivo. J Biol Chem 276(27): 24855–24861.

Yamaguchi N & Fukuda MN (1995) Golgi retention mechanism of beta-1,4-galactosyltransferase. Membrane-spanning domain-dependent homodimerization and association with alpha- and beta-tubulins. J Biol Chem 270(20): 12170–12176.

Yamamoto K, Fujii R, Toyofuku Y, Saito T, Koseki H, Hsu VW & Aoe T (2001) The KDEL receptor mediates a retrieval mechanism that contributes to quality control at the endoplasmic reticulum. EMBO J 20(12): 3082–3091.

Yamashita K, Totani K, Kuroki M, Matsuoka Y, Ueda I & Kobata A (1987) Structural studies of the carbohydrate moieties of carcinoembryonic antigens. Cancer Res 47(13): 3451–3459.

Yan Q & Lennarz WJ (1999) Oligosaccharyltransferase: a complex multisubunit enzyme of the endoplasmic reticulum. Biochem Biophys Res Commun 266(3): 684–689.

Yang JM, Byrd JC, Siddiki BB, Chung YS, Okuno M, Sowa M, Kim YS, Matta KL & Brockhausen I (1994) Alterations of O-glycan biosynthesis in human colon cancer tissues. Glycobiology 4(6): 873–884.

Yang RY, Rabinovich GA & Liu FT (2008) Galectins: structure, function and therapeutic potential. Expert Rev Mol Med 10: e17.

Yeaman C, Le Gall AH, Baldwin AN, Monlauzeur L, Le Bivic A & Rodriguez-Boulan E (1997) The O-glycosylated stalk domain is required for apical sorting of neurotrophin receptors in polarized MDCK cells. J Cell Biol 139(4): 929–940.

Yeh J & Cummings RD (1997) Differential recognition of glycoprotein acceptors by terminal glycosyltransferases. Glycobiology 7(2): 241–251.

Yoshihisa T, Barlowe C & Schekman R (1993) Requirement for a GTPase-activating protein in vesicle budding from the endoplasmic reticulum. Science 259(5100): 1466–1468.

Young WW,Jr (2004) Organization of Golgi glycosyltransferases in membranes: complexity via complexes. J Membr Biol 198(1): 1–13.

Yu LG (2007) The oncofetal Thomsen-Friedenreich carbohydrate antigen in cancer progression. Glycoconj J 24(8): 411–420.

Yu RC, Jahn R & Brünger AT (1999) NSF N-terminal domain crystal structure: models of NSF function. Mol Cell 4(1): 97–107.

145

Zeevaert R, Foulquier F, Jaeken J & Matthijs G (2008) Deficiencies in subunits of the Conserved Oligomeric Golgi (COG) complex define a novel group of Congenital Disorders of Glycosylation. Mol Genet Metab 93(1): 15–21.

Zhang GF, Driouich A & Staehelin LA (1996) Monensin-induced redistribution of enzymes and products from Golgi stacks to swollen vesicles in plant cells. Eur J Cell Biol 71(4): 332–340.

Zhao X, Lasell TK & Melancon P (2002) Localization of large ADP-ribosylation factor-guanine nucleotide exchange factors to different Golgi compartments: evidence for distinct functions in protein traffic. Mol Biol Cell 13(1): 119–133.

Zhao YY, Takahashi M, Gu JG, Miyoshi E, Matsumoto A, Kitazume S & Taniguchi N (2008) Functional roles of N-glycans in cell signaling and cell adhesion in cancer. Cancer Sci 99(7): 1304–1310.

Zolotarev AS, Shmukler BE & Alper SL (1999) AE2 anion exchanger polypeptide is a homooligomer in pig gastric membranes: a chemical cross-linking study. Biochemistry 38(26): 8521–8531.

Zolov SN & Lupashin VV (2005) Cog3p depletion blocks vesicle-mediated Golgi retrograde trafficking in HeLa cells. J Cell Biol 168(5): 747–759.

Zuber C, Lackie PM, Catterall WA & Roth J (1992) Polysialic acid is associated with sodium channels and the neural cell adhesion molecule N-CAM in adult rat brain. J Biol Chem 267(14): 9965–9971.

146

147

Original articles

I Rivinoja A, Kokkonen N, Kellokumpu I & Kellokumpu S (2006) Elevated Golgi pH in breast and colorectal cancer cells correlates with the expression of oncofetal carbohydrate T-antigen. J Cell Physiol 208:167–74.

II Rivinoja A, Hassinen A, Kokkonen N, Kauppila A & Kellokumpu S (2009). Elevated Golgi pH impairs terminal N-glycosylation by inducing mislocalization of Golgi glycosyltransferases. J Cell Physiol. 220:144–54.

III Rivinoja A, Kauppila A, Kvist A-P & Kellokumpu S (2009) AE2 Cl-/HCO3-

exchanger is involved in the regulation of Golgi steady state pH. Manuscript.

Reprinted with permission from John Wiley & Sons (I, II).

Original publications are not included in the electronic version of the dissertation.

148

A C T A U N I V E R S I T A T I S O U L U E N S I S

Distributed byOULU UNIVERSITY LIBRARY

P.O. Box 7500, FI-90014University of Oulu, Finland

Book orders:OULU UNIVERSITY PRESSP.O. Box 8200, FI-90014University of Oulu, Finland

S E R I E S A S C I E N T I A E R E R U M N A T U R A L I U M

520. Hilli, Anu (2009) The effect of crop quality and pre-treatment on germination inScots pine and Norway spruce seeds

521. Kreivi, Marjut (2009) Conservation genetics and phylogeography of endangeredboreoarctic seashore plant species

522. Riihijärvi, Jorma (2009) Tietojenkäsittelytieteiden koulutuksent yöe l ämäva s t a avuus . E s imerkk i t ap auksena Ou lun y l i op i s tontietojenkäsittelytieteiden laitoksen koulutusohjelma

523. Ilmonen, Jari (2009) Benthic macroinvertebrate and bryophyte assemblages inboreal springs: diversity, spatial patterns and conservation

524. Pujol, François (2009) Experiments on fatty acids chain elongation and glycanflipping in the ER membrane

525. Lampila, Satu (2009) The causes and consequences of population declines of twoboreal forest species. The case of the willow tit (Parus montanus) and the Siberianflying squirrel (Pteromys volans)

526. Anttila, Katja (2009) Swimming muscles of wild, trained and reared fish. Aspectsof contraction machinery and energy metabolism

527. Roppola, Katri (2009) Environmental applications of manometric respirometricmethods

528. Arhippainen, Leena (2009) Studying user experience: issues and problems ofmobile services. – Case ADAMOS: User experience (im)possible to catch?

529. Niemelä, Marika (2009) Biotic interactions and vegetation management on coastalmeadows

530. Sørensen, Louise Ilum (2009) Grazing, disturbance and plant soil interactions innorthern grasslands

531. Salo, Antti (2009) Expression of lysyl hydroxylases and characterization of a noveldisorder caused by mutations in the lysyl hydroxylase 3 gene

532. Vaismaa, Matti (2009) Development of benign synthesis of some terminal ?-hydroxy ketones and aldehydes

533. Meriläinen, Gitte (2009) Structural and enzymological studies of the thiolaseenzymes

534. Astorga, Anna (2009) Diversity patterns in marine and freshwater environments.The role of environmental and spatial factors across multiple scales

ABCDEFG

UNIVERS ITY OF OULU P.O.B . 7500 F I -90014 UNIVERS ITY OF OULU F INLAND

A C T A U N I V E R S I T A T I S O U L U E N S I S

S E R I E S E D I T O R S

SCIENTIAE RERUM NATURALIUM

HUMANIORA

TECHNICA

MEDICA

SCIENTIAE RERUM SOCIALIUM

SCRIPTA ACADEMICA

OECONOMICA

EDITOR IN CHIEF

PUBLICATIONS EDITOR

Professor Mikko Siponen

University Lecturer Elise Kärkkäinen

Professor Hannu Heusala

Professor Helvi Kyngäs

Senior Researcher Eila Estola

Information officer Tiina Pistokoski

University Lecturer Seppo Eriksson

University Lecturer Seppo Eriksson

Publications Editor Kirsti Nurkkala

ISBN 978-951-42-9268-2 (Paperback)ISBN 978-951-42-9269-9 (PDF)ISSN 0355-3191 (Print)ISSN 1796-220X (Online)

U N I V E R S I TAT I S O U L U E N S I SACTAA

SCIENTIAE RERUM NATURALIUM

U N I V E R S I TAT I S O U L U E N S I SACTAA

SCIENTIAE RERUM NATURALIUM

OULU 2009

A 535

Antti Rivinoja

GOLGI pH AND GLYCOSYLATION

FACULTY OF SCIENCE,DEPARTMENT OF BIOCHEMISTRY,UNIVERSITY OF OULU

A 535

ACTA

Antti R

ivinoja