Probing the plant endomembrane-secretory pathway using heterologous membrane protein markers.

Item Type text; Dissertation-Reproduction (electronic)

Authors Gong, Fangcheng.

Publisher The University of Arizona.

Rights Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.

Download date 30/09/2021 05:30:51

Link to Item http://hdl.handle.net/10150/187119 INFORMATION TO USERS

This manuscript ,has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.

The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.

In the unlikely. event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal SectiOllS with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book.

Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.

UMI A Bell & Howell Information Company 300 North Zeeb Road. Ann Arbor. MI48106·1346 USA 313/761-4700 800:521·0600

PROBING THE PLANT ENDOMEMBRANE-SECRETORY PATHWAY

USING HETEROLOGOUS MEMBRANE PROTEIN MARKERS

by

Fangcheng Gong

A Dissertation Submitted to the Faculty of the

DEPARTMENT OF PLANT SCIENCES

In Partial Fulfillment of the Requirements for the Degree of

DOCTOR OF PHILOSOPHY WITH A MAJOR IN AGRONOMY AND PLANT GENETICS

In the Graduate College

THE UNIVERSITY OF ARIZONA

I 995 UMI Number: 9531138

OMI Microform 9531138 Copyright 1995, by OMI Company. All rights reserved.

This microform edition is protected against unauthorized copying under Title 17, United States Code.

UMI 300 North Zeeb Road Ann Arbor, MI 48103 2

THE UNIVERSITY OF ARIZONA GRADUATE COLLEGE

As members of the Final Examination Committee, we certify that we have

read the dissertation prepared by_---"-F.:::a.!..:n.Qg~c.!..:h~e:..:.nc.g_G"'"o"'"n""gOl.______

entitled Probing the Plant Endomembrane-secretory Pathway

Using Heterologous Membrane Protein Markers

and recommend that it be accepted as fulfilling the dissertation

requirement for the Degree of _~D~o~c~t~o~r~o~f_P~h~i~l~o~s~o~p~h~y~ ______

Dr. David W. Galbraith Date lh~/I. f'~.' 1/61'1S- Dr. Brian A. Larkins Date

/ ! .. _./ /"1 /;'} .. ~ _<- \... . .ft 1 z· - -Dr. Date / (!If} Dr. Dean Della Penna D te

Date

Final approval and acceptance of this dissertation is contingent upon the candidate's submission of the final copy of the dissertation to the Graduate College.

I hereby certify that I have read this dissertation prepared under my direction and recommend that it be accepted as fulfilling the dissertation requirement. I/~ /v1) Dissertation Director Date I 3

STATEMENT BY AUTHOR

This dissertation has been submitted in partial fulfillment of requirements for an advanced dt!gree at The University of Arizona and is deposited in the University Library to be made available to borrowers under rules of the library.

Brief quotations from this dissertation are allowable without special permission, provided that accurate acknowledgment of source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the head of the major department or the dean of the Graduate College when in his or her judgment the proposed use of the material is in the interests of scholarship. In all other instances, however, permission must be obtained from the author.

SIGNED: I'

4

ACKNOWLEDGMENTS

I would like to express sincere thanks to my major advisor, Dr. David W. Galbraith, for his guidance, support, and assistance during the course of my graduate training and during the preparation of this dissertation. Dr. Galbraith not only served as a mentor for my dissertation project, but also provided me the valuable opportunity to practise science independently. I would also like to extend gratitude to my committee members, Drs. Brian A. Larkins, Hans 1. Bohnert, Carolyn A. Zeiher and Dean DellaPenna, for their valuable time, professional advice, and constructive criticism throughout this project. Although Dr. Bohnert did not administrate my final examination due to a conflict of schedule, he has contributed as much as other members to my graduate training and dissertation.

I wish to give my thanks to Drs. L. Andrew Staehelin, and Thomas H. Giddings, and Ms. Janet B. Meehl in Department of Molecular, Cellular and Developmental Biology, University of Colorado at Boulder for their excellent immuno-electron microscopic data.

I am very grateful to Dr. Carolyn E. Machamer (School of Medicine, John Hopkins University) for her generosity in providing plasmids and antibodies, and to Dr. Jean M. Wilson (Department of Anatomy, The University of Arizona) for her suggestions and discussions on immuno-electron microscopic work. I am also indebted to former and current members (Georgina M. Lambert, Ying-Lung Li and Daniel A. Coury, and Drs. Leslie E. Basel and Robert 1. Grebenok) of the Galbraith Laboratory for their help, support, encouragement, discussion, and suggestions at various stages of this project.

Last but not at all least, I wish to express my deep appreciation to my wife, Jingzhe Li, and my daughter, ZhanZhan, for their support and help. Above all, their patience and understanding, which comes from love, empowered them to endure my reduced family time, especially house chores, when I devoted my evening hours and weekends to this project. This could be summarized by my wife's philosophical remark "half of your dissertation belongs to me". 5

DEDICATION

To my wifc Jingzhc Li

and

To my daughtcr ZhanZhan 6

TABLE OF CONTENTS

LIST OF ILLUSTRATIONS ...... 11

LIST OF TABLES ...... 14

ABSTRACT...... 15

STATEMENT OF PENDING PATENT ...... 17

1. STRUCTURES AND SORTING MECHANISMS IN THE CONSTITUTIVE SECRETION OF PROTEINS ...... 18

INTRODUCTION ...... 18

ROUTES OF PROTEIN TRAFFICKING ...... 19

DEFAULT BULK FLOW, AND CONSTITUTIVE AND REGULATED SECRETION ...... 26

STRUCTURE AND FUNCTION OF THE VARIOUS COMPONENTS OF THE ENDOMEMBRANE-SECRETORY PATHWAy...... 28

THE ENDOMEMBRANE CONCEPT AND THE ENDOMEMBRANE-SECRETORY PATHWAy ...... 28

(1) THE ENDOPLASMIC RETICULUM AND THE OUTER NUCLEAR MEMBRANE ...... 29

(2) THE ER-GOLGI INTERMEDIATE COMPARTMENT ...... 33

(3) THE GOLGI ApPARATUS ...... 35

(4 ) VACUOLES AND LYSOSOMES ...... 38

(5) PROTEIN BODIES ...... 40

(6) THE PLASMA MEMBRANE ...... 43

(7) VESICLES ...... 44

MOLECULAR SORTING MECHANISMS INVOLVED IN THE CONSTITUTIVE SECRETION OF PROTEINS ...... 48 7

TABLE OF CONTENTS - Continued

ER TARGETING AND TRANSLOCATION ...... 48

ER RETENTION AND RETRIEVAL MECHANISMS ...... 56

ER Retention of Luminal Proteins from Default Secretion ...... 56

ER Retention of integral Membrane Proteins from Default Secretion ...... 61

TRANSPORT COMPETENCE FOR EXIT FROM THE ER ...... 64

RETENTION OF PROTEINS AT THE CIS-, MEDIAL- AND 7RANS-CISTERNAE OF THE GOlGI ... 65

RETENTION OF PROTEINS AT THE TRANs-GOlGI NETWORK ...... 72

SORTING OF PROTEINS AT THE TRANS-GOlGi NETWORK ...... 75

Trafficking of Cell Surface Proteins ...... 76

Sorting Proteins into Lysosomes ...... 80

Sorting Luminal Proteins into Vacuoles ...... 82

Sorting Membrane Proteins into Vacuoles, or An Alternative Default Pathway? ...... 85

EXPERIMENTAL APPROACHES FOR THE STUDY OF PLANT ENDOMEMBRANE-SECRETORYPATHWAY ...... 88

SUMMARY AND PERSPECTIVE ...... 95

2. GENERAL MATERIALS AND METHODS ...... 98

Generation and Maintenance of Axenic Plantlets of Tobacco ...... 98

Generation and Maintenance of Tobacco Callus and Suspension Cell Cultures ...... 99

Protoplast Isolation and Transfection ...... 101 8

TABLE OF CONTENTS - Continued

Agrobacterium Transformation of Tobacco Leaf-discs ...... 102

Measurement of Protein Content...... 104

Gel Electrophoresis and Western Blotting ...... 105

Preparation and Fractionation of Sucrose Gradients ...... 107

Chlorophyll Measurement ...... 108

NADH Cytochrome C Reductase Assay...... 108

Inosine Diphosphatase Assay ...... 109

Vanadate-sensitive ATPase Assay ...... 110

p-Glucuronidase Assay ...... 111

3. EXPRESSING HETEROLOGOUS MEMBRANE PROTEINS WITHIN PLANT CELLS FOR USE AS PROBES OF THE SECRETORY PATHWAY ...... 112

INTRODUCTION ...... 112

MATERIALS AND METHODS ...... 121

RESULTS ...... 133

EXPRESSION OF HETEROLOGOUS PROTEINS IN TRANSFECTED PROTOPLASTS ...... 133

EXPRESSION OF HETEROLOGOUS PROTEINS IN TRANSGENIC PLANTS ...... 144

SUBCELLULAR DISTRIBUTIONS OF HETEROLOGOUS PROTEINS ...... 150

DISCUSSION ...... 156

4. THE ZIPPERED-MEMBRANE, A NOVEL ARTIFICIAL ORGANELLE FOR EXPRESSING RECOMBINANT INTEGRAL MEMBRANE PROTEINS ...... 162 9

TABLE OF CONTENTS - Continued

INTRODUCTION ...... 162

MATERIALS AND METHODS ...... 165

RESULTS ...... 171

EXPRESSION OF IBY M-GUS IN TRANSFECTED PROTOPLASTS AND TRANSGENIC PLANTS ...... 171

ASSOCIATION OF IBY M-GUS WITH TilE ENDOMEMBRANE SYSTEM ...... 172

COSEGREGATION OF IBY M-GUS WITH THE GOLGI ELEMENTS ON ISOPYCNIC SUCROSE GRADIENTS ...... 175

KINETICS OF BIOSYNTHESIS AND ACCUMULATION OF THE IBY M-GUS CIIIMAERIC PROTEIN WITIHN THE PLANT SECRETORY PATlIWAY ...... 179

IN SITU FLUORESCENCE LABELING OF THE IBY M-GUS CHiMAERIC PROTEIN WITIIIN TRANSGENIC CELLS ...... 182

REVELATION OF NOVEL MEMBRANE STRUCTURES BY IMMUNO ELECTRON MiCROSCOPy...... ••• 187

DISCUSSION ...... 196

5. A SHIFT IN LOCATION OF THE Gml MEMBRANE PROTEIN, FROM THE MAMMALIAN GOLGI TO THE PLANT TONOPLAST: DOES THIS SUGGEST AN ALTERNATIVE DEFAULT SECRETORY PATHWAY IN PLANT CELLS'! ...... 204

INTRODUCTION ...... 204

MATERIALS AND METHODS ...... 207

RESULTS ...... 209

INHIBITION OF GM 1 GLYCOSYLA TION BY TUNICAMYCIN ...... 209

BIOSYNTHETIC KINETICS OF THE LARGER AND SMALLER ISOFORMS OF GM 1 ...... 210 10

TABLE OF CONTENTS - Continued

SUCROSE GRADIENT ANALYSES OF ENDOMEMBRANES BEARING GM 1 ...... 215

IN VIVO PROTEASE K INCUBATION OF GM 1 TRANSFECTED PROTOPLASTS ...... 218

ARRIVAL OF GM 1 AT THE TONOPLAST MEMBRANE ...... 220

PARTIAL DISSOCIATION OF THE SMALLER Iso FORM OF GM 1 FROM THE ENDOMEMBRANES ...... 222

DISCUSSION ...... 222

6. CONCLUSIONS AND PERSPECTIVES ...... 228

APPENDIX I: NTTO NUTRIENT STOCK (lOX) ...... 234

APPENDIX II: MINIMAL T MEDIUM WITH KANAMYCIN ...... 235

APPENDIX III: TOBACCO FEEDER-CELL PLATE MEDIUM ...... 236

APPENDIX IV: TOBACCO SHOOTING MEDIUM ...... 237

APPENDIX V: REAGENT A FOR MICRO-LOWRY ASSAy ...... 238

APPENDIX VI: FERROUS SULFATE-AMMONIUM MOLYBDATE SOLUTION ...... 239

REFERENCES ...... 240 !'

11

LIST OF ILLUSTRATIONS

Figure 1. 1. Diagram of the endomembrane-secretory pathway for proteins within the plant cell ...... 23

Figure 3. 1. Schematic maps of plasmid constructions for the transient gene expression cassettes ...... 123

Figure 3.2. Gel electrophoresis and Western blotting of the IBV M glycoprotein in transfected protoplasts ...... 135

Figure 3. 3. Transient expression of the !BV M-GUS chimaeric protein in tobacco leaf protoplasts ...... 136

Figure 3. 4. Transient expression of the !BV ml~-GUS chimaeric protein in tobacco leaf protoplasts ...... 138

Figure 3. 5. Gel electrophoresis, Western blotting, and detection of expression of the Gm 1 glycoprotein isoforms in transfected protoplasts ...... 139

Figure 3.6. Western blotting ofVSVG in transfected protoplasts ...... 141

Figure 3.7. Gel electrophoresis, Western blotting, and detection ofSYNVG in transfected protoplasts ...... 143

Figure 3.8. Western blotting of the GUS, mv M-GUS, and !BV ml~-GUS proteins synthesized within transgenic plants ...... 146

Figure 3.9. Western blotting of the Gm 1 protein synthesized in transgenic plants ...... 148

Figure 3.10. Expression ofVSVG and its mutants within transgenic plants ...... 149

Figure 3. 11. Expression of SYNVG within transgenic plants ...... 151

Figure 3. 12. Distribution of GUS activities between the subcellular fractions isolated from transfected protoplasts expressing GUS, IBV M-GUS and IBV ml~-GUS ...... 152 12

LIST OF ILLUSTRATIONS - Continued

Figure 3. 13. Subcellular distribution of the Gm 1 glycoprotein expressed in transfected protoplasts ...... 154

Figure 3. 14. Subcellular distribution of SYNVG in transfected protoplasts ...... 155

Figure 4.1. Schematic illustration of the topology of the IBV M-GUS chimaeric protein ...... 164

Figure 4. 2. Western blotting of the GUS and IBV M-GUS proteins in transfected protoplasts ...... 173

Figure 4. 3. Western blotting of the GUS and IBV M-GUS proteins in transgenic plants ...... 174

Figure 4. 4. Extractability of the IBV M-GUS chimaeric protein from endomembranes ...... 176

Figure 4. 5. Distribution of the IBV M-GUS chimaeric protein on isopycnic sucrose gradient ...... 177

Figure 4. 6. Isopycnic gradient distribution of the IBV M-GUS chimeric protein from transgenic plant pFG 14-23 ...... 180

Figure 4. 7. Kinetics ofIBV M-GUS biosynthesis and accumulation within compartments of the endomembrane-secretory pathway...... 181

Figure 4. 8. Relative accumulation rates ofIBV M-GUS cosegregating on sucrose gradients with the ER, the Golgi and the plasma membrane ...... 183

Figure 4. 9. Indirect immuno-fluorescence labeling ofmesophyll cells isolated from leaves of transgenic plants expressing IBV M-GUS ...... 185

Figure 4. 10. Electron micrographs of the pFG 14-23 transgenic suspension cells ... 189

Figure 4.11. Immuno-electron micrographs of the transgenic pFG 14-23 suspension cells ...... 192

Figure 4. 12. Diagram illustrating a possible mechanism of membrane-zippering of E-whorl by oligomerization of the IBV M-GUS chimeric protein ...... 199 13

LIST OF ILLUSTRATIONS - Continued

Figure 5. 1. Effect of in vivo tunicamycin incubation of protoplasts transfected with pFO 21 ...... 211

Figure 5. 2. Kinetics of Om 1 biosynthesis and degradation within transfected protoplasts ...... 212

Figure 5. 3. Relative rates of Om 1 biosynthesis and degradation ...... 214

Figure 5. 4. Distribution of Om 1 within isopycnic sucrose gradients ...... 216

Figure 5. 5. Protease K susceptibility of Om 1 within viable transfected protoplasts ...... 219

Figure 5. 6. Western blotting of proteins from intact vacuoles isolated from transgenic plants pFO 23-7 ...... 221

Figure 5.7. Partial dissociation of the Gm 1 smaller isoform from the endomembranes ...... 223 14

LIST OF TABLES

Table 3.1. Summary of the DNA plasmids employed for protoplast transfection and leaf-disc transformation of tobacco ...... 128

Table 3. 2. Primary antibodies, conjugated secondary antibodies and their usage in Western blotting ...... 132

Table 3. 3. Number of positive transformants and their percentages within the screened population ...... 145

Table 4. 1. Reported cases of membrane proliferation as a result of expressing integral membrane proteins ...... 200 I,

15

ABSTRACT

Using gene expression systems developed in Nicotiana tabacum, the transport of

various heterologous membrane proteins (SYNVG, IBV M, VSVG, and their

derivatives) was studied within the plant secretory pathway. The results of these

experiments indicated that the proteins accumulated in various endomembrane

compartments of plant cells, but not in those counterparts observed in their native hosts.

These results suggest that membrane proteins of mammalian and plant cells share a

well-conserved mechanism for ER targeting and translocation, but differences exist in

their machinery for retention and sorting within the secretory pathway.

Part of the project was aimed towards the immuno-isolation of the plant Golgi (or

its sub-cisternal elements) using heterologous expression ofIBV M and its derivatives.

The IBV M molecule was found to be undetectable in transfected pro top lasts or

transgenic plants. Two chimeric derivatives of IBV M, although stably expressed, were

found to be targeted to compartments other than the Golgi. Thus, immuno-isolation of

the plant Golgi could not be achievt:d using this approach.

IBV M-GUS (an IBV M derivative) was found to accumulate exclusively in a

novel membrane proliferation of the ER, termed an "E-whorl". E-whorls appear devoid

of ER-resident enzymes, and their production causes no developmental abnormalities in

transgenic plants. A model has been developed in which homo-oligomeric assembly

following biosynthesis, is sufficient to prevent forward movement of the IBV M-GUS 16 chimera to the Oolgi. Homo-oligomerization also serves to sequester the chimera away from ER resident proteins, and drives the generation of the novel membrane compartment. E-whorls might provide a useful tool for selective accumulation of important proteins that are unstable or insoluble within or toxic to host cells, as well as for amplification of the membrane compartment housing the proteins.

. Om 1 (another IBV M derivative) was found to be transported into the plant tonoplast, via a trafficking route from the ER and Oolgi to the vacuole. Om 1 has no lysosomal (or vacuolar) sorting information in either mammalian or plant cells; its targeting to the plant vacuole suggests that an alternative default secretion pathway might exist for the transport of membrane proteins, as reported in yeast. 17

STATEMENT OF PENDING PATENT

A patent application on a novel protein purification teclmique, which was developed based on the discovery ofE-whorls (Chapter 4), has been filed in the Patent

Office of the United States of America by the University of Colorado and the University of Arizona. Permission for commercial application of that patent must be obtained from the patent holders. 18

CHAPTER 1

STRUCTURES AND SORTING MECHANISMS IN

THE CONSTITUTIVE SECRETION OF PROTEINS

INTRODUCTION

Life constitutes a series of complex, controlled, and interrelated chemical reactions. These reactions require and produce diverse chemicals, whose properties range from oxidative to reductive, from inorganic to organic, and from alkaline to acidic.

Cellular chemical reactions must occur in highly-ordered temporal and spatial sequences, and must maintain favorable energy fluxes to sustain life within the context of a variable external environment. In order to accommodate this chemical diversity, regulate the individual reactions, and communicate with their surroundings, cells have evolved organized lipid membrane compartments (or organelles) to separate diverse reactions and sequester different compounds both spatially and temporally. These membrane compartments, in eUkaryotic cells, comprise the nucleus, the rough and smooth

Endoplasmic Reticulum (ER), the Golgi apparatus, mitochondria, various vesicles, microbodies (peroxisomes/glyxyosomes), Iysosomes in manlmalian cells, vacuoles

(lysosomal counterparts) in plant and yeast cells, plastids, oil bodies and protein bodies in plant cells. The organized membrane compartments provide the fundamental infrastructure underpinning diverse cellular functions. 19

ROUTES OF PROTEIN TRAFFICKING

In order to establish and maintain discrete membrane compartments, cells have

evolved two primary pathways. The first involves intracellular transport and targeting of

proteins from their sites of synthesis to their ultimate destinations. The second involves

the recycling of proteins from the plasma membrane to the appropriate intracellular

organelles. These two pathways are termed the biosynthetic and endocytotic protein

transport pathways respectively. Endocytic protein transport is specifically defined as

the selective movement of proteins from the plasma membrane to the Golgi apparatus,

lysosomes (vacuoles), or the ER. The massive body of literature on endocytosis,

extensively reviewed for both mammalian and plant cells (Goldstein et aI., 1985;

Gruenber and Howell, 1989; Hom et aI., 1989; Fowke et aI., 1991; Oparka et aI., 1991;

Anderson et aI., 1992; Schmid, 1992), is beyond the scope of this thesis. My work

concerns the pathway of biosynthetic protein transport, and broadly includes the

trafficking of all newly-synthesized polypeptides encoded by the nuclear genome, rather

than employing the narrow definition designated for exocytosis (Pfeffer and Rothman,

1987; Pryer et aI., 1992). Biosynthetic transport can be divided into three major

branched routes: retention of proteins in the cytosol, selective import of proteins from the cytosol into the nuclei, mitochondria, chloroplasts, and the other organelles, and exocytosis of proteins through membrane compartments via a mechanism of vesicle flow.

The first two routes will be briefly explained later. The third route (exocytosis) can be 20 further separated into the constitutive and regulated secretory pathways, of which the constitutive pathway will be extensively covered in this chapter. The other minor pathways of protein trafficking, including localization of proteins to the oil bodies and trafficking of proteins synthesized in mitochondria and chloroplasts, are beyond the scope of this review.

One of the features of protein biosynthesis is the existence of oligopeptide targeting signals, which reside in contiguous blocks of amino acid sequence or in specific three-dimensional arrangements of amino acids on the protein surface after polypeptide folding and assembly. Positive sorting, exemplified by import from the cytosol into the nucleus, mitochondria, chloroplasts, or the ER, represents the first targeting decision made during or after protein biosynthesis, and is based on the type of targeting signal present within the polypeptide. Retention of proteins in the cytosol has long been perceived as a passive default. Thus, most native isolated cytosolic proteins lack the targeting signals required for proteins to enter membrane-bound compartments.

However, recent reports suggest that some nuclear proteins may be retained in the cytosol through specific signals. This selective mechanism for nuclear entry or cytosolic retention is regulated by protein phosphorylation during the cell cycle (Hunt, 1989; Moll et aI., 1991; Jans et aI., 1991; Moll et aI., 1992; Hennekes et aI., 1993). In the case of the

Xenopus nuclear factor 7 (Xnt7), the cytoplasmic retention domain (CRD) is masked within the tertiary structure of the protein ifXnt7 is un- or de-phosphorylated. Under this condition, Xnt7 is targeted into the nucleus under the influence of an independent nuclear 21

localization signal (Li et aI., 1994). When Xnf7 is phosphorylated, the eRD becomes

exposed to the outside surface of the protein, and physically engages with a cytoplasmic anchor. Thus Xnf7 is actively retained in cytoplasm despite its nuclear localization signal

(Li et aI., 1994). It is not clear whether active retention of proteins in cytoplasm is universally present in all eukaryotic cells. At least in the cells of Xenopus embryos, passive and active retention systems coexist inside the same cells, in order to satisfy the demands of the developmental regulation and the constitutive house-keeping functions.

Selective import has been widely studied for proteins, encoded in the nuclear genome, destined to reside in the nucleus, mitochondria, chloroplasts, peroxisomes, and glyoxysomes. This group of proteins commonly possess either uncleavable or cleavable oligopeptide localization signals for the specific subcellular compartments within the primary sequences of the individual proteins. Most of these proteins move directly to their targets through diffusion in cytosol after translation. Some evidence suggests that cotranslational import exists for certain mitochondrial proteins (Kellems et aI., 1975;

Keith, 1993). Translocation across the lipid membrane bilayer is both receptor-mediated and energy-dependent. A large amount of information is available in this area (Keegstra and Olsen, 1989; Garcia-Bustos et aI., 1990; Restrepo et aI., 1990; Silver, 1991; Baker and Schatz, 1991; Swinkels et aI., 1992; Dingwall and Laskey, 1992; Glover and

Lindsay, 1992; Johansson and Forsman, 1992; Segui-Real et aI., 1992; Glick et aI., 1992;

Raikhel, 1992; Subramani, 1992 and 1993; Kaldi et aI., 1993; SuIter et aI., 1993;

Vanderklei et aI., 1993; Olsen et aI., 1993). 22

Exocytosis involves the process of protein trafficking in an anterograde manner along a specific series of cellular membrane compartments. The exocytotic pathway consists of the ER, the Golgi apparatus, the plasma membrane, various types of vesicular and/or tubular intermediates that comprise dynamic links between different membrane compartments, plus secretory granules and lysosomes in the mammalian cells, vacuoles in the yeast and the plant cells, and protein bodies in the plant cells. The overall direction of the traffic is from the ER to the plasma membrane via the Golgi apparatus

(Figure 1. 1). Proteins entering the exocytotic pathway are first synthesized at the rough

ERicytosol interface, then targeted into the ER via various signal sequences (Rapoport,

1986). Secreted and luminal proteins are completely translocated across the ER membrane after biosynthesis, whereas integral membrane proteins are partially translocated across membrane, and become anchored in the lipid bilayer. Membrane proteins are inserted in the lipid bilayer in four major ways: (l) Type I (a single transmembrane domain with N-terminus in lumen and C-terminus in cytosol), (2) Type II

(a single transmembrane domain but with the Nand C-termini in the opposite to the orientation of Type I proteins), (3) Type III (more than one transmembrane domain regardless of the membrane-sideness of the N- and C-termini, but excluding Type IV proteins) and, (4) Type IV (more than one transmembrane domain to form a permanent channel for water and solute transport across the membrane) (Singer, 1990). In the ER, 23

Figure 1. 1. Diagram of the endomembrane-secretory pathway for proteins within the plant cell. The discrete compartments in the pathway communicate with one other via secretory vesicles (solid thin-line circles). Membrane, luminal and secretory proteins are transported from the ER to the plasma membrane/cell surface or vacuole

(protein body) via the Golgi. Retention of luminal proteins in the ER is achieved through a retrieval mechanism operating between the ER and the salvage compartment

(this is still a hypothetical compartment in plant cell). The ER-derived protein body is formed through aggregation of storage proteins within the ER lumen. 24

Nucleoplasm

Ribosomes

Goigi

Cell Walland Extracellular Space 25

proteins undergo various co- and post-translational modifications (Lang et aI., 1984;

Kornfeld and Kornfeld, 1985; Berger and Schmidt, 1985; Elbein, 1987; Jones and

Robinson, 1989; Freedman, 1989). Only properly folded and assembled proteins can

exit the ER via a process of vesicle budding, and misfolded proteins are retained and

subsequently degraded, probably within the ER (Klausner et aI., 1990). The salvage

compartment, which is a still hypothetical compartment for plant cells, is currently

recognized as a structure functioning in the retrieval of ER resident proteins that have

escaped from the ER (Warren, 1987; Lippincott-Schwartz et aI., 1990). The Golgi

apparatus is a dynamic organelle that plays a pivotal role in controlling the bi-directional

flow of proteins and other membrane components, such as lipids, and sorting them to the

subsequent branched routes of the: exo- and endo-cytotic pathways (Farquhar, 1985). The

Golgi apparatus is also responsible for the carbohydrate modification of both proteins and

lipids as they pass through the cisternae. The trans Golgi network (TGN) comprises an

extensive tubular/vesicular network located at the trans surface of the Golgi apparatus. It

is the site of sorting and packaging for different types of membrane vesicles, which then exit from the Golgi stacks into a series of branched routes, including transport to the secretory granules and lysosomes of mammalian cells, vacuoles of plant and yeast cells,

and the different domains of the plasma membrane (Griffith and Simons, 1986; Rothman and Orci, 1992; Mellman and Simons, 1992). 26

DEFAULT BULK FLOW, AND CONSTITUTIVE AND REGULATED SECRETION

It is popularly believed that in most eukaryotic cells, upon entry into the ER, proteins are sorted and transported under the influence of bulk vesicular flow (Pryer et ai.,

1992). This concept has been confirmed to exist in the secretory pathway of plant cells

(Fowke et ai., 1991; Coleman et ai., 1991; Robinson et ai., 1991; Steer and O'Driscoll,

1991; Harris and Watson, 1991; Chrispeels, 1985 and 1991; Vitale and Chrispeels,

1992). Constitutive (unregulated) secretion involves the movement of proteins anterogradely along the pathway to the cell surface. This bulk flow movement is thought to occur by default (i. e. no sorting signal is required), whereas all diversions from the default pathway to specific compartments or branched routes are signal-mediated. These diversions include the recycling mechanism to retrieve ER resident proteins from the

Golgi, the retention scheme preventing Golgi resident proteins from moving out ofthe

Golgi by default bulk flow, and the sorting machinery required to divert soluble proteins into vacuole (Burgess and Kelly, 1987; Kornfeld and Mellman, 1989; Pelham, 1989;

Weisz et ai., 1993). It appears that no additional signals are required for the correct targeting of soluble proteins destined to the extracellular space. In constitutive secretion,

loaded vesicles constantly move from the TGN to fuse with the plasma membrane to release their contents, and this process does not require external stimuli. Constitutive secretion is closely coupled to the biosynthesis of proteins, in that proteins are secreted as fast as they are synthesized. This means that secretory vesicles have a short half-life. !.

27

Constitutive secretion with default bulk flow has served as the basic model for most

investigations of secretion through and retention within the exocytotic pathway.

Regulated secretion along the exocytotic pathway is currently viewed as a variant

of constitutive secretion (Kelly, 1985). It is typically found in specialized mammalian

cells, such as neurons, endocrine cells, exocrine cells, neutrophils and platelets.

Regulated secretion has been intensively sought in the plant kingdom, but its existence

has not been unambiguously established, even in cases for which regulated secretion

would appear likely, for example in the digestive glands of insectivorous plants (Robbins

and Juniper, 1980a, 1980b). In animal cells, the secretory products of regulated secretion

are accumulated to high concentrations within specialized secretory granules, which are

stored in the cytosol in close vicinity to the plasma membrane. Unlike secretory vesicles,

the secretory granules are inhibited from fusing with the plasma membrane until the

2 levels of specific cytosolic messengers (such as Ca +, cAMP and cGMP) are altered, or

until specific cell surface receptors bind to their hormones (Farquhar, 1985; Burgess and

Kelly, 1987). Regulated secretion can over a brief period release large amount of

proteins at a rate much higher than their rate of synthesis. In order to achieve rapid

transient rates of secretion, the secretory granules are packed with large amount of

newly-synthesized proteins, becoming morphologically distinguishable with a "dense

core" under electron microscope. It is clear that regulated and constitutive secretion

pathways can coexist within the same cells (Moore et aI., 1983; Kelly, 1985). Regulated

secretion puts another layer of complexity on already complicated constitutive secretion; 28 it also shows that various trafficking mechanisms exist in different types of cells to accommodate diverse biological functions.

STRUCTURE AND FUNCTION OF THE VARIOUS COMPONENTS

OF THE ENDOMEMBRANE-SECRETORY PATHWAY

THE ENDOMEMBRANE CONCEPT AND THE ENDOMEMBRANE-SECRETORY PATHWA Y

The endomembrane can be collectively defined as a structural, functional and developmental continuum of cellular membranes in eukaryotic cells, excluding the inner nuclear membrane and the membranes of mitochondria and chloroplasts. A central function of the endomembrane is the process of exo- and endo-cytosis of proteins. The endomembrane concept was suggested by MOITI~ et ai. (1971 and 1974), based on the ontogenetic, structural and functional continuity of the membranes seen within eukaryotic cells. The endomembrane includes all the aforementioned membrane components of the exocytotic pathway (also see Figure 1. 1), plus oil bodies, microbodies and the outer membrane of the nuclear envelope ( Mom! et aI., 1974; Bennett and Osteryoung, 1991).

The endomembrane-secretory pathway comprises a subset of the components of the endomembrane continuum. It represents an exocytic trail followed by various proteins, oligo saccharides, and lipids, and involves transport vesicles in their movement between 29 specific components of the endomembrane continuum. The oil bodies and the microbodies will not be discussed further as they do not fonn part of the endomembrane­ secretory pathway. Two other non-membrane components involved in the endomembrane-secretory pathway, the cytoskeleton and the motor proteins, will not be discussed here, as the mechanics of vesicle movement is beyond the scope of this review.

The endomembrane-secretory pathway can be functionally divided into seven areas: (1) the endoplasmic reticulum and the nuclear outer membrane, (2) the ER-Golgi intennediate compartment, (3) the Golgi apparatus, (4) vacuoles and Iysosomes, (5) protein bodies, (6) the plasma membrane and, (7) vesicles. The features of these components are as follows. It should be noted that most if not all of our infonnation comes from non-plant cell system, particularly mammalian and yeast systems.

(1) THE ENDOPLASMIC RETICULUM AND THE NUCLEAR OUTER MEMBRANE

On the basis of the presence of ribosomes, the ER can be morphologically described as being composed of the rough and smooth elements, which show structural continuity. Ribosomes are also found on the surface of some areas of the outer membrane of the nuclear envelope, and this membrane is frequently observed to be continuous with the rough ER (Newport and Forbes, 1987) (Figure 1. 1). In plant cells, the ER is additionally observed to foml a part of plasmodesmata that connect contiguous cells I,

30

(Robards and Lucas, 1990; Tilney et al., 1991). The structural continuum comprising the

ER-nuclear outer membrane can be separated into four distinct regions: the outer nuclear

envelope, the rough ER, the transitional elements, and the smooth ER; and each has

specific biosynthetic functions (Rose and Doms, 1988). The rough ER is mainly

involved in protein biosynthesis, the smooth ER in phospholipid biosynthesis, and the

transitional elements are involved in vesicle budding, fission, and secretion (Rose and

Doms, 1988; Mathews and van Holde, 1990a). It is 110t clear whether the outer nuclear

envelope intrinsically possesses specialized functions involved the secretory pathway, or

whether it has access to those functions as a consequence of its continuity with the rough

ER. Recent evidence suggests that the vesicles for the postmitotic assembly of the

nuclear envelope are derived from a ER-derived sUbpopulation, which is enzymatically

distinct from majority of ER vesicles within the secretory pathway (Wilson and Newport,

1988; Vigers and Lohka, 1991).

Whereas in the context of protein synthesis and secretion, the outer membrane of

the nuclear envelope is viewed as being part of the endomembrane system, it should be

noted that soluble nuclear proteins are actually localized to the nucleus through import

via nuclear pores rather than through the endomembrane-secretory pathway (Silver,

1991). Together with the inner membrane and the nuclear pore complexes, the outer

nuclear membrane forms a diffusion barrier separating nucleoplasm from cytoplasm

(Dingwall and Laskey, 1992). Although study on the specific functions of the outer

nuclear envelope are impeded by its continuity with the rough ER, it has been shown that 31 the outer nuclear membrane can perform at least some of the major functions of the rough

ER. For example, Puddington et al. (1985) showed that the synthesis and post-translational modification of vesicular stomatitis virus glycoprotein (VSVG) can be partially conducted on the outer nuclear membrane. Other functions of the outer nuclear membrane include serving as a specialized site for the selective insertion of certain proteins for the nuclear envelope (Rose and Doms, 1988). Further proposed functions of the outer nuclear membrane in the secretory pathway remain complicated by its continuity with the ER, and are supported by scant data, Thus any extending of these functions requires caution.

The ER is a primary site for the synthesis of phospholipids and sterols, which are major components of cytoplasmic membranes. In addition, the ER studded with ribosomes is the predominant site of synthesis of proteins entering the endomembrane-secretory pathway. The continuous biosynthetic activities of the ER means that it acts as an origin both of the membranes and of the proteins destined for the secretory pathway. Another synthetic ability of the ER is to manufacture the lipid-linked polysaccharide precursors for N-glycosylation of glycoproteins (Hirschberg and Snider,

1987).

The ER is responsible for the co- and post-translational modification of newly­ synthesized proteins. The ER contains signal peptidases to remove the cleavable signal peptides from preproteins (Alberts et aI., 1994). The ER also contains glycosyltransferases which progressively modify glycosylation of glycoproteins at I'

32

N-consensus sites (oligosaccharide chain are attached at Asn-X-Ser/Thr sites, where X

can be any amino acid except proline or aspartic acid) (Kornfeld and Kornfeld, 1985).

The ER, particularly in plant cells, also extensively modifies glycoproteins through

O-linked glycosylation of hydroxylproline, serine and threonine residues (Jones and

Robinson, 1989). Fatty acylation of cysteine occurs for a subset of membrane proteins in

the ER (Berger and Schmidt, 1985). Other enzymatic activities in the ER include

intranlOlecular disulfide bond formation by protein disulfide isomerase (Freedman, 1989;

Gething and Sambrook, 1992; Bardwell and Beckwith, 1993), proline hydroxylation of

protein in plant cells by prolyl-4-hydroxylase (Jones and Robinson, 1989), and folding,

oligomerization and assembly of proteins by the binding protein (BiP) (Gething and

Sambrook, 1992). Not all these posttranslational modifications are essential for the

correct protein transport. It appears that folding and oligomerization, which are

sequentially interrelated, are the only critical requirement for proteins to exit the ER

(Klausner et aI., 1990). In the absence of correct folding and oligomerization, proteins

fail to enter the default bulk flow, and are degraded (Klausner et aI., 1990).

The other responsibility of the ER is to execute events directly involved in the

traffic of membranes and proteins destined to the various endomembrane components,

including docking of signal recognition particles (SRP) and ribosomes, translocation of

proteins across the ER membrane, retention of the resident ER proteins, and sorting and

packaging of proteins into vesicles. 33

(2) THE ER-GOLGI INTERMEDIATE COMPARTMENT

It is difficult to delimit precisely the morphological and physical boundary of the

ER-Golgi intermediate compartment (ERGlC) in eukaryotic cells. At least three versions

ofERGIC have been proposed: (1) a specialized portion of the ER transitional elements

distal to the ER; (2) a luminally-interconnected tubular and cisternal structure of the

cis-Golgi and; (3) a distinct compartment between the ER and the Golgi, which is not

continuous with either of them, and that is amplified morphologically and structurally

under certain physiological conditions (Pelham, 1988 and 1989; Schweizer et aI., 1991;

Mellman and Simons, 1992). The confusion concerning the physical border is also

reflected by a variety of names that have been proposed: the post-ER compartment, the

cis-Golgi network, the pre-Golgi compartment, the salvage compartment, and ERGIC.

For convenience, the term ERGIC will be used in this review.

The ERGIC was originally described morphologically, as a group of large

vesicles formed between the ER and the cis-Golgi during the expression of Semliki

Forest virus membrane glycoprotein in BHK cells at 15°C (Saraste and Kuismanen,

1984). When the cells was shifted to higher temperatures, the effect of the low temperature in blocking the transport of newly-synthesized viral glycoproteins from the

ER to the Golgi stack is automatically reversed. The glycoproteins are rapidly transferred to the Golgi cisternae, and the amplified intermediate compartment disappears

(Saraste and Kuismanen, 1984). This phenomenon were later observed in other types of 34

mammalian cells including exocrine pancreatic cells (Saraste et aI., 1986; Tartakoff,

1986). Detailed morphological and biochemical studies revealed that the ERGle has

physical properties and resident proteins distinct from the ER and the cis-Golgi; it is not

affected by chemical agents that disrupt the ER or the Golgi, it lacks the resident proteins

characteristic of the ER and the Golgi, and mainly comprises smooth membrane vesicles

(Mellman and Simons, 1992). So far, no report have appeared concerning the existence

of the ERGle or an equivalent in plant cells.

Due to the poorly-defined boundary of the ERGle, its specific enzymatic

functions have sometimes changed, being attributed to either of its membrane neighbors

(the ER or the Golgi). Initially, the ERGle was proposed to function as a discrete

sorting site to retrieve escaped ER luminal proteins (Warren, 1987; Pelham, 1989).

Later, the ERGle has been suggested to be part of the cis Golgi network, and its

functions were consolidated into those of the Golgi apparatus (Lewis and Pelham, 1992;

Lippincott-Schwartz, 1993). The boundary issue of the ERGle has not been completely

settled (Hauri and Schweizer, 1992). Regardless of the issue on the morphological

boundary of the ERGle, the ERGle should also carry the functions reported previously

when it was considered as a discrete compartment, such as palmitolation of viral and cell surface proteins and phosphorylation of oligo saccharides attached to lysosomal enzymes

(Berger and Schmidt, 1985; Pelham, 1988). 35

(3) THE GOLGI ApPARATUS

The Golgi apparatus can be considered the hub of vesicle traffic in the

endomembrane- secretory pathway. In many eukaryotes, it displays an obvious polarity of organization from the cis to the trans side, having discrete disc-shaped cisternae (cis-, medial-, trans-cisternae) and a region termed the trans Golgi network (TGN), interconnected luminally with the trans-cisternae (Figure 1. 1). The polarity of the cisternae, coupled to the discrete and unequal enzyme distributions within the Golgi, allow it to efficiently process large quantities of secretory proteins passing through its various compartments in a unidirectional and sequential manner.

In general, The Golgi of mammalian and plant cells have similar ultrastructural features. However, they are different in many specific aspects. For example: (1)

Mammalian cells usually have only one Golgi, which is located in a juxtanuclear or perinuclear position. In contrast, plant cells contain many, even hundreds, ofGoigi, which are dispersed throughout the cytoplasm; thus, in most plant cells, the Golgi lack a fixed spatial relationship with the ER and the nucleus (Driouich et aI., 1993b). (2) In mammalian cells, the number of cisternal stacks in each Golgi is greater than that of its plant counterpart (Zhang and Staehelin, 1992). (3) In comparison with the well-developed cis-cisternae in mammalian cells, the Golgi of the plant cells have only one or two small, and less-discernible cis-cisternae (Zhang and Staehelin, 1992). It is not clear whether the less-developed cis-cisternae reflect a lower level of traffic of proteins from the ER or 36

reflect a large number of individual Golgi. (4) The TGN in the plant Golgi is also

different from that seen in mammalian cells. Not all plant cells, or the Golgi stacks within them have a distinctive TGN (Driouich et aI., 1993b). (5) The responses of the Golgi to

Brefeldin A (BFA) are different between mammalian and plant cells. In mammalian cells, BFA blocks protein secretion by causing a complete disassembly of the Golgi, the components of which redistribute into the ER. BFA treatment of the plant Golgi inhibits not only the processing and secretion of glycoproteins, but also the secretion of polysaccharides (Driouich et aI., 1993b). However, the plant Golgi stacks do not disassemble when treated with concentrations of BF A (2.5-7.5 flg/mL) that lead to disaggregation in mammalian cells. Instead, the plant Golgi cisternae aggregate to form

Golgi matrices containing multiple stacks. At higher concentrations of BF A (50-200 flg/mL), the plant Golgi stacks disaggregate into vesicles, but these vesicles accumulate as masses within the cytoplasm rather than redistributing back into the ER (Satiat­

Jeunemaitre and Hawes, 1992). The comparisons listed above support the idea that the function and general structure of the Golgi is conserved between mammalian and plant cells, but that the structural details are diversified to meet different biochemical demands.

A major function of the Golgi is to process oligosaccharide side chains ofN- and

O-linked glycoproteins. Newly-synthesized glycoproteins are sequentially exposed to a battery of glycan modification enzymes individually sequestered in the cis, medial. and

Irans cisternae of the Golgi. This converts N-linked high mannose glycans of the side-chains to complex type as glycoproteins move vee tori ally through the various 37 cisternae (Kornfeld and Kornfeld, 1985). Plant and animal Golgi posses common features for processing N-linked glycoproteins, but the type of glycosylation in plant cells is slightly different from that in animal cells. The differences in N-linked glycosylation between mammalian and plant cells emerge after glycoproteins are processed into a form (below) resistant to endo-p-N- acetylglucosaminidase H (Endo-H)

Aln-GlcNAc-GlcNAc-Man.>t~n I I 2 I 'M~n-GlcNAc by mannosidase II in the Golgi (Kornfeld and Kornfeld, 1985). The complex glycans of the side chains in plant cells are characterized by an additional xylose attached by a

P-I,2linkage to the core mmmose (Man I) through xylosyl transferase at the medial

Golgi, and by an additional fucose attached by an a.-1,3 linkage to the proximal

N-acetylglucosamine (GlcNAc I) through fucosyl transferase at the trans Golgi and the

TGN ( Faye et al., 1986, 1987 and 1989; Laine et aJ., 1991; Zhang and Staehelin, 1992).

In addition, the branched terminal chains in plant cells are different from those in mammalian cells. Some small complex glycan side chains in plant cells are characterized by a further removal of GlcNAc 3, leaving Man 2 and Man 3 as terminal residues (Driouich et aJ., 1993a; 1993b). Large complex glycans in plant cells are elongated by adding two more mannoses by P-I, 2 linkages to Man 2 and Man 3 respectively, and further elongated by adding GlcNAc, galactose, and fucose to both branched chains (Driouich et aJ., 1993a; 1993b). Completed complex glycans in plant 38

cells do not have the terminal sialic acid residues found on their mammalian counterparts

(Kornfeld and Kornfeld, 1985). The modification of O-linked glycoproteins in plant

cells is inferred to start at least from the rough ER where prolyl hydroxylase is located

(Jones and Robinson, 1989). Using a technique that immuno-Iabels terminal sugar

residues on the side chains of extensin I, Moore et al. (1991) have found that the plant

cis-Oolgi cisternae contain all of the arabinosyl transferases required to initiate and

complete O-linked glycosylation of hydroxyl proline-rich glycoproteins. It is still an open

question as to whether arabinosylation is confined to the cis-cisternae or extends further

into the medial and/or the trans cisternae. Another major function of the Oolgi apparatus

is to sort and pack various proteins and glycoproteins into specific vesicles destined to the

different cellular sites at the TON. The vesicles are then attached to various

microtube-based motor proteins which transport them to specific subcellular locations

(Farquhar, 1985; Vale, 1987 and 1992). In addition, the plant Oolgi has an additional,

highly significant role in synthesizing complex polysaccharides, sOlting and targeting

them to the correct cell wall domains; this role is unique to the plant Oolgi (Moore et aI.,

1986; 1991; Drouich et aI., 1993 b).

(4) VACUOLES AND LYSOSOMES

The majority of the volume of mature plant cells comprises amorphous vacuole(s), which are enclosed and separated from the cytoplasm by a single 39 lipid-bilayer membrane. Vacuoles consist of sap and the tonoplast, the sap being defined as the liquid-phase content inside the vacuole, and the tonoplast being the membrane structure that defines the border between the cytoplasm and the vacuole (Matile, 1978).

Vacuoles and lysosomes are viewed loosely as being plant and mammalian cell counterparts, since they share similar structures and some common functions.

Nonetheless, they exhibit many differences (Kornfeld and Mellman, 1989; Wink 1993).

Vacuoles or lysosomes are the final destinations for certain types of proteins in both biosynthetic and endocytotic pathways. Specialized vacuoles in plant cells serve as protein storage sites for reserve or defense proteins (see "PROTEIN BODIES" in the next section). This group of proteins are predominantly channeled into vacuoles by vesicles from the ER via the Golgi apparatus (Chrispeels, 1985), or are elaborated from the ER

(Larkins and HUl'kman, 1973).

Beside their functions in the endomembrane-secretory/endocytotic pathways, vacuoles have many other important functions, which will be briefly listed in this paragraph (for details, see Wink, 1993). Vacuoles contain hydrolytic enzymes and, in an analogous manner to lysosomes, they serve as terminal compartments for the degradative hydrolysis of nucleic acids, proteins and glycosides (Kornfeld and Mellman, 1989;

Wink, 1993). Vacuoles in many plant cells act as temporary storage sites for inorganic ions, sugars, amino acids; and transport mechanisms for these components exist in the tonoplast. Vacuoles also store a variety of toxic defense compounds and compounds that act as specific attractants. These are accumulated in tissue-specific and strategically- 40

favorable patterns (such as in the epidermal cells), both to prevent auto-toxicity and to

attract pollen carriers. Because the vacuolar sap has a high osmotic pressure, due to the

storage of polar metabolites and ions, a critical function of vacuoles is turgor regulation.

Finally, vacuole is involved in cellular pH homeostasis (Wink, 1993).

(5) PROTEIN BODIES

Protein bodies are organelles specialized for the storage of reserve proteins such

as albumins, prolamines, globulins and glutelins, and are particularly exemplified in the

endosperm and cotyledonary cells of seeds. They comprise protein aggregates

surrounded by a lipid membrane, and characteristically appear as semi-opaque or opaque

granules under the electron microscope. Protein bodies can be classified into two

categories based on their origination: ER-derived protein bodies, and vacuole-derived

protein bodies (or protein storage vacuoles).

The ER-derived protein bodies are typically found in the seeds of monocots, and

have been extensively studied in rice, maize, and sorghum. They are also termed

prolamine protein bodies as they are specialized to store prolamines (alcohol soluble

proteins). The prolamine protein bodies are directly formed from the rough ER as they are observed to be associated both with polyribosomes and NADH cytochrome-c reductase activity (Larkins and Hurkman, 1978; Lending et aI., 1989). Though prolamines lack typical ER retentional signals (such as KDEL and HDEL), they are 41 retained in the ER lumen, and subsequently aggregate into insoluble protein bodies.

BiPs are observed to be associated with prolamines at various stage of their synthesis, folding and assembly, and are postulated to facilitate the folding and assembly of prolamines into protein bodies (Li et al., 1993). The deposition of four different maize prolamines (a, p, y- and 8-zeins) was found to follow an ordered pattern during the development of protein bodies, suggesting that protein-protein interactions are required for the sequential aggregation of different zeins and subsequent protein body formation

(Lending and Larkins, 1989). The ER-derived protein aggregates, which closely resemble the prolamine protein bodies, can be artificially created in transgenic dicots by expressing vicilin (pea vacuolar storage protein) or a reporter protein fused with KDEL sequence (Denecke et al., 1992; Wandelt et al., 1992). Interestingly, expressions of a­ and p-zeins in transgenic tobacco leads to different fates; a-zeins are quickly degraded in the ER after synthesis, whereas p-zeins are accumulated in the protein storage vacuoles instead of ER-derived protein bodies (Hoffman et al., 1987; Ohtani et al., 1991).

The protein storage vacuoles are involved mainly in the storage of globulins, the salt-soluble storage proteins. Occasionally they also contain prolamines, such as in the seeds of wheat, barley, and oat (Krislman et al., 1986; Kim et al., 1988). The protein storage vacuoles are easily discriminated from the prolamine protein bodies by the absence of ribosomes on the enveloping membrane (Lending et al., 1989). The trafficking route of vacuolar storage proteins is still controversial. In most cases, the I'

42

proteins are synthesized in the rough ER, then are transported to vacuoles via the

Golgi-derived vesicles, the overall process being similar to that of the secretory proteins

(Chrispeels, 1985). However, recent evidence suggests that a substantial fraction of the

wheat storage proteins is aggregated into small protein bodies within the rough ER and is

then transported as intact small protein bodies to vacuoles by a novel route that does not

utilize the Golgi apparatus (Levanony et aI., 1992).

Regarding the biogenesis of the vacuole-derived protein bodies, a similar process

has been seen in dicots; thus, the central vacuoles of developing legume cotyledons were

observed to become subdivided into small compartments, and the storage proteins were

deposited into those small compartments and became aggregated into protein bodies

(Craig et aI., 1979). Protein-protein interactions are also reported to be necessary for the

assembly of the vacuole-derived protein bodies. The biosynthesis of lectins and storage

proteins follows a define temporal order (vicilin, lectin, convicilin, then legumin) in the

developing seeds of pea and other Leguminosae, with concurrent alteration of vacuolar

morphology (Wenzel et al. 1993). Thus, onset ofvicilin synthesis coincides with the

formation of regularly-shaped protein globules within vacuole; the globules flatten and

begin to adhere to the inner vacuolar membrane surface at the onset of lectin synthesis;

legumin only appears when vacuoles largely are replaced by smaller vesicles which

eventually develop into protein bodies (Wenzel et al. 1993). 43

(6) THE PLASMA MEMBRANE

For all pro- and eukaryotic cells, the plasma membrane is the fundamental

biological barrier that separates individual cells from their neighbors and from their

surrounding environment. It consists of a single lipid bilayer membrane. Proteins and

glycoproteins are embedded in the plasma membrane, or are associated with its surface

(Alberts et aI., 1994b). Under the electron microscope, the plasma membrane is

characteristically the thickest of the various cellular membranes (Mom! and Mollenhauer,

1974). In animal cells, the plasma membrane is formed as consequence of the steady

fusion of secretory vesicles derived from the TGN, balanced by a similar rate of

endocytosis. In plants, the experimental evidence for endocytosis is not quite so well

established. The plasma membrane, albeit not exactly an endo membrane, is included in

this discussion of the endomembrane system because of its relationship via exo- and

endo-cytosis with the endomembrane continuum (Mom! and Mollenhauer, 1974; Harris,

1986). A variety of transporters, receptors, and other specialized structures such as

plasmodesmata and channels on the plasma membrane serve as mechanisms whereby

cells interact with their neighboring cells and with environment (Chasan and Schroeder,

1992; Robards and Lucas, 1990).

Two major functions of the plasma membrane are transport of solutes, both into and out of each cell, and signal transduction (Sussman and Harper, 1989, Kjellbom et aI.,

1990). In plant cells, the plasma membrane is also responsible for the synthesis and 44 deposition of cell wall components, including cellulose microfibrils and various hydroxylproline rich glycoproteins and oligosaccharides (Delmer, 1987). From the viewpoint of protein trafficking, the plasma membrane is not only involved in protein secretion, but also in protein internalization. The plasma membrane is currently regarded as a default end-point for constitutive protein secretion. In other words, proteins destined to the plasma membrane do not in general require sorting signals for this journey, except for ER translocation signals for secreted proteins and ER translocation signals or transmembrane domains for the plasma membrane integral proteins (Farquhar, 1985).

However, this view is now being challenged bye the results of two types of experiments, one concerning the transport of membrane integral proteins to the vacuole in yeast, and the second concerning the targeting of proteins to apical and basolateral surfaces in polarized animal cells. Those exceptions are later discussed separately on pages 76 and 82.

(7) VESICLES

Membrane vesicles have long been known to provide means whereby materials are shuttled between discrete membrane compartments during exo- and the endo-cytosis

(Pfeffer and Rothman, 1987; Pelham, 1989; Rose and Doms, 1988). The concept of vesicle traffic, first proposed by George Palade in the study of exocytosis (1959), was further consolidated in the investigations of protein movement from the ER to the Golgi 45

elements, intra-Goigi flow of proteins, transport of plasma membrane glycoproteins,

and protein trafficking from Goigi to lysosomes (Jamieson and Palade, 1967; Palade,

1975; Rothman, et aI., 1980; Rothman et aI., 1984; Brown and Farquhar, 1984). The

existence of vesicles within plant cells is best evidenced by the studies on clathrin coated

vesicles, whose ultrastructure was extensively characterized and from which clathrin protein was purified and isolated (Coleman et aI., 1991).

Three classes of "coated" vesicles have been documented so far, two classes of clathrin coated vesicles (CCVs), and one class ofnon-clathrin coated vesicles (NCVs).

Of CCVs, one class is responsible for mediating endocytosis; the other is involved in exocytosis between the TGN and lysosome (or vacuole) (Goldstein et aI., 1979; Burgess and Kelly, 1987; Kornfeld and Mellman, 1989; Pearse and Robinson, 1990). In plant cells, CCVs have been described and isolated (Robinson, et ai. 1987; Coleman et aI.,

1991). CCV s have a uniform structure: three pairs of light and heavy chains of clathrin form a three-legged triskelion, which interweaves with other triskelions to form polygonal lattice cages on the cytoplasmic surface of coated vesicles (Pearse and

Robinson, 1990). Clathrin adapter protein complexes (HAl and HA2) mediate the attachment of the triskelions to the vesicle surface, through interactions with both clathrins and specific transmembrane receptors recruited into clathrin-coated pits (Pearse and Robinson, 1990). The HAl and HA2 complexes are heterotetramers; the former consists ofr-, W-adaptin, a 47, and a 19 KDa protein, and is responsible for protein exocytosis after the TGN; the latter comprises a-, p-adaptin, a 50, and a 17 KDa protein, 46

and is involved in internalization of materials from the plasma membrane (Pearse and

Robinson, 1990). In addition to clathrin, the adapter proteins, and receptors, other proteins can be associated with CCVs. For example, auxilin and AP180 are observed to be associated with CCVs isolated from brain, and are postulated to modulate and regulate the assembly of clathrin into polygonal lattices (Ahle and Ungewickell, 1986,

1990). The major functions of CCV s are to regulate budding and fission of vesicles from donor membranes, and to provide carriers for the unidirectional transport of cargo from one endomembrane compartment to the next. A further function of CCVs is to recognize, dock, and fuse with acceptor membranes (Rothman and Orci, 1992). CCVs are finally responsible for the signal-mediated sorting and packaging of specific proteins at the TGN (Pearse and Robinson, 1990).

NCVs were initially implicated in the intra-Golgi transport of proteins in mammalian cells. In contrast to CCVs, the surface ofNCVs are covered by coating proteins (COPs) rather than clathrin. NCVs have not been directly implicated in protein trafficking in plant cells, although some circumstantial evidence has emerged (Staehelin et aI., 1991; Palme et aI., 1992; Driouich et aI., 1993b). NCVs are implicated in default bulk flow (signal-independent secretory pathway). Specifically, they comprise shuttlc vesicles bctween the ER and the Golgi, and between the various Golgi sub-compartments

(Melancon et aI., 1987; Segev et aI., 1988; Melhotra et aI., 1989; Bacon et aI., 1989;

Baker et aI., 1990; Serafini et aI., 1991 a). In term of components, NCV s are less well characterized than CCV s (Malhotra et aI., 1989). The COPs consist of ct-, ~-, Y-, 8- 47

COP, and several smaller proteins comprising 30-40 KDa and -20 KDa (Rotlunan and

Orci, 1992). COPs form a complex of 13-14S with an apparent molecular weight of -

550 KDa. An abundant cytoplasmic non-clathrin coat protein complex (termed coatomer), comprising four COPs, is thought to have a function similar to that of the adapter protein complex in CCVs (Walters et aI., 1991). The N-terminal half of p-COP has significant homology with p-adaptin (Allan and Kreis, 1986). ADP- ribosylation factors are components of the COP-coat and may regulate association of the coat with the

ER or the Golgi membrane (Serafini et aI., 1991 b). The mechanism of the molecular assembly of the COP-coat at the membrane surface is still unknown. NCVs are only implicated in the non-selective or signal-independent "bulk flow" traffic of proteins. In this, they may selectively incorporate certain proteins and transport them to the cell surface via non-receptor mediated bulk flow (Rothman and Orci, 1992).

It has been postulated that coated vesicles other than the three types described above may exist within the total population of vesicles, but they have not been substantially characterized (Linppincott-Schwartz et aI., 1991; Klausner et aI., 1992). A number ofnon-"coated" vesicles have been identified as well; they lack the distinct protein coated structure of the "coated" vesicles (DeCamilli et aI., 1990). The best characterized one of them is the synaptic vesicle, which involves in regulation of neurotransmitter release from nerve endings. 48

MOLECULAR SORTING MECHANISMS INVOLVED IN

THE CONSTITUTIVE SECRETION OF PROTEINS

Molecular mechanisms of protein sorting here refer, in most general sense, to the

mechanisms that govern the distribution of proteins from their sites of synthesis to their

final destinations inside cells. In this review, emphasis is placed on the following

problem: how specific proteins can be sorted out from the non-selective default bulk

flow. The review will detail our current understanding of protein sorting signals and their

recognition machinery within the exocytotic pathway. In the interest of space, this

review will not cover the area regarding budding, fission, movement, docking, and fusion

of vesicles.

ER TARGETING AND TRANSLOCATION

ER targeting and translocation are the first sorting event encountered for all

proteins entering the endomembrane-secretory pathway. They occur concurrently at the rough ER and concomitantly with protein biosynthesis. The process, which is signal and energy dependent, and is mediated by a variety of protein complexes, can be separated into two distinct phases: ER targeting and translocation across the membrane (Rapoport,

1992). 49

Sorting proteins into the secretory pathway by signal sequences: Signal sequences, also called signal peptides, are defined as a contiguous block of amino acids that directs a particular protein to enter the endomembrane-secretory pathway at the ER.

From the analysis of a variety of signal sequences, consensus has emerged regarding its basic feature, a linear stretch of 6 to 20 apolar amino acid residues (Nothwehr and

Gordon, 1990). There are two types of signal sequences, those that are cleavable and those that are not. The first type is found at the N-termini of membrane, luminal and secretory proteins, their function being to target proteins to the ER, then to initiate their translocation across the ER membrane. During translocation the cleavable signal sequences are removed by signal peptidase, which is located inside the ER lumen. The second type of signal sequences usually serves as a transmembrane domain, and can be found at essentially any location within membrane proteins. As for non-cleavable signal sequences, besides targeting proteins to the ER and initiating translocation across the ER membrane, they also provide a translocation stop and membrane anchoring functions.

Signal sequences respectively with or without transmembrane domains result in two different translocation fates: complete translocation into the ER lumen for secretory and luminal proteins, and partial translocation across the ER for membrane proteins.

Protein complexes involved in ER targeting and translocation: Protein complexes identified as being involved in the translocation step currently include: the mRNA-nascent polypeptide-ribosome complex, the signal recognition particle (SRP), the 50

SRP receptor, the ribosome receptor, the translocation channel complex, signal peptidase,

BiP and other chaperons, and oligosaccharyltransferase. The mRNA-nascent polypeptide-ribosome complex has been well characterized (Mathews and van Holde,

1990). SRP is a ribonucleoprotein (RNP) composed of a RNA molecule of 300 nucleotides and six distinctive polypeptides, which are organized into two heterodimers

(SRP9/14 and SRP68172) and two monomers (SRPI9 and SRP54) (Siegel and Walter,

1988). SRP serves as an adapter between the cytoplasmic protein synthesis machinery and the membrane-bound protein translocation machinery. Specifically, it acts to target the nascent peptide to the ER membrane by interacting both with signal sequence and the translocation channel, and prevents the premature folding of signal sequence with the rest of the nascent chain (Rapoport, 1992). The SRP receptor is a membrane protein complex facing the cytoplasm, consisting of two subunits: SRcx. and SRP (Connolly and Gilmore,

1989; High et aI., 1991). It functions to recognize the SRP-mRNA-nascent polypeptide-ribosome complex and facilitate its docking with the protein channel and ribosome receptor. The ribosome receptor at present still remains elusive. Gorlich et al.

(1992b) have suggested that the yeast Sec61 protein and its mammalian homolog, likely are located in the translocation channel, also serve as ribosome receptors. Savitz and

Meyer (1993) have alternatively proposed that a mammalian p 180 protein acts as the ribosome receptor. The ribosome receptor is speculated to establish a tight seal between the translocation chmmel and the mRNA-nascent polypeptide-ribosome complex, and to maintain the permeability barrier for solutes between the ER lumen and cytoplasm I'

51

(Gilmore, 1993). The existence of the translocation channel has been recently confirmed

by electrophysiological results (Simon and Blobel, 1991). The channel is proposed to

comprise two major components: the Sec61 protein and TRAM (trans locating chain­

associating membrane protein) based on the results of various in vitro reconstitution

analyses of the proteins (Rothblatt et aI., 1989; Deshaies et aI., 1991; Gorlich et aI.,

1992b). The channels play the core function of ER translocation. SEC61 and TRAM

have ten and eight transmembrane domains, respectively (Rapoport, 1992). Signal

peptidase, oligosaccharyltransferases and BiP have been demonstrated to have

stimulatory effects on translocation of proteins into the ER. They are postulated to

prevent reversible translocation through the channel, and to facilitate unidirectional

translocation by binding emerging precursors within the lumen of the ER (Gilmore,

1993).

ER Targeting cycle ojSRP: The general steps in this cycle will be briefly

outlined here; for a complete review, see Rapoport (1992). The targeting cycle initiates

with elongation arrest and formation of the SRP-mRNA-nascent chain-ribosome complex

when the emerged signal sequence is recognized by a methionine-rich hydrophobic

pocket (M-domain) within SRP54. The next step involves docking the complex on the

SRP receptor through the interaction of SRP with SRa. The third step is an exchange of

a single molecule of GDP with a GTP located within a combined GTP binding site

provided by the interaction of SRP54, SRa, and SR~ in the docked complex. The GTP 52

binding site (G-domain) of SRP54 may function as a proofreading mechanism by

releasing, through the hydrolysis ofGTP, signal sequences erroneously-bound within the

M-domain pocket. The GTP binding site of SRa. appears to playa important role in the

formation of high affinity SRP-SRP receptor complexes during the targeting reaction. In contrast, the function of the GTP binding site in SRP is not understood. The fourth step is the release of the SRPISRP receptor complex from the signal sequence and the ribosomes, which involves occupation of the G-domain ofSRP54 by GTP. The nascent chain is transferred to the translocation channel, and the ribosomes with mRNA become membrane-bound by interaction with the ribosome receptor. The penultimate step is the complete release of SRP from its receptor, which occurs when the GTP bound to the

SRPISRP receptor complex is hydrolyzed. SRP and its receptor become free again in the cytoplasm and the membrane respectively, and are then available for the next cycle of

ER targeting. The final step is the actual translocation of the nascent chain across the ER membrane. This step does not require SRP and/or its receptor, and only starts after the

SRPISRP receptor complex becomes dissociated at the fourth step of the ER targeting cycle.

Sorting a/membrane proteins/rom /wnina/lsecret01Y proteins: After being sorted from proteins destined to other cellular organelles, the proteins entering the secretory pathway now face another round of sorting during translocation: this involves entering either the membrane or the luminal/secretion pathways. This sorting mechanism is 53

based on the availability of transmembrane domains, which have "start transfer" and

"stop transfer" signals at the beginning and the end of a stretch of apolar amino acids

(Alberts et aI., 1994). In the absence of a "stop transfer" signal, the nascent chain of protein is completely translocated across the ER membrane and into the lumen, and thereby become lumen or secretion destined. For membrane proteins, the translocation process is initiated either by their signal sequences or by internal transmembrane domains. "Stop transfer" signals then terminate the translocation of the nascent chains across the ER. "Stop transfer" signals have been hypothesized to be recognized by a receptor unit, which could be part of the channel protein or an independent protein in the vicinity of the channel. After recognition, the channel could open up laterally to allow integration of membrane proteins into the lipid (Gilmore, 1993). For membrane proteins with non-cleaved signal sequences, the signal for ER translocation is typically one of the transmembrane domains (Alberts et aI., 1994). The translocation event itselfmay be identical in both cases, the sorting being conducted after the translocation is initiated

(High, 1992).

Translocation across the ER membrane: The details of the ER translocation steps are less clear than those involved in ER targeting. Recent discoveries in this area have resolved a decade-long dispute as to whether the protein is translocated by passing through an aqueous channel or by inserting into membrane directly (Simon and Blobel,

1991; Gorlich et aI., 1992b; Gorlich and Rapoport, 1993). The Blobel model, involving 54

protein translocation across the ER membrane through a current-gated channel, now

appears to have been experimentally vindicated. Thus, if the rough microsomes (active

in protein synthesis) are caused to fuse with a planar lipid layer in the presence of an

osmotic gradient (created by adding either urea or sucrose) across the planar lipid layer,

An electrical current created by ion transport across the membrane can then be monitored using the patch-clamp method. The basal current is low, but spikes transiently after addition of puromycin (Simon and Blobel, 1991). This is consistent with the idea that puromycin unplugs the nascent polypeptide chain covering the channel by prematurely terminating chain elongation and subsequently detaching the chain from the ribosomes.

Other evidence in favor of the model is the observation that the current-gate can be opened by adding synthetic signal peptides to the ribosome-binding side of membrane, suggesting that signal sequences serve as ligands to open the channel (Simon and Blobel,

1991). The restriction barrier across the membrane has been confirmed to be tightly maintained during the translocation using fluorescence labeling and collision quenching techniques (Johnson, 1993; Crowley et aI., 1993).

Further evidence in favor of the existence of an aqueous channel has involved the isolation and characterization of SEC61, SEC62, SEC63, TRAM (Rothblatt et aI., 1989;

Deshaies et aI., 1991; Gorlich et aI., 1992a, 1992b). Reconstitution and functional analyses of the isolated channel candidates has demonstrated that SEC61 is essential for the translocation of membrane and secretory/luminal proteins, and that TRAM is required 55

for the translocation of certain proteins and has a stimulatory effect for others (Gorlich

and Rapoport, 1993). Gilmore (1993) proposed a model for the ER translocation

apparatus, in which SRP, the SRP receptor, SEC61, TRAM, and a still controversial

ribosome receptor, are directly involved the translocation reaction, and other factol's, such as signal peptidase, BiP, and oligosaccharyltransferase, improve the efficiency of translocation. The entry into the channel has been found to be GTP-dependent, but exit from the channel may be A TP-dependent. The conformations of SEC61 and TRAM, in the presence of signal sequence, are postulated to change to a status that favors the channel-gating so that the actual ER translocation can occur.

Most of the above progress has been achieved using yeast and mammalian cell systems. These mechanisms form the framework for the study of plant cells. The isolated SRP from plant cells resembles that of mammalian cells (Campos et aI., 1988;

Prehn et aI., 1987). In addition, the plant secretory proteins can be translocated both in vivo and in vitro into the ER of yeast and mammalian cells (Tague and Chrispeels, 1987;

Saalbach et aI., 1991; Vitale et aI., 1986; Simon et aI., 1990; Ceriotti et aI., 1991); similarly the mammalian secretory proteins can be sorted into the plant endomembrane secretory pathway (Hiatt et aI., 1989; Sijmons et aI., 1991); and glycoproteins of mammalian viruses can be incorporated into the secretory pathways of both mammalian and plant cells (Machamer and Rose, 1987, 1988; Galbraith et aI., 1992; also see

Chapter 3, 4 and 5 in this dissertation). The above observations suggest that ER 56

targeting and translocation, mediated by signal sequences, are processes that are well

conserved between the yeast, plant, and mammalian cells.

ER RETENTION AND RETRIEVAL MECHANISMS

After passing through two rounds of sorting during ER targeting and translocation, proteins in the ER lumen and membrane face a third round of sorting when they exit the ER to the next compartment (Golgi) under the influence of default bulk flow. The process of protein sorting now branches into two parallel pathways; from this point onwards, it operates in two different phases: a lipid bilayer phase for membrane proteins and an aqueous phase for all luminal and secretory proteins. The mechanisms underlying the sorting in two phases appears to be different, even if occurring within the same organelle.

ER Retention of Luminal Proteins from Default Secretion

ER retention signals for luminal proteins: This class of signals is one of the best­ characterized in the secretory pathway. The signals typically comprise tetrapeptide sequences located at the C-termini of soluble ER proteins. The ER retention sequences, which are well-conserved between yeast, plants, mammals, and other eukaryotes, are 57

LyslHis/Arg-Asp-Glu-Leu (KlHlRDEL) in mammalian and plant cells (Pelham, 1989;

Bednarek and Raikhel, 1992), and HlDDEL in yeast cells (Pelham et aI., 1988, Pelham,

1990). The tetrapeptide sequences have been found in luminal ER proteins of plant cells

such as BiP, auxin binding protein (ABP), protein disulfide isomerase, and ER-localized

heat shock proteins (Bednarek and Raikhel, 1992), and have been found to be both

necessary and sufficient to retain their cognate proteins within the ER lumen (Pelham,

1989; Chrispeels, 1991; Wandelt et aI., 1992). Any alternation to the tetrapeptide

sequence results in a partial or complete loss of ER retention (Bednarek and Raikhel,

1992). In some cases, the presence of a retention peptide is not sufficient to prevent

leakage from the ER in mammalian and plant cells (Zagouras and Rose, 1989; Herman et

aI., 1990; Jones and Herman, 1993). In the case of ABPs, most are located in the ER as would be expected from the presence of the retention peptide. However, some are found to be transported to the plasma membrane/cell wall via the Golgi (Jones and Herman,

1993). The degree of the leakage appears to be positively correlated with the severity of auxin deficiency in the medium surrounding the cells, suggesting that it may be necessary for some ABPs to reach the cell surface to have some function in binding auxin (Jones and Herman, 1993). It seems that the retention efficiency ofKDEL in ABPs may also be affected by the concentration of extracellular auxin. The retention efficiency of KDEL varies when different reporter proteins are translationally fused to the KDEL sequence

(Zagouras and Rose, 1989; Herman et aI., 1990). We can conclude that ER retention not 58

only is determined by the retentional sequences, but also may be influenced by

conformation changes that could affect the availability of the retentional sequences to

the receptors.

Receptor-mediated retrieval a/luminal ER protein escapees: Several mechanisms have been proposed for ER retention of luminal proteins, including "sorting and/or retention receptors" (Brands et aI., 1985; Munro and Pelham, 1987), "transport competence" (Gething et aI., 1986 and Klausner, 1989), "phase partitioning into an immobile matrix" and "selective and local aggregation" (Pfeffer and Rothman, 1987), and

"ER recycling" (Pelham, 1988). Prevailing opinion at present favors "ER recycling", thus that ER retention is achieved via continuous retrieval of luminal ER protein escapees from the putative ERGIC. Initial evidence for the retrieval model comes from the studies of lysosomal Cathepsin 0 fused with a KDEL sequence (Pelham, 1988). KDEL-bearing

Cathepsin D accumulated within the ER, as expected, but its oligosaccharide side chain was observed to be modified by N-acetylglucosaminyl-l- phosphotransferase residing in the cis-Golgi (Pelham, 1988).

The most recent evidence comes from isolation and characterization of the ERD2 gene from yeast (Semenza et aI., 1990) and its homolog from mammalian and plant systems (Lewis and Pelham, 1992; Tang et aI., 1993; Lee et aI., 1993). The ERD2 gene product is a receptor protein responsible for the retrieval of ER luminal proteins bearing 59

KDEL sequences from the ERGIC. Upon binding a KDEL-containing ligand, the ERD2 receptor moves to the ER, and releases the ligand (Tang et aI., 1993). The receptor accumulates in the ER under normal conditions, but is shifted to the ERGIC when mammalian cells are exposed to 15°C or BFA (Tang et aI., 1993). The receptor has a putative glycosylation site, seven transmembrane domains, and cytoplasmic N- and

C-termini (Semenza et al. 1990; Singh et aI., 1993). Amino acids 51-56 (DLL TFH in yeast, DLFTNY in mammals) have been found to be important in determining the binding specificity of the ERD2 receptor to the KDEL sequence (Semenza and Pelham,

1992). Ligand binding in vitro exhibits the same sequence specificity as seen in vivo, but is sensitive to pH, suggesting that changes in pH along the secretory pathway are involved in the binding of the KDEL peptide in the Golgi (slightly acidic), and its release in the ER (Wilson et al. 1993). Ligand binding is also dependent on charged residues within the transmembrane domains of the receptor (Townsley et al. 1993). Recycling of the occupied receptor to the ER is unaffected by its cytoplasmic loops, but is critically dependent upon an aspartic acid residue in the seventh transmembrane domain (Townsley et al. 1993), suggesting that the transmembrane domain has a role in transport of the loaded receptor back to the ER. Deleting the last twelve residues of the cytoplasmic

C-terminal tail results in the loss of ERD2 function in yeast (Semenza et aI., 1990), implying the C-terminus may require interaction with cytoplasmic factors for retrieval.

Although a reasonable outline for the retrieval mechanism is emerging, the details of its 60

operation are still unclear, such as how the occupied receptor moves back to the ER, what

machinery senses the loading/unloading of the receptor in the ER and the ERGle

respectively, and what cis-acting elements determine differential residency of the

receptor in the ER and the ERGle.

The ERD2 receptor in Arabidopsis is highly homologous to its counterpart in

yeast and mammals, with -50% identity of whole protein sequence, and an almost

identical ligand binding region (DLFTDY ) (Lee et aI., 1993). It, unlike the mammalian

receptor, can functionally complement the lethal erd2 mutant of yeast, and thus may

represent an evolutionary link between yeast and animals (Lee et aI., 1993). Further

functional analyses in plant systems are not available currently. It is possible that the

plant homolog operates using a mechanism similar to that in yeast and mammalian cells,

but its action site and mechanism details may tum out be significantly different. For

exanlple, plant cells have not been confirmed to contain an ERGle. In addition, the

spatial relationship between, and the distribution patterns of, the ER and the Golgi in

plant cells do not favor accommodation of the same retrieval machinery as found in

mammalian and yeast cells. A possible alternative is that the ERD2 receptors of plant

cells are located in the ER, perhaps in the vesicle packing and budding regions, and serve

as guards to detain potential escapees from the nonselective bulk flow of proteins, without departing from the ER. 61

ER Retention of Integral Membrane Proteins from Default Secretion

ER retention signals a/integral membrane proteins: So far there is no consensus

retentional sequence for ER membrane proteins. It is clear that certain motifs can play

critical roles in the retention of membrane proteins in the ER. For example, in

mammalian cells the retention within the ER of several cellular and viral Type I

membrane proteins (with a single transmembrane domain, having the N-terminus inside

the lumen and the C-terminus in the cytoplasm) has been shown to be dependent on a

C-terminal oligopeptide Lys-Lys-X-X (KKXX, X=any amino acid) or KXKXX (Hong

and Tang, 1993). This motifis also found in the cytoplasmic C-termini of two Type III

ER membrane proteins (having multiple transmembrane domains): 3-hydroxy-3-

methylglutaryl coenzyme A reductase (HMG-CoA reductase), a protein involved in the

early steps of ER translocation (Jackson et aI., 1990; Gorlich et aI., I 992a), as well as a

yeast nuclear envelope protein (Type I; Heesen et aI., 1991). Placing a double- arginine

(RR) or RXR motif at the cytoplasmic N-terminus maintains in the ER a human invariant

chain p33 (lip33, a Type II ER membrane protein) and the RR-bearing transferrin

receptor, a Type II plasma membrane protein (Schutze et aI., 1994). A cytoplasmic

C-terminal RKPRRE motif may be relevant to the retention in the ER of a human

homolog to canine calnexin membrane protein (David et aI., 1993). However, three rather surprising cases has been reported. HDEL sequences were found in the luminal

C-termini of two yeast ER membrane proteins: SEC20p and SED4p (Sweet and Pelham,

1992; Hardwick et aI., 1992). SEC20p seems to be recycled between the ER and the I.

62

ERGlC, as it has an O-linked oligosaccharide side chain, which is supposedly only

attached in the cis-Golgi (Sweet and Pelham, 1992). It is really a puzzle that the KKXX

motifis also found in the cytoplasmic C-terminus ofa yeast ERGlC-53, which is a

putative ERGIC membrane protein (Schindler et al., 1993).

Retention machinery jar ER membrane integral proteins: The mechanism for ER

retention of integral membrane proteins is still unknown. The fact that retrieval from the

ERGIC must occur has been shown for two ER membrane proteins (SEC20p and

SECI2p) which must penetrate as far as the cis-Golgi, since they contain the O-linked

glycosylation (Sweet and Pelhanl, 1992; Nishikawa and Nakano, 1993). Whether ER

membrane proteins use HDEL motif for the retrieval needs verification, since those

reports have not investigated the possible function of the cognate HDEL motif (Sweet

and Pelham, 1992; Hardwick et al., 1992). It seems likely that some ER membrane

proteins may employ, at least in part, the retrieval machinery for luminal ER proteins

since their HDEL motifs are oriented to the lumen where the ERD2 receptor would be

able to bind.

Nishikawa and Nakano (I 993} have beautifully demonstrated that the yeast

retention mechanism for ER membrane proteins is distinct from that for luminal proteins.

They used native and modified forms of the yeast Sec 12p gene product, which is an ER

membrane protein recycled between ER and Golgi, as a selection marker for mutants

defective in the retention of ER membrane proteins (Nishikawa and Nakano, 1993). 63

They identified a gene (red recessive mutant) that is responsible for the mislocalization

of the native SEC 12p, as well as two modified SEC 12p fusion proteins, to the late Golgi

and even to the cell surface (Nishikawa and Nakano, 1993). However, they found that

the recessive mutant (red) is not defective in retention of luminal BiP in the ER

(Nishikawa and Nakano, 1993). Considering that most retentional sequences of ER

membrane proteins are oriented towards the cytoplasm, it is reasonable to contemplate a

different ER retention mechanism for membrane proteins. Any model must involve

receptor proteins either completely or partially located in the cytoplasm, so that the ER

retentional sequences can be recognized and bound. Of course, it does not necessarily

exclude the possibility that the HDEL retrieval machinery can act in a supplemental

manner.

No published information is available concerning ER retention of membrane

proteins in plant cells. However, the genes of several ER membrane proteins have been

isolated from Arabidopsis. These include a homolog of the yeast ERD2 receptor (Lee et

aI., 1993), two isoforms ofHMG-CoA reductase (Enjuto et aI., 1994), and a homolog to

canine calnexin (Huang et aI., 1993). The plant ERD2 homolog has the sequence

KXKXX at its predicted cytoplasmic C-terminus. At the cytoplasmic C-terminus of the plant calnexin-like protein, there are three KKXX and two RR and RXR motifs. For the two isoforms of HMG-CoA reductase, both have the KKXX motifs in the middle of the cytoplasmic C-terminus and RRXX at the N-terminus. In addition, HMG2 has a KKXX 64

sequence at its N-terminus. Investigation of these sequences should shed some light on the mechanism of ER retention of membrane proteins in plants. One approach would be to assess those retention sequences by monitoring the extent of ER accumulation of the recombinant proteins comprising a known membrane protein from a later point in the secretory pathway (such as tonoplast intrinsic proteins, or vesicular stomatitis viral glycoprotein) fused N distally to a KKXX (KXKXX) or RRXX (RXR) motif. Another approach would be to evaluate anterograde movement of those ER-resident proteins in deletion and mutation analysis.

TRANSPORT COMPETENCE FOR EXIT FROM THE ER

"Transport competence" refers to a molecular mechanism of quality control that ensures proteins are assembled properly, and folded and oligomerized correctly before they are loaded into vesicles for anterograde transport. The process is not considered a true sorting event; however it achieves the similar if temporary effect of ER retention for misfolded and improperly-assembled proteins. Eventually these misfolded and improperly-assembled proteins either become processed correctly, or are degraded in the

ER (Burtley and Helenius, 1989). Chaperone proteins (i. e. BiP), protein disulfide isomerase and other ER proteins such as prolyl-peptidyl cis/trans isomerase are thought to facilitate the process of folding, assembly, and oligomerization. The details of the 65

have been thoroughly reviewed by Hurtley and Helenius (1989) for mammalian cells,

and Bednarek and Raikhel (1992) for plant cells.

RETENTION OF PROTEINS AT THE CIS-, MEDlAL- AND TRANS-CISTERNAE OF THE GOLGI

After the ER has sorted out and retained its residents, the remaining proteins, if

correctly assembled, are processed and packed into transport vesicles, and are forwarded

to the Golgi by nonselective bulk flow. Within the Golgi, they face further rounds of

sorting and retention. The signals and mechanisms governing protein retention within the

Golgi are not as well-understood as those in the ER. Most experiments are still at the

stage of defining protein signals that are involved in the Golgi retention (Paulson and

Collet, 1989). No clues have emerged concerning retention of luminal proteins in the

Golgi, including in the TGN (discussed in a later Section). A more serious problem is a

lack of luminal Golgi-resident proteins for study. Therefore the following discussion will

focus on the signals responsible for the Golgi retention of membrane proteins, and will

detail various hypotheses proposed about mechanisms.

Golgi retention signal::.'for integral membrane proteins: No data is available

from native resident proteins concerning their retention in the cis-Golgi. Studies on the localization of the IBV M glycoprotein, provide most of our understanding concerning retention within this sub-compartment. The group of Dr. Carolyn Machamer has 66 demonstrated that the first transmembrane domain (ml) of the IBV M glycoprotein is required and sufficient for targeting in animal cells to the cis-Goigi (Swift and Machamer,

1991). When the transmembrane domain of a plasma membrane protein (VSVG) was swapped with the ml domain, the chimaeric protein (Gml) was retained in the Golgi of transfected cells (Swift and Machamer, 1991). Conversely, IBV M was targeted to the plasma membrane when its ml domain was replaced by the transmembrane domain of

VSVG. The site-directed mutagenesis indicates that retention within the cis-Golgi requires four uncharged polar residues on one face of the predicted a-helix of the ml domain (Machamer et aI., 1993). Somewhat confusingly, the homolog of the IBV M glycoprotein from mouse hepatitis virus appears to require nearly all of the coding sequence for Golgi localization. Its ml domain is not sufficient to retain a VSVG chimaeric protein in the Golgi; instead, the molecule is transported to the cell surface

(Machamer et aI., 1993).

Retention signals for the medial-Golgi have been investigated using the native enzyme P-l,2- N-acetylglucosaminyltransferase I (GnTI). This is a Type II membrane protein, having a cytoplasmic N-terminus, a transmembrane domain, and a large catalytic luminal domain. The transmembrane domain and its flanking residues are sufficient to localize GnTI to the medial-Golgi, and will also target fusion proteins to this location, albeit with trace leakage to the cell surface (Burke et aI., 1992). Golgi localization of the fJsion proteins requires all three domains from GnTI (Burke et aI., 1994). The fusion proteins containing any two of the three domains were predominantly localized to the 67 medial-Golgi, but also leaked to the cell surface at low levels. Those.containing only the transmembrane or the luminal domain showed partial Golgi retention but substantial trafficking to the cell surface (Burke et aI., 1994). The cytoplasmic tail of GnTI independently was unable to retain the fusion protein, but was able to direct Golgi localization of either the luminal or transmembrane domain (Burke et al., 1994).

Furthermore, the fusion proteins containing GnTI were increasingly shifted to the cell surface in low salt buffer (Burke et aI., 1994), suggesting that protein-protein interaction, affected by ionic strength, may have a role in the retention of GnTI.

p-D-N-acetylglucosaminide p-I,4-galactosyltransferase (GT) is a Type II membrane protein anchored in the trans-cisternae of the Golgi. It has similar motifs to

GnTI, but with a distinct luminal stalk N-distal to the catalytic domains. Deleting either the cytoplasmic tail or the luminal stalk had no effect on GT retention (Aoki et aI., 1992).

However, the Cys-X-X-His motif in the transmembrane domain was found necessary for

Goigi retention ofGT. The Cys-X-X-His motif by itself does not confer Golgi retention on the transferrin receptor, as found via the swapping of transmembrane domains.

However, the cytoplasm-sided half of the transmembrane domain was found partially capable of retaining the transferrin receptor as a fusion in the Golgi, and the whole GT was fully capable (Aoki et al., 1992). A different scenario was observed for another trans-Golgi enzyme: p-galactoside a-2, 6-sialyltransferase (ST), a Type II membrane protein. The ST mutants containing only the stalk and catalytic domain are not rapidly I'

68

secreted, but are retained intracellularly, being predominantly localized to the Oolgi.

Deleting either the stalk or the cytoplasmic tail does not alter the Oolgi localization of ST

(Colley et ai., 1992). Further experiments revealed that specific sequences in the

transmembrane region are not essential for Golgi localization when they are entirely

replaced or partially altered in combination with or without the presence of most of the

stalk region (Dahdal and Colley, 1993). In chimaeric proteins, in which the cytoplasmic

and transmembrane domains of ST were fused with the extracellular domains of two

different plasma membrane proteins, poor Golgi retention was observed. However, one

of the chimeras showed significant Oolgi retention when two lysines were placed next to

the transmembrane domain on the luminal side. This suggests that the membrane

flanking sequences are critical elements for Oolgi retention of ST.

Speculative mechanisms for Golgi retention of integral membrane proteins:

"Golgi retention by oligomerization" has been proposed as a model by Dr. Machamer and

coworkers based on the uncovering of a relationship between Oolgi retention and

oligomerization status of the Oml fusion protein (Weisz et ai., 1993). Oml mutants,

having single amino acid substitutions in the m 1 domain and that were retained in the

Oolgi, formed SDS-resistant oligomers, whereas those mutant proteins rapidly released

to the plasma membrane did not. The luminal N-terminus of Om I was not required for

oligomerization, since its replacement with other polypeptides did not affect the

formation of the SDS-resistant oligomer; in contrast, the cytoplasmic C-terminus was 69

shown to facilitate oligomer formation. They suggested that the SDS-resistant oligomers

form when newly-synthesized GmI arrive at the Golgi and, for retention they may

interact (directly or indirectly) with an actin-based cytoskeletal matrix. However, as

Pelham and Munro have indicated (1993), it is not clear whether the my M glycoprotein

uses this mechanism for its Golgi retention, since oligomerization is a property of the Gm

1 fusion proteins, and not of the my M glycoprotein itself.

"Kin recognition" has been proposed by Nilsson et aL (1993) as a mechanism for retention of Golgi enzymes. The model is based on three tenets; first, that the Golgi enzymes are homodimers bound together by their luminal domains, consisting of a catalytic domain and a stalk; second, that their transmembrane domains can specifically bind to their kin within the same cisternae of the Golgi, but not to the proteins from the other cisternae, thus forming a large hetero-oligomer; third, that the oligomers are anchored in the intercisternal matrix through their cytoplasmic N-ternlini. Supporting evidence for the model is derived from the facts that, to date, all the cloned Golgi enzymes are type II membrane proteins, and that the luminal domains are capable of forming homodimers (Nilsson et aL, 1993). A more-recent study elegantly demonstrated that kin recognition can exist among Golgi enzymes within the same cisternae (Nilsson et aL, 1994). In this work, GnTI and mannosidase II (Mann II) of the medial-Golgi, and GT of the trans-Golgi were retained in the ER membrane when they were translationally fused at their N-termini with the additional cytoplasmic domain ofIip33 (sec "ER retention signals o/integral membrane proteins"). The endogenous Mann II was I. I

70

relocated to the ER in the cells expressing the Iip33/GnTI fusion proteins, as was the

endogenous GnT! in the cells expressing the Iip33/Mann II. Neither endogenous GnTI

nor Mann II became relocated to the ER in the cells expressing Iip33/GT. Both of these

observations fit the model of kin recognition, which could also well explain the different

results obtained from the mutational analyses of the resident Golgi enzymes: GnT!

(Burke et aI., 1992 and 1994), GT (Aoki et aI., 1992), and ST (Colley et aI., 1992;

Dahda1 and Colley, 1993).

"Lipid-mediated retention of Golgi enzymes" has been postulated by Bretscher

and Munro (1993). The model emphases the general properties of transmembrane

domains and their interactions with lipid membranes, rather than precise protein-protein

interactions required for retention. The model depends on two key characteristics: the

length difference of the transmembrane domains between the Golgi proteins and the

plasma membrane proteins, and the thickness difference between the Golgi cisternae and

the plasma membrane. In support of this, typical plasma membrane proteins are cited as

having an average of 21 amino acids within their transmembrane domains, with Golgi

enzymes having an average of only 17 amino acids in their transmembrane domains and

preponderance of phenylalanine residues (Pelham and Munro, 1993). The plasma

membrane has been observed to be the thickest cellular membrane in plant and

mammalian cells (Morre and Mollenhauer, 1974; Bretscher and Munro, 1993). In

mammals, the increased thickness of the plasma membrane is believed to be a

consequence of elevated concentrations of cholesterol and sphingolipids (Bretscher and 71

Munro, 1993). Embedding the shorter transmembrane domains of the Golgi enzymes would be less energetically favorable in the thicker plasma membrane than in the Golgi or

ER cisternae. This would also drive the Golgi enzymes to become naturally partitioned away from the area of vesicle budding, which are presumably high in cholesterol and sphingolipids. This would preferentially exclude Golgi enzymes from anterograde transport vesicles, and would favor their inclusion in retrograde transport vesicles

(Pelham and Munro, 1993). Key supporting evidence for this model is derived from mutational analyses on the transmembrane domain of ST (Munro, 1991). Replacing the transmembrane domain with 17 leucines did not alter Golgi retention of ST, but replacing the domain with 23 leucines released ST to the plasma membrane (Munro, 1991). As suggested by Pelham and Munro (1993), the model has advantages as well as disadvantages. On the plus side, it permits self-organization of the Golgi apparatus within the overall endomembrane flow without the need for large numbers of specific protein-protein interactions. On the minus side, it may not be sufficient to account for the precise, and short-distance segregation between the difTerent cisternae for the retention of

Golgi enzymes. Another issue concerning this model is that the length difference of the transmembrane domains does not actually exist between all Golgi enzymes and plasma membrane proteins (Machamer, personal communication).

The mechanisms whereby proteins are retained in the plant Golgi are not clear, since there are no protein probes available currently to dissect this important apparatus. 72

Two approaches have been explored in plant cells, as described for this project. One involves transplanting established probe systems from mammalian cells into plant cells.

Another involves isolating mutants that lack specific Golgi functions. By screening leaf extracts with an antiserum directed against complex glycans, Chrispeels' group was able to isolate an Arabidopsis thaliana mutant (cgl), which lacks GnTI and fails to convert glycoprotein glycans from high-mannose to complex form (von Schaewen et aI., 1993).

The mutant can be functionally complemented with a human cDNA encoding GnTI

(Gomez and Chrispeels, 1994). Unfortunately the CGL gene has not yet been isolated, since the mutant was selected from an EMS-mutated population. However, the subcellular localization of the human GnTI in plant cells was shown to occupy a position typical of the plant Golgi in sucrose gradients (Gomez and Chrispeels, 1994), implying that the human GnTI may be targeted to the plant Golgi. If so, this implies that the sorting signal and machinery for the Golgi retention are reasonably conserved between animals and plants.

RETENTION OF PROTEINS AT THE TRANs-GOLGI NETWORK

TGN retention signals/or integral membrane proteins: Signals for retention at the TGN have currently only been described for membrane proteins, more specifically for the mammalian TGN38 protein and its isoform TGN41 (Luzio et aI., 1990, Reaves et aI.,

1992), and for the Kexlp and Kex2p homo logs ofTGN38 and for a dipeptidyl 73

aminopeptidase A (DPAPA) in yeast (Nothwehr et aL, 1993; Cooper and Bussey, 1992;

Bryant and Boyd, 1993). TON38/41 are Type I integral membrane proteins with a

luminal N-terminus, a transmembrane domain, and a cytoplasmic C-tem1inus. They are

predominantly localized to the TON by signal sequences contained within their

cytoplasmic tails (Luzio et aL, 1990; Reaves et aL, 1992). Experiments involving

deletion, site mutagenesis, and protein fusions of TON41 13 8, indicated that a

cytoplasmic tail containing the critical sequence, Tyr-Oln-Arg-Leu (YQRL), is both

necessary and sufficient for TON localization (Bos et aL, 1993; Wong and Hong, 1993;

Humphrey et al., 1993). Interestingly, the YQRL sequence also contains a tyrosine motif

homologous to the endocytotic signal carried by the plasma membrane receptors,

although the endocytic signal alone is insufficient to confer TON residence (Bos et al.,

1993; Humphrey et al., 1993). These findings suggest that TON38 is maintained in the

TON by retrieval from the plasma membrane. Overexpression of the chimaeric proteins

containing TON retention signal resulted in their detection at the plasma membrane and

in the intracellular vesicles, consistent with saturation of the retrieval machinery

(Humphrey et al., 1993). The Kex 1p and Kex2p endopeptidases use TON retention

signals similar to those of TON38 (Cooper and Bussey, 1992; Bryant and Boyd, 1993).

DPAPA, a Type II membrane protein, employs a different, cytoplasmic retention signal

(Phe-X-Phe-X-Asp) for the efficient TON retention (Nothwehr et al., 1993).

Surprisingly, unlike their mammalian counterparts, Kexlp, Kex2p and DPAPA are released to the vacuolar membrane instead of to the plasma membrane when their 74

retention signals are deleted (Cooper and Bussey, 1992; Bryant and Boyd, 1993;

Nothwehr et ai., 1993); this implies a default membrane protein flow to vacuole.

Besides YQRL, an additional, non-overlapping signal from the transmembrane domain of

TON 38 has been also found to mediate TGN localization of the fusion proteins. The

different properties of these two signals suggest that two different mechanisms operate in

TON localization of one protein (Ponnambalam et ai., 1994).

Retrieval machinery/or TGN membrane proteins: A retrieval mechanism has

been proposed based on the observations of TGN38/41 recycling between the TON and

the plasma membrane (Ladinsky and Howell. 1992). Under steady-state conditions, the

majority ofTON38/41 is localized to the trans-most Oolgi cisternae and the TON, and

only a small amount of TON38/4 1 can be detected in prelysosomes (Ladinsky and

Howell, 1993). The natural distribution patterns of TON38/41 are changed when the

cycling TON38/4 I molecules become captured at the cell surface, for example by adding

specific antibodies conjugated to beads in the culture media (Ladinsky and Howell,

1993). Reaves et ai. (1993) demonstrated that the rate ofTGN38/41 recycling to the

TON from the cell surface can also be affected by Brefeldin A treatment, implying that the postulated retrieval machinery is similar to that for the retrieval of ER luminal proteins. The components of the machinery must reside in cytoplasm, since the TON retentional signals are located in the cytoplasmic C-termini. A peripheral 62-KDa (p6 I) protein and certain OTP-binding proteins have been identified to form a cytosolic complex that becomes associated with TON38/4 I. The binding of p6 I to TON38/4 I has I'

75

been implicated in the process of vesicle budding for TON38/41 transport, and the

dissociation of p61 from TON38/41 appears to be regulated by phosphorylation (Jones et

aI., 1993). However, the exact retrieval machinery for TON membrane proteins is far

from clear. It is also questionable whether the machinery of mammalian system is at all

applicable to TON retention in yeast, since yeast TON proteins are transported to the

vacuolar membrane in the absence of the retention signals. If TON membrane proteins,

as suggested by Ponnambalam et al. (1994), also use their transmembrane domains for

TON retention, this implies that two mechanisms operate for TON retention, and they

could complement each other. TON retention via transmembrane domain would greatly

reduce movement of TON-resident proteins from the TON to the plasma membrane,

reducing the workload for the TON retrieval machinery. On the other hand, the TON

retrieval machinery would recover escaped TON proteins from the plasma membrane,

increasing the retention rate of the proteins in the TGN.

SORTING OF PROTEINS AT TIlE TRANs-GOLGI NETWORK:

In all eukaryotic cells, proteins arriving at the TGN branch into two major routes,

leading to final destinations at the cell surface and within lysosomes (or vacuoles)

respectively. In mammalian cells, the TGN has been found to be a branch-point to an

additional major route, the pathway of regulated secretion. If we do not consider

retention/retrieval, protein sorting at the TGN represents a third round of true sorting 76

event, controlling the physical direction and routes of protein trafficking flow in the

exocytotic pathway; the other two events, targeting proteins to the ER and separating

membrane proteins from luminal/secretory ones, occur at the ERlcytosol interface.

Trafficking of Cell Surface Proteins

Proteins that reach the cell surface include both plasma membrane proteins and

secretory proteins. It is popularly believed that their transport from the TGN to the cell

surface is a default pathway, and does not require mediation of sorting signals (Pryer et

aI., 1992). However, considering that individual cells of eukaryotes (a notable exception

is yeast) typically are constituents of highly organized tissues within differentiated

anatomical structures, it is impossible to achieve precise organization of a specific tissue

without tightly controlling protein transport to the differentiated cell surfaces or

membrane domains. Recent advances in this area show that the transport of plasma

membrane proteins to the different cell surfaces is indeed signal-mediated, at least for

plasma membrane proteins in epithelial cells (Hong and Tang, 1993). This raises

question as to how the default pathway can operate in polarized cells.

Polarized transport of plasma membrane proteins in epithelial cells is thought to

be due to the presence of discrete targeting motifs found within their cytoplasmic domains. Targeting motifs to the apical cell surface have been identified within the cytoplasmic domains of glycosylphosphatidylinositol (GPI) anchored proteins, dipeptidylpeptidase IV, and the polymeric immunoglobulin (lg) receptor (Lisanti and I'

77

Rodriguez-Boulan, 1990; Casanova et aI., 1991; Mostov et aI., 1992; Mostov, 1993).

Motifs governing targeting to the basolateral cell surface have also been found within the

cytoplasmic domain of the low density lipoprotein (LDL) receptor (Matter et aI., 1992).

Mellman et al. (1993) surprisingly found the motifs are widely distributed, are not

specific to the proteins expressed in polarized cells, and also appear to function in

controlling polarized transport along both the exo- and endo-cytic pathways. The motifs

targeting proteins to the basolateral surface are similar to the signals used by the plasma

membrane receptors for accumulation at clathrin-coated pits, and often involve a critical

tyrosine residue (Matter et aI., 1992, 1993). The basolateral and apical transport motifs

are hierarchically arranged; a single protein may contain one or more motifs specifying

the basolateral transport, and deletion of these signals increases efficiency of the apical

transport (Prill et aI., 1993; Matter et ai., 1993; Mostov, 1993). Mellman et al. (1993)

hypothesized that non-polarized cells might contain the cognate apical and basolateral

pathways responsible for constitutive transport from the TON to the plasma membrane.

The availability of two pathways might confer a high degree of flexibility to

pre-differentiated cells, allowing them to rapidly initiate a polarized phenotype in

response to extracellular stimuli without need to synthesize epithelial cell-specific

transport machinery.

Polarized transport currently has been found to involve two different mechanisms

for targeting membrane proteins to the apical and basolateral cell surfaces. In the case of

the direct transport, separation of the apical and basolateral proteins occurs at the TON 78

via their incorporation into distinct apical and basolateral vesicles (Hong and Tang,

1993). The way this complex mechanism operates remains unclear. Through studies

using Madin-Darby canine kidney (MOCK) epithelial cells, it has been found that

dipeptidylpeptidase IV predominantly accumulates on the apical cell surface at a steady

rate, and the distribution is primarily due to fast vectorial transport from the TGN to the

apical surface (Casanova et aI., 1991). Isolation of distinct apical and basolateral

transport vesicles provides structural evidence for direct transport (Wandinger-Ness et aI.,

1991; Fiedler et aI., 1993). A speculative model has been proposed (Hong and Tang,

1993), in which the TGN has two distinct domains for protein transport to the plasma

membrane. One is a sorting domain to separate the proteins with the apical and

basolateral targeting motifs into different transport vesicles, and target them to their

respective cell surfaces. The second is a default domain located more distally from the

TGN than the sorting domain. It participates in the formation of default transport

vesicles for proteins without sorting motifs, and acts to deliver the vesicles to both the

apical and basolateral surface, without preference.

The other mechanism involved in polarized transport involves transcytosis of apical proteins. In this, apical proteins are first transported to the basolateral surface, and then are selectively retrieved into transcytotic vesicles for delivery to the apical surface.

Transcytosis is the predominant trafficking pathway for apical proteins in hepatocytes

(Casanova et aI., 1991; Zurzolo et aI., 1992). The same apical proteins in different epithelial cells may utilize two different mechanisms to get to the same final destination. 79

For example, dipeptidylpeptidase IV uses transcytosis in hepatocytes and direct transport

in MDCK cells. Zurzolo et ai. (1992) convincingly demonstrated that transcytosis of

dipeptidylpeptidase IV emerges and subsequently disappears during polarity

differentiation of the cell surface in an epithelial cell line.

Another model system for the study oftranscytosis involves polarized trafficking

of the polymeric Ig receptor, which contains several sorting motifs in its cytoplasmic tail.

A 17-residue oligopeptide domain closest to membrane acts as an autonomous motif that

is both necessary and sufficient for the basolateral sorting (Mostov, 1993). Polymeric Ig

receptor has two independent signals with critical tyrosine residues for rapid transcytosis,

which is signaled by phosphorylation of a particular serine in the receptor (Mostov,

1993).

The transcytotic vesicles and their associated proteins have been identified as

caveolae and cavcolinlcaveolin homologs respectively (Dupree, 1993, Sargiacoma, 1993;

Zurzolo et aI., 1994). Caveolin and its homolog VIP21 have been implicated in recruiting

GPI anchored proteins into glycosphingolipid complexes necessary for their apical

sorting, but additional factors are also required for accurate targeting (Zurzolo et aI.,

1994). Further molecular details of the transcytotic mechanism for apical targeting have not yet become clear.

In plant cells, it seems evident that polarized transport of cell surface proteins must exist for some situations, considering the polarized anatomical arrangement of specialized cells such as the epidermis. However, there is no direct evidence in support I'

80

of polarized protein transport in plants. It is highly possible that polarized plant cells

may not use the transcytotic pathway, but may use the direct transport pathway since

plant cells have less-developed endocytosis that their mammalian counterparts.

Sorting Proteins into Lysosomes

Sorting of both luminal and membrane proteins into lysosomes is a feature unique

to mammalian cells (Kornfeld and Mellman, 1989). The general outline of lysosomal

sorting will be briefly discussed here, since it provides a paradigm for our understanding

of the sorting of vacuolar proteins. Lysosomal sorting ofluminal proteins is

signal-receptor mediated process, the signal being identified as mannose-6-phosphate

(Man-6-P) on the N-linked polysaccharide side chains of glycoproteins (von Figura and

Hasilick, 1986). To date, this is the only signal that utilizes an oligosaccharide motiffor

protein targeting. The initial oligosaccharide is attached to lysosomal proteins ut

N-glycosylation consensus sites in the rough ER (Kornfeld and Kornfeld, 1985).

Mannose phosphorylation via a N-acetylglucosamine-l-phosphotransferase which

specifically recognizes lysosomal enzymes, occurs in the ERGIC (Pelham, 1988).

Two receptors have been isolated for the lysosomal enzymes. One is a cation­

independent Man-6-P receptor (CI-M6PR), and the other a cation-dependent

(CD-M6PR). Both are Type I membrane proteins with a cytoplasmic C-terminus

(Kornfeld, 1992). The actual M-6-P binding sites have not been identified for both

receptors. However, His13! and Argl37 ofCD-M6PR appear essential for ligand I ,

81

binding (Wendland et aI., 1991). In CI -M6PR, it contains 15 contiguous repeating

segments each approximately 147 amino acids in length. Each repeat is highly

homologous to the other repeats, with identities of amino acid sequences ranging from

16 to 38% (Kornfeld and Mellman, 1989). Repeats 1-3 and 7-11 have been shown to

bind to ligands containing M-6-P (Westlund et aI., 1991). It is well established that most

M6PRs do not reach the cell surface, instead moving directly from the TGN to join the

endocytotic route at prelysosomes, where the receptors predominantly accumulate in

steady state (Kornfeld and Mellman, 1989). Overall, the delivery of proteins to

lysosomes presumably occurs through both recycling routes: one from the TGN, and the

other from the plasma membrane (Hopkins, 1992). The lack of (lysosomal protein)

receptors in the lysosomes, along with high level of these receptors in the prelysosomes,

is interpreted to indicate that newly-synthesized lysosomal enzymes are first delivered to

prelysosomes rather than lysosomes. Then M6PRs release the lysosomal enzymes at

acidic prelysosomes, and recycle back to the TGN (Kornfeld, 1992). The regulation of

receptor trafficking between the TGN and prelysosome will be discussed in the next

paragraph, as they share characteristics with a pool of membrane proteins whose

cytoplasmic tails harbor tyrosine-containing signals for rapid internalization from the cell

surface.

The lysosomal sorting signals ofCI-M6PR and CD-M6PR are tyrosine-bearing

sequences similar to those of the LDL, polymeric Ig and transferrin receptors, lysosomal

acid phosphatase, and some other proteins (Hopkins, 1992). All contain the putative "

82

ligand binding sites to a-adaptin in the HA2 clathrin adapter protein complex, which is

involved in the internalization of proteins from the plasma membrane (Hopkins, 1992;

Kornfeld, 1992; also see "VESICLES" for the HA2 complex). However, only CI-M6PR

can interact, via low-affinity binding, with y-adaptin in the HA 1 complexes, which are

predominantly present in the traffic linking the TGN and the prelysosomes (Glickman et

aI., 1989; Hopkins, 1992). In addition, CI-M6PR can carry twice as much of the

lysosomal enzymes as can CD-M6PR, and has been shown to have two functional

binding sites for M-6-P (Repeats 1-3 and 7-11), rather than the one site in CD-M6PR

(Kornfeld, 1992). These results suggest that two classes of clathrin coats are capable of

differentiating the subset routes of protein trafficking after the TGN. y-adaptin may use

the advantage of low-affinity binding for the efficient and direct transport of large

quantities of cargo from the TGN to prelysosomes, whereas a-adaptin may use its higher

affinity to CI-M6PR to retrieve CI-M6PR leaked to the plasma membrane (Hopkins,

1992).

Sorting Luminal Proteins into Vacuoles

Sorting signals for soluble vacuolar proteins: Sorting signals are required for

soluble vacuolar proteins to be diverted into vacuoles from the default secretion pathway

(Chrispeels, 1991; Vitale and Chrispeels, 1992). However, the signals in either plant or

yeast cells are completely different from those involved in lysosomal sorting in

mammalian cells. They involve oligopeptide motifs within the primary protein ,.

83

sequences, rather than the M-6-P oligosaccharide signal. Furthermore, plant soluble

vacuolar proteins are found to utilize three types of targeting signals (internal composite

sorting signals, and N-, and C- terminal propeptide sorting signals), whereas yeast cells

appear only to employ N-terminal type signals (Chrispeels and Raikhel, 1992). So far

there is no consensus feature for all three types of plant vacuolar sorting signals. In yeast

and plant cells, the N-terminal propeptide signals become exposed after cleavage of the

ER targeting sequences; examples include sporamin and the Cathepsin 0 inhibitor of

potato, thio protease and aleurain of barley, carboxypeptidase and proteinase A of yeast

and various other vacuolar proteins (Chrispeels and Raikhel, 1992). Internal targeting

signals are found in a group of proteins that include 11 S legumin, patatin, and

phytohemagglutinin (PHA), which do not contain the cleavable ER targeting sequences;

the signals assemble and become functional from the elements scattered within the

primary sequences of the whole proteins during protein maturation (Bednarek and

Raikhel, 1992). The C-terminal type signals are found in osmotin, saporin, vacuolar

chitinase, and various lectins (Bednarek and Raikhel, 1992). Usually both N- and C­

terminal signals are proteolytically removed from the propeptides, before or during the

deposition of the mature proteins within vacuoles (Wilkins et ai, 1990; Holwerda et aI.,

1990 and 1992). Complete deletion of the sorting signals results in the secretion of the

protein into the extracellular space, and splicing of the N- or C-terminal propeptide

signals on the secretory proteins results in sorting of the proteins into vacuoles

(Chrispeels, 1991). Critical residues have been identified in each specific signal; any , •

84

change in these critical residues results in the complete or substantial loss of vacuolar

sorting function (Bednarek and Raikhel, 1992). The diverse sorting signals in plant cells

imply the presence of multiple sorting mechanisms for the targeting of vacuolar proteins.

Sorting machinelY for soluble vacuolar proteins: Sorting mechanisms have not

yet been elucidated for the soluble proteins of the plant vacuoles. Preliminary results

suggest the existence of receptors for soluble vacuolar proteins (Hinz et aI., 1993). A

soluble vacuolar protein (prolegumin) has been found to be transported in c1athrin coated

vesicles (CCV) during its trafficking to the storage vacuoles. Furthermore, it remains

associated with the components of CCV coat after the membrane is removed by detergent

(Hinz et aI., 1993). Double immunolabeling experiments (Bednarek and Raikhel, 1992)

have revealed that the non-conserved N- and C- terminal propeptide sequences target two

different proteins into the same vacuole. This suggests the existence of at least two, and

possibly more, independent sorting mechanisms. The other possibility is that a common

sorting mechanism may recognize some common physiochemical properties of different

signal sequences.

More evidence has been gathered for the sorting machinery from a collection of

vacuolar protein sorting (Vps) mutants in yeast. ArabS-like GTP-binding protein Vps21

is required for the correct sorting of many soluble vacuolar proteins, and possibly

functions in the targeting and/or fusion of transport vesicles that mediate the delivery of

proteins to the yeast vacuoles (Horazdovsky et aI., 1994). Vps 15 protein kinase and 85

Vps34 phosphatidylinositol (PI) 3-kinase are implicated in the sorting of soluble hydro lases to yeast vacuoles (Stack et aI., 1993). Vps15p is responsible for recruiting

Vps34 PI into a hetero-oligomeric protein complex associated with intracellular membranes, and its intact protein kinase domain is required for activation of Vps34 PI

3-kinase, suggesting that the sorting of the vacuolar proteins is regulated by a

Vpsl5p-mediated protein phosphorylation event (Stack et aI., 1993). Besides protein components, physiochemical conditions also play important roles in the sorting of yeast vacuolar proteins. It found that acidification of yeast vacuole is required for efficient sorting of soluble vacuolar proteins (Klionsky et aI., 1992). Compartment acidification may facilitate the release of receptors from bound ligands as described in the recycling of the M-6-P receptor for lysosomal enzymes. Morphological evidence indicates that yeast may have a prevacuolar compartment similar to prelysosome of mammalian cells, which is the site involved in the return of M-6-P receptors to the TON (Raymond et aI., 1992).

So far, no receptor has been isolated for shuttling of soluble vacuolar proteins in yeast.

Further investigations on other Vps mutants of yeast should provide a more complete understanding concerning the sorting machinery of soluble vacuolar proteins.

Sorting Membrane Proteins into Vacuoles, or An Alternative Default Pathway?

Recent analyses of the targeting domains of several yeast TON and vacuolar membrane proteins has revealed an astonishing result, in that the trafficking of the membrane proteins from the TON to vacuoles may operate as a default pathway of I'

86

secretion without any sorting signals (Roberts et aI., 1992; Cooper and Bussey, 1992;

Bryant and Boyd, 1993; Nothwehr et aI., 1993; Nothwehr and Stevens, 1994). DPAPA

is an integral TON protein (see "RETENTION OF PROTEINS AT TIlE TRANS-OOLGI NETWORK "),

whereas its sister DPAPB is a vacuolar membrane protein. The analyses ofDPAPB

targeting showed that no single domain is required for its delivery to vacuole (Roberts et

aI., 1992). Furthermore over-production of DPAPA, or removal of its cytoplasmic

retention tail resulted in its mislocation to vacuole (Roberts et aI., 1992). The delivery of

both wildtype DPAPB and mutant DPAPA to vacuoles is unaffected in the secl mutant,

which blocks vesicle transport from the TON to the plasma membrane; thus, transport to

the vacuole does not occur via the plasma membrane followed by endocytosis (Roberts et

aI., 1992). Furthermore, the deletion of cytoplasmic retention signals from the integral

TON protein Kex 1p and Kex2p also causes their mislocation from the TON to the

vacuole membrane (Cooper and Bussey, 1992; Bryant and Boyd, 1993).

Similar insight have emerged from studies of transport of plant vacuolar

membrane proteins, called tonoplast intrinsic proteins (TIPs). TIPs are a group of

abundant proteins that are present as different isoforms in embryos of seeds and

vegetative tissues (Johnson et aI., 1990; Hofte et aI., 1992). They have multiple

transmembrane domains and are suspected to form channels across membranes (Hofte

and Chrispeels, 1992). The studies on u-TIP postulated that one of the six

transmembrane domains may contains a vacuole targeting signal, based on the result that

deletion mutants in other regions all accumulated in the tonoplast. However, this 87

conclusion (Hofte and Chrispeels, 1992) is open to the criticism that a-TIP may be

directed to the tonoplast by default secretion. A later report provided biochemical

evidence that the TIPs and soluble vacuolar proteins are transported to vacuoles through

different paths (Gomez and Chrispeels, 1993). The transport of TIPs to vacuoles is faster

than that of soluble vacuolar protein PHA (Gomez and Chrispeeis, 1993). The transport

of TIPs is not blocked by BFA and monensin, which interfere with the ER to Golgi

transport of proteins and inhibit correct TGN sorting of proteins respectively (Gomez and

Chrispeels, 1993). However, either drug can block the transport of PHA to vacuoles and

alter the processing rate of the complex glycan side chain (Gomez and Chrispeels, 1993).

The question still remains open regarding whether TIPs use a default secretion pathway.

More conclusive results should come from the deletion analysis of retention signals in the

isolated ER membrane proteins from plants or the expression of heterologous membrane

proteins in plants. If the vacuole membrane is a default destination in yeast and plant

cells, it implies that the transport of membrane proteins from the TGN to the plasma

membrane should be mediated by sorting signals. Evidence is particularly scarce for

signal-mediated sorting to the plasma membrane in both yeast and plant cells. The yeast

Ste6p protein, a membrane protein involved in transport of the a-factor peptide mating

pheromone directly across the plasma membrane, is predominantly located in .,1

intracellular vesicle-like bodies (Kuchler et ai., 1993). Upon exposure of the yeast cells to the mating pheromone, the intracellular Ste6p shifted to the plasma membrane, 88

indicating that at least in this case the transport of St6p to the plasma membrane is

regulated by signals (Kuchler et aI., 1993).

EXPERIMENTAL APPROACHES FOR THE STUDY OF PLANT

ENDOMEMBRANE-SECRETORY PATHWAY

Plant cell biologists have adopted both ultrastructural and biochemical approaches to explore the plant endomembrane-secretory pathway. Ultrastructural investigation allows an overview of the subcellular compartments and domains involved in the pathway, and provides us with clues as to how the pathway operates biochemically. For example, the unambiguous identification of the TON in the plant Oolgi emphasizes the possibility that plant secretory proteins are sorted and packaged at the TON, and transported to the vacuoles and the plasma membrane in a manner similar to their mammalian counterparts (Staehelin et al., 1990). Another example is provided by the ultrastructural identification of c1athrin-coated vesicles in plant cells, and this eventually led to biochemical purification and isolation of plant c1athrin proteins (Coleman et aI.,

1991). Ultrastructural studies, albeit very useful, are technically demanding and laborious, and are limited to the components that can be detected by current equipment and technology. Further, in most situations, they reveal the steady, not the dynamic, "

89

status of the secretory pathway. Ultrastructural observations typically require

corroboration from counterpart biochemical studies.

Biochemical studies of the endomembrane-secretory pathway inevitably depend

on molecular markers. These markers can be oligosaccharides, lipids or polypeptides.

The key to successful biochemical studies is that the selected markers truly represent the

targets within the pathway for investigation and can be unambiguously detected in the

systems chosen for study. Various investigations have been conducted in plant cell

systems based on all three types of markers, but most published results involves the

investigations of systems that employ polypeptide markers. The usage of various

molecular markers, in conjunction with technique developments will be discussed below.

The presence of oligosaccharides in the plant secretory pathway derives from two

sources: (a) the glycosylation of polypeptides and lipids entering the pathway; (b) the

biosynthesis of polysaccharides destined to the cell wall. Using autoradiography in

conjunction with electron microscopy, it has been shown that the plant Golgi is involved

in the synthesis of oligosaccharides (reviewed by Mollenhauer, 1980; Harris, 1986).

Moore et al. (1991), using antibodies directed against various saccharides along with

electron microscopy, revealed that individual stacks of the plant Golgi process

glycoproteins and synthesize complex polysaccharides simultaneously, but separate the

synthesis and the processing events in different cisternae. The use of oligosaccharides as

markers to study the plant secretory pathway has been quite limited due to several factors:

(1) difficulties involved in the manipulation of their biosynthesis; (2) the laborious and 90

technically demanding nature of electron microscopy; (3) difficulties involved in the development of specific antibodies.

Lipid molecules comprise one of the major classes of biochemical molecules involved in the secretory pathway. Their movement between membrane compartments occurs via several different ways: (a) vesicle fission and fusion; (b) lateral diffusion between membrane compartments connected by membrane bridges; (c) transport oflipid monomers through the cytoplasm (Sleight, 1987; Bishop and Bell, 1988; Pagano, 1990).

Fluorescent lipid precursors and analogs such as N-[7-(4-nitobenzo-2-oxa-l,3-diazole)-E­

aminocaproyl sphingosine (C6-NBD-ceramide) have been used to study the transport of newly synthesized sphingolipids from the Golgi to the plasma membrane in mammalian

cells (Lipsky and Pagano, 1983). C6-NBD-ceramide and its derivatives have been subsequently developed as a vital stain for the Golgi apparatus in mammalian cells

(Lipsky and Pagano, 1985). Fluorescent ceramides have also been used as markers in the studies of plant endomembrane system. After analyzing the redistribution of photobleached ceramides in soybean root cells, Grabski et al. (1993) concluded that the

ER within plasmodesmata provides a dynamic link between contiguous cells. However, the ceramides have not been popularly used in the studies of the plant secretory pathway; this situation is perhaps due to the fact that they stain all cellular membranes indiscriminately in a very quick manner (less than 5 minutes for tobacco suspension cells;

Gong, unpublished data). "

91

In comparison with oligosaccharide and lipid markers, polypeptide markers offer

unparalleled advantages in the studies of the endomembrane-secretory pathway. These

advantages include: (I) the easy manipulation of the peptide sequences; (2) the easy

control of polypeptide transcription and expression; (3) the easy transfer of polypeptide

markers from one organism to another; (4) the availability of various molecular resources

(such as cDNAs and antibodies); (5) the fact that enzymatic activities of many

polypeptides can be used as detection systems; and (6) the availability or potential

availability of specific markers for every membrane domain or compartment. In a line of

pursuing polypeptide markers, two options have been adopted to harness the power of

polypeptide markers. The first involves the development of polypeptide markers within

their native (or homologous) environment. The other one adopts the strategy of

transferring established polypeptide markers from one organism to another (heterologous

environment). Homologous and heterologous polypeptide markers have been widely

used in the studies of the mammalian endomembrane-secretory pathway. However,

heterologous polypeptide markers have been essentially unused in studies of the plant

secretory pathway.

Various techniques involved in protein purification are used to isolate

homologous polypeptide markers. Many of the purified polypeptide markers have

already been discussed regarding their functions in the plant secretory pathway in the

various sections of this chapter, and detailed discussion of their isolation techniques is

beyond the scope of this project. Regardless of how successful those techniques are in 92

polypeptide isolation, they have failed in the isolation and purification of homologous polypeptide markers for the plant Golgi. Biochemical purification of the plant Golgi membrane has been extensively studied, typically including isopycnic and rate-zonal gradient centrifugation (Ray et aI., 1969; Ray, 1979) and, more recently, other methods such as phase-partitioning (Larsson et aI., 1987) and free flow electrophoresis (Morre et al.,1991). Those methods, although clearly effective in enrichment of the ER, plasma membrane, vacuoles and vesicles (Morre et aI., 1993), are unable to enrich the plant

Golgi to a point that allow the unambiguous identification of polypeptide markers for the various Golgi cisternae. Few attempts have been made to separate the plant Golgi into different cisternae. In one case, glycosyltransferases of the plant Golgi were reported being separated by centrifugation of a shallow sucrose or Percoll gradient (Ali et aI.,

1986). In contradiction, Sturms et al. (1987) reported that subcompartmentation of those enzymes was not achieved. Brummell et al. (1990) showed the separation of xyloglucan xylosyItransferase from fucosyItransferases using isopycnic gradit:nts, leading to the conclusion that synthesis ofxyloglucan initiates in the cis Golgi and is completed in the trans Golgi, and that the addition of fucosyl-residue to xyloglucan starts in the trans cisternae and terminates in the secretory vesicles. However, the group of Dr. Staehelin has provided convincing evidence that xyloglucan is synthesized and assembled exclusively within the trans cisternae and the TGN, and rhamnogalacturonan I only within the cis and medial cisternae (Moore et aI., 1991; Zhang and Staehelin, 1992). The differences between the results of the latter two groups may be reconciled by the 93

observation that the plant Golgi is highly tailored to accommodate the functions of

specific cell types and tissues. This type of tissue-specific heterogeneity of the plant

Golgi contributes tremendous theoretical and practical difficulties to its isolation using

the traditional methods discussed above.

Alternative techniques have also been developed to identify homologous polypeptide markers in the plant secretory pathway, including the utilization of the

Arabidopsis genetic system and the complementation of yeast secretory mutants. For example, by screening leaf extracts with an antiserum directed against complex glycans, von Schaewen et al. (1993) isolated an Arabidopsis thaliana mutant (cgl), which lacks

P-l,2-N- acetylglucosaminyltransferase I (a Golgi enzyme) and fails to convert glycoprotein glycans from high-mannose to complex form. Unfortunately the CGL gene and products have not yet been isolated because the mutant was selected from an

EMS-mutated population. Several successful examples have been reported from the systems using yeast complementation to identify the plant genes involved in the secretory pathway, including identification of plant homologues of the yeast KDEL receptor and

GTP binding proteins (Lee et aI., 1993; Palme et aI., 1992, 1993; d'Enfert et aI., 1993).

The success of these techniques depends on the availability of suitable screening protocols for specific Arabidopsis mutants, and the degree of convergence of the interesting genes between plants and yeast.

Once homologous polypeptide markers have been isolated, they can be used for the development of the corresponding detection systems (typically specific antibodies), 94 which are then employed to probe the subcellular location and kinetics of these markers in their native cellular environments. The main disadvantage of homologous polypeptide markers is that the detection systems only report the steady state distribution of the markers. Kinetic analysis requires radiolabeling under pulse-chase conditions and immuno-precipitation. Another minor issue is that the homologous cellular environment imposes some limitations to the studies on targeting motifs of polypeptide markers, especially in a situation that protein-protein interaction is a mechanism underlying the retention and sorting of proteins. A more-general solution to these problems is the adoption of heterologous polypeptide markers. The heterologous approach has become increasingly popular with the improvement of gene transfer and expression techniques.

The main advantage of heterologous polypeptide markers is that their detection is both easy and unambiguous; the other advantages and selection criteria of heterologous polypeptide markers are covered under the section of "DISCUSSION" in Chapter 3.

However, one caveat must be taken considered; heterologous markers may not represent the true characteristics of the secretory pathway for which they are intended to provide information. Therefore their behavior in the host secretory pathway requires full evaluation before they are used as probes to investigate specific steps in the pathway.

Heterologous polypeptide markers are generally used in two ways. The first way involves the direct transfer of makers from one organism to another; such examples have been widely reported across mammalian, plant and yeast cell systems. One of the best-used heterologous polypeptide markers is the glycoprotein of vesicular stomatitis 95

virus (Bergmann et aI., 1981; Rose and Bergmann et aI., 1982; Guan and Rose, 1984;

Machamer et aI., 1985; Machamer and Rose, 1988; Galbraith et aI., 1992; also discussed

under "INTRODUCTION" in Chapter 3). The second way involves the construction of

recombinant protein, in which a reporter molecule is fused translationally with a targeting

motif, which has been observed to be necessary and sufficient for the correct targeting of

the remainder of its native molecule. The function of the targeting motifis then

examined in vivo and in vitro using the reporter protein as a detection tool. Such way is

best exemplified by the studies of signal peptides of various secretory proteins (see "ER

TARGETING AND TRANSLOCATION" described previously in this chapter). In this project,

several heterologous membrane proteins were used in plant cells to evaluate their

feasibility as probes to study the cis-elements and the retention/sorting mechanisms of

membrane proteins. The results are analyzed and discussed later in the respective

chapters (from Chapters 3 to 5).

SUMMARY AND PERSPECTIVE

Molecular sorting and retention of proteins is basically a signal-receptor mediated

process in the endomembrane-secretory pathway, except for the retention of membrane

proteins in the three Golgi cisternae. Experimental evidence concerning the whole pathway is not evenly distributed; most of the detailed information is available in the 96

sorting and retention steps of proteins either before or after the Golgi apparatus, and

much less is known concerning retention in the Golgi apparatus. Fortunately, each discrete compartment in the endomembrane-secretory pathway has protein probes and/or mutants available for further investigations in mammalian and yeast cells.

Exploration of protein sorting in plant cells shows a similarly uneven distribution pattern of achievement. In essence, plant scientists are following the lead of mammalian and yeast cell systems, and are confirming the applicability of models from mammalian and yeast systems in plants. A certain amount of information has been accumulated in the molecular sorting of proteins in plant cells, but it far from provides a detailed understanding of the endomembrane-secretory pathway. Based on this limited body of evidence, our understanding of protein sorting in plant cells rests mostly on the concept of evolutionary convergence between yeast, plants, and mammals. Even in the two best studied areas of protein trafficking, the retention and the sorting of soluble ER and vacuolar proteins, most current plant investigations remain at the level of uncovering the cis-acting elements of proteins. Analysis of the sorting mechanisms of various membrane proteins in the secretory pathway is almost untouched in plant cell systems. Progress in plant cells is impeded by three factors: the scarcity of suitable protein probes for each compartment as described for mammalian cells, the lack of genetic systems of similar power to that of yeast, and the existence of technical challenges in the manipulation of plant cells imposed by their uniqueness. However, despite these unfavorable situations, the persistent effort of many workers in this field not only offers hope to better 97 understand the sorting mechanisms at a higher level, but also points to the particular areas that should be emphasized for progress in the near future. These areas are as follows: (1) the molecular retention machinery for ER luminal proteins; (2) the signals and mechanisms of retention of ER membrane proteins; (3) the development of means for isolating Golgi proteins or the establishment of heterologous markers for the Golgi; (4) the development of native marker systems from the currently-isolated membrane proteins; (5) the isolation of receptors for soluble vacuolar proteins and; (6) the clarification of the hypothesis that vacuolar membrane is a default destination for plant membrane proteins. 98

CHAPTER 2

GENERAL MATERIALS AND METHODS

The general experimental methods and techniques that were routinely used in this research are grouped together in this chapter for easy reference. Methods and techniques that were developed as part of work in a specific chapter are detailed within that chapter.

Standard sterile teclmique was employed for all work that required sterile conditions.

Generation and Maintenance of Axenic Plantlets of Tobacco

The procedure is based on that of Murashige and Skoog (1962) with slight modifications (Harkins et aI., 1990). Tobacco (Nicotiana tabacum L.) seeds are sterilized as follows: 100-300 seeds are placed in a sterile 15 mL conical tube, a 10 mL aliquot of

70% ethanol is added for 1 minute. The seeds are then collected by centrifugation in a table-top centrifuge at 1,000 rpm for 1 minute. The seeds are washed once with 10 mL of sterilized water, after decanting the ethanol. A 5 mL aliquot of 30% (v/v) Clorox bleach is then added for 15 seconds. The seeds are collected, and washed three times with 10 mL of sterilized water. The sterilized seeds are resuspended in a suitable amount of sterilized water (about 1 mL) for easy transfer to culture medium.

The sterilized seeds are pipeted out and evenly seeded onto the surface of solid 99 sterile MS culture medium in Magenta boxes. Solid MS culture medium contains 4.3 giL

MS medium powder of basal salts (Gibco), 30 giL sucrose, 1.0 mglL I-naphthaleneacetic acid (NAA), and 8 giL phytoagar, pH 5.8. For transgenic plantiets, 100 mg kanamycin sulfate (filter-sterilized stock solution) is added per 1000 mL medium, following cooling of the autoclaved medium to 50°C. After seeding, Magenta boxes are wrapped with parafilm and placed in a growth chamber (22°C; 2,000 lux of continuous white light).

Germinated plants are transplanted into Magenta boxes containing fresh medium (two per box) when they reach a height of2-3 cm. After 2-3 transplantings, the third to fifth leaves of most plantlets are well-expanded, and are suitable for protoplast isolation, leaf-disc transformation, or callus production.

Generation and Maintenance of Tobacco Callus and Suspension Cell Cultures

This procedure is modified from the protocols of Hall (1991 a and 1991 b).

Well-expanded leaves are excised from sterile plantiets and placed on the surface of solid

3 MS + medium, comprising solid MS culture medium (described above) supplemented with three additional ingredients (0.1 mg/L 6-benzylaminopurine, 0.4 mg/L thiamine-HC1, and 100 mg/L myo-inositol). The medium is contained in sterile plastic petri dishes 10 cm in diameter (20 mL medium per dish). The excised leaves are cut into 100

pieces (0.5 to 1 cm), leaf-margins and large vessel tissues are trimmed, and the pieces are placed upside down on the medium surface. Care is taken to ensure good contact with the medium surface. The petri dishes are wrapped with parafilm and placed in a temperature-controlled chamber (30°C, in darkness). Calli are transferred onto fresh medium when their diameters reach 0.3-0.4 cm. Calli will grow loosely after the first transfer, and can be maintained regularly thereafter. Transfer is required when the medium become exhausted (about 1.5 - 2 months). For generation and maintenance of

3 transgenic calli, kanamycin sulfate is added to solid MS + medium at a final concentration of 100 mg/L.

Suspension cell cultures are prepared from friable tobacco calli. Small cell aggregates are picked up by forceps from the periphery of calli, and are transferred into

3 100 mL of liquid MS + medium in a 500 mL Erlenmeyer flask. The flask mouths are stoppered with sterilized cotton balls, covered with aluminum foil. The flasks are shaken at a speed of 100 rpm in an incubator (25 oC, without illumination). Subculturing is needed every 2-3 weeks. Cells are harvested by pipeting 5 mL cell suspension (about 1 gram) into a pipet lacking a tip (this is broken-off by bending a 10 mL sterile serological plastic pipet within its individually-sealed package), and then pipeting medium out against the flask wall. The harvested cell clumps are then transferred into fresh medium by pipeting the medium up and down several times. For transgenic suspension cells, kanamycin is added to a final concentration of 100 mg/L. I'

101

Protoplast Isolation and Transfection

The procedures of protoplast isolation and transfection are essentially the same as

those described by Harkins et al. (1990), which are modified from the procedures of

Negrutiu et al. (1987). Well-expanded leaves are excised from the third to fifth leaf-node

of plantlets, and 3-4 of these are sliced perpendicularly to the midrib in 1 mm slices in 20

mL of sterile enzyme digestion solution (10,000 units/L cellulysin, 3,000 units/L

driselase, 3,000 units/L macerase, 7 mM CaCI2, 2 mM MES, 0.5 M mannitol, pH 5.8)

contained in a 10 cm sterile plastic petri dish. The leaf slices are incubated for 18 hours

in a growth chamber at 22°C with continuous illumination. The released protoplasts are

filtered through a double-layer of cheesecloth, collected, and diluted with an equal

volume ofW5 medium (154 mM NaCI, 125 mM CuCI2, 5 mM KCI, 5 mM glucose, pH

5.7). They are allowed to equilibrate for 10-15 minutes, and then pelleted using a

tabletop centrifuge (6 minutes, 150g). The pellet is gently resuspended in 20 mL of 25%

(w/v) sucrose solution (pH 5.7) containing 1 X NTTO nutrient stock (Appendix I),

overlaid carefully with 5 mL W5 medium, and centrifuged in a tabletop centrifuge at 150

g for 10 minutes. Viable protoplasts are collected with a sterile Pasteur pipette (wide

bore) from the interface between the W5 medium and 25% sucrose solution, diluted with

2-3 volumes of W5 medium, and counted using a haemocytometer.

The viable pro top lasts are divided into separate tubes for different transfection

treatments, and are pelleted using a tabletop centrifuge at 150 g for 5 minutes. The pellet I.

102

is then resuspended in an appropriate amount ofW5 medium to a final concentration of

about 106 protoplasts/mL. The protoplast suspension is mixed with specific plasmid

DNA and calf thymus carrier DNA to final concentrations of20 and 50 ).lg/mL

respectively. Plasmid DNA stock is prepared in TE buffer (10 mM Tris, 1 mM EDTA,

pH 8.0), and should be free of ethel' that can be carried over from the final step of large­

scale plasmid DNA isolation. Calf thymus carrier DNA stock is prepared by sonicating

calf thymus DNA solution (in TE buffer) using a Polytron sonicator with an autoclaved

tip, and filtering through a sterile 0.22 !-lm filter. The protoplast suspension is gently

swirled, and the tube tapped 2-3 times to achieve good mixing. To this is added 1.5

volumes of 40% (w/v) Polyethyleneglycol3750 (pH approx. 7.0) containing 0.4 M

mannitol and 0.1 M Ca(N03)AH20. The protoplasts are incubated at room temperature

for 25 minutes, then mixed with 8 volumes of IX NTTO nutrient stock (Appendix I),

0.154 M mannitol, 0.38 M glucose and 75 mg/L ampicillin (pH 5.7). The transfected

protoplasts are incubated in darkness at 22 oC for an appropriate amount of time (usually

20-24 hours). They are then diluted with an equal volume of W5 medium. Viable

transfected protoplasts are then isolated by sucrose floatation as described under

Protoplast Isolation.

Agrobacterium Transformation of Tobacco Leaf-discs

Plant transformation is carried out as described by Horsch and Klee (1986). I'

103

Using the triparental mating technique, pBIN 19-based plasmids containing the various

binary expression cassettes are transferred from E. coli DH5a. to the LBA 4404 strain of

Agrobactel'ium tumefaciens via a helper E. coli bacterium containing the RK. 2013

plasmid. The transformed agrobacteria are selected using Minimal T medium (Appendix

II) supplemented with 100 Ilg/mL kanamycin sulfate. Transformants are verified through

DNA restriction analysis of the binary plasmids, and are maintained for leaf-disc

transformation of tobacco, or stored as glycerol stocks at -80°C.

Leaves are selected from axenic tobacco plantlets, trimmed to remove the mid-rib

and leaf margins, and cut into discs of 0.5 - 1 em in diameter. The leaf-discs are soaked

briefly in medium containing log-phase agrobacteria, blotted clean on sterile filter paper,

and then placed upside down on the surface of tobacco feeder-cell plate medium

(Appendix III). Incubation is continued for 3 days in a growth chamber at 25°C with

continuous illumination (2,000 lux). The cocultivated leaf-discs are then transferred to

tobacco shooting medium (Appendix IV), supplemented with 200 Ilg/mL kanamycin

sulfate and 500 Ilg/mL carbenicillin. They are transferred to the same fresh medium

every seven days until calli emerge and grow into size of 0.4 em or more in diameter.

The calli emerging from different transformation sites on leaf explants are considered as

independent transformants, and maintained in separate containers. When shoots reach

about 1 em in length, they are excised from calli and transplanted into solid MS medium

containing 100 Ilg/mL kanamycin sulfate. Maintenance of transgenic plants is the same

as described under "Generation and Maintenance of Axenic Plantlets of Tobacco". 104

Measurement of Protein Content

Assay of protein content is based on the modified Micro-Lowry assay from

Peterson (1977). The procedure offers a simple, rapid, accurate analysis, and allows removal of materials from the reaction mixture that can interfere with the assay. The following steps are only necessary if samples are suspected to contain interfering materials or are too dilute. Duplicated sanlples of200 ~lL (estimated 1-20 ~g protein) are mixed with 20 ~L 0.15% (w/v) sodium deoxycholate, and incubated for 10 minutes at room temperature. The samples are then mixed with 20 ~L of 72% (v/v) trichloric acetic acid, vortexed briefly, and centrifuged in a microfuge for 15 minutes. The tubes are blotted upside down on a paper towel after removal of supernatants. The pellets are

dissolved in 200 ~L H20 for assay.

The 200 ~L duplicated samples and protein standards (bovine serum albumin,

BSA) are mixed with equal volumes of Reagent A (Appendix V), and incubated for 10 minutes at room temperature. The samples are then mixed with 100 ~L Reagent B (l part

Folin-Ciocalteu phenol reagent and 49 parts water), vortexed immediately, and incubated for 30 minutes at room temperature. Absorbance is measured at a wavelength of 750 nm and converted into protein concentrations using standard curve or Peterson's regression equation as below (Peterson, 1977).

Ln (Ano) = b Ln (~g protein/reaction tube) + a A7so: absorbance of protein sample at 750 nm. 105

)lg protein/reaction tube: protein concentration from either a known set BSA standards or from unknown samples: a, b: slope (b) and constant (a), calculated through regression on a known set of BSA concentrations and their corresponding absorbance (Am).

Gel Electrophoresis and Western Blotting

One-dimensional sodium dodecyl sulfate-polyacrylamide gel electrophoresis

(SOS-PAGE) is extensively used in this project. The procedure is derived from the protocol described by Laemmli (1970). Oepending on the expected molecular size of the expressed proteins, running gels are composed of 7.5, 10, 12.5, or 15% of monomer mixture (37.5 parts acrylamide: 1 part bisacrylamide), buffered with 0.375 M Tris (pH

8.8) and 0.1 % SOS. For gels containing more than 10% monomer, glycerol is added to a final concentration of 5% (v/v), in order to prevent gel cracking. Stacking gels are composed of 4% of the monomer mixture, buffered with 0.125 M Tris (pH 6.8) and 0.1 %

SOS. Gel solutions are degassed under vacuum, then polymerized with ammonium persulfate (0.75 Ilg/mL gel solution) and TEMEO (0.75 - 1.0 IlLlmL gel solution).

For each gel lane, 25 to 200 Ilg protein (or 105 to 106 protoplasts) are diluted with one volume of 2 X SOS-sample buffer (0.125 M Tris, 4% SOS, 20% glycerol, 10% l3-mercaptoethanol, and a trace amount of Bromophenol blue dye, pH 6.8). The sample mixtures are boiled in water bath for 5 minutes, and centrifuged in a microfuge for 3-5 minutes, and the supernatant loaded onto a gel. Gels are generally run at constant voltage

(100 Volts) for 1 - 1.5 hours or until the dye front reaches the bottom of the running gel. I'

106

After completion of electrophoresis, the gels are disassembled, soaked in cold

10% methanol, 10 mM CAPS (PH 10.5; Matsudaira transfer buffer) for 5 minutes, and

assembled into electroblotting cassettes containing polyvinylidene difluoride membrane

(Immobilon transfer membrane, 0.45 ).lm porosity; Millipore Corp., Bedford, MA, USA).

Proteins are electrophoretically transferred to the membranes in cold Matsudaira transfer

buffer under constant voltage (100 Volts) for 1 hour. The blotted membranes are washed

in deionized water briefly, then either sealed in plastic bags for storage in freezer (at -20

°C or below) or used for immediate antibody probing.

Antibody probing is conducted according to the protocol provided by Clonetech

(Palo Alto, CA, USA). The blotted membranes are washed 2 times (first for 60 minutes,

then for 30 minutes) with gentle shaking in TSW medium (10 mM Tris, 0.9% NaCl,

0.25% gelatin, 0.1% Triton X-IOO and 0.02% SDS, pH 7.4). The washed membranes are

then incubated in TSW medium plus primary antibodies from rabbit or mouse directed

against the specific antigens (the titers are varied for specific antibodies) for 1 hour at

room temperature. Two washings (15 minutes each) are then carried out in TSW

medium. The membranes are further incubated in TSW medium plus secondary

antibodies which are either goat anti- rabbit or mouse polyclonal Ig G (H + L) conjugated

with alkaline phosphatase (with a dilution of I: 10,000 for rabbit antibodies and 1:2,000

for mouse antibodies). The probed membranes are then washed two times (15 minutes

each) in TSW medium and one time (about 5 minutes) in color development solution

(100 mM Tris, 100 mM NaCl, 5 mM MgCI2 , pH 9.5). Final color development is 107 conducted in the color development solution plus 33 flg/mL 5-bromo-4-chloro-3-indolyl phosphate and 66 flg/mL nitroblue tetrazolium for 30 minutes or until protein bands appear on blots. The developed blots are briefly washed first with tap water for 2 or 3 times, then with deionzied water for 3 times (5 minutes each), and finally dried on dry clean tilter paper in darkness.

Preparation and Fractionation of Sucrose Gradients

Preparation of linear sucrose gradients is similar to that described by Galbraith et al. (1992). Linear sucrose gradients (20 - 50% w/w) are made from equal weights of20%

(w/w) and 50% (w/w) sucrose solutions buffered with 50 mM Tris (pH 8.0) and 1 mM

EDTA in a gradient mixing apparatus. The gradients are overlaid into centrifuge tubes containing 1110 volume of 50% sucrose solution as a cushion layer. Microsomal samples (usually about 1/10 volume of the gradient) prepared from differential centrifugation, are carefully overlaid onto the gradients. The gradients plus the samples and cushion (usually about 12 mL total volume per tube) are centrifuged at 4 °e and

35,000 rpm in a SW41 swinging-bucket rotor using a Beckman Ultracentrifuge (Model

L8-M) for 16 to 18 hours. The gradients are fractionated into 0.6-mL fractions using an

ISeO gradient fractionateI'. Fractions, placed on ice, are analyzed immediately for sucrose concentration (% w/w) with a refractometer, chlorophyll content, and the activities of labile NADH cytochrome c reductase. Other enzyme assays are later I'

108

conducted after fractions are stored at -80°C. For gel electrophoresis, 0.3 mL or more of

fresh fractions are diluted with three volumes of 50 mM Tris (pH 8.0) and 1 mM EDTA

solution, and centrifuged at 300,000g for 60 minutes. Membrane pellets are resuspended

in a suitable amount of 2X SDS-sample buffer and stored in at -80°C until use.

Chlorophyll Measurement

Chlorophyll measurement is conducted as described by Galbraith et al. (1992).

Fractions (50 I-1L or less) are mixed with 50 mM Tris (PH 8.0) and 1 mM EDTA solution

to a final volume of 200 I-1L in wells of a 96-well microtiter plate. Fluorescence

intensities are measured using a Titertek Fluoroskan II fluorometer (Marctronics, Vir.

USA) equipped with an excitation filter of 430 nm and an emission filter of 680 nm.

NADH Cytochrome C Reductase Assay

The assay is based on the protocol described by Donaldson et al. (1972). The

non-specific baseline activity is scanned prior to kinetic measurement ofNADH

dependent activity. Aliquots (100 I-1L) from each fraction of gradient are mixed quickly

with 314 I-1L of enzyme assay solution, which contains 0.2 mL K-phosphate buffer (0.2 M

,pH 7.5), 10 I-1L KCN (10 mM), 41-1L antimycin A (2 mg/mL in ethanol) and 0.1 mL

oxidized cytochrome C (2 mg/mL). Absorbance is recorded at a wavelength of550 nm 109

at room temperature for a period of 4 minutes. For determination ofNADH dependent

enzyme activities, 0.1 mL NADH substrate (2 mg/mL in water) are then added and

mixed. The readings are recorded as for the baselines. The kinetics of the enzymatic activities are calculated after subtracting the change due to baseline noise from the change of the absorbance in the presence of NADH.

Inosine Diphosphatase Assay

Assa.y is based on the protocol described by Gardiner and Chrispeels (1975). In order to determine the activities of latent inosine diphosphatase (IDPase), fractions are stored at -80°C for 2-3 days. A suitable amount (typically 50 ilL) of the fractions is mixed with IDPase buffer. The final reaction mixture contains 50 mM Tris (pH 7.2), 100 mM KCl, 5 mM MgCI2, plus 5 mM inosine diphosphate (to measure specific IDPase activity) or 5 mM p-nitrophenol phosphate (as a control to measure non-specific phosphatase activities). The reaction mixture is incubated at 35°C in a water bath for one hour or more. In order to ensure linearity of assay, aliquots are taken at defined time.

The reaction is stopped by adding trichloroacetic acid to a final concentration of 11.6%

(v/v). Supernatants are collected for assaying the released phosphate after insoluble materials are removed by centrifugation in a microfuge for 15 minutes.

The released inorganic phosphate is measured according to the method of Taussky and Shorr (1953). One volume of supernatant is added to 1.5 volume of ferrous 110

sulfate-ammonium molybdate solution (Appendix VI), and then incubated at room

temperature for 15 minutes or longer until development of a light-blue color. To ensure

linearity of assay, the reaction is done using various amounts of inorganic phosphate.

The phosphate released by specific IDPases in each fraction is calculated by subtracting the absorbance at 710 nm of the control from the absorbance of the IDPase reaction mixtures.

V anadate-sensitive ATPase Assay

The presence of the plasma membrane is determined by the activity of vanadate­ sensitive ATPase in fractions of the sucrose gradient using the protocol described by

Briskin et al. (1987). Fractions (50 ).lL or a suitable amount) are added to two sets of

reaction buffers with or without 0.5 mM Na3 V04 • The final concentration of the reaction buffer is 3 mM ATP-Tris, 30 mM MES-Tris (pH 6.5),50 mM KCl, 3 mM MgS04 ,0.1 mM NaN3 and 0.1 mM (NH4)6M07024' The reaction mixtures are incubated at 38°C in water bath for 60 minutes or more. It is difficult to confirm the linearity of each assay routinely due to the limited amount of samples in each fraction. The optimal incubation period is empirically estimated according to the amount of microsomes loaded onto sucrose gradient and the volume of the fraction. The reaction is then stopped by adding trichloroacetic acid to a final concentration of 11.5% (v/v). Supernatants are collected for measuring the released inorganic phosphate after removal of the insoluble materials by 111

centrifugation (room temperature) in a microfuge for 15 minutes. The released inorganic

phosphate is measured according to the method described by Taussky and Shorr (1953);

also see "Inosine Diphosphatase Assay" .

p-Glucuronidase Assay

The assay is conducted by following the protocol of Jefferson et al. (1987). An appropriate amount of the samples is mixed with 1 X GUS extraction buffer, containing

50 mM NaH2P04 (pH 7.0),10 mM EDTA, 0.1% Triton X-IOO and 10 mM

p-mecaptoethanol. The reaction is initiated by adding methyl umbelliferone glucuronide

(MUG, final concentration 1 mM) at 37°C. After various incubation periods (typically

0, 15,30,60 and 90 minutes), the reaction is stopped by adding Na2C03 solution to a final concentration of 0.18 M. Fluorescence of the hydrolyzed MUG is measured in a

Titertek Fluoroskan II fluorometer equipped with an excitation filter of 365 nm and an emission filter of 455 nm (Marctl'Onics, VA, USA). 112

CHAPTER 3

EXPRESSING HETEROLOGOUS MEMBRANE PROTEINS WITHIN

PLANT CELLS FOR USE AS PROBES OF THE SECRETORY PATHWAY

INTRODUCTION

One of the major protein trafficking routes within eukaryotic cells is the endomembrane-secretory pathway, which comprises the outer nuclear envelope, the endoplasmic reticulum (ER), the Golgi apparatus, the plasma membrane and cell surface, plus lysosome (vacuoles), and various secretory, intermediate, and storage compartments

(Harris, 1986; Burgess and Kelly, 1987; Kukuruzinska et aI., 1987; Bennett and

Osteryoung, 1991; Herscovics and Orlean, 1993). The structures and mechanisms underlying the operation of the endomembrane-secretory pathway have been intensively investigated at the molecular and cellular levels, most notably using mammalian and yeast cells (Pfeffer and Rothman, 1987; Pelham, 1989; Hurtley and Helenius, 1989; Pryer et aI., 1992; Rothman and Orci, 1992; Hong and Tang, 1993). Advances have been driven in yeast largely due to its ease of manipulation as a genetic system, and in mammals mostly by the availability of sophisticated protein probe systems to unravel biochemical steps in the pathway. A reasonable amount of knowledge concerning the pathway in plant cells has accumulated via studies on the retention ofER luminal proteins, and the sorting and transport of soluble vacuolar and storage proteins 113

(Chrispeels, 1991; Bednarek and Raikhel, 1992; Satiat-Jeunemaitre and Hawes, 1993).

However, work on the retention/sorting of plant membrane proteins in the pathway has been greatly impeded by a scarcity of suitable protein probes, and a lack of efficient means to screen desirable mutants. To circumvent these deficiencies, expressing heterologous mcmbrane proteins within plant cells represents a feasible approach to dissect the molecular retention/sorting mechanisms govcrning the targeting of membrane proteins in the plant secretory pathway.

There is considerable interest in the mechanisms that control sorting of membrane proteins in the endomembrane-secretory pathway (Hopkins, 1992; Luzio and Banting,

1993; Pelham and Munro, 1993; Nothwehr and Stevens, 1994). Protein-protein interactions are clearly important. At least two cis-acting elements are thought to be involved in protein sorting. The first comprises specific signals contained in the primary structure of the proteins. These signal sequences interact with cytoplasmic or, perhaps, luminal proteins, and thereby permit specific targeting onwards through the endomembrane-secretory pathway, retention within a specific compartment, or retrograde recycling, such as the targeting of mannose-6-phosphate receptors to lysosomes

(Kornfeld, 1992). The second element involves alterations to the tertiary or quaternary structure of the proteins, resulting in the production of oligomers or aggregates. The molecular size of the unit experiencing sorting is altered, and this can affect the compartment and mechanisms available and accessible to this unit. For example, oligomerization of various Golgi-resident enzymes is required for their retention in the 114

Golgi (Nilsson et ai., 1993 and 1994) as is also true for the chimaeric VSV -IBV M

molecule (Weisz et ai., 1993). Obviously, our understanding on the retention and sorting

of membrane proteins requires exploration of the cis-elements within the proteins, before

proceeding to the study of the Irans-acting factors involved in retention and sorting.

Several membrane proteins cloned from plant sources have been actually reported or implied to be transported to specific destinations in the endomembrane- secretory pathway, these include a KDEL receptor homologue, a 3-hydroxy-3-methyglutaryl­

CoA reductase and a calnexin homologue in the ER membrane (Lee et ai., 1993; Enjuto et ai., 1994; Huang et aI., 1993), the tonoplast intrinsic proteins (TIPs) (Hofte el al.,

1991), and the plasma membrane I-t-ATPase and K+ -channel proteins (Pardo and

Serrano, 1989; Schachtman et ai., 1992). The KDEL receptor homologue from

Arabidopsis can functionally complement yeast mutants that are deficient in the retention of luminal ER proteins, but details regarding the retention signal and machinery in the plant ER remain unclear (Lee et aI., 1993). TIPs were further investigated with respect to retention/sorting within the plant secretory pathway (Hofte and Chrispeels, 1992). The sorting signal of a.-TIP was proposed to reside within the transmembrane domains, acting to divert a.-TIP from a default pathway to the plasma membrane. However, the studies could not exclude the possibility that a.-TIP may be transported to vacuoles by a default pathway (Hofte and Chrispeels, 1992; Gomez and Chrispeels, 1993).

In this project, a heterologous probe approach was adopted to investigate the cis-elements and possible Irans-acting factors involved in the retention/sorting of 115

membrane proteins in plant cells. Several heterologous proteins and their derived

recombinant proteins were engineered, and their expression evaluated in plant cells in

terms of their utility as targeted membrane protein markers. The proteins were selected

for study based on their predicted destinations in the plant endomembrane-secretory

pathway. Sonchus yellow net viral glycoprotein (SYNVG) was predicted to be targeted

to the inner nuclear membrane, coronavirus avian infectious bronchitis virus M

glycoprotein (IBV M) and its derivatives to the Golgi apparatus, and vesicular stomatitis

viral glycoprotein (VSVG) to the plasma membrane. These protein probes cover the

major membrane compartments in the main branch of the secretory pathway, and this

type of coordinated attack employing a variety of different elements was thought to

potentially provide a broad view on the operation of the retention and sorting of

membrane proteins within the plant secretory pathway.

I. SYNVG. This protein is the major envelope component found in SYNV, a minus-sense RNA plant rhabdovirus (Jackson, 1978). After synthesis in virally-infected cells, SYNVG is transported within the plant endomembrane-secretory pathway, as shown by the observation that infection in the presence oftunicamycin results in a faster mobility ofSYNVG in gel electrophoresis through an inhibition of the in vivo addition of

N-linked oligosaccharide side chain (Jones and Jackson, 1990). Maturation and budding of enveloped SYNV particles occurs at the inner nuclear membrane (Christie et aI., 1974;

Jackson et aI., 1987). Assembly of viral particles is also interrupted by the presence of tunicamycin, and this results in the accumulation of non-enveloped protein cores of the 116

viral particles at the nucleoplasmic side of the inner membrane (van Beek et aI., 1985).

The deduced amino acid sequence indicates that SYNVG contains a cleavable signal

peptide, a large luminal N-terminus with several potential glycosylation sites, a

transmembrane domain and a small cytoplasmic C-terminal tail (Goldberg et aI., 1991).

Motifs within the primary sequence include a conserved Lys-Lys-Lys-Arg sequence, a

typical motif involved in the nuclear localization of soluble proteins within the

C-terminal region next to the transmembrane domain (Goldberg et aI., 1991).

Surprisingly, in the middle of this cytoplasmic region is also found a Lys-Lys-X-X (X = any amino acid) motif, which is reported to be the consensus ER retention signal for several mammalian and viral proteins (Jackson et aI., 1990).

The fact that SYNVG is synthesized at the ER membrane implies that, for viral assembly to occur, it must then be targeted to the inner nuclear membrane (Jones and

Jackson, 1990; Jackson et aI., 1987). Nothing is known about the traffic of membrane proteins from the ER to the inner nuclear membrane in any system, with the possible recent exception of reports on the lamin B receptor (Smith and Blobel, 1993; Soullam and Worman, 1993; Ye and Worman, 1994). SYNVG may use one of the three following possible routes to reach the inner nuclear membrane. It may move into the inner membrane by diffusion within the plane of the nuclear membranes through the outer periphery of the nuclear pore complex. In this model, its traveling from the site of synthesis to the nuclear pores may involve the nuclear localization machinery conventionally used with soluble proteins. Another possible pathway to the inner I'

117

nuclear membrane would be similar to the translocation of chloroplast proteins encoded

by the nuclear genome (Keegstra and Olsen, 1989). In this model, the inner and the

outer nuclear membranes would become appressed under the mediation of receptor-like

protein(s) for SYNVG. A third possible pathway for SYNVG traffic could be dependent

on the cell cycle. Thus, SYNVG would be retained in the ER by its putative retention

signal (Lys-Lys-X-X), then transported to the inner membrane during the postmitotic

assembly of the nuclear envelope. This pathway is lent credence by the observation that

the nuclear envelope is assembled from a distinct ER-derived subpopulation of vesicles

(Wilson and Newport, 1988; Vigers and Lohka, 1991).

The intriguing life cycle ofSYNV provides further impetus to the study of the

targeting of its individual proteins. This life cycle involves infection and replication in

both plants (Sonchus oleraceus) and insects (Aphis cOl'eopsidis) (Christie et al., 1974;

Jackson and Christie, 1979). In plants, the viruses mature at the inner nuclear membrane,

whereas in insects they mature in the cytoplasm (Christie et aL, 1974; Jackson and

Christie, 1979). This implies a difference in the protein sorting/retention mechanisms in

plant and insect cells, most likely in the recognition and targeting of the SYNV protein

components, this difference enabling sorting to radically different subcellular locations

within the different host cells. Uncovering the possible trafficking route of SYNVG will

certainly extend our understanding beyond the current traffic routes of membrane proteins

(ER-Golgi-Plasma membrane/vacuole/ lysosome) in the secretory pathway, as well as on

the morphogenesis of SYNV and the possible amelioration of this viral disease. 118

2. IBV M glycoprotein (formerly termed El). This glycoprotein has been

successfully used as a probe to study protein retention mechanisms in the mammalian

Golgi. IBV M is one of the major viral envelope proteins. It is restricted to the

endomembrane system of infected cells, only reaching the cell surface as a component of

the budded virion at maturation (Sturman and Holmes, 1983; Armstrong et aI., 1984;

Rottier et aI., 1985, 1986). IBV M glycoprotein is an integral membrane protein with a

molecular weight of25 to 32 KDa, depending on the status of its N-Iinked glycosyJation

(Rottier et aI., 1985, 1986). It consists of a short luminal N-terminus with two putative

N-Iinked glycosylation sites, three transmembrane domains, and a long cytoplasmic

C-terminus (Armstrong et aI., 1984; Rottier et aI., 1985, 1986). In virally-infected or transfected mammalian cells, the first transmembrane (m 1) domain has been shown to act as a signal patch for ER translocation and as a retention signal for accumulation in the cis cisternae of the Golgi (Rottier et aI., 1985; Machamer and Rose, 1987; Machamer et aI.,

1990; Swift and Machamer, 1991). A recombinant membrane protein (Gm 1), constructed by swapping the VSVG transmembrane domain with the ml domain, IS retained in the cis-Golgi (Swift and Machamer, 1991). Oligomerization of Gm 1 is postulated to present an important molecular mechanism underlying the retention of

Golgi-resident proteins in mammalian cells (Weisz et aI., 1993).

Expressing IBV M and its derivatives within plant cells was intended to determine whether it might provide a unique protein marker for the plant Golgi apparatus. As well as expressing full-length IBVM in plant cells, the expression of a series of additional I,

119

chimaeric constructions was also evaluated. These constructions encode translational

fusions (IBV M-GUS and IBV mlL\-GUS), in which the whole or partial IBV M

glycoprotein sequence forms the N-terminal, and p-glucuronidase (GUS) the C-terminal

portion of the fusion proteins. GUS is the protein encoded by the uidA gene of

Escherichia coli (Jefferson, 1985; Jefferson et al., 1986). Native GUS enzyme is

synthesized and retained in the bacterial cytosol. The active enzyme is a homotetramer

with an apparent monomeric molecular weight of 73 KDa (Jefferson, 1985; Jefferson et

aI., 1986). GUS has been widely exploited as an expression marker for the functional

analysis of eukaryotic promoter regions. It is well expressed in plant cells, and its facile

and sensitive assay is not complicated by endogenous activity within non-transgenic

cells (Jefferson et al., 1987; Harkins et al., 1990). As GUS activity is strictly localized

in the cytosol of plant cells, GUS enzymes have also been employed to a limited extent

in probing the retention/sorting motifs of plant secretory proteins (Farrell and Beachy,

1990).

3. VSVG. This is the envelope spike glycoprotein ofVSV, an animal rhabdovirus

in structure similar to SYNV. The primary sequence ofVSVG contains a signal peptide,

a large luminal N-terminus with two N-linked glycosylation sites, a transmembrane

domain, and a short cytoplasmic C-terminal domain (Rose and Gallione, 1981). In

animal cells, the glycoprotein is synthesized at the ERlcytosol interface, and is inserted

and anchored in the rough ER membrane by the signal peptide and the transmembrane

domain, respectively. Two high-mannose side chains are co-translationally attached to I'

120

Asn-X-Thr/Ser sites of VSVG in the ER, then converted into complex glycans in the ER

and Golgi sequentially. The membrane-bound VSVG moves through the Golgi to the

plasma membrane, where the viral particles assemble, mature, and are released by

budding (Katz et a1., 1977). VSVG has been extensively studied in terms of its plasma

membrane targeting (Bergmann et a1., 1981; Rose and Bergmann, 1982; Guan and Rose,

1984; Machamer et a1., 1985; Machamer and Rose, 1988), which has been later found to

be a signal-independent process (Kelly, 1985; Burgess and Kelly, 1987; Pryer et a1.,

1992). The extensive knowledge gathered from VSVG has been used as a paradigm for

work on the targeting and transport of other mammalian membrane proteins. The VSVG

system has also been transplanted from mammalian into plant cells (Galbraith et a1.,

1992). This work indicated that VSVG is transported through the plant endomembrane­

secretory system as described in the mammalian cells, but two isoforms differing by

approximate 9 KDa are detected in transfected protoplasts (Galbraith et a1., 1992). The

larger isoform co-localizes predominantly with the plasma membrane, to a lesser extent,

with the ER and Golgi (Galbraith et a1., 1992). The smaller isoform is associated with

membrane fractions of low isopycnic densities, which are not identical to those of the ER

membrane, and are more similar to those of the tonoplast (Galbraith et a1., 1992). The

results suggest that the larger isoform is transported from the ER to the plasma membrane

via the Golgi (Galbraith et a1., 1992). However, it is not clear whether the smaller

isoform is internalized from the cell surface after the larger one has been transported to,

and proteolytically cleaved at, the cell surface, or whether the smaller one is derived from 121

the proteolysis of a subset of the larger one that does not reach the plasma membrane. As

the VSVG system has been already established in the plant cell system (Galbraith et ai.,

1992), it was also intended to provide as a reference system in this project, particularly

for the analyses of Gm 1 targeting.

This chapter documents the development of a series of heterologous membrane protein systems for the plant secretory pathway, and describes the establishment and the preliminary analyses of the individual proteins. This aimed to provide step-stones towards investigating the function and organization of the specific compartments in the plant secretory pathway. Most importantly, the systems will serve as workbenches to probe the molecular mechanisms regarding membrane protein retention and sOlting within the plant secretory pathway. The resulting knowledge could also be compared with those achieved in animal cell system to view the evolutionary relationship of protein trafficking between plant and animal cells.

MATERIALS AND METHODS

Plant materials: Tobacco (Nicotiana tabacum L. cv Xanthi) plantlets in axenic culture, either transgenic or non-transgenic, were used for analyses. Their generation and maintenance are described in Chapter 2.

DNA manipulation and Bene c\oninB: All manipulation of DNA was conducted using the protocols as described by Sambrook et al. (1989) unless indicated otherwise. "

122

1. Construction ofthe expression cassettes for IE V M glycoproteins and its

derivatives (lEV M-GUS, Gm1, lEV m1 !J.-GUS). The expression cassette for IBV M was

constructed by excising the complete coding sequence from pSV/IBV EI (Machamer and

Rose, 1987) using Xho I, and ligating it into the Sal I site of modified pJIT 117.

Modified pJIT 117 was constructed from pJIT 117 (Guerineau et aI., 1988) by end­

blunting and self-ligating after excision ofthe chloroplast transit peptide sequence using

Hind III and Sph I. The resultant plasmid (pFG 8) provides regulated transcription for

inserted coding sequences by a doubled Cauliflower Mosaic Virus 35S promoter (CaMV

35S-35S), and by the 3' non-translated polyadenylation signal (Poly N terminator) of

the same gene (Figure 3.1). The construction of the IBV M-GUS fusion protein was

done via several intermediate steps. First, the entire M coding region was amplified

using the polymerase chain reaction (PCR) with pSV/IBV E1 as the template (Machamer

and Rose, 1987), and two synthetic primers (5'-primer: 5'-GCG CGT CGA CCG ACC

ATG TCC AAG GAG ACA AAT-3', 3'-primer: 5'-GGC CCC CAT GOT GTA AAG

ACT ACT TCC-3'). The 5'-primer provides a unique Sal I site upstream of the

translational start site. The 3'-primer has the effect of eliminating the stop codon from the

open reading frame ofM protein, and also inserts a unique Nco I site. Second, a 1.8 kb

Hind III/EcoR I fragment from pRAJ 275 containing the complete GUS coding sequence

(Clonetech, CA, USA) was ligated with the larger segment of Hind III/EcoR I cut, I,

123

Figure 3. 1. Schematic maps of plasmid constructions for the transient gene

expression cassettes. Circled G symbols indicate the consensus N-linked glycosylation

sites. Abbreviations "ml ", "m2", "m3", "TM" and "tm" represent the transmembrane

domains ofIBV M, VSVG and SYNVG respectively. VSVG-TAl, VSVG-TA2 and

VSVG-TAl,2 are the VSVG mutants lacking the first, the second, or both N-linked

glycosylation sites. Hind III, EcoR I, Xho I, Sal I, Nco I, BamH I, Kpn I, Sma I, Sst I,

and Pst I are the restriction digestion sites of the appropriate restriction enzymes.

"AmpR" designates the p-Iactamase gene responsible for bacterial resistance to

ampicillin. Designation "ORl" with arrow shows the orientation of the origin of

replication. The maps are only schematic, and are not proportional to the actual sizes of

the various inserts and vectors. 124

..... fuM GUS

IBVM hill~·'II~NJI~~~lIlIlr_' ______~

,." IBVM-GUS 50' IBV mIA-GUS

hi ~I GmI I,: .. ,.... "."',.\Ii, :VSVG' ,,,·1.,1,·,·· I ------/

a.tI I(h I a.H VIto I VSVG F~,\Ii VSVG' 1"'1· I ------,,.----{ VSVG-TAI VSVG-TA2

SYNVG

IIIncII)----- 1'111 SIll BuIll !mil ("''' Jpli 1111>1 I'

125

gene-cleaned pJIT 117 (Guerineau et ai., 1988). The resultant plasmid (pBAP 15) has

unique Hind III, Sal I and Nco I sites between the CaMV 35S-35S promoter and the GUS

coding region, and provides the poly A + terminator downstream from the GUS coding

sequence. Third, a polylinker (Hind III-SuI I-Xho I) from pBluScript SK(-) was excised

using Hind III and Xho I, then inserted into Hind III/Sal I cut, ClAP-treated pBAP 15 to

create an intermediate plasmid (pFG 10). This has the effect of separating the Sal I and

Nco I sites in pBAP 15, as these sites overlap and can't be used simultaneously for

cloning. Finally, the PCR product of the IBV M coding sequence was digested with Sal I

and Nco I, and inserted into Sal IlNco I cut, CIAP- treated pFG 10. The resultant

construct (pFG II A) displays a perfect translational fusion at the junctions of IBV M and

GUS; overall expression is regulated by the CaMV 35S-35S promoter and the Poly N

terminator (Figure 3. 1).

The IBV ml.1-GUS expression cassette (pFG 20) was constructed using a strategy

identical to that described for pFG l1A, but with a different 3'-primer for PCR (5'-GGC

CCC CAT GGT ATA AAT AAC CTT ACT TCT-3'). This primer amplifies the DNA

sequence for the first 49 amino acids in the N-terminal portion ofIBV M containing the

luminal N-terminus, the m 1 domain and the partial cytoplasmic loop between the m 1 and

the second transmembrane domains (Figure 3.1). The cloning ofGml into the plant

transient expression vector was straight-forward (pFG 21; Figure 3. 1). A 1.67 kb Xho I

fragment was excised from pT7/Gm 1(Swift and Machamer, 1991), inserted into the Sal I

site of the modified pJIT 117 (Guerineau et ai., 1988), and clones with correctly oriented 126

inserts were identified. For the conditional expression of the IBV M-GUS chimaeric

protein, the CaMV 35S-35S promoter ofpFG IIA was replaced with a soybean hsp6871

promoter (Baumann et aI, 1987). The complete expression cassette was inserted into

pBIN 19, giving the resultant plasmid pFG 13 (Table 3. 1).

2. Construction ofthe expression cassette for SYNVG: It is thought that the native sequence upstream of the coding sequence of SYNVG can be recognized by the transcriptional and translational machinery of E. coli, and that expression of SYNVG in bacteria is toxic (Goldberg-Scholthof, personal communication, 1992). In order to bypass this problem during gene cloning, a 1.97 kb fragment lacking the leader sequence was

PCR amplified using pGT 573 as template (Goldberg et a!., 1991) and two primers

(5'-primer: 5'-CCC TCG AGA TGT CTC ATA TAA TGA ACC TT-3' and 3'-primer:

5'-GGG GTA CCT TCA GAT GTC GTT CAG AA-3'). The amplified fragment contains an Xho I site prior to the initiation codon, and a Kpn I site after the stop codon. The fragment was digested using Xho I and Kpn I, and inserted into the 3.75 kb Xho I / Kpn I cut pDG 711 transient expression vector (Galbraith et a!., 1992; see Figure 3. 1).

Expression of the resultant plasmid (pFG 15, Figure 3. 1) is under the control of a single CaMV 35S promoter and a nopaline-synthetase (nos) terminator.

3. Construction of the expression cassettes for VSVG and its mutants: The plasm ids for the wild-type and mutant VSVG (pDG 71, pDG 712, pDG 715, and pDG

716) were constructed as described (Galbraith et a!. 1992; Figure 3. 1). They are under the control of the single CaMV 35S promoter and the nos terminator. A plasmid (pFG 5) 127

was also assembled for the wild-type VSVG under the c'ontrol of the CaMV 35S-35S

promoter and Poly N terminator (Figure 3.1). A 1.6 kb Bam HI fragment was excised

from pDG 71, and inserted into the Bam HI site within the modified pJIT 117. Clones

with correctly-oriented inserts were identified by restriction mapping.

Protoplast isolation and trans feet jon: Protoplasts were prepared from leaf tissues

of axenic ally cultured tobacco plantlets, and were transfected as detailed under

"Protoplast Isolation and Transfection" in Chapter 2. Tobacco pro top lasts were transfected with plasmid DNA containing the various transient expression cassettes (pFG

5, pFG 8, pFG l1A, pFG 15, pFG 20, pFG 21, pBAP 15, pFG 71, pDG 712, pDG 715, and pDG 716, see Table 3.1) using sonicated thymus calf DNA as carrier, in the presence of polyethylene glycol 4000, and were cultured at 25°C in darkness prior to analysis. Controls were transfected with carrier DNA but without plasmid DNA under the same conditions.

Transformation of tobacco and selection of transgenic plants: Details are described under "Agrobacterium Transformation of Tobacco Leaf-discs" in Chapter 2.

To mobilize the recombinant genes into N. tabacum cv. Xanthi, the Pvu II, Kpn I, Hind

III/EcoR I or Sst IlXho I fragments, containing the complete expression cassettes from pFG 8, pFG l1A, pFG 15, pFG 20, pFG 21, pDG 712, pDG 715, and pDG 716, were ligated with the corresponding enzyme-cut, ClAP-treated pBIN 19. The resultant 128

Table 3.1. Summary of the DNA plasmids employed for protoplast transfection and

leaf-disc transformation of tobacco.

Transient Binary Vectors

Gene Constructions Expression for

Cassettes Transformation

CaM V 3SS-35S-GUS-CaM V Poly N pBAP IS A not available

CaM V 3SS-GUS-nos terminator pBI 2211J pBI 1211J

CaM V 3SS-3SS-IBV M-CaM V Poly N pFG 8 pFG 19

CaM V 3SS-3SS-IBV M-GUS-CaM V Poly N pFG IIA pFG 14

hsp6871-IBV M-GUS-CaM V Poly N not available pFG 13

CaM V 3SS-3SS-IBV ml il-GUS-CaM V Poly N pFG 20 pFG22

CaM V 3SS-3SS-Gm I-CaM V Poly N pFG 21 pFG23

CaM V 3SS-3SS-VSVG-CaM V Poly N pFG S not available

CaM V 3SS-VSVG-nos terminator pDG 71 c pDG 714D

CaM V 3SS-VSVG-TAl-nos terminator pDG 71Sc pFG2

CaM V 3SS-VSVG-TA2-nos terminator pDG 716c pFG 3

CaM V 3SS-VSVG-TAl,2-nos terminator pDG 712c pFG 1

CaM V 3SS-SYNVG-nos terminator pFG IS pFG 18

Note A Pierson et aI., 1994

IJ Harkins et aI., 1990 c Galbraith et aI., 1992

D constructed by Dr. Carolyn A. Zeiher I,

129

recombinants are pFG 1, pFG 2, pFG 3, pFG 14, pFG 18, pFG 19, pFG 22, and pFG 23

(Table 3. 1), and were transferred into Agrobacterillm llI111efaciens LBA 4404 via

triparental mating (Horsch and Klee, 1986). Xanthi leaf-discs were inoculated with the

agrobacteria containing the various pBlN based plasmids, and were cultured under the

kanamycin selection (1 00 ~lg/mL). Shoots were regenerated and plants were maintained

in the presence of kanamycin (100 f.lg/mL). Independently transformed plants were

screened for the expression of the gene products (see "Gel Electrophoresis and Western

Blotting" and "p-Glucuronidase Assay" in Chapter 2).

Sample homogenization and differential centriful.!ation: Viable transfected

protoplasts were purified by sucrose flotation, washed with W5 medium and

pelleted as previously described (Harkins el al., 1990, also see "Protoplast Isolation and

Transfection" in Chapter 2). They were then resuspended in ice-cold 1 X GUS extraction

buffer as described (Jefferson et al., 1987, see "p-Glucuronidase Assay" in Chapter 2), or

resuspended in 2 X SDS-sample buffer for gel electrophoresis (see "Gel Electrophoresis

and Western Blotting in Chapter 2), or resuspended in ice-cold HBST medium (Meyer et

ai., 1988; 50 mM Tris, 1 mM EDTA, 300 mM sucrose, 1 mM PMSF, 5 mM ascorbic

acid, 3.6 mM cysteine, and 100 f.lg/mL BHT, pH 8.0) for subcellular fractionation. For

subcellular fractionation, protoplasts were lysed in microfuge tubes by several passages

through a 26 gauge, one-half inch hypodermic syringe needle. Leaf samples of transgenic

and non-transgenic plants were homogenized in either 1 X extraction buffer or HBST 130

medium for various assays. All procedures were conducted at 4 oC unless analyses

required different temperatures.

Subcellular fractionation was conducted through differential centrifugation as

described by Galbraith et al. (1992). Lysed protoplasts in HBST medium were sUbjected to two rounds of centrifugation. The first was at 1000 g for 20 minutes to pellet nuclei and heavy debris, designated as nuclei. The second round was conducted in a Beckman

TL-IOO Ultrafuge at 300,000g for 60 minutes to separate membranes from the cytosolic fraction, designated respectively as microsomes and cytoplasm. For analyses of transgenic plants, leaf homogenates in HBST medium were also separated into supernatant and microsomal fractions as described for protoplasts, but employing

13,SOOg for the first centrifugational step.

Isolation and sortin~ of nuclei: Crude nuclei, isolated from transfected or transgenic protoplasts as described above, were resuspended in HBST medium. They were filtered through a layer of nylon mesh (size 15 /lm). A chopping procedure, developed by Galbraith et al. (1983), was also used. Leaves of transgenic plants were chopped on a cold ceramic tile (kept on ice) using a razor blade in the presence of lysis

buffer (45 mM MgCl2 ,30 mM Na citrate, 20 mM MOPS, and 1 mg/mL Triton X-IOO, pH 7.0). The chopped lysate was filtered through nylon mesh (size 15 /lm). The filtrate, containing crude nuclei, was incubated for 10 minutes in the presence of propidium iodide (final concentration, 100 mg/L). The nuclei were subjected to flow analysis and sorting using a Coulter Elite flow cytometer as described by Fox and Galbraith (1990). 131

The sorted nuclei were pelleted by centrifugation at 1000 g for 5 minutes in a tabletop

centrifuge, and were resuspended in a suitable amount of2X SDS-sample buffer for gel

electrophoresis.

Detection methods: Expression of the various gene constructs was evaluated by

enzyme assay, gel electrophoresis, and Western blotting (see "p-Glucuronidase Assay"

and "Gel Electrophoresis and Western Blotting" in Chapter 2). Table 3. 2 lists primary

antibodies, conjugated secondary antibodies, and their condition for usage in Western

blotting.

Before use, affinity preabsorption was conducted for the various primary antibodies, including rabbit anti GUS polyclonal Ig G (Molecular Probes, CA, USA), rabbit anti IBV EI polyclonal antibodies (provided by Dr. Machamer), rabbit anti VSV serum (provided by Dr. Machamer), and mouse anti SYNV ascites (provided by Dr.

Goldberg-Scholthof). First, 107 non-transgenic tobacco leaf protoplasts were resuspended in 5 ml of a solution comprising 0.2 M NaCI, 60 mM Tris, 1% Triton Xl 00,

2 mM EDTA, and 1 mM PMSF (PH 6.8), and were subjected to one round of freeze-thawing followed by incubation at 4 °C for 2-3 hours. The lysate was centrifuged at 300,000 g for 1 hour at 4 °C to remove insoluble material. The supernatant was collected, exchanged to a buffer of 0.2 M NaCI and 0.1 M MOPS (pH 7.5) using an

Amicon Centricon PIO concentrator (Boston, MA, USA), and protein content was measured (Peterson, 1977, see "Measurement of Protein Content" in Chapter 2). Second, preparation and coupling of gel support matrix (BioRad Affigel 10) was done according 132

Table 3. 2. Primary antibodies, conjugated secondary antibodies and their usage in

Western blotting.

Alkaline Phosphatase Antigens Primary Antibodies Conjugated Polyclonal Ig G (dilution and suppliers) (dilution and suppliers)

GUS polyclonal rabbit anti-GUS Ig G goat anti-rabbit Ig G (H + L) (1: 1,000; Molecular Probe) (1: 10,000; Sigma)

IBVM polyclonal rabbit anti-El antibodies goat anti-rabbit Ig G (H + L)

(l: 1,000; Machame~) (1: 10,000; Sigma)

IBV M-GUS polyclonal rabbit anti-GUS Ig G goat anti-rabbit Ig G (H + L) (1: 1,000; Molecular Probe) (1: 10,000; Sigma)

IBV ml ~-GUS polyclonal rabbit anti-GUS Ig G goat anti-rabbit Ig G (H + L) (1: 1,000; Molecular Probe) (1: 10,000; Sigma)

Gm 1 polyclonal rabbit anti-VSV serum goat anti-rabbit Ig G (H + L)

ll (l :2,500; Machamer ) (l: 10,000; Sigma)

VSVG & Mutants polyclonal rabbit anti-VSV serum goat anti-rabbit Ig G (H + L) (l : 1,600; Moore) (1 :2,000; Bio-Rad)

SYNVG mouse anti-SYNV ascites goat anti-mouse Ig G (H + L) (1 :2,500; Goldberg-Scholthof) (1 :2,000; Bio-Rad)

Note: A Machamer and Rose, 1987

Il Weisz et al., 1993 133 to the manufacturer's protocol for aqueous coupling. Residual active sites were blocked via incubation (22°C, 1 hour with gentle shaking) in the presence of an appropriate amount of glycine ethyl ester. The binding capacity of the column was calculated after the column was thoroughly washed with phosphate-buffered saline (pH 7.4). Third, purification of the anti IBV M, GUS, VSV, and SYNV antibodies was achieved by their passage three times over the affinity column. Non-bound antibodies were collected, and concentrated using a Centricon PI 0 to the original antibody volume prior to absorption.

Bound antibodies on the column were discarded during regeneration of the affinity column.

Densitometeric analyses of Western blots were performed for accurate determination of band intensity. The blots were digitized into TIF formatted files using a

256 grayscale either with a Stratascan™ 7000 Scanning System (Stratagene, La Jolla,

CA, USA) or a Star I Digital CCD Camera System (Photometries, Tucson, AZ, USA), and analyzed with Stratascan ™ 20 Densitometry Software.

RESULTS

EXPRESSION OF HETEROLOGOUS PROTEINS IN TRANSFECTED PROTOI'LASTS

To save labor and time, expression of the various foreign proteins in plants was first evaluated in transfected protoplasts, before undertaking the long and tedious process I'

134

of producing transgenic plants.

1. Expression oflEV M and its derivatives in protoplasts. The IBV M

glycoprotein was barely detected in transfected protoplasts (Figurc 3. 2), and positive

results required many attempts. The molecular size ofIBV M in the protoplasts was

found to be approximately 26 KDa, which is consistent with the predicted size of the

non-glycosylated form ofIBV M. It was found that the antibodies directed against the

IBV M glycoprotein also non-specifically cross-reacted with several protein bands even

after absorption using the affinity column (Figurc 3. 2, control). The difficulty that was

experienced in detecting IBV M using Western blotting prompted an alternative

approach. In this, IBV M was fused N-distally to GUS, in order to utilize both the

enzymatic activity of GUS and commercially-available anti-GUS antibodies for

monitoring the biosynthesis and targeting of the IBV M-GUS chimaeric protein inside

plant cells. A perfect translational fusion between IBV M and GUS (Figur'c 3. 1) was

achieved through several PCR and cloning steps described under "MATERIAL AND

METHODS" in this Chapter. It was found that GUS activity of the IBV M-GUS

chimaeric proteins was easily detected through enzymatic analysis oftransfected

protoplasts (Figurc 3.3). In the absence of the plasmid DNA (negative control), GUS

activities in transfected protoplasts were at baseline levels. With plasmid pFG 11 and

pBAP 15, GUS activities were significantly higher than that of the negative control

(Figurc 3. 3). However, comparison of the total GUS activities in the protoplasts 135

I -., -- --...... , -26KDa

Figure 3.2. Gel electrophoresis and Western blotting of the IBV M glycoprotein in transfected protoplasts. Viable transfected protoplasts, 24 hours after incubation in darkness, were purified by sucrose floatation, and were pelleted and l'esusj:'ended in 2X

SDS-sample buffer. Protein samples (equi valent to 15 x 104 protoplast!,) from both treatments of the control and the pFG 8 (IBV M) plasmids, were loaded onto a one-dimensional SDS-PAGE gel (4% stacking gel, 12.5% running gel). After protein transfer, preabsorbed polyclonal rabbit anti-IBV M antibodies were used at a dilution of

1: 1,000. Secondary antibody was goat anti-rabbit IgG (H+L) alkaline phosphatase conjugate. and was employed at a dilution of 1: 10,000. 136

13320 I >. .-:= .:= (j -<- 13280 ~ 1 00 ;::l ...... e,:, .-::: I -t:: e 13240/ -Q "";;:; Eo-... ~ 16 -!l.I - 1201 :- .!::! -a 80 l e1.0 Q 40 Z l 0 Control IBVM-GUS GUS

Figure 3. 3. Transient expression of the IBV M-GUS chimaeric protein in tobacco leaf protoplasts. Control: negative control (transfection without expression vector). GUS: positive control (transfection using GUS construct pBAP 15 containing the wild-type GUS coding sequence). Viable protoplasts were purified through sucrose floatation 20 hours after transfection, and were lysed in IX extraction buffer for GUS assay. Specific GUS activities of the samples (arbitrary fluorescence units) were determined in a 96-well plate using a Fluoroskan II fluorometer (excitation at 365 run, emission at 455 run). Five time points (0, 15,30,45,60 minutes), each with 2 replications, were sampled to provide kinetics data. Total GUS activities were normalized based on equal amounts of total protein in each treatment. 137

transfected with the chimaeric IBV M-OUS and the non-chimaeric (wild-type) GUS

constructions indicated that inclusion of the IBV M sequences considerably depressed

enzymatic activity (Figure 3. 3). IBV m I ~-OUS showed the same pattern of enzymatic activity as that ofIBV M-OUS, but the GUS enzymatic activity was depressed to a lesser degree than that seen for the IBV M-OUS chimaeric protein (Figure 3. 4).

Expression of Om I in transfected protoplasts was also easily detected using

Western blotting (Figure 3. 5). Interestingly, the Om 1 chimaeric protein showed two isoforms, as seen for expression ofVSVO in transfected protoplasts; these isoforms had apparent molecular sizes of61 and 52 KDa, respectively. Although the antibodies from

Dr. Machamer's laboratory were found to cross-react non-specifically with a protein band located in the region of 52 KDa (Figure 3.5, contra!), the band intensity of the Om I smaller isoform was sufficiently more conspicuous than that of the non-specific band in the control to affirm its existence (the control and the Om I treatments were equalized to the same number of viable transfected protoplasts prior to gel electrophoresis; Figure 3.

5). Quantitative intensity analyses of the 52 KDa bands in the control and the Om 1 treatments indicated that, following transfection with the Om I plasmid, about three times as much of the 52 KDa isoform is produced as compared to the non-specific protein band in the control (Figure 3.5). Further definitive evidence confirming the existence of the

Om I 52 KDa isoform was obtained in expression experiments in vivo using tunicamycin

(see Chapter 5). 138

14000 >. I :E- -C.I -< 13000 [ ("JJ ~ i 0 - I:: .~ 12000/: - ;:;; / -Q / e- ~ l -~ .:::! -- -I:: 1000 i:= ;z:Q 0 Control IBY ml.1.-GUS GUS

Figure 3.4. Transient expression of the IBV mIll-GUS chimaeric protein in tobacco leaf protoplasts. Control: negative control (transfection without expression vector plasmid DNA). GUS: positive control (transfection using GUS construct pBAP 15).

Viable protoplasts were purified through sucrose floatation 24 hours after transfection, and were lysed in IX extraction buffer for GUS assay. Specific GUS activities were determined as described in Figure 3. 3. I.

139

; , :lit', ~ 1 I '''''', , ".~.

- 61 KDa 1- 52 KDa

Figure 3. 5. Gel electrophoresis, Western blotting, and detection of expression of

the Gm 1 glycoprotein isofonns in transfected protoplasts. Viable transfected protoplasts,

14 hours after incubation in darkness, were purified by sucrose floatation, pelleted, and

resuspended in 1X SDS-sample buffer. Protein (equivalent to 15 x 104 protoplasts),

from both treatments of the control and the pFG 11 (Gm 1) plasmid, was loaded onto a

one-dimensional SDS-PAGE gel (4% stacking gel, 7.5% running gel). After protein

transfer, preabsorbed polyclonal rabbit anti-VSV antibodies were used at a dilution of

1:1,500. Secondary antibody, goat anti-rabbit IgG (H+L) alkaline phosphatase

conjugate, was used at a dilution of 1: 10,000. 140

2. Expression of VSVG and its mutant forms in transfected pl'otoplasts.

Expression ofV8VG and its mutants in transfected protoplasts has been previously documented by Galbraith et aI., (1992). Repeating their expression in this project was done in order to allow comparison between the strengths of the CaMV doubled 358 and single 35 8 promoters, and allow use ofV8VG as a control for the analyses of Gm 1 expression and targeting (Gm 1 and VSVG differ only in the transmembrane domain;

8wift and Machamer, 1991). Western blotting (Figure 3. 6) revealed that two isofornls are detectable in transfected protoplasts and their measured sizes correspond to those previous reported for V8VG (Galbraith et aI., 1992). Rabbit polyclonal anti-V8V antibody obtained from Dr. H.-P. H. Moore was particularly specific in Western blotting and probing, unlike the antibody we currently employ for Gm 1 and V8VG (Figure 3.

6). Unfortunately, that antibody is no longer available.

Promoter strength was assessed by densitometric analysis of band intensities. The combined intensities of the larger and smaller isoforms synthesized under the regulation of the CaMV 358-358 promoter was found to be about three times that observed for the

CaMV 358 promoter. Comparisons of the two isofornls within each treatment indicated that the larger isoform ofV8VG predominated for the single promoter, whereas the larger and the smaller isoforms had roughly equal intensities for the double promoter, when the incubation of the transfected protoplasts only lasted 18 hours (Figure 3. 6). This suggests that the biosynthesis and conversion of the larger isoform occurs more quickly when transcription occurs under the regulation of the double 358 promoter. 141

1 2 3

---61KDa ---52 KDa

Figure 3. 6. Western blotting of VSVG in transfected protoplasts. Viable transfected protoplasts, 18 hours after incubation in darkness, were purified via sucrose floatation, pelleted and resuspended in 2X SDS-sample buffer. Protein samples equivalent to 3.0 x lOs protoplasts, from the treatments of the control, VSVG plasmids with a single (pDG 71) and a double (pFG 5) CaM V 35S promoter (lanes 1,2, and 3, sequentially), were loaded onto a one-dimensional SDS-PAGE gel (4% stacking gel,

7.5% running gel). After protein transfer, polyclonal rabbit anti-VSV antibody (from Dr.

H.-P. H. Moore) was used at a dilution of 1: 1,600. Secondary antibody. goat anti-rabbit

IgG (H+L) alkaline phosphatase conjugate, was used at a dilution of 1:2,000. , ,

142

3. Expression ojSYNVG in tl'ansjected protoplasts. SYNVG was found to be

well-expressed in transfected protoplasts (Figure 3. 7). The detennined molecular mass

ofSYNVG was around 76 KDa, the band being of identical mobility to that of the native

glycoprotein solubilized from purified viral particles taken from the plant host (Figure 3.

7, SYNV). Native SYNVG, from viral particles isolated from the plant host, is

susceptible to digestion by endoglycosidase H (Endo H; Basel, personal communication,

1993). Endo H susceptibility is used as an index for the degree of conversion ofN-linked

oligosaccharides from high mannose (sensitive) to complex fonn (resistant). This

provides an indication of the passage ofN-linked glycoproteins through the medial

cisternae of the Golgi apparatus. The results imply that the glycoproteins from both the

viral particles and the transfected protoplasts are transported into the endomembrane­

secretory pathway, but not carried as far as to the medial Golgi cisternae. The antibodies

also cross-reacted with a non-specific band ofprotein(s) at about 68 KDa, even after

preabsorption using affinity column (Figure 3.7, control). This problem only occurs in

the samples of transfected protoplasts employed for gel electrophoresis, and is not seen

when using samples of transgenic plants. The cause is not clear; it may be due to

transfection- induced expression of non-specific protein(s), 01' due to the cytoplasmic

localization of non-specific protein(s), because the analyses of transgenic plants were, in

the most part, conducted on microsomal samples. I'

1~3

-= ~ -76KDa

Figure 3. 7. Gel electrophoresis. Western blotting, and detection of SYNVG in

transfected protoplasts. Viable transfected protoplasts, 15 hours after incubation in

darkness, were purified via sucrose floatation, and were peUeted and resuspended in 2X

SDS-sample buffer. Protein samples equivalent to 2.0 x lOs protoplasts, from both

treatments of the control and the SYNVG (pFG 15) plasmids (lanes 1 and 3,

respectively), were loaded onto a one-dimensional SDS-PAGE gel (4% stacking gel,

7.5% running gel). Lane 5 was loaded with the lysate of SYNV particles resuspended in

2X SDS sample buffer. After protein transfer, preabsorbed mouse anti-SYNV ascites

was used at a dilution of 1: 1,000. Secondary antibody, goat anti-mouse IgG (H+L)

alkaline phosphatase conjugate, was used at a dilution of 1:2,000. 144

EXPRESSION OF HETEROLOGOUS PROTEINS IN TRANSGENIC PLANTS

Transgenic tobacco plants expressing the recombinant proteins were screened

using either the fluorometric GUS assay (pFG 13, pFG 14, and pFG 22) or Western

blotting (pFG 1, pFG 2, pFG 3, pFG 18, pFG 19, and pFG 23). The results are detailed

in Table 3.3. With the exception of the transgenic plants containing the pFG 19 cassette,

all transgenic plants were found to express the respective proteins. Failure to detect IBV

M in transgenic plants containing pFG 19 cassette may be due to rapid protein

degradation after synthesis, as IBV M was barely detectable in the transient expression

system (Figure 3. 2). Alternatively, it may be due to transcriptional problems.

When screening of transgenic plants was based on the GUS assay, a high

percentage of positive plants were identified within the transformed population, reaching

90% (Table 3.3). When Western blotting was used as a screening method,

positively-identified plants comprised about 50% of the transforn1ed popUlation (Table

3.3).

1. Expression oflEV-GUS and lEV ml /::,.-GUS in transgenic plants. Western blotting (Figure 3.8) indicated that the IBV M-GUS chimaeric protein was readily detectable in transgenic plants. It has a molecular mass of about 95 kilodaltons, as would be predicted from the size of the chimaeric coding sequences in the recombinant construct. Interestingly, two bands of the chimaeric protein were observed in the extracts 145

Table 3. 3. Numbers of identified positive transformants and their percentages within

the screened population.

Binary Vectors for Total Transformants Positive Percentage (%) of

Transformation Screened Transformants Positive Plants

pFG 19 23 NA NA

pFG 14 19 17 (IS) 89.5

pFG 13 17 13 (12) 76.5

pFG22 20 14 (10) 70.0

pFG23 11 6 (4) 54.5

pFG 1 3 1 33.3

pFG 2 3 2 66.7

pFG 3 2 1 50.0

pFG 18 24 11 (5) 45.8

Note: The numbers in parentheses indicate the plants that highly express the

engineered genes. 146

Figure 3.8. Western blotting of the GUS. IBV M-GUS, and IBV mIll-GUS proteins synthesized within transgenic plants. Leaf microsomes were isolated from the non-transgenic plant (control), IBV M-GUS transgenic plants with expression regulated by the heat-shock promoter (pFG 13-8) and by the CaM V 35S-35S promoter (pFG

14-23). an IBV mIll-GUS transgenic plant with expression regulated by the CaM V

35S-35S promoter (pFG 22-41), and transgenic plant expressing native GUS (PBI 121).

Proteins were extracted from microsomes using IX extraction buffer for GUS assay. The native GUS from the pBI 121 plant was concentrated from the supernatant of the leaf homogenate using an Amicon Centricon P-30. Proteins (100 iJ.g) from each sample were subjected to SDS-PAGE gel electrophoresis (4% stacking gel, 7.5% running gel).

Antibody probing was done using preabsorbed polyclonal rabbit anti-GUS Ig G (I: 1,000) and goat anti-rabbit IgG(H.,-L) alkaline phosphatase conjugate (1: I 0,000). 147 from the transgenic plant containing a CaMV doubled 35S promoter, whereas only one band, oflesser intensity, was seen in the extracts from the transgenic plant in which the

IBV M-GUS chimaeric coding sequence was under the transcriptional control of a heat shock promoter (Figure 3. 8). Heat treatment of leaves was conducted at 42 oC for 2 hours in a water bath, followed by incubation at 22 oC for 20 hours prior to homogenization of leaves. Two bands of the IBV ml~-GUS chimaeric protein were observed in the transgenic plants following Western blotting. The apparent molecular size of the larger band is around 77 KDa, consistent with the DNA sequence of the chimaeric gene. The smaller bands ofIBV M-GUS and mv ml~-GUS possibly represent proteolytic products.

2. Expression ofGm 1 in transgenic plants. Transgenic plants expressing the Gm

1 chimaeric protein showed a pattern similar to that observed in transfected protoplasts

(Figure 3. 9 and Figure 3. 5). The molecular sizes of the two isoforms were estimated as 61 and 52 KDa. The smaller isoform showed a lesser intensity, but a more diffuse pattern, than the larger isofonn (Figure 3. 9).

3. Expression ofVSVG and its mlltants in transgenic plants. Transgenic plants expressing VSVG and its mutants (pFG 1, pFG 2 ,and pFG 3) displayed two isofOlms

(Figure 3. 10). The shifts in mobility on the SDS polyacrylamide gel reflect their

N-linked glycosylation status, confirming the report of Galbraith et al. (1992). In these, as also observed by Galbraith et al. (1992), the band intensities of the smaller isoform 148

61 KDa 52KDa

Figure 3.9. Western blotting of the Gm I protein synthesized in transgenic plants. Microsomes were isolated (as described in "MATERlALS AND METHODS") from the leaves of a non-transgenic plant (control) and a Gm 1 transgenic plant with CaM

V 35S-35S promoter (pFG 23-7). Proteins were extracted by resuspending microsomes in IX extraction buffer for GUS assay. Protein (200 Ilg) from each sample was loaded onto a one~dimensional SDS-PAGE gel (4% stacking gel, 7.5% running gel). After protein separation and transfer, preabsorbed polyclonal rabbit anti-VSV serum was used at a dilution of I :2.500. Secondary antibody, goat anti-rabbit IgG (H+L) alkaline phosphatase conjugate, was used at a dilution of I: 10,000. 149

61 KDa---.. 52 KDa--'"

Figure 3.10. Expression ofVSVG and its mutants within transgenic plants.

Microsomes were isolated (as described in "MATERIALS AND METHODS" ) from the leaves ofa non-transgenic plant (control) and VSVG transgenic plants expressing VSVG,

VSVG-TA2, VSVG-TAI and VSVG-TAI,2. Proteins were extracted by resuspending microsomes in IX extraction buffer for GUS assay. Protein (200 flg) from each sample were loaded onto a one-dimensional SDS-PAGE gel (4% stacking gel, 7.5% running gel). Antibody probing was done using preabsorbed polycIonal rabbit anti-VSV serum

(1: 1,600; from Dr. H.-P. H. Moore) and goat anti-rabbit IgG (H+L) alkaline phosphatase conjugate (1 :2,000). 150 became more diffuse and less intense than the larger isoform, if the proteins were extracted from the microsomal fraction (Figure 3. 10).

4. Expression ojSYNVG in transgenic plants. SYNVG was strongly expressed in the transgenic plants containing pFG 18 (Figure 3. 11), having an apparent molecular size of76 KDa. This is the same as that of the native viral glycoprotein, extracted from the viral particles purified from infected plant hosts. No non-specific bands were observed on Western blotting for the transgenic plants (Figure 3.7). The large molecular aggregates seen in the SYNV lane are probably due to formation of SYNVG oligomers that are resistant to SDS; this behavior is commonly observed both for SYNVG and

VSVG (Goldberg-Scholthof, Machamer, personal communications).

SUBCELLULAR DISTRIBUTIONS OF THE HETEROLOGOUS PROTEINS

I. lEV M-GUS and lEV mll1-GUS. Most of the GUS activity of pro top lasts transfected with the mv M-GUS plasmid construct was located in the endomembrane fraction (about 70%), whereas 99% of the GUS activity was found in the cytosolic fraction of pro top lasts transfected with the wild-type (non-targeted) GUS plasmid construct (Figure 3. 12). The distribution pattern of mv mll1-GUS within plant cells was found to be similar to that of mv M-GUS (Figure 3. 12). A small portion of GUS activity was always detected in the cytoplasmic fractions in protoplasts expressing mv

M-GUS and mv mll1-GUS (ranging from 20% to 25% of total). This result may be 151

...... 0 ~ o""- ~

-76KDa

Figure 3. 11. Expression of Syp,TVG within transgenic plants. Microsomes were

isolated (described under "iviA TERIALS AND METHODS" in this chapter) from the

leaves of a non-transgenic plant (Control), and transgenic plants bearing pFG 18-23

(SYNVG). Samples (40 !lg protein) were loaded onto a one-dimensional SDS-PAGE gel (4% stacking gel, 7.5% running gel). Viral particles (S'J.'NV) isolated from infected plant hosts were also loaded on the gel as a standard. After protein transfer, preabsorbed mouse anti-SYNVG ascites was used at a dilution of 1: 1,000. Secondary antibody, goat anti-mouse IgG (H-'-L) alkaline phosphatase conjugate, was used at a dilution of 1:2,000. III I .S:! [ :E- ~ Cytosol []JID! Nuclei ~ Microsomes j CJ ~ -<- I 1 en 100 ;:J Ir I I I - I111 " 1111 " I '- r '"'Q r 0-~ r i -= so '- - .S! L -...- I 'i: r .:a- Q a ~ GUS mv M-GUS IBV mIll-GUS

Figure 3. 12. Distribution of GUS activities between the subcellular fractions isolated from transfected protoplasts expressing GUS, IBV M-GUS, and IBV ml~-GUS.

Viable transfected protoplasts were lysed in ice-cold HBST medium 20 hours after transfection. Differential centrifugation was conducted at 200 g for 20 minutes to pellet nuclei and heavy debris, designated as nuclei. The supernatant was subjected to further centrifugation at 300,000 g for 60 minutes to separate fractions designated as microsomes and cytosol. Protein concentrations and specific GUS activities were measured for each fraction. The percentage of GUS activity in each fraction was calculated as a proportion of total GUS activities in each treatment. 153

partly due to multiplication of the higher baseline noise of GUS assay with the larger

amount of protein in the cytoplasmic fractions (over 90% of total cellular protein), and

partly due to release of the chimaeric proteins from membranes during sample

preparation. The possibility that a small amount of the chimaeric protein might be

released in vivo from the membranes before sample processing could not be excluded.

2. VSVG and Gm 1. Since the subcellular distribution of VSVG has been demonstrated to be mainly in the endomembrane fractions (Galbraith et aI., 1992), VSVG and its mutants were not analyzed in this aspect. However, protoplasts expressing the closely related protein Gm I were subfractionated and analyzed (Figure 3. 13). The larger isoform of Gm 1 was predominantly present in the microsomal fraction, with a lesser amount in the nuclear/ heavy membrane fraction. The larger isoform was not found in the soluble cytoplasmic fraction (Figure 3. 13). The smaller isoform was observed in all three subcellular fractions (Figure 3. 13), suggesting that certain amount of the smaller isoform was released from membranes into cytoplasm. The major, non-specific protein band at 52 KDa, observed in control transfections, was mainly associated with the microsomal fraction. The association of Gm 1 with the nuclear fraction may be due to the fact that the outer nuclear membrane fomls a continuum with the rough ER.

3. SYNVG. Analyses of the subcellular localization ofSYNVG showed it to be located predominantly in the nuclear and microsomal fractions in transfected protoplasts

(Figure 3.14). As SYNVG would be expected to be ultimately targeted into the inner 154

,.... ,.... ::: - 2 -.-.t: ,...... -r- -r- - .... --...... ::: ::: 0 - c- - c ,.- u 0 u 0 u '-'

____ 61 KDa ----52 KDa

Figure 3. 13. Subcellular distribudon of the Gm 1 glycoprotein expressed in transfected protoplasts. Viable protoplasts, 24 hours after transfection, were lysed in ice-cold HBST medium. The lysate was centrifuged at 200 g for 20 minutes to pellet nuclei and heavy debris. designated as nuclei. The supernatant was further centrifuged at

300,000 g for 60 minutes to separate the fractions designated as microsomes and cytosol.

The cytosolic fraction was concentrated through an ..<\micon Centricon P-30. The nuclear, cytosolic. and microsomal fractions (equivalent to 4 x I OS protopla~ts, in

SDS-sample buffer.) from the control and the Gm I-expressing protoplasts were loaded onto u one-dimensional SDS-PAGE gel (4% stacking gel, 7.5% running gel). Antibody probing was done using preabsorbed polyclonal rabbit anti-VSV serum (1 :2,500) and goat anti-rabbit IgG (H"t-L) alkaline phosphatase conjugate (1: I 0,000). 155

-76KDa

Figure 3.14. Subcellular distribution of SYNVG in transfected protoplasts.

Viable protoplasts, 20 hours after transfection, were lysed in ice-cold HBST medium.

The lysate was centrifuged at 200 g for 20 minutes to pellet nuclei and heavy debris, which were further purified (see "Nuclei isolation and sorting"). The supernatant was centrifuged again at 300,000 g for 1 hour to separate the fractions designated as microsomes and cytosol. The cytosolic fraction was concentrated through a Centricon

P-30. The nuclear, microsomal, and cytosolic samples (equivalent to 106 protoplasts. in

SOS-sample buffer) were analyzed on a SOS-PAGE gel (4% stacking gel, 7.5% running gel). The first lane was a protein sample taken from SYNV particles. Antibody probing was done using preabsorbed mouse anti-SYNV ascites (1: 1,000) and goat anti-mouse Ig

G (H+L) alkaline phosphatase conjugate (1 :2,000). 156

nuclear membrane, the nuclei oftransfected protoplasts with SYNVG were isolated by

flow cytometry, and were subjected to gel electrophoresis and Western blotting.

SYNVG was only detected in the nuclear fraction when nuclei were isolated in media

lacking detergent (Figure 3.14). However, if isolation of nuclei was conducted using

the chopping buffer described by Galbraith et al. (1983), SYNVG was not subsequently

detected in the isolated nuclei (data not shown).

Analyses of the subcellular distribution ofSYNVG also provided a clue

concerning the presence of a non-specific protein band in the samples prepared from

intact transfected protoplasts but not in the samples from the transgenic plants (compare

Figure 3.7 and Figure 3. 11) The non-specific protein band was found exclusively in

the cytoplasmic fraction oftransfected protoplasts (Figure 3.14), and was not seen in the

blot of the transgenic plants, which was done using microsomal samples. This suggests that microsomal isolation can be used to routinely exclude cytoplasmic cross-reacting proteins during SOS-PAGE and Western blotting.

DISCUSSION

Investigating the behavior of heterologous membrane proteins within plant cells should provide new perspectives on plant protein trafficking at the molecular and cellular level. The heterologous probe method has been well established for dissecting secretory pathways in animal and plant cells (Garoff, 1985; Machamer et al., 1987, Galbraith et aI., 157

1992). The use of heterologous markers offers such advantages as simplifying kinetic

studies, providing an ability for the functional analysis of retention/sorting signal

sequences without interference from endogenous proteins, and allowing the use both of

the well- established body of knowledge from other systems and of various molecular

resources (antibodies, cDNAs and etc.) for monitoring the movement of the probes.

Expression of heterologous probes also avoid potential problems inherent to

over-expression of homologous probes, particularly if they are involved in vital

physiochemical functions.

For heterologous probes to be useful, the following criteria should be satisfied: 1)

The probe should be a monomeric and stable molecule coded by a single, well defined

gene, and its expression and assay in plant cells should be possible both in vivo and in

vitro. 2) Its detection should be sensitive and accurate. 3) Most importantly, its

expression and targeting behavior in plant cells should be both controllable and

reproducible. The probes and their derivatives selected for this work meet most of these requirements. With the exception of the native mv M glycoprotein, these heterologous probes are well expressed in both transfected protoplasts and transgenic plants.

In order to efficiently establish specific heterologous probe systems, we adopted a systematic, linear protocol. The first step comprised the construction of the transient expression vectors, and was followed by analysis of these vectors in transfected protoplasts. Those constructions that were successfully expressed in protoplasts were 158

then used in the construction of binary expression cassettes for plant transformation. This

was finally followed by plant transformation and analysis of expression in transgenic

plants. The first two steps are obviously critical in the whole experimental strategy;

however, it should be noted that protoplast isolation and transfection are particularly

labor-intensive, and it is quite difficult to routinely achieve sufficient transfected

protoplasts for successful data analysis. The obvious advantages offered by protoplast

transfection are (i) the easy and uniform nature of transfection treatments, (ii) the ability

to analyze the early kinetics of probe biosynthesis and traffic, (iii) the fact that

subcellular fractionation is relatively simple, and, above all (v) the fact that the

procedures provide data rapidly. In this project, this sequential order was followed with

two exceptions. In those two cases, transient expression and analysis failed to detect the expected protein. Construction of transformation vectors and production of transgenic plants was therefore de-emphasized for those markers; One is the case of the native IBV

M glycoprotein, and the other one the glycosylinositol phospholipid anchored protein

(Thy-l ).

In order to quickly address the potential targeting activities inherent to IBV M, an alternative strategy was devised. This involved producing chimaeric gene fusions between the coding sequences of IBV M and the GUS marker gene. Major advantages of this approach are that (a) GUS can be sensitively and quickly assayed following transfection and; (b) that anti-GUS polyc1onal antibodies are available commercially, and are suitable for Western blotting analysis. I'.

159

Some disadvantages were identified. Fusion of the whole or partial IBV M

sequences N-distal to GUS considerably depressed the enzymatic activities observed after

transfection (Figure 3.3 and Figure 3.4). This phenomenon may either be due to effects

on transcription/translation, or due to a direct effect of the presence of the M glycoprotein

on GUS activity or on assembly of the GUS holoenzyme. Inhibition is probably not due

to glycosylation of GUS at its putative N-linked glycosylation site, since that site in the

chimaeric GUS protein resides at the cytoplasmic face of the endomembrane

compartment harboring the IBV M transmembrane sequences. The data from

transfection experiments involving IBV ml~-GUS (Figure 3.4) confirmed that a smaller

N-terminal insert results in a significant increase in total GUS activities, about ten times

as much as that observed for the IBV M-GUS chimaeric protein.

The band of the Gm 1 smaller isoforms showed a lesser intensity when

microsomal samples rather than total protoplasts were used in Western blotting (Figure

3. 9). Furthermore, the Gm I smaller isoform was found to be released from the

endomembrane fraction into the cytosolic fraction (Figure 3. 13). As an explanation, it

is possible that the native VSVG and derivatives (or the larger isoforms) are transported

to, and become proteolytic ally processed at, a downstream site within the secretory

pathway, as postulated by Galbraith et al. (1992). This results in two isoforms with

different subcellular localization; the smaller isofonns are exposed to further

proteolytical cleavage, are released from the membrane, and are subsequently subjected

to rapid degradation. I.

160

Although Gm 1 behaves in fashion similar to VSVG and its mutants in the

preliminary analyses described above, the results raise a serious question about the

subcellular location of Gm1. Gm I was originally intended as a marker to assess the

Golgi retentional signal in the ml domain of the IBV M glycoprotein within plant cells.

It differs from the native VSVG only through replacement of the transmembrane domain

ofVSVG with the ml domain. If Gm I is transported and proces5ed in vivo in the same

manner as reported for VSVG (Galbraith et aI., 1992), it is highly possible that the

mammalian Golgi retention signal in the ml domain is not functional within plant cells,

and Gm I becomes transported to the plasma membrane by the default secretory pathway.

If Gm 1 does not reach the plasma membrane, it will be very interesting to ascertain its

subcellular location as well as the site at which it becomes proteolytically cleaved.

The preliminary results obtained with SYNVG also raise issues concerning its

expected final destination. SYNVG was detected in the nuclei oftransfected protoplasts

when nuclei were isolated in media lacking detergent Triton X-tOO (Figure 3.14). It

was not detected if detergent was included in the nuclear isolation medium. Inclusion of

detergent in the medium is intended to solubilize membranous debris associated with

nuclei prior to sorting in the flow cytometer. It also serves to solubilize the outer nuclear

membrane and any ER associated with nuclei. The inner nuclear membrane is relatively

resistant to low concentrations of Triton X-I 00 at low temperatures, and is responsible

for maintaining nuclei intact in the presence of Triton X-I 00 (Galbraith et aI., 1983). The

absence ofSYNVG in the nuclei isolated in the presence of Triton X-IOO suggests that 161

SYNVG is not transported into the inner nuclear membrane as originally hypothesized.

If SYNVG alone can't be targeted to the inner nuclear membrane, what is its subcellular

location? and what other viral protein(s) might be required for its transport during virus

assembly? or is its transport into the inner nuclear membrane dependent on specific stage of the cell division cycle? These questions remain to be resolved by further experiments. "

162

CHAPTER 4

THE ZIPPERED-MEMBRANE, A NOVEL ARTIFICIAL ORGANELLE

FOR EXPRESSING RECOMBINANT INTEGRAL MEMBRANE PROTEINS

INTRODUCTION

The plant Golgi apparatus exhibits a general similarity to its mammalian

counterpart, in that it is responsible for the post-translational modification, retention,

sorting, and transport of proteins that pass through its lumen (Farquhar, 1985;

Chrispeels, 1985, 1991). In contrast to the animal Golgi, the plant Golgi is involved

more heavily in the processing ofO-linked glycoproteins and, particularly, in the

synthesis of polysaccharides for the cell wall (Driouich et aI., 1993). Little is currently

known about the molecular sorting signals and retentional machinery regarding proteins

within the plant Golgi. Neither native Golgi enzymes nor protein markers have been

isolated or established for studies on protein retention. Establishment of a working

system in plants is therefore of prime importance for investigation of the molecular

mechanisms of retention and sorting in the Golgi.

To circumvent the problems faced in direct isolation of plant Golgi enzymes,

alternative approaches can be devised. One such approach would be to establish an

Arabidopsis genetic system to select mutants that lack specific Golgi enzymes (von

Schaewen et aI., 1993). A cgl mutant, deficient in N-acetylglucosaminyltransferase, was 163

identified within an EMS-mutated Arabidopsis population, but the cg/ gene and its gene

product have not yet been isolated (von Schaewen et aI., 1993). An alternative way,

albeit not reported yet, would be to exploit the powers of yeast genetics through

selection of plant cDNA sequences that can functionally complement yeast mutants

defective in retention and sorting of resident Golgi proteins. A final alternative would be

to use heterologous protein probes whose targeting has been established in mammalian or

yeast cells.

We chose to attempt transplanting a well-studied heterologous system from

mammalian cells into plants when this dissertation project was initiated about three years

ago. We found, surprisingly, that expressing one of the chimaeric protein probes in

tobacco, coronavirus avian infectious bronchitis virus M glycoprotein (IBV M) fused

N-distally to p-glucuronidase (GUS), created an artificial novel cellular organelle. This

chapter documents work done on the expression and characterization of the IBV M-GUS

chimaeric proteins in plant cells and the identification of the novel organelle.

The first element of the IBV M-GUS chimaeric protein (Figure 4. 1) comprises

the complete primary sequence of the IBV M glycoprotein. IBV M is required for the

formation of the viral membrane envelope, and is restricted to the internal membrane of

the infected mammalian cells. It only reaches the cell surface as part of the budded

virion at maturation (Sturman and Holmes, 1983; Armstrong et aI., 1984; Rottier et aI.,

1985, 1986). It is an integral membrane glycoprotein of 225 amino acids, and has a molecular mass of 25 to 32 KDa dependent on its glycosylation status (Rottier et aI., 164

~ IBV M glycoprotein lumen

cytoplasm

GUS subunit

Figure 4.1. Schematic illustration of the topology of the IBV M-GUS chimaeric protein. 165

1985, 1986). IBV M is a type IlIa membrane protein, according to the classification of

Singer (1990). Its short luminal N-terminus, with two putative N-glycosylation sites, is followed by three transmembrane domains and a large cytoplasmic C-terminus

(Armstrong et ai., 1984; Rottier et ai., 1985 and 1986; Figure 4. 1). The first transmembrane (ml) domain serves as a signal patch for ER translocation and as a retentional sequence for the accumulation ofIBV M and its derivatives in the cis compartments of the mammalian Golgi (Rottier, 1985; Machamer and Rose, 1987;

Machamer et ai., 1990; Swift and Machamer, 1991). The surface charge of the predicted helix of the ml domain was found to be required for the cis-Golgi retention of the various proteins (Machamer et ai., 1990). Oligomerization caused by the ml domain is postulated as a molecular mechanism underlying the retention of proteins in the mammalian Golgi (Weisz et ai., 1993). The second element of the IBV M-GUS chimaeric protein comprises the complete GUS coding sequence from the uidA locus of

E. coli (Jefferson, 1985; Jefferson et a\., 1986). Detailed information concerning GUS is described under "INTRODUCTION" in Chapter 3.

MATERIALS AND METHODS

Plant materials: Experiments were conducted using axenically cultured plantlets and suspension cultured cells of transgenic and non-transgenic tobacco (Nicotiana 166

tabaclim L. cv Xanthi). Their generation and maintenance are described in Chapter 2

(Harkinsetal., 1990; Meyeret aI., 1988; Hall, 1991a, 1991b).

DNA manipUlation and plant transformation: DNA manipulation was

conducted as described by Sambrook et al. (1989), unless indicated otherwise. Fusion of

the IBV M and GUS coding sequences was achieved through use of the polymerase

chain reaction, and involved several intermediate cloning steps (described under

"MATERIALS AND METHODS" in Chapter 3). The chimaeric protein is under the

control of a doubled 35S promoter and a 3' non-translated poly A + signal from the CaMV

35S gene. The completed expression cassette (pFG 11A) was used for transient

expression. The binary transformation cassette (pFG 14) was used for production of

transgenic plants via leaf-disc transformation with Agrobacterillm (Horsch and Klee,

1986, also see Chapter 2 and the "MATERIALS AND METHODS" of Chapter 3).

Erotoplast isolation and transfectjon: Protoplasts were prepared from leaf tissues ofaxenically cultured tobacco plantlets. and were transfected using chemical procedures

(Harkins et aI., 1990; Galbraith et aI., 1992; the details are described in "Protoplast

Isolation and Transfection" in Chapter 2). Tobacco protoplasts were transfected with plasmid pFG llA and carrier DNA (sonicated thymus calf DNA) in the presence of polyethylene glycol 4000, and were cultured at 25°C in darkness before analysis. The control was transfected with carrier DNA, but without plasmid DNA, under the same conditions as described for plasmid pFG llA. "

167

Sample homo~enization, differential and isopycnic ~radient centrifugation:

Viable transfected protoplasts were purified by sucrose flotation, washed with W5

medium and pelleted as previously described (Harkins et al., 1990, see "Protoplast

Isolation and Transfection" in Chapter 2), then resuspended in ice-cold 1 X GUS

extraction buffer described by Jefferson et al. (1987; see "p-Glucuronidase Assay" in

Chapter 2), or resuspended in 2 X SDS-sample buffer for gel electrophoresis (Galbraith et

aI., 1992; also see "Gel Electrophoresis and Western Blotting" in Chapter 2), or

resuspended in ice-cold HBST medium (50 mM Tris, 1 mM EDTA, 300 mM sucrose, 1

mM PMSF, 5 mM ascorbic acid, 3.6 mM cysteine and 100 J.lg/mL BHT, pH 8.0) for

subcellular fractionation and isopycnic sucrose gradient centrifugation. Protoplasts were

lysed in microfuge tubes by passage through a hypodermic syringe needle several times.

Leaf samples of transgenic and non-transgenic plants were homogenized in either I X

extraction buffer or HBST medium for various assays. All further procedures were

conducted at 4 °C unless analyses required different temperatures.

Subcellular fractionation of membranes and cytoplasm was conducted via

differential centrifugation as previously described by Galbraith et al. (1992). Lysed

protoplasts in HBST medium were subject to two rounds of centrifugation. The tirst

round (1,000 g for 20 minutes) served to pellet nuclei and heavy debris, and resultant

pellet was designated as nuclei. The second round was conducted in a Beckman TL-I 00

Ultrafuge (300,000g for 60 minutes) to separate the membranes from the cytosoIic

fractions, designated respectively as microsomes and cytoplasm. 168

The supernatants from the first round of differential centrifugation were loaded

onto a 10 mL 20-50% linear sucrose gradient (Galbraith et aI., 1992; see "Preparation and

Fractionation of Sucrose Gradients" in Chapter 2) to separate the membrane fractions

based on their densities. After centrifugation at 35,000 rpm (Beckman L8-M

Ultracentrifuge and SW41 rotor) for 16-18 hours, sucrose gradients were fractionated

into 0.6 mL samples for protein assays, for marker enzyme assays of membranes, and for

GUS assays. For gel electrophoresis, fresh fractions were diluted with three volumes of

TE buffer (50 mM Tris, I mM EDTA, pH 8.0), and were pelleted by centrifugation at

300,000 g for 60 minutes; the pellets were then resuspended in 2 X SDS-sample buffer.

The location of the various endomembranes and organelles were determined by the following marker enzyme assays: NADH cytochrome C reductase for the ER and mitochondria, latent inosine diphosphatase (IDPase) for the Golgi, vanadate-sensitive

W -ATPase for the plasma membranes, and autofluorescence of chlorophyll for the thylakoid membrane (Donaldson et aI., 1972; Gardiner and Chrispeels, 1975; Briskin el al., 1987; Galbraith el al., 1992, also see Chapter 2). Sucrose concentration of fractions was determined by refractometry.

GUS assays. gel electrophoresis and Western blotting: The expression of the IBV

M-GUS chimaeric protein was evaluated by the GUS enzyme assay, and by gel electrophoresis and Western blotting (Harkins et aI., 1990; Galbraith et aI., 1992; the details are described in "p-Glucuronidase Assay" and "Gel Electrophoresis and Western

Blotting" of Chapter 2). Primary antibody employed for probing was a rabbit polyclonal 169

anti-GUS Ig G (H+L) from Molecular Probes, Inc. (Eugene, OR, USA). Before use, the

anti-GUS antibody was absorbed using an affinity column coupled with total proteins

prepared from non-transfected protoplasts (to eliminate non-specific antibodies), and was

concentrated to the original volume using a Centricon P-30 concentrator (Ami con, MA,

USA) as described in Chapter 3. The anti-GUS antibody was used at a dilution of

1: 1,000 (equivalent to the dilution using the original stock). Goat anti-rabbit Ig G (H+L) alkaline phosphatase conjugate was purchased from Sigma Chemical Co. (St. Louis, MO,

USA), and used at a dilution of 1: 10,000.

Immuno fluorescence labelini:. flow cytometric analysis and confocal microscopy: Protoplasts were isolated as described from leaves of non-transgenic and pFG 14-23 (IBV M-GUS) transgenic plants (Harkins et aI., 1990; Galbraith et aI., 1992; the details are described under "Protoplast Isolation and Transfection" of Chapter 2).

Protoplasts were washed by resuspension in 1 X K-medium (0.35 M KCl, 10 mM CaCI2,

3 mM MES, pH5.7), then were immediately recovered by centrifugation as described by

Galbraith et al. (1992), and were fixed in 2% formaldehyde buffered with 1 X K-medium at 4 °C overnight. After fixation, the pro top lasts were subjected to three cycles of washing and centrifugation using ice-cold 1 X K-medium (each wash step involved 30 minutes of incubation in K-medium with gentle shaking). Permeabilization of the fixed protoplasts was carried out on ice in TSW medium (Galbraith et aI., 1992; also in "Gel

Electrophoresis and Western Blotting" of Chapter 2) for 1 hour. After the treated protoplasts were washed two times for 15 minutes each in modified TSW medium 170

(supplemented with 0.02% Triton X-I 00, but lacking SDS), they were subjected to

primary antibody labeling in modified TSW medium containing preabsorbed rabbit

polyclonal anti-GUS antibodies (1:100) at 22°C with gentle shaking for 1 hour. After

two washings for 15 minutes each in modified TSW medium, the labeled protoplasts

were mixed with goat anti-rabbit BODIPY conjugate antibody (diluted 1:50; Molecular

Probes Inc.) in modified TSW medium at 22°C with gentle shaking for 1 hour. The final

two washings for 10 minutes each were in modified TSW medium. Fluorescence

analysis of the labeled pro top lasts were conducted on Coulter Elite flow cytometer with a

laser excitation at 488 nm, with being measured emission at 525 nm, as previously

described (Harkins et aI., 1990).

For microscopic examination of fluorescence-labeling patterns, the labeled

protoplasts were mounted on glass slides after mixing with two volumes of

Aqua-Poly/Mount mounting medium (Polysciences, Inc.). Examination was conducted

using a Zeiss Standard Microscope equipped with epi-fluorescence illumination

(excitation bandpass filter at 485±8 nm, bean1 splitter at 510 nm, and emission bandpass

filter at 543±25 nm) or on a Bio-Rad MRC-600 confocal microscope equipped with dual

wavelength excitation (for red and green fluorescence respectively).

Immuno electron microscopic examination: This work was conducted by Dr.

Staehelin's group in the University of Colorado at Boulder. Suspension culture cells of

non-transgenic and transgenic (pFG 14-23) plants were used because they are well-suited to the process of high-pressure freezing, freeze-substitution, resin filtration, and antibody 171

labeling required for immuno electron microscopy. The suspension cells were harvested

and resuspended in fresh culture medium containing 0.2 M sorbitol as a cryogen

protective agent as described by Zhang and Staehelin (1992). They were cryofixed in a

Balzer HPN 010 high pressure freezing apparatus, freeze-substituted in 2% osmium

tetraoxide in acetone at -80°C for three days, and then warmed to -20 °C, 4 °C and room

temperature for 2 hours each (Craig and Staehelin, 1988). After rinsing in acetone, the

samples were embedded in Epon-Araldite. The procedures of resin embedding, san1ple sectioning and staining, antibody probing, and examination were done as described by

Zhang and Staehelin (1992).

RESULTS

EXPRESSION OF IBV M-GUS IN TRANSFECTED PROTOPLASTS AND TRANSGENIC PLANTS

Expression of the chimaeric protein was first evaluated in transfected protoplasts isolated from tobacco leaves. Without plasmid DNA, GUS activities in transfected protoplasts were at baseline levels, whereas significant enzymatic activities were observed following transfection with pFG llA (see Figure 3. 3 in Chapter 3).

The molecular weight of the IBV M-GUS chimaeric protein was determined through SDS-PAGE analysis using protein samples extracted from transfected protoplasts and transgenic plants (Figure 4. 2 and Figure 4. 3). The IBV M-GUS chimaeric protein 172

was detectable in all extracts, and had an apparent molecular mass of 95 KDa. This size

is very close to the predicted molecular mass of the chimaeric protein (828 amino acids,

93.6 KDa) and is consistent with a lack of glycosylation (Figure 4.2 and li'igure 4.3).

Tunicamycin treatment in vivo and endoglycosidase H treatment in vitro of extracts did not alter the mobility of the chimaeric protein on gels (data not shown). Interestingly, two bands of the protein were observed in extracts from transgenic plants (Figure 4.3), whereas only one band was seen in extracts made from transfected protoplasts or from leaf tissue expressing mv M-GUS transiently (Figure 4.2 and Figure 3.8). The smaller band of the chimaeric protein may be a result of proteolytic cleavage at the N-terminal region, which is located within the lumen of the endomembrane system, because the smaller band was still associated with the microsomal membrane fraction in transgenic plants (Figure 4. 3 and Figure 3. 8).

ASSOCIATION OF mv M-GUS WITH THE ENDOMEMDRANE SYSTEM

Enzymatic GUS activity of pro top lasts transfected with the mv M-GUS construct was found to be mostly localized in the membrane particulate fraction (about 70%), as found by differential centrifugation of cell homogenates. In contrast, 99% of the GUS activity was found in the cytosolic fraction for protoplasts expressing the native GUS

(lacking topogenic sequences; Figure 3. 12 in Chapter 3). Identification ofIBV 173

- 95KDa -73 KDa

Figure 4. 2. Western blotting of the GUS and IBV M-OUS proteins from

transfected protoplasts. Viable transfected pro top lasts (2 x lOs) from treatments without expression vector (control), with pFO 11 A (IBV M-OUS), and with pBAP 15 (OUS), were resuspended in a suitable amount of 2 X SDS-sample buffer, and were loaded onto a one-dimensional SDS·PAOE gel (4% stacking gel, 7.5% running gel). The native

GUS enzyme (L:-GUS: 2.5 u, about 0.5 )..lg protein; Sigma Chemical Co. ) was employed as a positi\'c control. Western blotting was conducted as described under

"MA TERlALS A~D :VIETHODS" of this Chapter. 174

___ 95KDa ::=:::::-- 92 KD a ---73 KDa

Figure 4. 3. Western blotting of the GUS and!BV M-GUS proteins from

transgenic plants. Microsomes were isolated from non-transgenic (control) and

transgenic pFG 14-23 (!BV M-GUS) plants through homogenization and differential

centrifugation (described in "MATERIALS AND METHODS"), with the exception that the first centrifugational step was at 13.500 g for 15 minutes to pellet debris. Proteins were extracted by dissolving the microsomes in 1 X GUS extraction buffer. For the transgenic plant expressing soluble GUS from pBI 121 (GUS), the supernatant was collected from leaf homogenates. and was concentrated using a Centricon P-30 concentrator (Amicon, MA. USA). Proteins from each sample (200 ).lg) were loaded onto a one-dimensional SDS-PAGE gel (4% stacking gel, 7.5% running gel). Western blotting was conducted as described in "MATERIALS AND METHODS". 175

M-GUS as a membrane intrinsic protein is further supported by the results of various

extraction treatments (Figure 4.4). The data show that over 99% of the GUS activity

remained associated with the membrane after treatment with 0.4 M NaCI, but GUS

activity was completely released from membranes by 1 % Triton X-IOO.

COSEGREGATION OF mv M-GUS WITH THE GOLGI ELEMENTS ON ISOPYCNIC SUCROSE GRADIENTS

To detennine which compartment within the endomembrane-secretory pathway

houses the chimaeric protein, transfected protoplasts were used to analyze the early stages

of protein biosynthesis, targeting and accumulation. Twenty-four hours after transfection

with pFG 11 A, the predominant membrane elements carrying GUS activities (Figure 4.

5, panel C) were observed to be distinct from those containing chlorophyll (Figure 4. 5,

panel A), NADH cytochrome c reductase (Figure 4.5, panel B) and vanadate-sensitive

ATPase (Figure 4. 5, panel E), but coincided with the membrane fractions carrying

latent inosine diphosphatase (IDPase), a marker enzyme for the plant Golgi (Figure 4.5,

panel D). A smaller peak of GUS activity cosegregated with the marker enzyme of the

ER membrane fraction (Figure 4. 5, panels C and B), implying that the chimaeric

protein is synthesized at the ER. When accumulation patterns were analyzed using leaf

microsomes prepared from transgenic plant pFG 14-23, which expresses mv M-GUS

constitutively, a large and broad peak ofIBV M-GUS activity was observed on sucrose gradient, and the distinct small peak associated the ER fractions disappeared 176

.~ :E- D Membranes ~ Supernatant -Col 100 CI.l< ;:::J ~ c,., Q 'if.- ~ 50 .S:- .Q= 'i:

Figure 4. 4. Extractability of IBV M-GUS fusion protein from the

endomembranes. The endomembranes were isolated from leaves of transgenic plant

pFG 14-23. The leaves were ground in ice-cold HBST medium. The homogenate was

filtered and centrifuged, first at 13,000 g for 15 minutes to pellet debris, then at 300,000 g for 60 minutes to pellet microsomes. Microsomes were resuspended in HBST medium

(control), HBST with 0.4 M NaCI, and 1% Triton X-IOO, respectively. After a

30-minute incubation at 22°C, the particulate material was pelleted at 300,000 g for 60 minutes, and was solubilized in 1 X GUS extraction buffer. GUS activities and their percentage of distributions were measured and calculated as described by Jefferson et al.

(1987). 177

Figure 4. 5. Distribution of the IBV M-GUS chimaeric protein on isopycnic

6 sucrose gradients. Viable protoplasts (2xl0 ), purified 24 hours after transfection with pFG 11 A plasmid, were lysed in ice-cold HBST medium. After centrifugation at 1000 g at 4°C for 20 minutes, the lysate supernatant was collected, adjusted to a sucrose concentration of 10-12%, and loaded onto a 10 mL 20-50% linear sucrose gradient

(buffered with 10 mM tris, 1 mM EDTA, pH 8.0, 1 mL of 50% sucrose being placed at the bottom as a cushion). After centrifugation at 240,000 g at 4 °C for 18 hours, the gradient was fractionated into 20 fractions of 0.6 mL. Chlorophyll content (arbitrary fluorescence units/0.05 mL fraction), the activities of antimycin A insensitive NADH cytochrome c reductase (ilAssof4 mini mL fraction), and sucrose (% w/w) were measured immediately after fractionation (Donaldson et aI., 1972; Galbraith et aI., 1992). The activities of GUS (arbitrary fluorescence unit/hr/O.l mL fraction), latent IDPase

(A7IO/hr/mL fraction), and Vanadate-sensitive ATPase (A710 /hr/mL fraction) were measured 2-3 days later after fractions were stored at -80°C (Gardiner and Chrispeels,

1975; Jefferson et aI., 1987; Briskin et aI., 1987) 178

875 -;>-, -a 625 Q I. ~- 375 o 125

c:.l ~ 1.05 B 'tj .; 0.75 c:.l ~ C.J 0.45 Q ~ 0.15 ~~~~~+W~~~ .f' 500 c 1\ .~ 'tj -< 300 I - (JJ _ i \ ;:l \.!) 100 /\ l -\ I,.. - I, ,,1:;0.---­ ,e. 0.21 :E 'tj 0.15 -< c:.l ~ 0.09 ~ S 0.03 ~WQ~~~~~~~ E 0.30 E .~ C.J -<- 0.20 c:.l

(Figure 4. 6, lower panel). This broad peak still overlapped the peak of latent IDPase, but its range was generally displaced to lighter densities (Figure 4. 6, middle and lower panels). This suggests that the IBV M-GUS chimaeric protein does not exactly colocalize with the Golgi membrane fractions. The above observations are consistent with the idea that the IBV M-GUS chimaeric protein is first synthesized on ribosomes of the rough ER membranes, then is translocated across and anchored in the ER membrane, and finally accumulates within a more-dense membrane compartment.

KINETICS OF BIOSYNTHESIS AND ACCUMULATION OF THE IBV M-GUS CHIMAERIC PROTEIN WITHIN THE PLANT SECRETORY PATHWAY

When the distribution ofIBV M-GUS activities on sucrose gradients was examined as a function of time after transfection, IBV M-GUS was observed to be first detected in the ER membranes that contained antimycin A-insensitive NADH cytochrome c reductase (Figure 4. 7). At subsequent time points, there was a progressive accumulation of GUS activity within heavier membrane fractions, having isopycnic densities of 1.12-1.17 g/mL. The position of this heavier membrane fraction was similar to that of the Golgi membrane fraction in the sucrose gradients (Figure 4. 5, panel C). As incubation periods were increased, the accumulation of GUS activity within the heavier membrane fractions were progressively shifted toward a lower position in the sucrose gradients, but not as far as the positions of the plasma membrane fraction (1.18-1.22 180

Co) III c: 0.15 -CJ "0= Co) ~ 0.15 U I 0 .... 0.05 U- .... :~- 0.15 -<~ Co) c:III ~ 0.05 -Q ...... 5000 :~ -<-CJ 3000 CIJ ~ 0 1000

1.00 1.05 1.10 1.15 1.20 1.25 Sucrose Density (g/mL)

Figure 4. 6. Isopycruc gradient distribution of the IBV M-GUS chimeric protein from transgenic plant pFG 14-23. Microsomes were isolated as described under

"MATERIALS AND METHODS", but the first centrifugational step was at 13,500g for

15 minutes. The remaining steps were the same as conducted in Figure 4. 5. I'

181

500

<>-<> 24.0 hours 400 6-6 16.7 hours 0-0 12.0 hours .e- 0-0 8.0 hours :~ 300 'V-'V 4.0 hours ....y < rJl ;;J 200 - ~

100

o ~-L~~-L~~~-L~~-L~~-L-L-I 1.00 1.05 1.10 1.15 1.20 1.25 Sucrose Density (g/mL)

Figure 4.7. Kinetics of IBV M-GUS biosynthesis and accumulation within

6 compartments of the endomembrane secretory pathway. Viable protoplasts (1 x 10 ) were

purified 4, 8, 12, 16.7, and 24 hours after transfection with plasmid pFG 11 A, and were

analyzed as described for the sucrose gradient in Figure 4. 5. 182

g/mL, Figure 4.5, panel C). When the accumulation rates oflBV M-GUS were

compared between membrane fractions labeled by the marker enzymes of the ER, the

Goigi and the plasma membrane (Figure 4. 8), it seems unlikely the plasma membrane

was the site for the ultimate mv M-GUS accmnulation, since this rate of accumulation

was essentially zero after protein synthesis for at least 16 hours. The rate of accumulation

in the ER remained constant (Figure 4. 8). It seems more likely that the membrane

fractions of similar density to the Goigi membrane represent the main accumulation sites for the chimaeric protein, as the rate of accumulation was dramatically increased over 16 hours of protein synthesis (Figure 4. 8). Taken together, the data from transfected protoplasts point to the Goigi membrane as a possible candidate for the accumulation of

IBV M-GUS, since the peak of accumulation ofIBV M-GUS overlapped with that of the

Goigi marker enzyme (Figure 4. 5, panel C and D). However, the data from the transgenic plants provide some suggestion that this hypothesis was incorrect. Further elucidation of this point was then obtained through the following cytological analyses.

IN SITU FLUORESCENCE LABELING OF THE IBV M-GUS CHiMAERIC PROTEIN WITHIN TRANSGENIC CELLS

The above biochemical analyses indicated that the IBV M-GUS chimaeric protein predominantly accumulates in heavy membrane fractions cosegregating with the Golgi 183

>. :~- 1600 y -<- CIJ C-:J cosegregated with Goigi marker ~ c.:l 1200 c~:) cosegregated with ER marker c­ c'----o cosegregated with PM marker o

.~= .s- 800 e= (J= y -< (J .~ == 'ii- ~ 4 8 12 16 20 24 Hours after Transfection

Figure 4. 8. Relative accwnulation rates ofIBV M-GUS cosegregating on sucrose gradients with the ER, the Golgi, and the plasma membrane. The total GUS activities on each sucrose gradient in Figure 4. 7 were classified into three groups of membrane fractions according to the marker enzyme positions of the ER(::51.12 mglmL), the Golgi (1.12 -1.17 mglmL), and the plasma membrane (PM; ~ 1.17 mglmL). The relative total GUS activities from each group were plotted against the time of incubation, after subtracting non-specific activity (baseline noise). 184

elements (Figure 4.5, panel C and Figure 4.7). This is consistent with the suggestion

that the chimaeric protein was targeted to the plant Oolgi apparatus, as originally

expected. However, to investigate the subcellular location of the protein more thoroughly, indirect immunofluorescence labeling was employed with protoplasts isolated from leaves of transgenic plant pFO 14-23. Two types of analyses were then done; in the first, the labeled protoplasts were subjected to flow cytometry. Results indicated that the fluorescent BODIPY probes strongly labeled the protoplasts from transgenic plant pFO 14-23, but did not labeled the non-transgenic control (Figure 4. 9, panel A). Although flow cytometer detected very faint fluorescence on the control protoplasts in the presence of both the primary and secondary antibodies (median and peak channels of 6 and 6 respectively), this fluorescence was not visible under the epifluorescence microscope. Its intensity was significantly lower than that displayed by their transgenic counterparts (median and peak channels of 22 and 33 respectively;

Figure 4.9, panel A). Correspondingly, the BODIPY-Iabeled transgenic protoplasts were easily visible under epifluorescence microscope. However, spatial resolution for the image provided by the conventional epifluorescence microscope was severely compromised by out-of-focus fluorescence, due to the spherical shape and large size of the protoplasts. In order to provide greater spatial resolution, the BODIPY-Iabeled protoplasts were then examined using a confocal microscopy. A composite picture 185

Figure 4. 9. Indirect immuno-fluorescence labeling of mesophyll cells isolated from transgenic plants expressing!BV M-GUS. Protoplasts were processed for immuno-fluorescence labeling as described under "MATERIALS AND METHODS".

(A) Flow cytometric analysis of the labeled protoplasts with (+) or without (-) primary antibody labeling (10 ab) and with (+) secondary antibody labeling (2° ab). 45,000 protoplasts for each labeling treatment were analyzed using the flow cytometer.

"M-GUS" represents the transgenic protoplasts expressing mv M-GUS. "Xanthi" represents the non-transgenic protoplasts as control. "peak" and "mean" indicate the median and the mean fluorescence intensities on a log detection scale ("Channel") of the flow cytometer. (B) Confocal microscope composite of a BODIPY-tagged protoplast from transgenic plant pFG 14-23 expressing the mv M-GUS chimeric protein.

Mounting and examination of the BODIPY -labeled protoplasts were conducted as described in "MATERIALS AND METHODS". 186

;;::p*!F

...... c - ::J Cl)O UU 187

(Figure 4. 9, panel B) was constructed by stacking digital images of twenty optical

sections collected from one labeled transgenic protoplast. The pattern of fluorescence

indicated that the BODIPY probes were associated with a number of globular structures,

estimated to be 2-8 f.lm in diameter (Figure 4. 9 panel B). These structures were distinct

them from chloroplasts, in that the pattern of red autofluorescence due to chlorophyll did

not superimpose the green BODIPY fluorescence (data not shown). The fluorescence

pattern also did not represent morphological characteristics typical of known components

in the plant endomembrane-secretory system (for example, a reticulate network). Nor did

it correspond to conventional histological descriptions of nuclei and mitochondria. We

speculated that the fluorescence pattern might represent a new cellular structure, possibly

formed to accommodate the expression of the IBY M-GUS chimaeric protein in

transformed plants of Nicotiana tabacum. This speculation was substantiated through

further histochemical investigations at the electron microscope level.

REVELATION OF NOVEL MEMBRANE STRUCTURES BY IMMUNO ELECTRON MICROSCOPY

To define the subcellular location of the IBY M-GUS chimaeric protein more precisely, transgenic suspension-cultured cells were processed using high-pressure freezing and freeze-substitution techniques. These provided excellent preservation of 188

morphology and excellent antigen labeling. The most obvious difference between the

cells of the control (Xanthi) and those expressing IBV M-OUS, was the presence of

onion-like structures within the transgenic cell cytoplasm (Figure 4. 10, C and D).

Some of them were observed in association with polyribosomes (Figure 4. 10, C). The

novel subcellular structures were composed of membrane bilayers wrapped straticulately

around inner cores. They have been named E-whorls as they appear to originate from the

ER. The Oolgi in both the control and pFO 14-23 cells were normal in appearance, and

showed recognizable cis, medial and trans cisternae (Figure 4. 10, A and B). Besides

E-whorls, some fibrous bundles were occasionally seen in the transgenic pFO 14-23

suspension cells, but not in the cells of the control (Figure 4. 10, B).

Following immuno-labeling, gold particles were exclusively associated with the conspicuous E-whorls on the sections of the transgenic cells, whereas the remaining cellular components were not labeled (Figure 4. 11, C, D, E and F). The sections of the non-transgenic suspension cells (Xanthi) displayed no labeling (Figure 4. 11, A). The

Oolgi apparatus was not labeled in the suspension cells ofpFO 14-23 (Figure 4. 10. B).

Fibrous bundles, at an early stage of membrane zippering, were labeled with gold particles at the appressed end, whereas the normal-looking rough ER was not labeled

(Figure 4. 11, D). The fibrous structures appeared to represent appression of the adjacent ER cisternae (Figure 4. 10, B and Figure 4. 11, D). I.

189

Figure 4.10. Electron micrographs of the pEG 14-23 transgenic suspension cells.

Thin sections were cut from the samples embedded in Epon resin after various sample

treatments, and were observed under the electron microscope. (A) Control (Xanthi)

suspension cells. A Golgi complex was shown with characteristic morphology and

polarity. (B) pEG 14-23 trans~enic suspension cells. Besides the recognizable Golgi, a

fibrous membrane stack is observed (indicated. by arrow). (C) and (D) Structures of

E-whorls jn the pEG 14-23 suspension cells. The darker circular areas within the

E-whorls represent a thin layer of cytoplasm located between the appressed ER

membranes. The lighter circular areas are derived from the lumen of the ER. Ribosomes

were sometimes found to be associated with E-whorls (labeled by arrows). I'

190

I:~ ,.~, ____ ••• _ 191 "

192

Figure 4.11. Immuno-electron micrographs of the transgenic pFG 14-23

suspension cells. Thin sections embedded in Epon were labeled with anti-GUS/gold

antibodies after treated with sodium metaperiodate to remove osmium and unblock

antigenic sites. (A) Xanthi suspension cells (control). (B) Suspension cells ofpFG 14-23

(magnification: x 31,000). (C) Three immuno-gold labeled E-whorls (magnification: x

15,500). (D) Immuno-gold labeled E-whorl and the appressed ER at the early stage

(magnification: x 15,500). (E) and (F) Anti-GUS gold labeled E-whorls. Magnification:

x 21,000x and x 15,500 respectively. 193 194 195 196

DISCUSSION

Under normal circumstances, integral membrane proteins entering the secretory

pathway are partially translocated across and anchored in the ER membrane in various

ways (Singer, 1990). After co- and post-translational modifications, only properly folded

and assembled proteins are able to exit the ER, and they are then transported

anterogradely to various further compartments under the general aegis of default

secretion (Pfeffer and Rothman, 1987; Pelham, 1989; Hurtley and Helenius, 1989; Pryer

et aI., 1992; Rothman and Orci, 1992; Hong and Tang, 1993). Misfolded proteins are

retained in the ER, and are subsequently degraded (Hurtley and Helenius, 1989;

Klausner et aI., 1990). It has been found that many recombinant proteins are vulnerable

to trapping and degradation within the secretory pathway, particularly in experiments

involving the study of protein trafficking (Garoff, 1985). For example, subunits of the T

cell receptor for surface antigen were found to be trapped and degraded in the ER after

their synthesis (Lippincott-Schwartz, 1988; Bonifacino et aI., 1990). In the case

described here, a contrast is seen; the IBV M-GUS chimaeric protein, although trapped within the early endomembrane-secretory pathway, appears to resist degradation and becomes concentrated in discrete subcellular structures (E-whorls).

The fom1ation of whorls can be explained as follows: the cytoplasmic domain of the chimaeric proteins contains the GUS coding region. GUS in its native state is a tetrameric homo-oligomer, and therefore has a strong tendency to homo-oligomerize. For "

197

the chimeric IBV M-GUS molecules, which are tethered to the membrane,

homo-oligomers of the GUS domains can be envisaged as occurring between molecules

that are either in the same membrane sheet or within different membranes of close

proximity. Evidence of the occurrence of oligomerization is directly indicated by

detection of substantial amounts of GUS activity from the chimaeric proteins (Figure 4.

5, Figure 4. 6 and Figure 4. 7). Homo-oligomeric IBV M-GUS within the ER may

then be too large to permit exit from the ER. Such a situation has been reported to be one

of the true retentional mechanisms that act to maintain membrane proteins in the ER. For

example, ribophorins and other proteins involved in ribosome binding and membrane

translocation, are thought to form large suprarnolecular complexes which facilitate their

ER residency (Hortsch and Meyer, 1985). A more recent example elegantly

demonstrated that engineered Golgi enzymes bearing ER retention sequence can retain

their native kin (lacking ER retention motif) in the ER through the formation of

oligomers (Nilsson et aI., 1994). However, it remains unclear regarding the critical size

that renders oligomers incompetent for exit from the ER.

Oligomerization between kindred molecules would also have the effect of limiting

the exposure of proteins to other enzymes in the ER and cytosol. The fact that E-whorls

become specialized ER membranous compartments that exclude other ER resident

proteins is demonstrated by a lack of ER marker enzymes in the peak of the IBV M-GUS

activity on sucrose gradients (Figurc 4. 5, panels Band C; Figurc 4. 6, upper and lower

panels). A hypothetical model for the formation of E-whorl is diagrammed in Figure 198

4.12. In this model, molecular interaction between the GUS subunits is the driving

force for the formation of E-whorls. This could be tested by analyzing endomembrane

association of native and engineered GUS (with or without a signal peptide, but lacking

the transmembrane domain) when co-expressed with the IBV M-GUS chimeric protein.

We would predict that native and engineered GUS lacking transmembrane domain would

become associated with membrane, and that association would be extrinsic.

Reports of novel organelles have sporadically appeared over the last decade. The cause underlying their formation remains unclear, and no consensus has emerged. Those studies were based on diverse systems and cell types, for example, including the effect of inhibition of native membrane enzymes in mammalian cells (Chin et aI., 1982; Singer et aI., 1984; Orci et aI., 1984; Pathak et aI., 1986), and the effect of overexpressing recombinant membrane proteins in mammalian, yeast, and bacterial cells (Weiner et aI.,

1984; von Meyenburg et aI., 1984; Jingami et aI., 1987; Wright et aI., 1988; Schunck et aI., 1991; Vergeres et ai., 1993; Calafat et ai., 1994). In terms of identifying a common property governing the formation of novel organelles, it seems that all reported cases involved proteins that either are multimeric integral membrane proteins, or that are required to interact directly or indirectly with other integral membrane proteins as part of their normal cellular function (Table 4. 1). Alternative hypotheses have been advanced in some cases (Vergeres et aI., 1993), for example that the formation of novel organelles is regulated by some type of "sensor" that monitors the biophysical characteristics of membranes (protein density, membrane fluidity, and surface charge), Even in that case, 199

r------ERLumcn IBV M-GUS Zippers

:rYN ••••• Monomer & Ollgomers Polyrlbosomes of IBV M-GUS

Figure 4. 12. Diagram illustrating a possible mechanism of membrane-zippering ofE-whorls mediated by oligomerization of the mv M-GUS chimeric protein. 200

Table 4. 1. Reported cases of membrane proliferation as a result of expressing integral

membrane proteins.

Year Hosts Expressed Proteins Multi-subunits

1984 E. coli Fumarate Reductase Yes

1984 E. coli ATP Synthase Yes

1987 Chinese Hamster Ovary Cells HMO CoA Reductase Yes

1988 S. cerevisiae HMO CoA Reductase Yes

1991 S. cerevisiael/< Cytochrome P-450 NA

1993 S. cerevisiael/< Cytochrome bs Yes

1994 Human Embryonic Kidney Cells MHC II Molecules Yes

Note: I/< = heterologous host

NA = information is not available

MHC = major histocompatibility complex. 201

the alternative possibility of protein-protein interactions leading to zippering can not be

excluded. Our results lead to the hypothesis that molecular oligomerization among

kindred membrane proteins is the driving force underlying novel organelle formation.

Since our system is simple, and non-deleterious to the cells in vivo, it can be easily tested

to determine whether organelle engineering can be achieved across different eukaryotic

cell systems; this will be particularly important in mammalian cells, in which mv M

glycoprotein is extensively studied, and in yeast, for which genetic methods are well

developed.

E-whorls may also find important applications in biotechnology. The production

of many types of foreign proteins within transgenic hosts can be hindered by a variety of

problems including (i) that foreign proteins become unstable, or insoluble, or lose activity

in host cells, and (ii) some important types of proteins can be toxic to the host. E-whorls

can offer a possible solution to these problems, by sequestering foreign proteins away

from the cellular contents immediately after biosynthesis. This would limit their

exposure to degradative enzymes, avoid their aggregation into insoluble particles leading

to loss of activity, and prevent their interference with normal cellular functions of the

host.

To achieve selective accumulation of transgenic proteins, the satisfaction of several basic criteria would be required for the engineered proteins: (a) a membrane insertion signal, (b) one or more transmembrane domains, (c) a cytoplasmic or luminal domain and, (d) a strong oligomerization tendency by anyone of these domains. The 202

oligomerized structures must also be large enough to prevent entry into transport vesicles.

This production concept is applicable to all transformable eukaryotic host cells having an

endomembrane-secretory pathway.

In our case the observation that, of 19 independently transformed plants, 17

plants stably and constitutively express the active IBV M-GUS chimaeric proteins

without any visible deleterious effects, suggests that the host cells adjust to the failure of

homeostatic regulation of protein concentrations within the secretory pathway (which

would typically be achieved via extracellular secretion, protein degradation, and/or

down-regulation of expression), by up-regulating lipid biosynthesis to maintain normal

cellular endomembrane compartments and functions, and at the same time to

accommodate the requirement of the chimaeric protein assemblies for a lipid bilayer

within which to accumulate. This adjustment could be coordinated with protein

biosynthesis, since the ER is also the site of lipid synthesis. The idea of coordinate

adjustment is further supported by evidence that overexpression of native membrane

proteins through addition of enzyme inhibitors can result in the production of novel subcellular membrane structures (Chin et aI., 1982; Singer et al., 1984; Orci et aI., 1984;

Pathak et aI., 1986). Establishing transgenic plants with regulated over-expression of

IBV M-GUS, using inducible systems, would provide a valuable experimental tool for studying the process of plant membrane biogenesis.

The project initiated as a feasibility study ofIBV M-GUS as a probe of molecular mechanisms underlying retention of Golgi-resident proteins, but has progressed with 203

finding of a novel zippered-membrane organelle. The implication of the finding is three­

fold. It reveals that certain physical criteria are required for the less-studied steps of

protein degradation in the early secretory pathway, and provides evidence in favor of kin recognition as a mechanism underlying the retention of resident proteins in specific endomembrane compartments (Nilsson et al., 1993, 1994). The finding also has biotechnological potential for the selective accumulation of recombinant proteins in non-degraded form. Furthermore, it can be utilized as a simple model system to investigate the biogenesis of membrane organelles, and as a tool to chart a novel and untouched field - the artificial engineering and amplification of membrane organelles. 204

CHAPTERS

A SHIFT IN LOCATION OF THE Gmt MEMBRANE PROTEIN,

FROM THE MAMMALIAN GOLGI TO THE PLANT TONOPLAST:

DOES THIS SUGGEST AN ALTERNATIVE DEFAULT

SECRETION PATHWAY WITHIN I)LANT CELLS?

INTRODUCTION

It is commonly held that proteins entering the endomembrane-secretory pathway are sorted and subsequently transported to their final destinations via default bulk flow, unless they are retained in the specific compartment or sorted at branch points according to specific signals (Pryer et al., 1992). The default endomembrane-secretory pathway is viewed as being initiated from the endoplasmic reticulum (ER), proceeding via the Golgi apparatus, and terminating at the plasma membrane. All diversions from the default pathway are therefore considered as being-signal mediated, for example the retrieval of the luminal ER proteins "escapees" bearing KlHDEL sequences, the retention of resident

Golgi proteins via a process of oligomerization, and the sorting of soluble lysosomal

(vacuolar) proteins into lysosomes (vacuoles) (Burgess and Kelly, 1987; Kornfeld and

Mellman, 1989; Pelham, 1989; Chrispeels, 1991; Weisz et al., 1993; Nilsson et al., 1993 and 1994). Except in the case of polarized cells, no additional signals appear necessary 205

for sorting of either membrane or soluble proteins that are destined to the plasma

membrane/cell surface. However, recent studies ofthe targeting domains of several

yeast membrane proteins residing in the 'rans-Golgi network (TGN) and vacuoles

reveals an iconoclastic result, in that trafficking of these proteins from the TGN to the

vacuole appears to operate via an alternative default pathway (Roberts et aI., 1992;

Cooper and Bussey, 1992; Bryant and Boyd, 1993; Nothwehr et aI., 1993; Nothweh.r

and Stevens, 1994). Thus, dipeptidyl aminopeptidase A (DPAPA), which is an integral

TGN protein (type I), mislocates to the yeast vacuole when over-produced, or when its

cytoplasmic retention tail is removed (Roberts et aI., 1992). Transport of the mutated

DPAPA (lacking cytoplasmic retention signal) to vacuole is unaffected in the sec I

mutant of yeast, which is defective in vesicle transport between the TGN and the plasma

membrane. Thus transport of the mutated DPAPA to vacuole can't be via the plasma

membrane, followed by endocytosis (Roberts et aI., 1992).

Studies on the a-tonoplast intrinsic protein (a-TIP) of plant cells have led to the

suggestion that one of the six transmembrane domains may contain a vacuolar sorting

signal. This was because mutants in which the other regions were deleted all

accumulated in the tonoplasts. This conclusion did not exclude the possibility of default

secretion of a-TIP (Hofte and Chrispeels, 1992). A recent report from the same

laboratory indicated that TIPs and soluble vacuolar proteins are transported to vacuoles through different paths (Gomez and Chrispeels, 1993). The transport of TIPs to vacuoles 206 was found to be faster than that of the soluble vacuolar protein phytohemagglutinin

(PHA) under normal conditions, and was not blocked by Brefeldin A or monensin

(Gomez and Chrispeels, 1993). In contrast, transport of PHA was blocked by either drug; processing of high mannose side chain to complex form was both retarded and incomplete in the presence of either drug, and missorting and secretion of PHA occurred in the presence of monensin (Gomez and Chrispeels, 1993). It is unclear, therefore, as to whether an alternative default pathway might exist in plant cells for some membrane proteins. Preliminary results obtained concerning the transport of Gm 1 imply that such an alternative default pathway might transport plant membrane proteins to the tonoplast.

The recombinant protein Gm 1 was constructed by swapping the transmembrane domain of vesicular stomatatis viral glycoprotein (VSVG) with the first transmembrane

(ml) domain of coronavirus infectious bronchitis virus M glycoprotein (Swift and

Machamer, 1991). VSVG has been reported to be transported from its ER site of biosynthesis to the plasma membrane by a default pathway in the mammalian cells (Guan and Rose, 1984; Machamer and Rose, 1988; Rose and Doms, 1988; Pryer et aI., 1992).

The ml domain was revealed to contain information necessary for the retention of proteins in the cis compartment of the mammalian Golgi (Machamer et aI., 1990; Swift and Machamer, 1991). Weisz et al. (1993) suggested that the mechanism underlying the retention ofGm 1 within the mammalian Golgi involves its oligomerization. Our expressing Gm 1 in plant cells was initially intended to assess whether the m 1 domain "

207

conferred Oolgi retention within plant cells, and to compare the fate of Om 1 with those

of other membrane proteins bearing ml domains (such as IBV M, IBV M-OUS, and IBV

miLl-OUS). The results presented in Chapter 3 indicated that the Om I glycoprotein is

well-expressed in transfected protoplasts and in transgenic plants, and is targeted into the

plant endomembrane-secretory pathway. However, the targeting behavior of Om I

within plant cells raised questions concerning its expected Oolgi localization (Chapter 3).

This Chapter reports the results of further dissection of the subcellular location of Om 1,

and discusses the implication of these results with respect to the transport of the plant

vacuolar membrane proteins.

MATERIALS AND METHODS

Plant materials and DNA manipulation. These were as described in Chapter 3.

Plasmids pFG 21 and pFG 23 were respectively used for protoplast transfection and

transformation of tobacco leaf-discs.

Protoplast isolation and transfection. This was done as reported in Chapter 3.

Sample homogenization. and differential and isopycnic gradient centrifugation.

These procedures were the same as described in Chapter 4.

In vjvo incubation with tunicamycin and protease K. The transfected protoplasts

were incubated in a medium containing 1 x NTTO nutrient medium (Appendix 1), 0.154 208

M mannitol, 0.38 M glucose, 75 mg/L ampicillin, pH 5.7, and plus 20 Ilg/mL

tunicamycin or 1 mg/mL protease K for various time intervals.

Tonoplast isolation. Vacuoles were isolated from the protoplasts by osmotic and

6 heat shock as described by Gomez and Chrispeels (1993). The protoplasts (- 5 x 10 ) were washed with IX K mediunl (Galbraith et aI., 1992), and were resuspended in 5 mL prewarmed (42°C) lysis medium (0.2 M mannitol, 10% Ficoll-400, 20 mM EOTA, 2 mM OTT, 5 mM HEPES and 1 mM PMSF, pH 8.0) by pipeting twice up-and-down using a wide-bore Pasteur pipet. The mixture was placed on ice after incubation at room temperature for 5 minutes. The mixture was overlaid sequentially with 3 mL of an ice-cold solution comprising of 1 volume lysis medium and 1.5 volumes of vacuole buffer, and then with 0.5 mL of vacuole buffer (0.6 M betaine, 10 mM HEPES, pH 7.5, containing 1 Ilg/mL each of leupeptin, aprotinin, and pepstatin A). The gradient was centrifuged at 1,500 g for 20 minutes at 10°C. Intact vacuoles were collected in a minimal volume at the interface between the first and second gradient layers. The yield of vacuoles was determined by haemocytometry after staining with neutral red, which involved the addition of 10 ilL of 1% neutral red (dissolved in vacuole buffer, filtered through a 0.22 Ilm Millipore filter) to 1 mL of the vacuole sample. The vacuole samples were stored at -70°C until use. Contamination of the isolated vacuoles by the ER was estimated using Western blotting with anti-BiP (an ER-resident protein) antibodies.

Gel electrophoresis and Western blotting. Protein samples were analyzed using one-dimensional SOS-PAGE gel electrophoresis (4% stacking gel, 7.5% running gel), 209

and were probed with preabsorbed polyc1onal rabbit anti-VSV antibodies at a dilution of

1:2,500 (obtained from Dr. Machamer). Pre-absorption of primary antibody was done as

described in Chapter 2. Secondary antibody for Western blotting, goat anti-rabbit IgG

(H+L) alkaline phosphatase conjugate (Sigma Chemical Co., MO, USA), was used at a

dilution of 1: 10,000.

Densitometric analyses. Band intensities in Western blots were determined using

the following equipment and software. The blots were digitized into TIF-fonnatted files,

using a 256-level grayscale, either with a Stratascan™ 7000 Scanning System

(Stratagene, CA, USA) or a Star I Digital CCD Canlera System (Photometries, AZ,

USA). The band intensities were analyzed using Stratascan™ 20 Densitometry Software.

RESULTS

INHIBITION OF GM 1 GLYCOSYLATION BY TUNICAMYCIN

Expression of Gm 1 was detected in transfected protoplasts as well as in transgenic plants (Figurc 3. 5. and Figurc 3. 9.). The existence of a smaller isofonn, having an apparent molecular weight of 52 KDa, was deduced by comparison of the band intensities from transfected protoplasts with the control. This was complicated by the fact that the antibodies directed against Gm 1 also cross-reacted with a non-specific protein band at 52 KDa (Chapter 3). To confinn the existence of the smaller isofonn, 210

experiments were done in which an N-linked glycosylation inhibitor, tunicamycin (final

concentration, 20 jlg/mL), was included in the culture medium of the transfected

protoplasts. This caused equal shifting of both the larger and the smaller isoforms to

lower molecular sizes on the gel (Figure 5. 1). The mobility shifting of both Gm 1

isoforms, consistent with an inhibition ofGm 1 glycosylation, definitively confirmed the

existence ofthe 52 KDa Gm 1 isoform, as its location was clearly moved from beneath the non-specific protein band by the inclusion oftunicamycin (Figure 5. 1). This result confirmed the preliminary conclusion of Chapter 3, that Gm 1 is transported through the endomembrane secretory pathway, and becomes glycosylated at its putative

Asn-X-Thr/Ser sites as reported for native VSVG in plant cells (Galbraith et ai., 1992).

It also incidentally suggests that the non-specific protein is N-glycosylated.

BIOSYNTHETIC KINETICS OF THE LARGER AND SMALLER ISOFORMS OF GM 1

The rate of biosynthesis of the Gm 1 isoforms were analyzed as a function of incubation time after protoplast transfection, in order to examine the rate of conversion of the larger isoform to the smaller one. This conversion was quicker (less than 8 hours) than that reported for VSVG (Galbraith et ai., 1992), and that described for IBV M-GUS

(Chapter 4). Within 8 hours after transfection, both the larger and the smaller isoforms had already accumulated to near maximal amounts (Figure 5. 2). Interestingly, the 211

Tunicamycin + +

61 KDa~- 52 KDa:+--

Figure 5. 1. Effect of in vivo tunicamycin incubation of protoplasts transfected with pFG 21. Viable transfected protoplasts, were incubated for 24 hours in the presence oftunicamycin (20 )lg/mL), were purified via sucrose floatation, and were pelleted and resuspended in 2X SDS-sanlple buffer. Protein samples, equivalent to 2.0 x lOs protoplasts, were loaded onto a one-dimensional SDS-PAGE gel (4% stacking gel,

7.5% running gel). For antibody probing, preabsorbed polyclonal rabbit anti-VSV antibody was used at a dilution of I :2,500. Goat anti-rabbit IgG (H+L) alkaline phosphatase conjugate was used as secondary antibody at a dilution of I: I 0.000. 212

o 8 17 24 48 72 98 hours

---61 KDa '-52 KDa

Figure 5. 2. Kinetics of Om 1 biosynthesis and degradation within transfected protoplasts. Viable transfected protoplasts, 0, 8, 17,24,48, 72, and 98 hours after incubation, were pelleted by centrifugation. and resuspended in 2X SDS-sample buffer.

Protein samples, nomlalized to an equivalent of 2.0 x lOS protoplasts, were loaded onto a one-dimensional SDS-PAOE gel (4% stacking gel, 7.5% running gel). Antibody probing was the same as described under "MATERIAL AND METHODS". 213

non-specific 52 KDa protein band was only detected at a trace amounts at the zero hour time point (a sample of the isolated protoplasts that had not been transfected). This suggests that the non-specific protein band is normally present in plant cells at trace levels, but increases to a substantial amount only after the polyethyleneglycol treatment

(Figure 3. 5, Figure 5. 2, and Figure 5. 1). The larger isoform of Gm 1 disappeared completely only after 72 hours incubation. This stability is similar to that of the VSVG larger isoform (71 hours; Galbraith et aI., 1992). In order to correct for the interference between the non-specific protein band and the smaller isoform, a partitioning ratio between the smaller isoform (75%) and the non-specific protein band (25%) was derived from the Western blot loaded with protein samples normalized to each well (Figure 3. 5).

The amount of the smaller isoform was then estimated and plotted in Figure 5. 3 following subtraction of an intensity value corresponding to the amount of the non­ specific protein band. The accumulation rates of both the larger and the smaller isoforms essentially overlapped up to 17 hours after transfection (Figure 5. 3). This suggests that the conversion from the larger isoform to the smaller one is very quick and is not delayed as seen for VSVG (Galbraith et aI., 1992). However, the highest accumulation of the

Gm I smaller isoform occurred at 24 hours after transfection, which is 7 hours later than that of the larger one (Figure 5. 3). The degradation rate of the smaller isoform was also observed to be slower than that of the larger one. The smaller isoform was still present after 72 hours of incubation, but disappeared around 98 hours after I'

214

>. 'iii- =~ -= -"0 co:= =: 15 "0 ~ .!:! C; ...e Q Z 10

o 8 17 24 48 72 98 Hours after Transfection

Figurc 5. 3. Relative rates of Gm 1 biosynthesis and degradation. The band

intensities of the Western blot in Figurc 5. 2 were digitized, and nonnalized using the

band with minimal intensity as baseline. In this plot, the relative intensities of the Gm 1

smaller isofonn were calculated after subtraction of the intensities of the non-specific

protein band at 52 KDa region. 215

transfection (Figure 5. 3). It did not exhibit the long stable lifetime (over 118 hours) as

reported for VSVG within transfected protoplasts (Galbraith et aI., 1992).

SUCROSE GRADIENT ANALYSES OF ENDOMEMBRANES BEARING GM 1

To assess the sites of localization ofGm 1 within the endomembrane system, the protoplast homogenates (24 hours after transfection) were subjected to differential and sucrose gradient centrifugation. The different membrane fractions were identified by marker enzyme assay. The ER membrane fractions were assayed using antimycin

A-sensitive NADH cytochrome c reductase (Figure 5. 4, NADH Cyto-c). The Golgi membrane fractions were assayed using latent inosine diphosphatase (Figure 5. 4,

IDPase). The membrane fractions containing chlorophyll were detected by chlorophyll autofluorescence (Figure 5. 4, Chlorophyll). The Western blot in Figure 5. 4 clearly demonstrates a lesser intensity of the smaller isoform than that observed in the blots derived from separation of proteins from whole cell homogenates (Figure 5. 1 and

Figure 5. 2). This observation suggests that the smaller isoform may be partially released from the endomembrane. Because of the interference of the non-specific protein band, it is difficult to analyze the distribution pattern of the smaller isoform in sucrose gradients. On the other hand, the distribution of the Gm 1 larger isoform can be 216

Figure 5. 4. Distribution ofGm 1 within isopycnic sucrose gradients. Viable

6 pro top lasts (4.S x 10 ), 24 hours after transfection with plasmid pFG 21, were purified and lysed in 1 mL ice-cold HBST medium. After centrifugation at 1000 g at 4 °C for 20 minutes, the supernatant was collected, adjusted to a sucrose concentration of 10-12%, and loaded onto a 10 mL 20-S0% linear sucrose gradient (buffered with 10 mM Tris, 1 mM EDTA, pH 8.0; cushioned with 1 mL of SO% sucrose at the bottom). After centrifugation at 240,000 g at 4 °C for 18 hours, the gradient was fractionated into seventeen fractions (0.7 mL). The chlorophyll content (arbitrary fluorescence units/O.OS mL fraction), the activities of antimycin A insensitive NADH cytochrome c reductase

(6.Assol4 mini 0.05 mL fraction), and sucrose concentration (% w/w) were measured immediately after fractionation (Galbraith et aI., 1992; Donaldson et aI., 1972). The

activities of latent IDPase (A7l o1hrl 0.1 mL fraction) were measured 2-3 days later, the fractions being stored at -80°C (Gardiner and Chrispeels, 1975). For SDS-PAGE analysis, 0.4 mL aliquots of each fraction were centrifuged for 60 minutes at 4 oC and

60,000 rpm (Beckman TL 100 Ultrafuge with TL 100.2 rotor) after dilution with 1.2 mL of a solution comprising SO mM Tris and 1 mM EDTA, pH 8.0. The membrane pellets were dissolved in 30 j.lL of2X SDS-sample buffer, and were analyzed using one-dimensional SDS-PAGE gels (4% stacking gel, 7.S% running gel). The antibody probing was the same as described in "MATERIALS AND METHODS". 217

~... 1.225 -= 1.175 -...... ,en ~ 1.125 CI) E 1.075 ...C.I CI) 1.025 -

~0 .=.. d

C.I :b 0.125 ->. U 0.075 -..... -<: 0.025 Z

0.5

~ CI) 0.4 ~ c...... - 0.3 0.2 - 3... i: ~

CI) -~ 3: 0 5 10 15 Fraction Number 218

unambiguously examined. The larger isoform clearly cosegregated with membrane

fractions having densities of 1.08 -1.11 g/mL and 1.13 - 1.18 g/mL, which coincide with

the locations of the ER and Golgi marker enzymes, respectively (Figure 5. 4 Western

blot). However, the distribution of Gm 1 larger isoform in the low density range of the

gradient was not identical with that of the ER fractions. The larger isoform was also

present in fractions having sucrose densities of 1.05 - 1.07 g/mL, a likely location for

tonoplast vesicles (Figure 5. 4, Western blot). These analyses are consistent with the

idea that Gm 1 is synthesized on and translocated across the rough ER, trafficked

through the Golgi, then transported to vacuole.

IN VIVO PROTEASE K INCUBATION OF GM 1 TRANsFEcrED PROTOPLASTS

From previous results, it appears that Gm 1 only penetrates as far as the Golgi membrane factions on the sucrose gradient (Figure 5. 4). To evaluate whether Gm 1 reaches the plasma membrane, in vivo protease K incubation of the transgenic protoplasts was also conducted. This relies on the principle that the plasma membrane of viable protoplasts is impermeable to protease K. Thus, Gm 1 expressed by viable protoplasts should only be accessible to exogenous protease K if it is transported to the plasma membrane and its N-terminus (about 80% of whole protein sequence) thereby becomes exposed to the extracellular milieu. The results indicated that Gm 1 within the viable 219

Triton X-IOO - + Protease K + + +

_---61 KDa -"'51 KDa

Figure 5. 5. Protease K susceptibility of Gm 1 within viable transfected protoplasts. Viable protoplasts were purified 24 hours after transfection with pFG 21 plasmid, and incubated with or without protease K (1 mglmL) for 2 hours at 37 °c. Each treatment contained 1.5 x 106 protoplasts. A control treatment (protease K plus 1%

Triton X-I 00) was done on the same number of the protoplasts expressing Gm 1. The viable pro top lasts were purified after the 2-hour incuhation, and resuspended in 2X

SDS-sample buffer. The normalized samples were analyzed in one-dimensional

SDS-PAGE gel (4% stacking gel, 7.5% running gel) and probed with the antibodies as described in "MATERIALS AND METHODS". 220 transgenic protoplasts was not affected by the presence of protease K in the medium

(Figure 5. 5). On the other hand, Om I was completely degraded when protease K was added under conditions in which the integrity of the plasma membrane was compromised by the presence of the detergent Triton X-IOO.

ARRIVAL OF OM I AT THE TONOPLAST MEMBRANE

The above analyses suggest that Om I is transported to the vacuole via the Oolgi directly, rather than through a route involving movement to the plasma membrane followed by endocytosis to the vacuole. To directly ascertain the presence of Om I in vacuole, intact vacuoles of high purity were isolated through sucrose step gradient centrifugation as described by Oomez and Chrispeels (1993). Oel electrophoresis and

Western blotting of the isolated tonoplast membranes showed that Om I was present

(Figure 5. 6, lejl upper panel). The isolated tonoplasts were demonstrated to be free from contamination by ER membranes, since analysis of the blots using an antibody directed against BiP (an ER-resident protein) did not reveal the presence of BiP signals in the isolated tonoplasts (Figure 5. 6, left IOlVer panel). Intact vacuoles were easily detected under the ordinary light microscope after staining with neutral red (Figure 5. 6, right upper panel). 221

II o 2- :...- --r- ,....,... -,.... ,....,- ".... ~ -,.... - Wv""""1"\ w'-' 0

:----61 KDa IIII 1------52 KDa

Figure 5. 6. Western blotting of proteins from intact vacuoles isolated from transgenic plant pFG 13-7. Intact vacuoles and protoplasts were isolated as described under "MATERIALS AND METHODS" (right upper and lower panels. respectively).

The intact pro top lasts and vacuoles were resuspended in 2X SDS-sample buffer.

SDS-PAGE gels (5% stacking gel over 7.5% running gel) were separately loaded with the same amount of the materials from the protoplast and the vacuolar samples. One gel was probed for the presence of Gm 1 (left upper panel). The second (left lower panel) was probed with rabbit anti-BiP polycJonal antibody (1 :5,000) and goat anti-rabbit Ig G alkaline phosphatase conjugate (1: 10,000). 222

PARTIAL DISSOCIATION OF THE SMALLER ISOFORM OF OM 1 FROM THE ENDOMEMBRANES

Previous results indicated a lesser amount of the smaller isoform of Om 1 was

associated with the endomembrane fractions than of the larger isoform. This suggests

that the smaller isoform might have become partially released from the membranes. In

order to further examine this phenomenon, transfected protoplasts expressing Om I was

subfractionated through differential centrifugation, and were analyzed by gel

electrophoresis and Western blotting (Figure s. 7). The blot shows that the larger

isoform of Om 1 is predominantly present in the microsomal fraction, with a lesser

amount in the crude nuclear fraction; it is not found in the soluble cytoplasmic fraction

(Figure S. 7). The smaller isoform was observed in all three subcellular fractions

(Figure S. 7), suggesting that certain amount of smaller isoform had became released

from the endomembrane secretory system. The non-specific protein band found at 52

KDa was not released from the endomembranes (Figure 5. 7). The association of Om 1

with crude nuclei fraction is presumably due to the association of the rough ER

membrane with the nuclei.

DISCUSSION

In plant cells, transport of proteins to vacuole is currently perceived as a process dependent on the presence of a sorting signal. A large amount of evidence, almost all - 2~ 2- 2-- "-',... ,- "-' "-',... ,... - .... c:~ -,­ c- ,...... ~ -.... sa uu uv

----61 KDa ------52 KDa

Figure 5.7. Partial dissociation of the Gm 1 smaller isoform from the endomembranes. Viable transfected protoplasts, 24 hours after incubation, were purified and lysed in ice-cold HBST medium. The lysate was centrifuged at 200 g for 20 minutes to pellet nuclei and heavy debris, designated as nuclei. The supernatant was further centrifuged at 300,000 g for 60 minutes to separate fractions designated as micro somes and cytosol. The nuclear and microsomal pellets were resuspended in 2X SDS-sample buffer. The cytosolic fraction was concentrated through an ArTIicon Centricon P-30 and mixed with an equal volume of 2X SDS-sample buffer. The nuclear, cytosolic, and microsomal fractions (equivalent to 4 x 105 protoplasts), from control and Gm 1 plants, were run on one-dimensional SDS-PAGE gel (4% stacking gel, 7.5% running gel), and after blotting, were probed with the antibodies as described in liMA TERIALS AND

METHODS". 224

from the studies on the transport of soluble vacuolar proteins, supports this concept

(Chrispeels, 1991; Bednarek and Raikhel, 1992; Vitale and Chrispeels, 1992). Vacuolar

sorting signals for plant cells appear quite diverse and do not show consensus. They

either are short oligopeptide sequences, which can be located at any point within the

particular protein, or are composite signals which are assembled from patches scattered

throughout the whole protein sequences, during the process of protein folding (Bednarek

and Raikhel, 1992; Chrispeels and Raikhel, 1992). Sorting signals from vacuolar

membrane proteins have not been studied in great detail, as cloned genes for this class of

proteins are few. One study, involving a-TIP, has raised the issue that sorting of

vacuolar membrane proteins may not be signal-mediated (Hofte and Chrispeels, 1992).

Yeast and plant cells share some common features concerning the molecular

mechanisms governing protein transport to vacuoles. Both plant and yeast utilize short

peptide domains rather than mannose-6-phosphate glycans to divert soluble vacuolar

proteins into vacuoles (Chrispeels and Raikhel, 1992). Yeast mutants defective in the

retention of luminal ER proteins can be complemented by an Arabidopsis receptor for

KDEL-bearing proteins, but not by the mammalian counterpart (Lee et aI., 1993). From

maize and Arabidopsis, two cDNA (yptml and yptm2) were found to be highly homologous to the genes of yeast GTP binding proteins (Ypt), which are involved in vesicle transport and binding in the early secretory pathway (Palme et aI., 1992, 1993).

Functional homologues to yeast Sarlp (GTP binding protein) and Sec12p (membrane protein), involved in the control of vesicle transport from the ER to the Golgi, have been "

225

isolated from Arabidopsis (d'Enfert et aI., 1992). A homologue to yeast Vps 34, a

phosphatidylinositol 3-kinase involved in the sorting of soluble hydrolase to the yeast

vacuole, has also been isolated from Arabidopsis (Welters et aI., 1993, Stack et aI.,

1993). In yeast, recent biochemical and molecular analyses clearly demonstrate that

sorting of soluble and membrane proteins into the yeast vacuole occurs through two

distinct molecular mechanisms. Soluble proteins are transported through a

signal-dependent process, whereas membrane proteins end up at the vacuole through a

signal-independent default pathway (Seeger and Payne, 1992; Roberts et aI., 1992;

Cooper and Bussey, 1992; Bryant and Boyd, 1993; Nothwehr et aI., 1993; Nothwehr

and Stevens, 1994). The question as to whether plants also employ a default pathway to

transport vacuolar membrane proteins remains open. However, some reports indicate

that two different pathways exist to transport soluble and membrane-bound vacuolar

proteins respectively (Gomez and Chrispeels, 1993).

VSVG, which differs from Gm 1 only in the transmembrane domain, was

postulated to arrive at the plant plasma membrane, at which it becomes converted into

two isoforms (G6\ and G52 ) by selective proteolysis (Galbraith et aI., 1992). Gm 1

expressed in plant cells (Figure 5. 1) also displays two isoforms, which are very similar

to those ofVSVG in plant cells (Galbraith et aI., 1992). Gm 1 expression in plant cells is

different to that (a single protein band is detected) in the transfected mammalian cells

(Swift and Machamer, 1991). However, since Gm 1 was shown not to reach the plasma

membrane using two separate experimental methods (Figure 5. 4 and Figure 5. 5), this I.

226

argues that the conversion of the larger to the smaller Gm 1 isoform occurs elsewhere,

either in the ER or in the vacuole. In this regard, both compartments are known to be rich

in proteolytic enzymes. Furthermore, comparison of the degradation rates between Gm

1 smaller isoform and Gn ofVSVG revealed that the Gm 1 smaller isoform was less

stable within plant cells and became released from the endomembranes (Figure 5. 3 and

Figure 5. 7). Gn of VSVG maintains a constant level within plant cells over at least 118

hours (Galbraith et aI., 1992), whereas the Gm 1 smaller isoform showed a constant rate

of degradation, and disappeared almost completely within 100 hours (Figure 5. 3). In

addition, Gm 1 was detected in the highly purified tonoplasts (Figure 5. 6), but not in

the plasma membrane (Figure 5. 4 and Figure 5. 5). All these results suggest that Gm 1

transport is different from that ofVSVG. It is possible that transport ofGm 1 to the

vacuole might be through an ER-Golgi-vacuole route, rather than the ER-Golgi-plasma

membrane/endocytosis route postulated for VSVG (Galbraith et al. 1992).

The transmembrane domain of Gm 1 differs from that of plant TIPs, except for

having similar hydrophobicity. Based on the conventional model of bulk flow, we would

expect Gm 1, if not retained in the plant Golgi, to be transported to the plasma

membrane. However, our data suggest a different scenario; Gm 1 appears to be targeted

to the tonoplasts, implying that an alternative default pathway might exist for the

transport of plant membrane proteins to vacuoles. Confirmation of this alternative default

pathway in plant cells will require use of native membrane protein probes isolated from

the compartments of the early secretory pathway in plant cells. The best candidates "

227

would be membrane proteins residing in the ER or Golgi membranes, such as the KDEL

receptor, HMG-CoA reductase, or calnexin (Lee et aI., 1993; Huang et aI, 1993; Enjuto

et aI., 1994). Mutational analyses of those membrane proteins to define their retentional

sequences should clearly confirm or disprove the existence of an alternative default

pathway for plant membrane proteins. 228

CHAPTER 6

CONCLUSIONS AND PERSPECTIVES

This dissertation project was originally designed with two major objectives in

mind. The first objective was to improve our understanding of how membrane proteins

are sorted within the plant endomembrane-secretory system. In recent years, sorting and

retention of membrane proteins has become an important research area in the study of

protein targeting, particularly in mammalian and yeast systems. Recent discoveries

have significantly revised our understanding of protein traffic within the endomembrane

secretory pathway (Hopkins, 1992; Luzio and Banting, 1993; Pelham and Munro, 1993;

Nothwehr and Stevens, 1994). Similar investigations within plant cells have been quite

limited, and only a few examples appear in the literature (Hofte and Chrispeels, 1992;

Galbraith et a!., 1992; Gomez and Chrispeels, 1993). This project would further extend

the scope of investigations on plant membrane protein transport using heterologous

probes, as established by Galbraith et a!. (1992).

The second objective of the project was to investigate the possibility of establishing a heterologous marker system for the plant Golgi, with the idea of using this marker for the immuno-purification of the plant Golgi. This would enable us to analyze the function of the Golgi apparatus at the molecular level, this being fundamental to our understanding of molecular and cellular mechanisms underlying the biosynthesis, modification, and sorting of glycoproteins and polysaccharides within plant cells. Very 229

little is known about the molecular composition of the plant Golgi. None of the

Golgi-resident enzymes or proteins have been isolated and characterized from plant cells.

Little information is available about the molecular organization of Golgi enzymes, or

about the biosynthesis, assembly, and targeting of complex polysaccharides for cell

wall. This situation is partly due to difficulty in obtaining highly-purified Golgi in

sufficient quantity, and an inability to subfractionate the plant Golgi into its various

subcompartments. The second objective was specifically designed to solve those

problems and more emphasized the development of novel techniques and methods for

plant cell biology.

A heterologous probe approach was adopted in the project to achieve these two

objectives. Investigations were specifically conducted on the expression and localization

of SYNVG, VSVG, the IBV M glycoprotein, and their derivatives within plant cells.

The project resulted in three important findings concerning the plant endomembrane­

secretory pathway: (1) the artificial creation of a novel membrane organelle by expressing

the IBV M-GUS chimaeric protein, (2) the possible existence of an alternative default

pathway for membrane proteins within plant cells, and (3) the observation that plants and

mammals share similar mechanisms regarding ER targeting and insertion for membrane

proteins, but they diverge concerning the retention of membrane proteins within specific

compartments of the endomembrane-secretory system. These findings clearly advance our understanding concerning mechanisms that govern membrane protein transport within plant cells. "

230

Over the period of three years occupied in pursuit of the plant Golgi, my

investigations have indicated that the immuno-purification of the plant Golgi, or its

sub compartments is not feasible as originally proposed. The IBV M glycoprotein itself

was difficult to detect when expressed within plant cells (Chapter 3). By way of contrast,

two derivatives of the IBV M glycoprotein, IBV M-GUS and Gm 1, were well expressed

and could be easily detected within plant cells. However, they were not targeted to the

expected destination (the Golgi apparatus or its subcompartments). One (IBV M-GUS)

heavily accumulated within extension of the plant ER membranes, resulting in a novel

organelle (Chapter 4). The other (Gm 1) was efficiently and unexpectedly transported

into the plant tonoplast (Chapter 5).

Artificial organelles have been reported from mammalian, yeast, and bacterial

cells, typically occurring when certain membrane proteins are over-expressed (Weiner et

aI., 1984; von Meyenburg et al.. 1984; Jingami et aI., 1987; Wright et aI., 1988; Schunck

et aI., 1991; Vergeres et aI., 1993; Calafat et aI., 1994). To the best of our knowledge,

our discovery of a novel artificial organelle (E-whorl) is the first case involving plant

cells. This discovery provides an important clue concerning the formation of artificial

organelles. Clearly, molecular interactions amongst membrane proteins is the most­

likely cause for the formation of the novel organelle. Our observations also include that

proteins are not automatically degraded when they are trapped in the early secretory

pathway. Degradation may require a specific physical confirmation of proteins, for

example exposure of serines or other amino acid residues to proteases. The potential 231 commercial usage of E-whorl could involve production of important recombinant proteins, that otherwise unstable within, or toxic to, host cells. Other possible applications of E-whorls include investigation of organelle biogenesis and artificial engineering of novel organelles.

Since all of the heterologous membrane proteins employed in this study were shown to be associated with the plant endomembrane secretory system, this suggests that mammalian and plant cells use a well-conserved molecular mechanism to target membrane proteins to the ER membrane and to anchor them within the membrane.

Conservation of this function is supported by the following evidence. Isolated plant signal recognition particles (SRPs) resemble their mammalian counterparts (Prehn et at.,

1987; Campos et ai., 1988). The recognition of either signal peptides or transmembrane domains by the methionine-rich domains within SRPs is not dependent on the specific amino acid sequences, but rather on the hydrophobicity of the sequences (Rapoport,

1992). The expression of the soluble secretory proteins between different hosts, for example, plant secretory proteins in yeast and mammalian cells, or mammalian secretory proteins in plant cells, demonstrates that ER targeting and translocation of secretory proteins is well conserved among yeast, plant and mammalian cells (Bednarek and

Raikhel, 1992). In addition, our results also indicate that targeting of membrane proteins to specific secretory compartments may be not well conserved between mammals and plants. Thus, similar molecular mechanisms may operate for the retention and sorting of 232

membrane proteins in mammalian and plant cells, whereas they appear to diverge at the

level of specific details.

In the near future, the field of membrane protein retention and sorting in plant

cells will be wldoubtedly advanced by the analyses on the cis-elements of

retention/sorting for the recently-isolated integral membrane proteins, by the

complementation analyses of plant cDNA in the ever-increasing stocks of the yeast

secretory mutants, and by the development of an Arabidopsis genetic system for the

secretory pathway. In the long term, work should focus on the development of the in

vitro reconstitution systems for plant retention and sorting mechanisms. So far in plant

cells, successful in vitro assays, such as ER targeting and translocation, and

reconstitution of vesicle trafficking between donor and acceptor compartments, are quite

limited. More successful and sophisticated in vitro assays will enhance our

understanding of the requirements for the retention and sorting of proteins, and will

further expand our ability to explore protein transport in the plant

endomembrane-secretory pathway. For eXanlple, in manlmalian cells, a cell-free system

was successfully used to isolate non-clathrin coated vesicles and their coat proteins responsible for intra-Golgi transport (Malhotra et aI., 1989; Serafini et aI., 1991 b).

Powerful cell-free systems offer advantages, such as anlplifying the transient elements that are almost undetectable in vivo, and exploring the roles of specific factors in a transport step (i.e. cytosolic factors and membrane constituents in vesicle formation and 233

docking). In vitro reproduction of complex molecular and cellular events is one of the most important goals of cell biologists. 234

APPENDIX I: NTTO NUTRIENT STOCK (lOX)

NH4N03 8.2500 giL KN03 9.5000 gIL CaCl2 2Hp 2.2000 gIL MgS04 7H20 1.8500 gIL KH2P04 0.8500 gil ZnS04 7Hp 0.0178 gIL H3B03 0.0100 gIL Inositol 1.0000 gIL Calcium Pantothenate 0.0100 gIL Nicotinic Acid 0.0100 giL Pyrodoxine-HCI 0.0100 gIL Thiamine-HCI 0.0100 gIL NITO Salt Stock (lOX) 1.000 mLIL Biotin Stock (0.1 mg/mL) 1.000 mLlL

NTTO Salt Stock (lOX)

CuSO" 5Hp 300.0 mg/L

AICl3 300.0 mglL NiCl3 HP 300.0 mg/L KI 100.0 mg/L MnSO"Hp 758.0 mg/L

Note: The Biotin Stock is made by dissolving the biotin first in a few mL of 0.1 N NaOH, then adjusting to the final volume with deionized water. I'

235

APPENDIX II: MINIMAL T MEDIUM WITH KANAMYCIN

Minimal T Salt Stock (20X) 50 mLiL Minimal T Buffer Stock (20X) 50 mLiL Sucrose 5 giL Bactoagar 16 gIL Kanamycin Sulfate Stock (100 mg/mL) 1 mLIL (added after autoclaving)

Minimal T Salt Stock (20X)

NH4CI 20.000 giL MgS04 7HP 4.000 giL MnCI 2 4H20 0.060 gIL CaCI 2 0.020 giL FeS04 7Hp 0.100 gIL

Minimal T Buffer Stock (20X)

210 gIL 90 giL 236

APPENDIX III: TOBACCO FEEDER-CELL PLATE MEDIUM

W38 Solid Medium (sterile) 100 mL W38 Cells -4 g (the log-phase grown tobacco suspension cells are collected and mixed quickly with the liquid medium maintaining at 35-42 0c).

Pour 10 mL of the above mixture into a 15 x 50 mm plate before the medium is solidified. The feeder-cell plates are wrapped with parafilm and are incubated in the growth incubator (with light) for 48 hours before use.

W38 Solid Medium (PH 5.8)

MS medium powder (Gibco) 4.3 giL Sucrose 30.0 giL myo-Inositol 0.1 giL Thiamine-HCI Stock (0.4 mg/mL) 1.0 mLiL 6-Benzylaminopurine Stock (0.1 mg/mL) 1.0 mLiL 1-Naphthaleneacetic Acid (1.0 mg/L) 3.0 mLiL Low Melting Point Agarose 8.0 giL (using phytoagar for normal plates)

Note: W38 solid medium is essentially the same as the solid MS3+ medium (described under "Generation and Maintenance of Tobacco Callus and Suspension Cell Cultures" in Chapter 2; except with a higher concentration of I-naphthaleneacetic acid. 237

APPENDIX IV: TOBACCO SHOOTING MEDIUM

MS mediwn powder (Gibco) 4.30 giL Sucrose 30.00 giL 6-Benzylaminopurine Stock (0.1 mg/mL) 9.00 mLiL 1-Naphthaleneacetic Acid (1.0 mg/L) 0.09 mLiL B-5 Vitamin Stock 1.00 mLiL Phytoagar 8.00 giL pH 5.8

B-5 Vitamin Stock

myo-Inositol 10.000 giL Thiamine-HCl 1.000 giL Pyridoxine 0.100 giL Nicotinic Acid 0.100 giL 238

APPENDIX V: REAGENT A FOR MICRO-LOWRY ASSAY

10% Sodium Dodecyl Sulfate 1 part O.S N NaOH 1 part eTe Stock 1 part Hp 1 part

eTe Stock

euso4 SHP 1.0 giL K2e 2H40 6 1I2Hp (neutral) 2.0 gIL N~eo3 100.0 gIL

Note: eTe Stock is made by dissolving the first two chemicals to double-strength concentrations, then adding an equal volume of double-strength N~eo3 solution. eTe Stock is stable for at least two months at room temperature.

Reagent A is stable for at least two weeks at room temperature. 239

APPENDIX VI: FERROUS SULFATE-AMMONIUM MOLYBDATE SOLUTION

This solution is freshly made before use.

Ammonium Molybdate Stock (10%) 0.10 mLlmL FeS047Hp 0.05 g/mL

Ammonium Molybdate Stock (10%)

10.0 g of (NH4)6M02024 4H20 are added to about 60 mL 5 M H2S04 under the constant stirring. When the ammonium molybdate is completely dissolved, the final volume is adjusted to 100 mL with HP in a volumetric flask. 240

REFERENCES

Ahle S. and Ungewickell E. (1986). Purification and properties of a new clathrin assembly protein. EMBO J. 5: 3143-3149.

Ahle S. and Ungewickell E. (1990). Auxilin, a newly identified clathrin-assocaited protein in coated vesicles from bovine brain. J. Cell Bioi. 111: 19-30.

Alberts B., Bray D., Lewis J., Ratt M., Roberts K. and Watson J.D. (1994). Molecular Biology of the Cell (3rd ed.). Chapter 12. Intracellular Compartments and Protein Sorting. pp. 551-598. Garland Publishing, Inc., New York.

Ali M.S., Mitsui T. and Almzawa T. (1986). Golgi-specific localization of transglycosylases engaged in glycoprotein biosynthesis in suspenson-cultured cells of sycanlore (AceI' pseudoplatanus L.). Arch. Biochem. Biophy. 251: 421-431.

Allan V.J. and Kreis T.E. (1986). A microtube-binding protein associated with membranes of the Goigi apparatus. J. Cell Bioi. 103: 2229-2239.

Armstrong J., Niemann H., Smeekens S., Rottier P. and Warren G. (1984). Sequence and topology of a model intracellular membrane protein, E 1 glycoprotein, from a coronavirus. Nature 308: 751-752.

Anderson R.G.W., Kamen B.A., Rothberg K.G. and Lacey S.W. (1992). Potocytosis­ sequestration and transport of small molecules by caveolae. Science 255: 410-411.

Aoki D., Lee N., Yamaguchi N., Dubois C. and Fulmda M.N. (1992). Goigi retention of a trans-Golgi membrane-protein, galactosyltransferase, requires cysteine and histidine residues within the membrane-anchoring domain. Proc. Natl. Acad. Sci. USA 89: 4319-4323.

Bacon B.A., Slllminen A., Ruohola H., Novick P. and Ferro-Novick S. (1989). The GTP-binding protein ypt 1 is required for transport in vitro: The Golgi apparatus is defective in ypt 1 mutants. J. Cell Bioi. 109: 1015-1022.

Baker D., Wuestshube L.J., Schekman R., Botstein D. and Segev N. (1990). 2 GTP-binding ypt 1 protein and Ca + function independently in a cell-free protein transport reaction. Proc. Natl. A cad. Sci. USA 87: 355-359.

Baker K.P. and Schatz G. (1991). Mitochondrial proteins essential for viability mediate protein import into yeast mitochondria. Nature 349: 205-208. I'

241

Bardwell J.C.A. and Beckwith J. (1993). The bonds that tie: catalyzed disulfide bond formation. Cell 74:769-771

Baumann G., Raschke E., Bevan M. and Schoffl F. (1987). Functional analysis of sequences required for transcriptional activation of soybean heat shock gene in transgenic tobacco plants. EMBO J. 6: 1161-1166.

Bednarek S. Y. and Raikhel N. V. (1992). Intracellular trafficking of secretory proteins. Plant Mol. Bioi. 20: 133-150.

Bennett A.B. and Osteryoung K.W. (1991). Protein transport and targeting within the endomembrane system of plants. in "Plant Biotechnology. Vol. 1. Plant Genetic Engineering" D. Grierson ed. pp. 199-237. Blackie & Son Ltd., London.

Berger M. and Schmidt M.F. (1985). Protein fatty acyltransferase is located in the rough endoplasmic reticulum. FEBS Lett. 187: 289-294.

Bharathan G., Lambert G. and Galbraith D.W. (1994). Nuclear DNA content of monocotyledons and related taxa. Am. J. Bot. 81: 381-386.

Bishop W.R. and Bell R.M. (1988). Assembly of phospholipids into cellular membranes: biosynthesis, transmembrane movement and intracellular translocation. Ann. Rev. Cell Bioi. 4: 579-610.

Bonifacino J.S., Cosson 0., and Klausener R.D. (1990). Colocalized transmembrane determinants for ER degradation and subunit assembly explain the intracellular fate of TCR chains. Cell 63: 503-513.

Bos K, Wraight C. and Stanley KK. (1993) TGN38 is maintained in the Trans-Golgi Network by a tyrosine-containing motif in the cytoplasmic domain. EMBO J. 12: 2219-2228.

Brands R., Snider M., Yukinobu H., Park S., Gelboin H. and Rothman J. (1985). Retention of membrane proteins by the endoplasmic reticulum. J. Cell Bioi. 101: 1724-1734.

Bretscher M.S. and Munro S. (1993). Cholesterol and the Golgi apparatus. Science 261 : 1280-1281.

Briskin D.P., Leonard R.T. and Hodges T.K. (1987). Isolation of the plasma membrane: membrane markers and general principles. Method. Enzymol. 148: 542-559. I'.

242

Brown W.J. and Farquhar M.G. (1984). Accumulation of coated vesicles bearing mannose 6-phosphate receptors for lysosomal enzymes in the Golgi region of I cell fibroblasts. Proc. NaIl. Acad. Sci. USA 81: 5135-5139.

Brummell D.A., Camirand A. and Maclachlan G.A. (1990). Differential distribution ofxyloglucan glycosyltransferases in pea Goigi dictyosomes and secretory vesicles. J. Cell Sci. 96: 705-710.

Bryant N.J. and Boyd A. (1993). Immunoisolation ofkex2p-containing organelles from yeast demonstrates colocalization of 3 processing proteinases to a single Goigi compartment. J. Cell Sci. 106: 815-822.

Burgess T.L. and Kelly R.B. (1987). Constitutive and regulated secretion of proteins. Annu. Rev. Cell Bioi. 3: 243-293.

Burke J., Pettitt J. M., Schachter H., Sarkar M. and Gleeson P.A. (1992). The transmembrane and flanking sequence of beta-I, 2-N-acetylglucosaminyltransferase I specify medial-Golgi localization. J. Bioi. Chem. 267: 24433-24440

Burke J., Pettitt J.M., Humphris D. and Gleeson P.A. (1994). Medial-Golgi retention ofN-acetylglucosaminyltransferase-I - contribution from all domains of the enzyme. J. Bioi. Chem. 269: 12049-12059.

Busch M.B. and Sievers A. (1993). Membrane traffic from the endoplasmic-reticulum to 2 the Golgi-apparatus is disturbed by an inhibitor of the Ca +-ATPase in the ER. Proloplasma 177: 23-31.

Calafat J., Nijenhuis M., Janssen H., Tulp A., Dusseljee S., Wulbbolts R. and Neefjes J. (1994). Major histocompatibility complex class II molecules induce the formation of endocytic MIIC-like structures. J. Cell Bioi. 126: 967-977.

Campos N., Palau .J., Torrent M. and Ludevid D. (1988). Signal recognition particles are present in maize. J. Bioi. Chem. 263: 9646-9650.

Casanova J.E., Mishumi Y., Ikehara Y., Hubbard A.L. and Mostov K.E. (1991). Direct apical sorting of rat-liver dipeptidylpeptidase IV expressed in madin-darby canine kidney-cells. J. Bioi. Chem. 266: 24428-24432.

Ceriotti A., Pedrazzini E., Fabbrini M.S., Zoppe M. Bollini R. and Vitale A. (1991). Expression of the wild-type and mutated vacuolar storage protein phaseolin in Xenopus oocytes reveals relationships between assembly and intracellular transport. Eur. J. 243

Biochem. 202: 959-968.

Chasan R. and Schroeder J.I. (1992). Excitation in plant membrane biology. Plant Cell 4: 1180-1188.

Chin D.J., Luskey K.L., Anderson R.G.W., Faust J.R., Goldstein J.L. and Brown M.S. (1982). Appearance of crystalloid endoplasmic reticulum in compactin-resistant Chinese hamster cells with a 500-fold increase in 3-hydroxy-3-methylglutaryl-coenzyne A reductase. Proc. Natl. Acad. Sci. USA 79: 1185-1189.

Chrispeels M.J. (1985). The roles of the Golgi apparatus in the transport and post-translational modification of vacuolar (protein body) proteins. Oxford SlIr'veys of Plant Mol. & Cell. Bioi. 2: 43-68.

Chrispeels M.J. (1991). Sorting of proteins in the secretory system. Annll. Rev. Plant Physiol. Plant Mol. Bioi. 42: 21-53.

Chrispeels M.J. and Raikhel N.V. (1992). Short peptide domains target proteins to plant vacuoles. Cell 68: 613-616.

Christie S.R., Christie R.G. and Edwardson J.R. (1974). Transmission ofa bacilliform virus of sowthistle and Bidens pilosa. Phytopathol. 64: 840-845.

Coleman J.O.D., Evans D.E., Horsley D. and Hawes C.R. (1991). The molecular structure of plant clathrin and coated vesicles. in "Endocytosis, exocytosis and vesicle traffic in plants", (Ist ed.) C.R. Hawes, J.O.D. Coleman and D.E. Evans eds. pp. 41-64. Cambridge University Press, Cambridge.

Colley K.J., Lee E.U. and Paulson J.C. (1992). The signal anchor and stem regions of the beta-galactoside alpha-2, 6-sialyltransferase may each act to localize the enzyme to the Golgi-apparatus. J. Bioi. Chem. 267: 7784-7793.

Connolly T. and Gilmore R. (1989). The signal-recognition particle receptor mediated the GTP-dependent displacement ofSRP from the signal sequence of the nascent polypeptide. Annll. Rev. Cell 57: 699-610.

Cooper A. and Bussey H. (1992). Yeast kexlp is a Golgi-associated membrane-protein­ deletions in a cytoplasmic targeting domain result in mislocalization to the vacuolar membrane. J. Cell Bioi. 119: 1459-1468.

Coughlan S.J. (1993). Molecular characterization of plant endoplasmic-reticulum. Plant Physiol. 102: 134-134. 244

Craig S., Goodchild D.J. and Hardham A.R. (1979). Structural aspects of protein accumulation in developing pea cotyledons. I. Qualitative and quantitative changes in parenchyma cell vacuoles. Aust. J. Plant Physiol. 6: 81-92.

Craig S. and Staehelin L.A. (1988). High pressure freezing of intact plant tissues. Evaluation and characterization of novel features of the endoplasmic reticulum and associated membrane systems. Eur. J. Cell BioI. 46: 80-93.

Crowley K.S., Reinhart G.D. and Johnson A.E. (1993). The signal sequence moves through a ribosomal tUImel into a noncytoplasmic aqueous environment at the ER membrane early in translocation. Cell 73: 1101-1115.

Dahams N.M., Lobel P. and Kornfeld S. (1989). Mannose 6-phosphate receptors and lysosomal targeting. J. BioI. Chern. 264: 12115-12118.

Dahdal R.Y. and Colley K.J. (1993). Specific sequences in the signal anchor of the beta-galactoside alpha-2, 6-sialyltransferaseare not essential for Golgi localization - membrane flanking sequences may specify Golgi retention. J. BioI. Che,n. 268: 26310-26319.

David V., Hoehstenbach F., Rajagopalan S. and Brenner M.B. (1993) Interaction with newly synthesized and retained proteins in the endoplasmic-reticulum suggests a chaperone function for human integral membrane-protein ip90 (calnexin). J. BioI. Chern. 268: 9585-9592.

De Camilli P., Bcnfenati F. Valtorta F. and Grecngard P. (1990). The synapsins. Annu. Rev. Cell. BioI. 6: 433-460

Delmer D.P. (1987). Cellulose biosynthesis. Annu. Rev. Plant Physiol. 38: 259-290.

Denecke J., Deryeke R. and Botterman J. (1992). Plant and mammalian sorting signals for protein retention in the endoplasmic-reticulum contain a conserved epitope. EMBO J. 11: 2345-2355.

Denfert C., Gensse M. and Gaillardin C. (1992). Fission yeast and a plant have functional homologs of the sari-protein and sec 12-protein involved in ER to Golgi traffic in budding yeast. EMBO.1. 11: 4205-4211.

Deshaies R.J., Sanders S.L., Feldheim, D.A. and Schckman R. (1991). Assembly of Yeast Sec proteins involved in translocation into the endoplasmic reticulum into a membrane-bound multi subunits. Nature 349: 806-808. "

245

Ding B., Turgeon R. and Parthasarathy M. V. (1991). Microfilament organization and distribution in freeze substituted tobacco plant-tissues. Protoplasma 165: 96-105.

Dingwall C. and Laskey R. (1992). The nuclear-membrane. Science 258: 942-947.

d'Enfert C., Gensse M. and Gaillardin C. (1992). Fission yeast and a plant have functional homologues of the SarI and Secl2 proteins involved in ER to Golgi traffic in budding yeast. EMBO J. 11: 4205-4211.

Doms R.W., Russ G. and Yewdell J.W. (1989). Brefeldin A redistributes resident and itinerant Golgi proteins to the endoplasmic reticulum. J. Cell Bioi. 109: 61-72.

Donaldson R. P., Tolbert N. E. and Schnarrenberger C. (1972). Acomparison of microbody membranes with microsomes and mitochondria from plant and animal tissue. Arch. Biochem. Biophy. 152: 199-215.

Driouich A., Zhang G.F. and Stachelin L.A. (1993a). Effect of Brefeldin-A on the structure of the Golgi apparatus and on the synthesis and secretion of proteins and polysaccharides in sycamore maple (AceI' pselldoplatanlls) suspension-culture cells. Plant Physiol. 101: 1363-1373.

Driouich A., Faye L. and Staehelin L.A. (1993b). The plant Golgi-apparatus - a factory for complex polysaccharides and glycoproteins. Trends Biochem. Sci. 18: 210-214.

Drucker M., Hinz G. and Robinson D.G. (1993). ATPases in plant coated vesicles. J. Experimtl. Bot. 44: 283-291.

Elbei A.D. (1987). Inhibitors of the biosynthesis and processing ofN-linked oligosaccharide chains. Annu. Rev. Biochem. 56: 497-534.

Enjuto M., Balcells L., Campos N., CacHes C., Arro M. and Boronat A. (1994). Arabidopsis thaliana contains 2 differentially expressed 3-hydroxy-3-methylglutaryl-CoA reductase genes, which encode microsomal forms of the enzyme. Proc. Natl. Acad. Sci. USA 91: 927-931.

Farquhar M.G. (1985). Progress in unraveling pathways of Golgi traffic. Annll. Rev. Cell Bioi. 1:447-488.

Farrell L.B. and Beachy R.N. (1990). Manipulation of p-glucuronidase for use as a reporter in vacuolar targeting studies. Plant Mol. Bioi. 15: 821-825. 246

Faye L., Sturm A., Bollini R., Vitale A. and Chrispeels M.J. (1986). The position of the oligosaccharide side-chains of phytohemagglutinin and their accessibility to glycosidases determines their subsequent processing in the Golgi. Eur. 1. Biochem. 158: 655-661.

Faye L. and Chrispeels M.J. (1987). Transport and processing of the glycosylated precursor of Concanavalin A in jackbean. Planta 170: 217-224.

Faye L., Johnson K.D., Sturm A. and Chrispeels M.J. (1989). Structure, biosynthesis, and function of asparagine-linked glycans on plant glycoproteins. Physiol. Plant. 75: 309-314.

Fiedler K. Kobayashi T. Kurzehalia T.V. and Simons K. (1993). Glycosphingolipid­ enriched, detergent-insoluble complexes in protein sorting in epithelial-cells. Biochem. 32: 6365-6373.

Fowke L.C., Tanchak M.A. and Galway M.E. (1991). Ultrastructural cytology of the endocytotic pathway in plants. in "Endocytosis, exocytosis and vesicle traffic in plants", (1st ed.) C.R. Hawes, J.O.D. Coleman and D.E. Evans eds. pp. 15-40. Cambridge University Press, Cambridge.

Fox M. H. and Galbraith D. W. (1990). Application of flow cytometry and sorting to higher plant systems. in "Flow Cytometry and Sorting" (2nd ed.) M.R. Melamed, T. Lindmo and M.L. Mendelsolm eds. pp.633-650. Wiley-liss Inc., New York.

Freedman R.B. (1989). Protein disulfide isomerase: multiple roles in the modification of nascent secretory proteins. Cell 57: 1069-1072.

Galbraith D.W., Harkins K.R., Maddox J.M., Ayres N.M., Sharma D.P. and Firoozabady E. (1983). Rapid flow cytometric analysis of the cell cycle in intact plant tissues. Science 200: 1049-1051.

Galbraith D.W., Zeiher C.A., Harkins K.R. and Afonso C.L. (1992). Biosynthesis, processing and targeting of the G-protein of vesicular stomatitis-virus in tobacco protoplasts. Planta 186: 324-336.

Garcia-Bustos J., Heitman J. and Hall M.N. (1991). Nuclear protein localization. Biochim. Biophys. Acta 1071: 83-101.

Gardiner M. and Chrispeels M.J. (1975). Involvement of the Golgi apparatus in the synthesis and secretion of hydroxyproline-rich cell wall glycoprotein. Plant Physiol. 55: 536-541. I'

247

Garoff H. (1985). Using recombinant DNA techniques to study protein targeting in the eucaryotic cell. Annu. Rev. Cell Bioi. 1: 403-445.

Gething M.-J., McCammon K. and Sambrook J. (1986). Expression of wild-type and mutant forms of influenza hemagglutinin: the role offolding intracellular transport. Cell 46: 3600-3607.

Gething M.-J. and Sambrook J. (1992). Protein folding in the cell. Nature 355:33-45.

Gilmore R. (1993). Protein translocation across the endoplasmic reticulum: A tunnel with toll booths at entry and exit. Cell 75: 589-592.

Glick B.J., Beasley E.M. and Schatz, G. (1992). Protein sorting in mitochondria. Trends Biochem. Sci. 17: 453-459.

Glickman J.N., Conibear E. and Pearse B.M.F. (1989). Specificity of binding of clathrin adapters to signals on the mannose 6-phosphate/insulin-like growth factor II receptor. EMBO J. 8: I 041-1047.

Glover L.A. and Lindsay J.G. (1992). Targeting proteins to mitochondria - a current overview. Biochem. J. 284: 609-620.

Goldberg K.-B., Modrell B., Hillman B.I., Heaton L.A., Choi T.-J. and Jackson A.O. (1991) StlUcture of the glyco ... rotein gene of sonchus yellow net virus, a plant rhabdovilUs. Viro/. 185: 32-38

Goldstein J.L., Anderson R.G.W. and Brown M.S. (1979). Coated pits, coated vesicles, and receptor-mediated endocytosis. Nature 279: 679-685.

Goldstein J.L., Brown M.S., Anderson R.G.W., Russell D.W. and Schneider W.J. (1985). Receptor-mediated endocytosis: concepts emerging from the LDL receptor system. Annu. Rev. Cell Bioi. I: 1-39.

Gomez L. and Chrispeels M.J. (1993). Tonoplast and soluble proteins are targeted by different mechanisms. Plant Cell 5: 1113-1124.

Gomez L. and Chrispeels M.J. (1994). Complementation of an Arabidopsis-thaliana mutant that lacks complex asparagine-linked glycans with the human cDNA-encoding N-acetylglucosaminyltransferase-1. Proc. Nat!. Acad. Sci. USA 91:1829-1833.

Gorlich D., Hartmann E., Prehn S. and Rapoport T.A. (1992a). A protein of the 248

endoplasmic-reticulum involved early in polypeptide translocation. Nature 357: 47-52.

Gorlich D., Prehn S., Hartmann E., Kalies KU. and Rapoport T.A. (1992b). A mammalian homolog of sec61 p is associated with ribosomes and nascent polypeptides during translocation. Cell 71 : 489-503.

Gorlich D. and Rapoport T.A. (1993). Protein translocation into proteoliposomes reconstituted from purified components of the endoplasmic-reticulum membrane. Cell 75: 615-630.

Grabski S., de Feijter A.W. and Schindler M. (1993). Endoplasmic reticulum forms a dynamic continuum for lipid diffusion between contiguous soybean root cells. Plant Cell 5: 25-38.

Griffiths G. and Simons K. (1986). The Trans-Golgi Network: sorting at the exit site of the Golgi complex. Science 234: 438-443.

Gruenberg J. and Howell K.E. (1989). Membrane traffic in endocytosis: insight from cell-free assays. Annll. Rev. Cell Bioi. 5: 453-482.

Guan J.-L. and Rose J.K (1984). Conversion of a secretory protein into a transmembrane protein results in its transport to the Golgi complex but not to the cell surface. Cell 37: 779-787.

Guerineau F., Woolston S., Brooks L. and Mullineaux P. (1988). An expresson cassette for the targeting foreign proteins into chloroplasts. NlIc:leic Acids Res. 16: 11380.

Hall R.D. (1991 a). The initiation and maintenance of callus Cultures of carrot and tobacco. in "Plant Tissue Culture Manual: Fundamentals and Applications" K. Lindsey ed. pp. A2: 1-19. Kluwer Academic Publishers, Dordrecht, Netherlands.

Hall R.D. ( 1991 b). The initiation and maintenance of plant cell suspension culture. in "Plant Tissue Culture Manual: Fundamentals and Applications" K. Lindsey ed. pp. A3: 1-21. Kluwer Academic Publishers, Dordrecht, Netherlands.

Hardwick KG., Boothroyd J.c., Rudner A.D. and Pelham H.R.B. (1992) Genes that allow yeast-cells to grow in the absence of the HDEL receptor. EMBO J 11: 4187-4195.

Harkins KR., Jefferson R.A., Kavanagh T.A., Bevan M.W. and Galbraith D.W. (1990). Expression of photosynthesis-related gene fusions is restricted by cell type in transgenic plants and transfected protoplasts. Proc. Nail. Acad. Sci. USA 87: 816-820. 249

Harris N. (1986). Organization of the endomembrane system. Annu. Rev. Plant Physiol. 37: 73-92.

Harris N. and Watson M.D. (1991). Vesicle transport to the vacuole and the central role of the Golgi apparatus. in "Endocytosis, exocytosis and vesicle traffic in plants", (1st ed.) C.R. Hawes, J.O.D. Coleman and D.E. Evans eds. pp. 143-164. Cambridge University Press, Cambridge.

Hauri H.-P. and Schweizer A. (1992). The endoplasmic reticulum - Golgi intermediate compartment. Curro Opinion Cell Bioi. 4: 600-608.

Heesen S., Raubut R., Aebersold R., Abelson J., Aebi M. and Clark M.W. (1991). An essential 45 KDa yeast transmembrane protein reacts with anti-nuclear pore antibodies: purification of proteins, immunolocalization and cloning of the gene. Eur. J. Cell Bioi. 56: 8-18.

Hennekes D., Peter M., Weber K. and Nigg E.A. (1993). Phosphorylation on protein

kinase C sites inhibits nuclear import oflamin B2• J. Cell Bioi. 120: 1293-1304.

Herman E.M., Tague B.W., Hoffman L.M., Kjemtrup S.E. and Chrispeels M.J. (1990). Retention of phytohemagglutinin with carboxyterminal tetrapeptide KDEL in the nuclear envelope and endoplasmic reticulum. Planta 182: 305-312.

Herscovics A. and Orlean P. (1993). Glycoprotein biosynthesis in yeast. FASEB.1. 7: 540-550.

Hiatt A., Cafferkey R. and Bowdish K. (1989). Production of antibodies in transgenic plants. Nature 342: 76-78.

High S. (1992). Membrane protein insertion into the endoplasmic reticulum - another channel tunnel. BioEssay 14: 535-540

High S., Flint N. and Dobberstein B. (1991). Requirements for the membrane insertion of signal-anchor type proteins. J. Cell. Bioi. 113 :25-34.

Hinz G., Hoh B., and Robinson D.G. Strategies in the recognition and isolation of storage protein receptors. J. Experimtl. Bot. 44:351-357. 1993

Hirschberg C.B. and Snider M.D. (1987). Topography of glycosylation in the rough endoplasmic reticulum and Golgi apparatus. Annu. Rev. Biochem. 56: 63-87. 250

Hoffman L.M., Donaldson D.O., Bookland R., Rushka K. and Herman E.M. (1987). Synthesis and protein deposition of maize 15-KD Zein in transgenic tobacco seeds. EMBOJ. 6: 3213-3221.

Hofte H., Faye L., Dickinson C., Herman E.M. and Chrispeels M.J. (1992). The protein body proteins phytohemagglutinin and tonoplast intrinsic protein are targeted to vacuoles in leaves of transgenic tobacco. Planta 184: 431-437.

Hofte H. and Chrispeels M.J. (1992). Protein sorting to the vacuolar membrane. Plant Cell 4: 995-1004.

Holwerda B.C., Galvin N.J., Baranski T.J. and Roger J.e. (1990). In vitro processing ofaleurain, a barley vacuolar thio protease. Plant Cell 2: 1091-1106.

Holwerda B.C., Padgett H.S. and Rogers J.e. (1992). Proaleurain vacuolar targeting is mediated by ShOlt contiguous peptide interactions. Plant Cell 4: 307-318.

Hong W.J. and Tang B.L. (1993). Protein trafficking along the exocytotic pathway. BioEssays 15: 231-238.

Hopkins C.R. (1992). Selective-membrane protein trafficking - vectorial flow and filter. Trends Biochem. Sci. 17: 27-32.

Horazdovsky B.F., Busch G.R. and Emr S.D. (1994). Vps21 encodes arabS-like GTP-binding protein that is required for the sorting of yeast vacuolar proteins. EMBO J. 13: 1297-1309.

Horn M.A., Heinstein P.F. and Low P.S. (1989). Receptor-mediated endocytosis in plant cells. Plant Celli: 1003-1009.

Horsch R. and Klee H. (1986). Rapid assay of foreign gene expression in leaf discs transformed by Agl'obactel'ium tumefaciens: Role ofT-DNA borders in the transfer process. Proc. Natl. Acad. Sci. USA 83: 4428-4436

Hortsch M. and Meyer 0.1. (1985). Immunochemical analysis of rough and smooth microsomes from rat liver. Segregation of docking protein in rough membrane. EliI'. J. Biochem. 150: 559-564.

Huang L.-Q., Franklin A.E. and Hoffman N.E. (1993). Primary structure and characterization of an Arabidopsis thaliana calnexin-like protein. J. Bioi. Chem. 268: 6560-6566. I.

251

Humphrey J.S., Peters P.J., Yuan L.C. and Bonifacino J.S. (1993). Localization of TGN38 to the Trans-Golgi Network - involvement of a cytoplasmic tyrosine-containing sequence. J. Cell Bioi. 120: 1123-1135.

Hunt T. (1989). Cytoplasmic anchoring proteins and the control of nuclear localization. Cell 59: 949-951.

Hurtley S.M. and Helenius A. (1989). Protein oligomerization in the endoplasmic reticulum. Annll. Rev. Cell Bioi. 5: 277-307.

Jackson A.O. (1978). Partial characterization of the structural proteins of sonchus yellow net virus. Virol. 87: 172-181.

Jackson A.O. and Christie S.R. (1979). Sonchus yellow net virus. in " Descriptions of Plant Virus". No.205. Commonw. Mycol. Inst., Assoc. Appl. BioI., Kew, Surrey, England.

Jackson A.O., Francki R.I.B. and Zuidema D. (1987). Biology structure and replication of plant rhabdoviruses. in" The Rhabdoviruses". R.R. Wagner ed. pp. 427-508. Plenum Press, New York.

Jackson M. R., Nilsson T. and Peterson P. (1990). Identification of a consensus motif for retention of transmembrane proteins in the endoplasmic reticulum. EMBO J. 9: 3153-3162.

Jamieson J.D. and Palade G. (1967). Intracellular transport of secretory proteins in the pancreatic exocrine cell. I. Role of the peripheral elements of the Golgi complex. J. Cell Bioi. 34: 577-596.

Jans D., Ackermann M., Bischoff J., Beach D. and Peters R. (1991). p34cdc2 mediated phosphorylation at TI24 inhibits nuclear import of SV -40 T antigen proteins. J. Cell Bioi. 115: 1203-1212.

Jefferson R.A. (1985). PhD Thesis. University of Colorado.

Jefferson R.A., Burgess S.M. and Hirsh D. (1986). p-glucuronidase from Escherichia coli as gene-fusion marker. Proc. Natl. Acad. Sci. USA 83: 8447-8451.

Jefferson R.A., Kavanagh T.A. and Bevan M.W. (1987). GUS fusions: p-glucuronidase as a sensitive and versatile gene fusion marker in higher plants. EMBO J. 6: 3901-3907. 252

Jingami H., Brown M.S., Goldstein J.L., Anderson R.G.W. and Luskey K.L. (1987). Partial deletion of membrane-bound domain of 3-hydroxy-3-methyl-glutaryl coenzyme A reductase eliminates sterol-enhanced degradation and prevents formation of crystalloid endoplasmic reticulum .•1. Cell Bioi. 104: 1693-1704.

Johansson I. and Forsman C. (1992). Processing of the chloroplast transit peptide of pea carbonic anhydrase in chloroplasts and in Escherichia coli: identification of two cleavage sites. FEBS Lett. 314: 232-236.

Johnson A.E. (1993). Protein translocation across the ER membrane: a fluorescent light at the end of the tunnel. Trends Biochem. Sci. 18: 456-458.

Jones A.M. and Herman E.M. (1993). KDEL-containing auxin-binding protein is secreted to the plasma-membrane and cell-wall. Plant Physiol. 101 :595-606.

Jones R.W. and Jacl

Jones R.L. and Robinson D.G. (1989). Protein secretion in plants. New Phytol. 111: 567-597.

Jones S.M. (1993). A cytosolic complex ofp62 and rab6 associates with TGN38/41 and is involved in budding of exocytic vesicles from the Trans-Golgi Network. J. Cell Bioi. 122: 775-788.

Kaldi K., Diestelkotter P., Stenbeck G., Auerbach S., Jakie V., Magert H.J., Wieland F.T. and Just W.W. (1993). Membrane topology of the 22-kda integral peroxisomal membrane-protein. FEBS Lett. 315: 217-222.

Katz F. N., Rothman J.E., Lingappa V. R., Blobel, G. and Lodish H. F. (1977). Membrane assembly in vitro: synthesis, glycosylation, and asymmetric insertion of a transmembrane protein. Proc. Natl. Acad. Sci. USA 74:3278-3282.

Keegstra K. and Olsen L.J. (1989). Chloroplastic precursors and their transport across the envelope membranes. Annu. Rev. Plant Physiol. Plant Mol. Bioi. 40: 471-501.

Keith V. (1993). Co-translational protein import into mitochondria: an alternative view. Trends Biochem. Sci. 18: 366-371.

Kellems R.E., Allison V.F. and Butow R.A. (1975). Cytoplasmic type 80S ribosomes associated with yeast mitochondria [Saccharomyces cerevisiae] IV. Attachment of ribosomes to the outer membrane of isolated mitochondria. J. Cell Bioi. 65: 1-14. I,

253

Kelly R.B. (1975). Pathway of protein secretion in eukaryotes. Science 230: 25-32.

Kim W.K., Franceschi V.R. Krishnan H.B. and Okita T.W. (1988). Formation of wheat protein bodies: involvement of the Golgi apparatus in gliadin transport. Planta 176: 173-182.

Kjellbom P., Chory J. and Lamb C.J. (1990). Molecular biology of the plasma membrane - perspectives. in "The Plant Plasma Membrane: Structure, Function and Molecular Biology". C. Larsson and I. M. Moller eds. pp. 377-402. Springer-Verlag, Berlin.

Klausner R.D. (1989). Sorting and traffic in the central vacuolar system. Cell 57: 703-706.

Klausner R.D., Lippincott-Schwartz J. and Bonifacino J.S. (1990). The T-cell antigen receptor: insights into organelle biology. Annll. Rev. Cell Bioi. 6: 403-431.

Klausner R.D., Donaldson D. J. and Lippincott-Schwartz J. (1992). Brefeldin A - insights into the control of membrane traffic and organelle structure. J. Cell. Bioi. 116: 1071-1080.

Klionsky D.J., Nelson H. and Nelson N. (1992). Compartment acidification is required for efficient sorting of proteins to the vacuole in Saccharornyces-cerevisiae. J. Bioi. Chern. 267: 3416-3422.

Kornfeld R. and Kornfeld S. (1985). Assembly of asparagine-linked oligosaccharides. Annll. Rev. Biochern. 54: 631-664.

Kornfeld S. and Mellman I. (1989). The biogenesis of lysosomes. Annll. Rev. Cell Bioi. 5: 485-525.

Kornfeld S. (1992). Structure and function of the mannose 6-phosphate/insulinlike growth factor II receptors. Annll. Rev. Biochern. 61 :307-330.

Krishnan H.B., Franceschi V.L. and Okita T.W. (1986). Immunochemical studies on the role of the Golgi complex in protein-body formation in rice seeds. Planta 169: 471-480.

Kuchler K., Dohlman H. and Thorner J. (1993). The a-factor transporter (ste6 gene-product) and cell polarity in the yeast Saccharomyces-cerevisiae. J. Cell Bioi. 120: 1203-1215. 254

KuKuruzinska M.A., Bergh M.L.E. and Jackson B.J. (1987). Protein glycosylation in yeast. Annu. Rev. Biochem. 56: 915-944.

Ladinsky M. S. and Howell K. E. (1992). The Trans-Goigi Network can be dissected structurally and functionally from the cisternae of the Golgi-complex by Brefeldin-A. Eur. J. Cell Bioi. 59:92-105

Ladinsky M. S. and Howell K. E. (1993). An electron-microscopic study ofTGN38/41 dynamics. J. Cell Sci. S17: 41-47.

Lacmmli U.K. (1970). Cleavage of structural protein during the assembly of the head of bacteriophage T4. Nature 227: 680-685.

Laine A.C., Gomord V. and Faye L. (1991). Xylose-specific antibodies as markers of subcompartmentation of terminal glycosylation in the Goigi apparatus of sycamore cells. FEBS Lett. 295: 179-184.

Lang L., Reitman M., Tang J., Roberts R.M. and Kornfeld S. (1984). Lysosomal enzyme phosphorylation. Recognition of a protein-dependent determinant allows specific phosphorylation of oligosaccharides present on lysosomal enzymes. J. Bioi. Chem. 259: 14663-14671.

Larkins B.A. and Hurkman W.J. (1978). Synthesis and deposition of Zein in protein bodies of maize endosperm. Plant Physiol. 62: 256-263.

Larsson C., Widell S. and Sommarin M. (1988). Inside-out plasma membrane vesicles of high purity obtained by aqueous polymer two-phase partitioning. FEBS Lett. 229: 289-292.

Lee H.I., Gal S., Newman T.C. and Raikhcl N.V. (1993). The Arabidopsis endoplasmic reticulwn retention receptor functions in yeast. Proc. Nat!. Acad. Sci. USA 90: 11433-11437.

Lending C.R., Chesnut R.S., Shaw K.L. and Larkins B.A. (1989). Immunolocalization of avenin and globulin storage proteins in developing endosperm of Avena sativa L. Planta 178: 315-324.

Lending C.R. and Larkins B.A. (1989). Changes in the Zein composition of protein bodies during maize endosperm development. Plant Cell 1: 1011-1023.

Levanony H., Rubin R., Altschuler Y. and Galili G. (1992). Evidence for a novel route 255

of wheat storage proteins to vacuoles. J. Cell BiD!. 119: 1117-1128.

Lewis M.J. and Pelham H.R.B. (1992). Ligand-induced redistribution of a human KDEL receptor from the Golgi complex to the endoplasmic reticulum. Cell 68: 353-364.

Li X., Shou W., Kloc M., Reddy B.A. and Etkin L.D. (1994). Cytoplasmic retention of Xenopus nuclear factor 7 before the mid blastula transition uses a unique anchoring mechanism involving a retention domain and several phosphorylation sites. J. Cell BioI. 124: 7-17.

Li X.-X., Wu Y.-J., Zhang D.-Z., Gillikin J.W., Boston R.S., Franceschi V.n.. and Okita T.W. (1993). Rice prolamine protein body biogenesis: A BiP-mediated process. Science 262: 1054-1056.

Lippincott-Schwartz J., Bonifacino J.S., Yuan L.C. and Klausner R.D. (1990). Degradation from the endoplasmic reticulum: Disposing of newly synthesized proteins. Cell 54: 209-229.

Lippincott-Schwartz J., Donaldson J.G., Schweizer A., Berger E.G., Hauri H.P., Yuan L.C. and Klausner R.D. (1990). Microtube-dependent retrograde transport of proteins into the ER in the presence of Brefeldin A suggests a ER recycling pathway. Cell 60: 821-836.

Lippincott-Schwartz J., Yuan L.C., Tipper c., Amherdt M., Orci L. and Klausner R.D. (1991). Brefeldin A's effects on endosomes, lysosomes and the TGN suggest a general mechanism for regulating organelle structure and membrane traffic. Cell. 67: 601-616.

Lippincott-Schwartz J. (1993). Membrane cycling between the ER and Golgi apparatus and its role in the regulation ofbiosynthetic transport. in "Subcellular Biochemistry. Vol. 21: Endoplasmic Reticulum". N. Borgese and J.R. Harris eds. pp. 95-119. Plenum Press, New York.

Lipsky N.G. and Pagano R.E. (1983). Sphingolipid metabolism in cultured fibroblasts: microscopic and biochemical studies employing a fluorescent ceramide analogue. Proc. Natl. Acad. Sci. USA 80: 2608-2612.

Lipsky N.G. and Pagano R.E. (1983). A vital stain for the Golgi apparatus. Science 228: 745-747.

Lisanti M.P. and Rodriguesz-Boulan E. (1990). Glycophospolipid membrane anchoring provides clues to the mechanism of protein sorting in polarized epithelial cell. Trends 256

Biochem. Sci. 15: 113-118.

Luzio J.P., Brake B., Banting G., Howell KE., BragheUa P. and Stanley K. (1990). Identification, sequencing and expression of an integral membrane protein of the Trans-Golgi Network (TGN38). Biochem. J 270,97-102

Luzio J.P. and Banting G. (1993). Eukaryotic membrane traffic: retrieval and retention mechanisms to achieve organelle residence. Trend\' Biochem. Sci. 18: 395-398.

Machamer C.E. and Rose J.K. (1987). A specific transmembrane domain of a coronavirus E1 glycoprotein is required for its retention in the Golgi region. J Cell Bioi. 105: 1205-1214.

Machamer C.E. and Rose J.K (1988). Influence of new glycosylation sites on expression of the vesicular stomatatis virus G protein at the plasma membrane. J Bioi. Chern. 263: 5948-5954.

Machamer C.E., Mentone S.A., Rose J.K. and Farquhar M.G. (1990). The E 1 glycoprotein of an avian coronavirus is targeted to the cis Golgi complex. Proc Natl. Acad Sci. USA 87: 6944-6948.

Machamer C.E., Grim M.G., Esquela A., Chung S.W., Rolls M., Ryan K and Swift A.M. (1993). Retention of a cis Golgi protein requires polar residues on one face of a predicted alpha-helix in the transmembrane domain. Mol. Bioi. Cell 4: 695-704.

Malhotra V., Serafini T., Orci L., Shepherd J.C. and Rothman .J.E. (1989). Purification of a novel class of coated vesicles mediating biosynthetic protein transport through the Golgi stack. Cell 58: 329-336.

Sambrook J., Fritsch E.F. and Maniatis T. (1989). Molecular cloning: A laboratory manual. 2nd ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York.

Mathews C. K. and van Holde K. E. (1990). Biochemistry. Chapter 27 - Information Decoding: Translation. pp. 954-995. Benjamin/Cummings Publishing Company Inc., New York.

Matile P. (1978). Biochemistry and function of vacuoles. Annu. Rev. Plant Physiol. 29: 193-213.

Matter K., Hunziker W. and Mellman I. (1992). Basolateral sorting of LDL receptor in MOCK cells the cytoplasmic domain contains 2 tyrosine-dependent targeting determinants. Cell 71 :741-753. 257

Matter K., Whitney J.A., Yamamoto E. M. and Mellman I. (1993). Common signals controllow-density-lipoprotein receptor sorting in endosomes and the Golgi-complex of MOCK cells. Cell 74: 1053-1064.

Melancon P., Glick B.S., Malhotra V., Weidman P.J., Serafini T., Gleason M.L., Orci L. and Rothmnn J.E. (1987). Involvement of GTP binding "G" protein in transport through the Golgi stack. Cell 51: 1053-1062.

Mellman I. and Simons K. (1992). The Golgi complex: in vitro veritas? Cell 68: 829-840.

Mellman I., Yamamoto E., Whitney J.A., Kim M., Hunziker W. and Matter K. (1993). Molecular sorting in polarized and non-polarized cells - common problems, common solutions. J. Cell Sci. S 17: 1-7.

Meyer D.J., Afonso C.L. and Galbrnith D. W. (1988). Isolation and characterization of monoclonal antibodies directed against plant plasma membrane and cell wall epitopes: Identification of a monoclonal antibody that recognizes extensin and analysis of the process of epitopes biosynthesis in plant tissues and cell cultures. J. Cell Bioi. 107: 163-175.

Moll T., Tebb G., Surana V., Robitsch H. and Nasmyth K. (1991). The role of phosphorylation and the CDC28 protein kinase in cell-cycle regulated nuclear import of the S. cerevisiae transcription factor SWI5. Cell 66: 743-758.

Moll V.M., Riou G. and Levine A. (1992). Two distinct mechanisms alter p53 in breast cancer: mutation and nuclear exclusion. Proc. Natl. Acad Sci. USA 89: 7262-7266.

Mollenhauer H.H. (1980). Isolation of the Plant Golgi. in "The Biochemistry of Plants" R. Stumpf and R. Conn eds. pp.437-488. Academic Press, New York.

Moore P.J., Darvill A.G., Albershein P. and Staehelin L.A. (1986). Immunogold localization ofxyloglucan and rhamnogalacturonan I in the cell walls of suspension-cultured sycamore cells. Plant Physiol. 82: 787-794.

Moore P.J., Swords K.M.M., Lynch M.A. and Staehelin L.A. (1991). Spatial organization of the assembly pathways of glycoproteins and complex polysaccharides in the Golgi apparatus of plants. J. Cell Bioi. 112: 589-602.

Moore H.P.H., Walker M.D., Lee F. and Kelly R.B. (1983). Expressing a human proinsulin eDNA in a mouse ACTH-secreting cell. Intracellular storage, proteolytic 258

processing and secretion on stimulation. Cell 35: 531-538

Morrc D.J., Mollenhauer H.H. and Bracker C.E. (1971). The origin and continuity of Golgi apparatus. in "Results and Problems in Cell Differentiation. II. Origin and Continuity of Cell Organelles". T. Reinert and H. Ursprung eds. pp. 82-126. Springer-Verlag, Berlin.

Morrc D.J. and Mollenhauer H.H. (1974). The endomembrane concept: a functional integration of endoplasmic reticulum and Golgi apparatus. in "Dynamic Aspects of Plant Ultrastructure" A.W. Robard ed. pp. 84-137. MacGraw-Hill Book Company (UK) Ltd., England.

Morre D.J., Penel C., Morrc D.M., Sandelius A.S. Moreau P. and Andersson B. (1991). Cell-free transfer and sorting of membrane lipids in spinach. Donor and acceptor specificity. Protoplasma 160: 49-64.

Morrc D.J., Keenan T.W. and Morre D.M. (1993). Golgi apparatus isolation and use in cell-free systems: A perspective. Protoplasma 172: 12-26.

Mostov K., Apodaca G., Aroeti B. and Okamoto C. (1992). Plasma-membrane protein sorting in polarized epithelial-cells. 1. Cell Bioi. 116: 577-583.

Mostov K. (1993). Protein traffic in polarized epithelial cells - the polymeric immunoglobulin receptor as a model system. 1. Cell Sci. S 17: 21-26.

Munro S. and Pelham H.R.B. (1987). A C-terminal signal prevents secretion of luminal ER proteins. Cell 48: 899-907.

Munro S. (1991). Sequences within and adjacent to the transmembrane segment of alpha-2,6-sialyltransferase specify Golgi retention. EMBO 1. 10: 3577-3588.

Murashige T. and Skoog F. (1962). A revised medium for rapid growth and bioassay with tobacco tissue cultures. Physiol. Plant. 15: 473-497.

Nakamura K. and Matsuoka K. (1993). Protein targeting to the vacuole in plant-cells. Plant Physiol. 10 1: 1-5.

Nakamura K., Matsuoka K., Mukumoto F. and Watanabe N. (1993). Processing and transport to the vacuole of a precursor to sweet-potato sporamin in transformed tobacco cellline-by-2.1. Experimtl. Bot. 44: 331-338.

Negrutiu I., Shillito R., Potrykus I., Biasini G. and Sala F. (1987). Hybrid genes in the 259 analysis of transformation conditions. Plant Mol. Bioi. 8: 363-373.

Newport J.W. and Forbes D.J. (1987). The nucleus: structure, function, and dynamics. Annu. Rev. Biochem. 56: 535-565.

Nilsson T., Slusarewicz P., Hoe M.H. and Warren G. (1993). Kin recognition - a model for the retention of Golgi enzymes. FEBS Lett. 330: 1-4.

Nilsson T., Hoe M.H., Slusarewicz P., Rabouille C., Watson R., Hunte F. Watzele G. Berger E.G. and Warren G. (1994). Kin recognition between medial Golgi enzymes in HeLa cells. EMBO J. 13: 562-574.

Nishikawa S. and Nakano A. (1993) Identification of a gene required for membrane-protein retention in the early secretory pathway. Proc. Natl. Acad. Sci. USA. 90: 8179-8183

Nothwehr S.F. and Gordon J.I. (1990). Targeting of proteins into the eukaryotic secretory pathway: signal peptide structure/function relationships. BioEssays 12: 479-484.

Nothwehr S.F., Roberts C.J. and Stevens T.n. (1993). Membrane-protein retention in the yeast Golgi apparatus - dipeptidyl aminopeptidase-A is retained by a cytoplasmic signal containing aromatic residues. J. Cell Bioi. 121: 1197-1209.

Nothwehr S.F. and Stevens T.H. (1994). Sorting of membrane proteins in the yeast secretory pathway . .1. Bioi. Chem. 269: 10185-10188

Ohtani T., Galili G., Wallace J.C., Thompson G.A. and Larkins B.A. (1991). Normal and lysine-containing zeins are unstable in transgenic tobacco seeds. Plant Mol. BioI. 16: 117-128.

Olsen L.J., Ettinger W.F., Damsz B., Matsudaira K., Webb M.A. and Harada J.J. (1993). Targeting of glyoxysomal proteins to peroxisomes in leaves and roots of a higher plant. Plant CellS: 941-952.

Oparka K.J., Col, L., Wright K.M., Hawes c.R., Evans D.E. and Coleman J.O.D. (1991). Fluid-phase endocytosis and the subcellular distribution of fluorescent probes in plant cells. in "Endocytosis, exocytosis and vesicle traffic in plants", (1st ed.) C.R. Hawes, J.O.D. Coleman and D.E. Evans eds. pp. 81-102. Cambridge University Press, Cambridge.

Orci L., Brown M.S., Goldstein J.L., Garcia-Segura L.M. and Anderson R.G.W. 260

(1984). Increase in membrane cholesterol: A possible trigger for degradation of HMO CoA reductase and crystalloid endoplasmic reticulum in UT-l Cells. Cell 36: 835-845.

Pagano R.E. (1990). Lipid traffic in eukaryotic cells: mechamisms for intracellular transport and organelle-specific enrichment of lipids. Curl'. Opin. Cell Bioi. 2: 652-663.

Palade G. (1959). Functional changes in the structure of cell components. in "Subcellular Particles", T. Hayashi ed. pp. 64-80. Ronald, New York.

Palade G. (1975). Intracellular aspect of the process of protein synthesis. Science 189: 347 - 358.

Palmc K., Diefenthal T., Vingron M., Sander C. and Schell J. (1992). Molecular cloning and structural analysis of genes from Zea mays (L) coding for members of the ras-related ypt gene family. Proc. Nat!. A cad. Sci. USA 89: 787-791.

Palme K., Diefenthal T. and Moore I. (1993). The ypt gene family from maize and Arabidopsis, structure and function analysis. 1. Experimtl. Bot. 44S: 183-195.

Pardo J.M. and Serrano R. (1989). Structure of a plasma membrane W -ATPase gene from the plant Arabidopsis thaliana. 1. Bioi. Chem. 264: 8557-8562.

Pathak R.K., Luskey K.L. and Anderson G.W. (1986). Biogenesis of the crystalloid endoplasmic reticulum in UT-l cells: evidence that newly formed endoplasmic reticulum emerges from the nuclear envelope. 1. Cell Bioi. 102: 2158-2168.

Paulson J.C. and Colley K.J. (1989). Olycosyltransferases: structure, localization, and control of cell type-specific glycosylation. 1. Bioi. Chem. 264: 17615-17618.

Pearse B.M.F. and Robinson M.S. (1990). Clathrin, adapters, and sorting. Annu. Rev. Cell Bioi. 6: 151-171.

Pcdrazzini E., Giovinazzo G., Bollini R., Ceriotti A. and Vitale A. (1994). Binding of BiP to an assembly-defective protein in plant-cells. Plant 1. 5: 103-110.

Pelham H.R.B. (1988). Evidence that luminal ER proteins are sorted from secreted proteins in a post-ER compartment. EMBO 1. 7: 913-918.

Pelham H.R.B., Hardwick K.G. and Lewis M.J. (1988). Sorting of soluble ER proteins in yeast. EMBO 1. 7: 1757-1762.

Pelham H.R.B. (1989). Control of protein exit from the endoplasmic reticulum. Annu. 261

Rev. Cell Bioi. 5: 1-23.

Pelham H.R.B. (1990). The retention signal for luminal ER proteins. Trends Biochem. Sci. 15: 483-486.

Pelham H.R.B. and Munro S. (1993). Sorting of membrane-proteins in the secretory pathway. Cell 75: 603-605.

Peterson G.L. (1977). A simplification of the protein assay methods of Lowry et al. which is more generally applicable. Anat. Biochem. 83: 346-356.

Pfeffer S.R. and Rothman J.E. (1987). Biosynthetic protein transport and sorting by the endoplasmic reticulum and Golgi. Annu. Rev. Biochem. 56: 829-852.

Pierson E., Gong F.-C., Afonso C.L., Lambert G. and Galbraith D.W. (1995). Analysis of nuclear targeting in Nicotiana tabacum: characterization of an efficient nuclear localization signal in protoplasts and in transgenic plants. In preparation.

Ponnambalam S., Rabouille C., Luzio J.P., Nilsson T. and Warren G. (1994). The TGN38 glycoprotein contains 2 nonoverlapping signals that mediate localization to the trans-golgi network. J Cell Bioi. 125: 253-268.

Prehn S., Wiedmann M. and Rapoport T.A. (1987). Protein translocation across wheat germ microsomal membranes requires an SRP-like component. EMBO J 6: 2093-2097.

Prill V., Lehmann L., von Figura K. and Peters C. (1993). The cytoplasmic tail of lysosomal acid-phosphatase contains overlapping but distinct signals for basolateral sorting and rapid internalization in polarized MOCK cells. EMBO J. 12: 2181-2193.

Pryer N.K., Wuestshube L.J. and Schekman R. (1992). Vesicle-mediated protein sorting. Annu. Rev. Biochem. 61: 471-516.

Puddington L., Lively M.D. and Lyles D.S. (1985). Role of the nuclear envelope in synthesis, processing, and transport of membrane glycoproteins. J Bioi. Chem. 260: 5641-5647.

Raikhel N. V. (1992). Nuclear targeting in plants. Plant Physiol. 100: 1627-1632.

Rambourg A. and Clermount Y. (1990). Three-dimensional electron microscopy: structure of the Golgi apparatus. EliI'. J. Cell Bioi. 51: 189-200. 262

Rapoport T.A. (1986). Protein translocation across and integration into membranes. CRC Cri. Rev. Biochem. 20: 73-137.

Rapoport T. A. (1992). Transport of Proteins across the endoplasmic reticulum membrane. Science 258: 931-936.

Ray P.M., Shininger T.L. and Ray M.M. (1969). Isolation of p-glucan synthetase particles from plant cells and identification with Oolgi membranes. Proc. Natl. Acad. Sci. USA 64: 605-612.

Ray P.M. (1979). "Plant Organelle" E. Reid ed. pp. 135-146. Horwood, Chichester

Raymond C.K., Howaldstevenson I., Vater C.A. and Stevens T.H. (1992). Morphological classification of the yeast vacuolar protein sorting mutants - evidence for a prevacuolar compartment in c1ass-e vps mutants. Mol. Bioi. Cell 3: 1389-1402.

Reaves B. Wilde A. and Banting G. (1992) Identification, molecular characterization and immunolocalization of an isoform of the Trans-Oolgi-Network (TON) - specific integral membrane-protein TON38. Biochem. J 283: 313-316.

Reaves B. Horn M. and Banting G. (1993). TON38/41 recycles between the cell-surface and the TON - Brefeldin-A affects its rate of return to the TON. Mol. Bioi. Cell 4: 93-105.

Restrepo M.A., Freed D.O. and Carrington J.C. (1990). Nuclear transport of plant potyviral proteins. Plant Cell 2: 987-998.

Robards A.W. and Lucas W.J. (1990). Plasmodesmata. Anml. Rev. Plant Physiol. Plant Mol. Bioi. 41: 369-419.

Robbins R.J. and Juniper B.E. (1980a). The secretory cycle of Dionaea muscipula Ellis. II. Storage and synthesis of the secretory proteins. New Phytol. 86: 297-311.

Robbins R.J. and Juniper B.E. (1980b). The secretory cycle of Dionaea muscipula Ellis. I. The fine structure and the effect of stimulation on the fine structure of the digested gland cells. New Phytol. 86: 279-296.

Roberts C.J., Nothwehr S.F. and Stevens T. H. (1992). Membrane-protein sorting in the yeast secretory pathway - evidence that the vacuole may be the default compartment. J Cell Bioi. 119: 69-83.

Robinson D.G., Andreae M., Depta H., Hal·tman D. and Hillmer S. (1987). Coated I'

263

vesicles in plants: Characterization and function. in "Plant Membrane: Structure, Function, Biogenesis", (1st ed.) C. Leaver and H. Sze eds. pp. 341-358. Alan R. Liss, Inc. New York.

Robinson D.G., Balusel{ K., Depta H., Hoh B. and Holstein S.E.H. (1991). Isolation and characterization of plant coated vesicles. in "Endocytosis, exocytosis and vesicle traffic in plants", (1st ed.) C.R. Hawes, J.O.D. Coleman and D.E. Evans eds. pp. 65-80. Cambridge University Press, Cambridge.

Rose J.K. and Doms R.W. (1988). Regulation of protein export from the endoplasmic reticulum. Annu. Rev. Cell Bioi. 4: 257-288.

Rose J.K. and Gallione C. J. (1981) Nucleotide sequences of the mRNA's encoding the vesicular stomatitis virus G and M proteins determined from cDNA clones containing the complete coding regions. J. Virol. 39: 519-528

Rothblatt J.A., Deshaies R.J., Sanders S.L. Daum G. and Schekman R. (1989). Multiple genes are required for proper insertion of secretory proteins into the endoplasmic reticulum in yeast. J. Cell Bioi. 109: 2641-2652.

Rothman J.E. and Orci L. (1992). Molecular dissection of the secretory pathway. Nature 355: 409-415.

Rothman J.E., Miller R.L. and Urbani L.J. (1984). Intercompartmental transport in the Golgi complex is dissociative process: Facile transfer of membrane protein between two Goigi populations. J. Cell Bioi. 99: 260-271.

Rothman J.E., Pettegrew H.G. and Fine R.E. (1980). Transport of the membrane glycoprotein of the vesicular stomatitis virus to the cell surface in two stages by clathrin-coated vesicles. J. Cell Bioi. 86: 162-171.

Rottier P., Armstrong J. and Meyer D.1. (1985). Signal recognition particle-dependent insertion of coronavirus E 1, an intracellular membrane glycoprotein. J. Bioi. Chem. 260: 4648-4652.

Rottier P., Welling G.W., Welling-Wester S., Niesters H.G.M., Lenstra J.A. and Van der Zeijst B.A.M. (1986). Predicted membrane topology of the coronavirus protein E1. Biochem. 25: 1335-1339.

Saalbach G., Jung R., Kunze G., Saalbach I., Adler K. and Muntz K. (1991). Different legumin protein domains act as vacuolar targeting signals. Plant Cell 3: 145-147. 264

Saraste J. and Kuismanen E. (1984). Pre- and post-Golgi vacuoles operate in the transport of Semliki Forest virus membrane glycoproteins to the cell surface. Cell 38: 535-549.

Saraste J., Palade G. and Farquhar M.G. (1986). Temperature-sensitive steps in the transport of secretory proteins through the Golgi complex in exocrine pancreatic cells. Proc. Natl. Acad. Sci. USA 83: 6425-6429.

Satiat-Jeunemnitre D. and Hawes c.R. (1993). Insights into the secretory pathway and vesicular transport in plant-cells. Bioi. Cell 79: 7-15.

Satoh S., Sturm A., Fujii T. and Chrispeels M.J. (1992). eDNA cloning of an extracellular dermal glycoprotein of carrot and its expression in response to wounding. Planta 188: 432-438

Savitz A.J. and Meyer D.I. (1993). 180-kd ribosome receptor is essential for both ribosome binding and protein translocation. J. Cell Bioi. 120: 853-863.

Schachtman D.P., Schroeder J.I., Lucas W .• f., Anderson J.A. and Gaber R.F. (1992). Expression of an inwards-rectifying potassium channel by the Arabidopsis KATl eDNA. Science 258: 1654-1658.

Schindler R., Hin C., Zerial M., Lottspeich F. and Hauri H.-P. (1993) ERGle-53, a membrane-protein of the ER-Golgi intermediate compartment, carries an ER retention motif. EliI'. J. Cell Bioi. 61: 1-9.

Schmid S.I. (1992). The mechanism of receptor-mediated endocytosis - more questions than answers. BioEssays 14: 589-596.

Schunck W.-H., Vogel F., Gross D., Kal'gel E., Mauersberger S., Kopke K., Gengnagel C. and Muller H.-G. (1991). Comparison of two cytochromes P-450 from Candida maltosa: primary structures, substrate specificities and effects of their expression in Saccharomyces cerevisiae on the proliferation of the endoplasmic reticulum. EliI'. J. Cell Bioi. 55: 336-345.

Schutze M.P., Peterson P.A. and Jackson M.R. (1994). An N-terminal double-arginine motif maintains type-II membrane- proteins in the endoplasmic-reticulum. EMBO J. 13: 1696-1705.

Schweizer A. Matter K. Ketcham C.M. and Hauri H.-P. (1991). The isolated ER-Golgi intermediate compartment exhibits properties that different from ER and 265

cis-Golgi. J. Cell Bioi. 113: 45-54.

Schwob E., Choi S.Y., Simmons C., Migliaccio F., lIag L., Hesse T., Palme K. and Soli D. (1993). Molecular analysis of 3 maize 22 kda auxin-binding protein genes - transient promoter expression and regulatory regions. Plant J. 4: 423-432.

Seeger M. and Payne G.S. (1992). Selective and immediate effects of clathrin heavy chain mutation on Goigi membrane protein retention in Saccharomyces cerevisiae .•1. Cell Bioi. 118: 531-540.

Segev N., Mulholland J. and Botstein D. (1988). The yeast GTP-binding yptl protein and a mammaliari counterpart are associated with the secretory machinery. Cell 52: 915-924.

Segui-Real B., Stuart R.A. and Neupert W. (1992). Transport of proteins into the various subcompartments of mitochondria. FEBS Lett. 313: 2-7.

Semenza J.C., Hardwick K.G., Dean N. and Pelham H.R.B. (1990). ERD2, a yeast gene required for the receptor-mediated retrieval of luminal ER proteins from the secretory pathway. Cell 61: 1340-1357.

Semenza J.C. and Pelham H.R.B. (1992). Changing the specificity of the sorting receptor for luminal endoplasmic-reticulum proteins. J. Mol. Bioi. 224: 1-5.

Serafini T., Stenbeck G., Brecht A., Lottspeich F., Orci L., Rothman J.E. and Wieland F. T. (1991a). A coat subunit of Golgi-derived non-clathrin-coated vesicles with homology to the clathrin-coated vesicle coat protein beta-adaptin. Nature 349: 215-220.

Serafini T., Orci L. Amherdt M., Brunner M., Kahn R.A. and Rothman J. E. (1991 b). ADP-ribosylation factor is a subunit of the coat of Golgi-derived COP-coated vesicles: a novel role for a GTP-binding protein. Cell 67: 239-253.

Siegel V. and Walter P. (1988). Functional dissection of the signal recognition particle. Trends Biochem. Sci. 13: 314-316.

Sijmons P.C., Dekker B.M.M, Schrammaijer B., Vcnvocrd T.C., van dacn Elzem P .J.M. and Hoekema A. (1990). Production of correctly processed human serum albumin in transgenic plants. Bio/tecJmology 8: 217-221.

Silver P.A. (1991). How proteins enter the nucleus. Cell 64: 489-497. 266

Simon R., Altschuler Y., Rubin R. and Galili G. (1990). Two closely related wheat storage proteins follow a markedly different subcellular route in Xenopus laevis ooocytes. Proc. Nat!. Acad Sci. USA 88: 834-838.

Simon S.M. and Blobel G. (1991). A protein-conducting channel in the endoplasmic reticulum. Cell 65: 371-380

Singh P., Tang B.L., Wong S.H. and Hong W.J. (1993). Transmembrane topology of the mammalian KDEL receptor. Mol. Cell. Bioi. 13: 6435-6441.

Singer S.J. (1990). The structure and insertion of integral proteins in membranes. Annu. Rev. Cell Bioi. 6: 247-296.

Singer 1.1., Kawka D.W., Kazazis D.M., Alberts A.W., Chen J.S., Huff J.W. and Ness G.C. (1984). Hydroxymethylglutaryl-coenzyme A reductase-containing hepatocytes are distributed periportally in nomal and mevinolin-treated rat livers. Pmc. Natl. Acad Sci. USA 81: 5556-5560.

Sleight R.G. (1987). Intracellular lipid transport in eukaryotes. Ann. Rew. Physiol. 49: 193-208.

Smith S. and Blobel G. (1993) The first membrane spanning region of the lamin-B receptor is sufficient for sorting to the inner nuclear-membrane. J. Cell Bioi. 120: 631-637.

Soullam B. and Worman H.J. (1993) The amino-terminal domain of the lamin-B receptor is a nuclear-envelope targeting signal. J. Cell Bioi. 120: 1093-1100.

Stack J. H., Herman P.K., Schu P.V. and Emr S.D. (1993). A membrane-associated complex containing the vps 15 protein-kinase and the vps34 pi 3 .. kinase is essential for protein sorting to the yeast lysosome-like vacuole. EMBO J. 12: 2195-2204.

Staehelin L.A., Driouich A., Giddings T.H.and Zhang G.F. (1993). Organization and dynamics of the glycoprotein and polysaccharide synthetic pathways in the plant Golgi-apparatus. J. Cell. Biochem. S 17:4-4.

Staehelin L.A., Giddings T.H., Kiss J.Z. and Sack F.D. (1990). Macromolecular differentiation of Golgi stacks in root tips of Arabidopsis and Nicotiana seedlings as visualized in high pressure frozen and freeze-substituted samples. Pl'otoplasma 157: 75-91.

Staehelin L.A., Giddings T.H., Levy S., Lynch M.A., Moore P.J. and Swords K.M.M. 267

(1991). Organization of the secretory pathway of cell wall glycoproteins and complex polysaccharides in plant cells. in "Endocytosis, exocytosis and vesicle traffic in plants", (1st ed.) C.R. Hawes, J.O.D. Coleman and D.E. Evans eds. pp. 182-198. Cambridge University Press, Cambridge.

Steer M.W. and O'Driscoli D. (1991). Vesicle dynamics and membrane turnover in plant cells. in "Endocytosis, exocytosis and vesicle traffic in plants", (1st ed.) C.R. Hawes, J.O.D. Coleman and D.E. Evans eds. pp. 129-142. Cambridge University Press, Cambridge.

Sturbois B., Moreau P., Manetapeyret L., Morrc D.J. and Cassagne C. (1994). Cell-free transfer of phospholipids between the endoplasmic reticulum and the Golgi-apparatus ofleek seedlings. Biochimi. Biop/1ys. Acta - Biomembr. 1189: 31-37.

Sturm A., Johnson K.D. Szumilo T., Elbein A.D. and Chrispeels M.J. (1987). Subcellular localization of glycosidases and glycosyltransferases involved in the processing ofN-linked oligosaccharides. Plant Physiol. 85: 741-745.

Sturman L.S. and Holmes K.V. (1983). The molecular biology of coronavirus. Advances. Virus Res. 28: 35-111.

Subramani S. (1992). Targeting of proteins into the peroxisomal matrix. J. Membr. Bioi. 125: 99-106.

Subramani S. (1993). Protein import into peroxisomes and biogenesis of the organelle. Annu. Rev. Cell Bioi. 9: 445-478.

Suiter G.J., Waterham n.R., Vrieling E.G., Goodman J.M., Harder W. and Veenhuis M. (1993). Expression and targeting of a 47-kda integral peroxisomal membrane-protein of Candida boidinii in wild-type and a peroxisome-deficient mutant of Hansenula polymOlpha. FEBS Lett. 315: 211-216.

Sussman M.R. and Harper J.F. (1989). Molecular biology of the plasma membrane of higher plants. Plant Cell I: 953-960.

Sweet D.J. and Pelham H.R.B. (1992). The Saccharomyces-cerevisiae sec20 gene encodes a membrane glycoprotein which is sorted by the HDEL retrieval-system. EMBO J. 11: 423-432.

Swinkels B. W., Gould S.J. and Subramani S. (1992). Targeting efficiencies of various permutations of the consensus C-terminal tripeptide peroxisomal targeting signal. FEBS Lett. 305: 133-136. 268

Swift A.M. and Machamer C.E. (1991). A Golgi retention signal in a membrane-spanning domain of coronavirus E 1 protein. J. Cell BioI. 115: 19-30.

Sze H., Ward J.M., Lai S. and Perera I. (1992). Vacuolar-type W-translocating A TPases in plant endomembranes - subunit organization and multi gene families. J. Experimtl. BioI. 172: 123-135.

Tague B.W. and Chrispeels M.J. (1987). The plant vacuolar protein, phytohemagglutinin, is transported to the vacuole of transgenic yeast. J. Cell BioI. 105: 1971-1979.

Taiz L. (1992). The plant vacuole. J. Experimtl. BioI. 172: 113-122.

Tang B.L., Wong S.H., Qi X.L. Low S.H. and Hong W. (1993). Molecular-cloning, characterization, subcellular-localization and dynamics ofp23, the mammalian KDEL receptor. J. Cell BioI. 120: 325-338.

Tartakoff A.M. (1986). Temperature and energy dependence of secretory protein transport in the exocrine pancreas. EMBOJ. 5: 1477-1482.

Taussky H. H. and Shorr E. (1953) A microcolorimetric method for the determination of inorganic phosphorus. J. BioI. Chem. 245: 1090-1100.

Tilney L.G., Cooke T.J., Connelly P.S. and Tilney M.S. (1991). The structure of plasmodesmata as revealed by plasmolysis, detergent extraction, and protease digestion. J. Cell. BioI. 112: 739-747.

Townsley F.M., Wilson D.W. and Pelham H.R.B. (1993). Mutational analysis of the human KDEL receptor - distinct structural requirements for Golgi retention, ligand-binding and retrograde transport. EMBO J. 12: 2821-2829.

Vale R.D. (1987). Intracellular transport using microtuble-based motors. Annu. Rev. Cell BioI. 3: 347-378.

Vale R.D. (1992). Microtubule motors - many new models off the assembly line. Trends Biochem. Sci. 17: 300-304. van Beck N.A.M., Lohuis D., Dijkstra J. and Peters D. (1985). Morphogenesis of sonchus yellow net virus in cowpea protoplasts. J. Ultrastuct. Res. 90: 294-303.

Vanderklei I.J., Faber K.N., Keiz~rgunnink I., Giet) c., Harder W. and Veenhuis M. I'

269

(1993). Watermelon glyoxysomal malate-dehydrogenase is sorted to peroxisomes of the methylotrophic yeast, Hansenula-polym01pha. FEBS Lett. 334: 128-132.

Verge res G., Benedict T.S., Aggeler J., Lausier J. and Waskell L. (1993). A model system for studying membrane biogenesis: Overexpression of cytochrome b5 inyeast results in marked proliferation of the intracellular membrane . .J. Cell Sci. 106: 249-259.

Vigers G.P. and Lohka M.J. (1991). A distinct vesicle population targets membrane and pore complex to the nuclear envelope in Xenopus eggs . .J. Cell Bioi. 112: 545-556.

Vitale A., Strum A. and Bollini R. (1986). Regulation of processing of a plant glycoprotein in the Golgi-complex: a comparative study using Xenopus oocytes. Planta 169: 108-116.

Vitale A. and Chrispeels M.J. (1992). Sorting of proteins to the vacuoles of plant cells. BioEssays 14: 151-160.

von Figura K. and Hasilik A. (1986). Lysosomal enzymes and their receptors. Annu. Rev. Biochem. 55: 167-193.

von Meyenburg K., Jorgensen B.B. and van Deurs B. (1984). Physiological and morphological effects of overproduction of membrane-bound ATP synthase in Escherichia coli K-12. EMBOJ. 3: 1791-1797.

von Schaewen A., Sturm A., Oneill J., and Chrispeels M.J. (1993). Isolation of a mutant Arabidopsis plant that lacks N-acetylglucosaminyltransferase-I and is unable to synthesize Golgi-modified complex N-linked glycans. Plant Physiol. 102: 1109-1118.

von Schaewen A. and Chrispeels M.J. (1993). Identification of vacuolar sorting information in phytohemagglutinin, an unprocessed vacuolar protein. J. Experimtl. Bot. 44S: 339-342.

Walters M.G. Seratini T. and Rothman J.E. (1991). 'Coatomer': a cytosolic protein complex containing subunits of non-clathrin-coated Golgi transport vesicles. Nature 349: 248-251.

Wandelt C.I., Khan M.R.I. Craig S., Schroeder H.E., Spencer D. and Higgins T.J.V. (1992). Vicilin with carboxyl-terminal KDEL is retained in the endoplasmic reticulum and accumulates to high levels in the leaves of transgenic plants. Plant J. 2: 181-192.

Wandinger-Ness A., Bennett M.K., Antony C. and Simons K. (1990). Distinct transport vesicles mediate the delivery of plasma membrane proteins to the apical and 270

basolateral domains of MDCK cells. J. Cell Bioi. Ill: 987-1000.

Warren G. (1987). Signals and salvage sequences. Nature 327: 17-18.

Weisz O.R., Swift A.M. and Machamer C.E. (1993). Oligomerization of a membrane protein correlates with ins retention in the Golgi complex. J. Cell Bioi. 122: 1185-1196.

Wendland M., Waheed A., von Figura K. and Pohlmann R. (1991). Mr 46,000 mannose 6-phosphate receptor. The role of histidine and arginine residues for binding of ligand. J. Bioi. Chem. 266: 2917-2923.

Weiner J.H., Lemire B.D., Elmes M.L., Bradley R.D. and Scraba D.G. (1984). Overproduction of fumarate reductase in Escherichia coli induces a novel intracellul~r lipid-protein organelle. J. Bacterial. 158: 590-596.

Wenzel M., Gersbar H., Schimpl A. and Rudiger H. (1993). Time-course of lectin and storage protein biosynthesis in developing pea (Pisum sativum) seeds. Bioi. Chem. Hoppe-Seyler 374: 887-894.

Westers P., Emr S.D. and Chl'ispeels M.J. (1993). Isolation of an Arabidopsis thanliana gene homolog to a yeast gene involved in vacuolar protein sorting. .1. Cell Biochem. 17C: 42

Westlund B., Dahms N.M. and Kornfeld S. (1991). The bovine mannose 6-phosphate/insulin-like growth factor II receptor. Localization of mannose 6-phosphate binding sites to domain 1-3 and 7-11 of the extracytoplasmic rgion. J. Bioi. Chem. 266: 23233-23239.

Wilkins T.A., Bednarek S. and Raihhel N.V. (1990). Role of pro peptide glycan in post-translational processing and transport of barley lectin to vacuoles in transgenic tobacco. Plant Cell 2: 301-313.

Wilson K.L. and Newport J. (1988). A trypsin-sensitive receptor on membrane vesicles is required for nuclear envelope formation in vitro. J. Cell Bioi. 107: 57-68.

Wilson D.W., Lewis M.J. and Pelham H.R.B. (1993). pH-dependent binding ofKDEL to its receptor in vitro. J. Bioi. Chem. 268: 7465-7468.

Wink M. (1993). The plant vacuole - a multifunctional compartment. J. Experimtl. Bot. 44S: 231-246.

Wong S.H. and Hong W.J. (1993). The SXYQRL sequence in the cytoplasmic domain 271

ofTGN38 plays a major role in Trans-Golgi Network localization. J. Bioi. Chern. 268: 22853-22862.

Wright B., Basson M., D'Ari L. and Rinc J. (1988). Increased amounts ofHMG-CoA reductase induce "Karmellae": A proliferation of stacked membrane pairs surrounding the yeast nucleus. J. Cell Bioi. 107: 101-114.

Ye Q. and Worman H.J. (1994). Primary structure-analysis and lamin-B and DNA-binding of human LBR, an integral protein of the nuclear-envelope inner membrane. J. Bioi. Chern. 269: 11306-11311.

Zagouras P. and Rosc J.K. (1989). Carboxy-terminal SEKDEL sequences retard but do not retain two secretory proteins in the endoplasmic reticulum. J. Cell Bioi. 109: 2633-2640.

Zhang G.F. and Staehclin L.A. (1992). Functional compartmentation of the Golgi-apparatus of plant cells - immunocytochemical analysis of high-pressure frozen-substituted and freeze-substituted sycamore maple suspension-culture cells. Plant Physiol. 99: 1070-1083.

Zhang G.F., Driouich A. and Staehelin L.A. (1993). Effect ofmonensin on plant Golgi - reexamination of the monensin-induced changes in cisternal architecture and functional activities of the Golgi-apparatus of sycamore suspension-cultured cells. J. Cell Sci. 104: 819-831.

Zurzolo C., Lebivic A., Quaroni A., Nitsch L. and Rodriguez-Boulan E. (1992). Modulation of transcytotic and direct targeting pathways in a polarized thyroid-cell line. EMBO J. 11: 2337-2344.

Zurzolo C., VanthofW., Vanmeer G. and Rodriguez-Boulan E. (1994). Vip211caveolin, glycosphingolipid clusters and the sorting of glycosylphosphatidyinositol anchored proteins in epithelial-cells. EMBO J. 13: 42-53.