Citation
A novel fucosylation pathway in the cytosol of Dictyostelium Discoideum

Material Information

Title:
A novel fucosylation pathway in the cytosol of Dictyostelium Discoideum
Creator:
Gonzalez-Yanes, Beatriz, 1964-
Publication Date:
Language:
English
Physical Description:
xvi, 166 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Cytosol ( jstor )
Digestion ( jstor )
Enzymes ( jstor )
Gels ( jstor )
Glycopeptides ( jstor )
Glycoproteins ( jstor )
In vitro fertilization ( jstor )
Lectins ( jstor )
Oligosaccharides ( jstor )
Radioactive decay ( jstor )
Cytosol ( mesh )
Department of Anatomy and Cell Biology thesis Ph.D ( mesh )
Dictyostelium ( mesh )
Dissertations, Academic -- College of Medicine -- Department of Anatomy and Cell Biology -- UF ( mesh )
Fucose ( mesh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1991.
Bibliography:
Includes bibliographical references (leaves 156-165).
Additional Physical Form:
Also available online.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Beatriz Gonzalez-Yanes.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
026886073 ( ALEPH )
26208727 ( OCLC )

Downloads

This item has the following downloads:


Full Text











A NOVEL FUCOSYLATION PATHWAY IN THE CYTOSOL
OF DICTYOSTELIUM DISCOIDEUM

















By

BEATRIZ GONZALEZ-YANES


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1991































A mis padres
(To my parents)














ACKNOWLEDGEMENTS

First of all I would like to acknowledge my advisor,

Dr. Christopher M. West, for his time, training, and example

throughout these studies. I would also like to thank the

members of my committee, Drs. Robert Cohen, William Dunn,

and Carl Feldherr, for their time and guidance regarding

this project.

I would like to acknowledge and thank Dr. Michael Ross

for his interest and support of my graduate education.

Part of the work presented in Chapter II is reproduced

from the journal of Developmental Biology, 1989, volume 133,

pages 576-587 by copyright permission of Academic Press,

Inc.

I am very grateful to the following researchers and the

members of their laboratories for sharing equipment and/or

reagents with us: Drs. Gudrun Bennett, Ross Brown and Lina

Gritzali, William Dunn, Carl Feldherr, Michael Humphreys-

Beher, K.J. Kao, and Gillian Small.

In general, I would like to express my gratitude to the

faculty and staff in the Department of Anatomy and Cell

Biology for their support and interest in my advancement as

a scientific investigator. I would like to acknowledge Kari

Eissinger for her help in photographing and printing some of


iii








the figures contained herein. Particularly, I would like to

thank Drs. Kelly Selman and Gillian Small for the advice and

friendship they have provided during these years. I am also

grateful to have been in contact with other students, past

and present, which have made the time spent in the

department more enjoyable. I am grateful to Dan Tuttle for

reviewing my dissertation. I would also like to acknowledge

all the people that have worked in the laboratory of Dr.

West, including Scherwin Henry, for making the time spent in

the laboratory more rewarding and Mel Fields for his

assistance in some experiments.

I would like to acknowledge my professors at the

University of Puerto Rico, who encouraged me to pursue a

career in biological research.

I appreciate all the friends I have made in Gainesville

in the past years and would like to acknowledge their

importance in my education. Especially, I thank Hans van

Oostrom for his moral support, encouragement, and for

teaching me about the wonders of the electronic world and

those rare creatures called computers.

Finally, and most importantly, I would like to thank my

family. I thank my grandparents, especially Abuela Emma and

Abuela Yuya, for the faith they have showed in me and for

their prayers, which have helped me through my graduate

education. My brothers German, Omar, and Carlos have been

great sources of happiness and pride. Lastly, I thank my








parents, Beatriz and German, for everything. I firmly

believe that the education received during my first twenty-

one years of life in their company is ultimately responsible

for every achievement in my life.















TABLE OF CONTENTS


ACKNOWLEDGEMENTS . . . . . . . . . iii

LIST OF TABLES . . . . . ........ ...... .. viii

LIST OF FIGURES . . . . . . . . . .. ix

LIST OF ABBREVIATIONS . . . . . . . .. xi

ABSTRACT . . . . . . . . . . . xiv

CHAPTERS

I HISTORICAL REVIEW AND BACKGROUND . . . . 1

Introduction . . . . . . . . . 1
Fucosylated Macromolecules . . . . . . 3
Fucose-Binding Proteins . . . . . .. 13

II CHARACTERIZATION OF A FUCOSYLATION MUTANT . . 18

Introduction . . . . . . . . .. 18
Materials and Methods . . . . . . .. 19
Results . . . . . . . . . . 24
Discussion . . . . . . . . . . 40

III IDENTIFICATION OF A CYTOSOLIC FUCOPROTEIN . . 45

Introduction . . . . . . . . .. 45
Materials and Methods . . . . . . .. 46
Results . . . . . . . . . .. 54
Discussion . . . . . . . . . .. 83

IV EVIDENCE FOR A CYTOSOLIC FUCOSYLTRANSFERASE . 90

Introduction . . . . . . . . .. 90
Materials and Methods . . . . . . .. 92
Results . . . . . . . . . .. 98
Discussion . . . . . . . . .. 135

V SUMMARY AND CONCLUSIONS . . . . . .. 144

Summary of Results . . . . . . .. 144
Future Studies . . . . . . . . . 149

vi








REFERENCES . . . . . . . . . .. .156

BIOGRAPHICAL SKETCH . . . . . . . .. .166


vii















LIST OF TABLES


Table Page

2-1. Fucose content of normal and mutant spores and
vegetative cells . . . . . . . .. 29

2-2. Effect of time on conversion of GDP-mannose to
GDP-fucose . . . . . . . . . .. 34

2-3. Specific activities of fucose . . . . .. .39

3-1. Distribution of protein and radioactivity in S100
and P100 fractions . . . . . . . . 56

3-2. Radioactivity recovered in the second S100 after
different P100 treatments . . . . .. .. 65

3-3. Distribution of markers among S100 and P100
fractions . . . . . . . . . .. 66

4-1. Effect of different treatments on the cytosolic
fucosyltransferase activity . . . . . .. .102

4-2. Failure to sediment S100 fucosyltransferase
activity . . . . . . . . . .. 104

4-3. Comparison between the S100 and P100
fucosyltransferase activities . . . . .. .112

4-4. Fucosyltransferase activity in Ax3 S100 fraction 122

4-5. Utilization of 8-methoxycarbonyloctyl synthetic
acceptors by cytosolic fucosyltransferase activity
from Ax3 and HL250 . . . . . . . .. 128

4-6. Evaluation of the suitability of p-nitro-phenyl
glycosides as acceptors for cytosolic fucosyl-
transferase activities in HL250 and Ax3 S100 . 132

4-7. Reduction of fucosylation of acceptor type I analog
by purified FP21 . . . . . . . .. 134

4-8. In vitro fucosyltransferase activity of HL250 and
Ax3 slug extracts . . . . . . . .. 136


viii















LIST OF FIGURES


Figure Page

2-1. Localization of SP96 in prespore and spore cells 27

2-2. Biosynthesis of GDP-L-fucose . . . . .. .32

2-3. Conversion of GDP-mannose to GDP-fucose by normal
and mutant strains . . . . . . . .. 37

3-1. Incorporation of [3H]fucose into macromolecular
species of the S100 and P100 . . . . .. .58

3-2. Proteinaceous nature of FP21 . . . . .. 60

3-3. Comparison of S100 and releasable P100 components 63

3-4. Gel filtration chromatography of FP21
glycopeptides . . . . . . . . . . 71

3-5. Gel filtration chromatography of PNGase F digests 74

3-6. Gel filtration chromatography of FP21
oligosaccharides . . . . . . . .. 77

3-7. Gel filtration chromatography of P100
glycopeptides . . . . . . . . . 81

3-8. Incorporation of [3H]fucose into macromolecular
species of slug stage cells . . . . . .. .85

4-1. Fucosylation of endogenous acceptors by S100
fraction . . . . . . . . . .. 101

4-2. SDS-PAGE profile of endogenous acceptors
fucosylated in vitro . . . . . . .. 107

4-3. BioGel P-4 gel filtration chromatography of in
vitro labelled FP21 oligosaccharide . . . .. .110

4-4. Effect of pH on S100 and P100 fucosyltransferase
activities in the presence of Tween-20 . . . 115








4-5. Effect of GDP-fucose concentration on S100 and P100
fucosyltransferase activities in the presence of
Tween-20 . . . . . . . . . .. 117
4-6. Effect of GDP-fucose concentration on intact S100
and P100 fucosyltransferase activities . . .. .120

4-7. Fucosylation of mutant FP21 by Ax3 S100 fraction 126

4-8. Haworth projections of the structures of the synthetic
glycolipid acceptors . . . . . . .. 130















LIST OF ABBREVIATIONS

APA Asparagus Pea Agglutinin

ATP Adenosine-5'-triphosphate

BSA Bovine Serum Albumin

14C Radioactive Carbon

cDNA Complementary Deoxyribonucleic Acid

cm Centimeter(s)

Con A Concanavalin Agglutinin

cpm Counts per Minute

CHO Chinese Hamster Ovary

dH2O Distilled Water

dpm Disintegrations per Minute

EDTA Ethylenediaminetetraacetic Acid

EtOH Ethanol

GlcNAc N-acetyl Glucosamine

GDP Guanosine-5'-diphosphate

h Hour(s)
3H Tritiated

HMG High Mobility Group

HPLC High Performance Liquid Chromatography

kD Kilodaltons

K Michaelis Constant
1 Liter(s)
1 Liter(s)









M

mCi

mAb

MES

mg

ml

min

mm

mM

MW

pCi

PM

nm

p

PAGE

pmol

PMSF

PNGase F

Rev

RNAse B

SDS

TCA

Tris

U

UEA-I

Ve


Molar Concentration

Millicurie(s)

Monoclonal Antibody

2-(N-Morpholino)ethanesulfonic Acid

Milligram(s)

Milliliter(s)

Minutes)

Millimeter(s)

Millimolar

Molecular Weight

Microcurie(s)

Micromolar

Nanometers

probability

Polyacrylamide Gel Electrophoresis

Picomole(s)

Phenylmethylsulfonyl Fluoride

Peptide N-glycosidase F

Relative Elution Coefficient

Ribonuclease B

Sodium Dodecyl Sulphate

Trichloroacetic Acid

Tris(hydroxymethyl)aminomethane

Unit(s)

Ulex europaeus Agglutinin

Elution Volume


xii









Vi

V
max

vo

v/v

WGA

w/v


xiii


Inclusion Volume

Maximal Velocity

Void volume

Volume per Volume

Wheat Germ Agglutinin

Weight per Volume















Abstract of Dissertation Presented to the Graduate School
of the University Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

A NOVEL FUCOSYLATION PATHWAY IN THE CYTOSOL
OF DICTYOSTELIUM DISCOIDEUM

By

Beatriz GonzAlez-Yanes

December 1991

Chairman: Dr. Christopher M. West
Major Department: Anatomy and Cell Biology

The existence of a fucosylation pathway in the cytosol

of Dictyostelium discoideum was investigated in the

glycosylation mutant strain, HL250, and the normal parental

strain, Ax3. HL250 was characterized as a conditional

mutant that cannot convert GDP(guanosine-5'-diphosphate)-

mannose to GDP-fucose, resulting in a lack of macromolecular

fucosylation unless grown in the presence of extracellular

fucose. HL250 or Ax3 cells were metabolically labelled with

[3 H]fucose, filter-lysed, and fractionated by high speed

centrifugation into sedimentable (P100) and soluble (S100)

fractions. The fractions exhibited unique profiles of

fucoconjugates as analyzed by gel electrophoresis. The

major acceptor in the S100 was a 21 kilodalton molecular

weight protein, FP21. Analysis of FP21 oligosaccharide by

mild alkaline hydrolysis and gel filtration chromatography

xiv








revealed that fucose was incorporated into an 0-linked

oligosaccharide with an average size of 4.8 glucose units.

FP21 appeared to be endogenous to the cytosol, based on the

failure to release FP21 from P100 vesicles by sonication,

and the absence of FP21-like glycopeptides derived by

pronase digestion of the P100.

To determine if FP21 was fucosylated in the cytosol,

S100 and P100 fractions from HL250 were assayed for

fucosyltransferase activity, measured as ability to transfer

[14C] from GDP-[14C]fucose to endogenous acceptors.

Fucosylation of FP21 by the S100 was time- and protein

concentration-dependent. The cytosolic activity was

distinguished from the bulk P100 activity by its absolute

divalent cation dependence, alkaline pH sensitivity, and

very low apparent K for GDP-fucose. Fucosyltransferase

activity was not detectable in Ax3 cytosol. However,

activity was reconstituted by addition of purified mutant

FP21, suggesting FP21 was already fucosylated in living

cells, and Ax3 possessed a cytosolic fucosyltransferase

equivalent to the HL250 enzyme. S100 fractions from Ax3 and

HL250 were able to fucosylate an al,4fucosyltransferase-

specific acceptor glycolipid analog, but not analogs capable

of being modified by al,2 and al,3 fucosyltransferases.

Since fucosylation of the analog was reduced by addition of

purified FP21, the same enzyme appears to be responsible for

fucosylation of both molecules. Thus it is proposed that









FP21 is synthesized and fucosylated in the cytosol by an

al,4fucosyltransferase.


xvi















CHAPTER I
HISTORICAL REVIEW AND BACKGROUND


Introduction

Glycoprotein synthesis and localization have been

subjects of intense study in the last few decades. Much has

been learned about glycosylation and some excellent reviews

are available (Kornfeld and Kornfeld, 1985; Hirschberg and

Snider, 1987). Carbohydrate moieties are usually

categorized as N and/or 0-linked depending on whether the

carbohydrate moiety is attached to an asparagine by an amide

glycosidic linkage or to a serine or threonine,

respectively. Glycosylation is generally considered to be a

modification restricted to macromolecules that pass through

the secretary pathway, starting in the rough endoplasmic

reticulum where N-linked glycosylation is initiated

(Kornfeld and Kornfeld, 1985), but in some instances 0-

linked glycosylation also takes occurs (Spielman et al.,

1988). Glycosylation then continues in the Golgi apparatus

where further processing of N-linked oligosaccharides may

occur and 0-linked glycosylation takes place (Abeijon and

Hirschberg, 1987). In the case of fucosylation, L-fucose

has been shown to be added as a terminal modification to

either N-linked or 0-linked oligosaccharides in the Golgi








2

apparatus utilizing GDP-fucose as the sugar nucleotide donor

(Bennett et al., 1974; Hirschberg and Snider, 1987).

However, there are a few exceptions to these remarks, since

fucose has been found to be attached directly to serine and

threonine, although the site for this modification has not

been identified (Hallgreen et al., 1975), and to be present

in homopolymers, as in fucoidans (Flowers, 1981). It has

been suggested that fucosylation also occurs in the

endoplasmic reticulum as well as in the Golgi apparatus in

thyrotrophs under different physiological conditions (Magner

et al., 1986). There are two pathways for the biosynthesis

of GDP-fucose. The main source of GDP-fucose is the

conversion pathway of GDP-mannose to GDP-fucose (Yurchenco

et al., 1978; Flowers, 1985). Alternatively, synthesis of

GDP-fucose may occur by the fucose salvage pathway in the

presence of extracellular L-fucose (Yurchenco et al., 1978;

Flowers, 1985; Ripka and Stanley, 1986; Reitman et al.,

1980).

Glycosylation is carried out by diverse specific

glycosyltransferases which have been located in the

endoplasmic reticulum and Golgi apparatus. One underlying

assumption is that the acceptor macromolecule must

colocalize with the transferase enzyme in order to be

glycosylated. However, there have been occasional reports

of glycoproteins in compartments topologically discontinuous

with the lumen of the rough endoplasmic reticulum and Golgi








3

apparatus. These findings contradict the dogma that

glycosylation is strictly an endoplasmic reticulum- and

Golgi apparatus-dependent event. Two models could account

for the existence of glycoproteins outside the realms of the

secretary pathway; one postulates that lumenal glycosylated

proteins translocate across the membrane back to the

cytosolic space, the other that glycosyltransferases are

localized outside of the lumen of the endoplasmic reticulum

or Golgi apparatus. Evidence has been accumulating in the

past decade that supports the latter model. In this chapter

I shall be concerned with the presence of fucosylated

proteins in compartments topologically discontinuous with

the lumen of the endoplasmic reticulum and Golgi apparatus.

Subsequently, I shall analyze some of the studies that

suggest the presence of fucosyltransferases that would co-

compartmentalize with such fucoproteins. In light of the

results presented in this dissertation, a review focussing

on the presence of fucoproteins and fucosyltransferases

outside the secretary pathway will be useful. An excellent

review is available that examines nuclear and cytosolic

glycosylation in general (Hart et al., 1989a).



Fucosylated Macromolecules

The presence of fucosylated proteins in compartments

discontinuous with the secretary pathway has been documented

since the 1970's. Various techniques have been employed in








4

these reports, including the direct analysis of carbohydrate

content, use of lectins, use of radiolabelled fucose and/or

a combination of biochemical and morphological techniques.



Nuclear Fucosylated Macromolecules

In contrast to the prevailing dogma, fucosylated

macromolecules have been detected in the nucleus. Such

studies demonstrated lectin binding to nuclear membranes,

chromatin, nuclear proteins, and the nuclear matrix. These

studies generally involved localization by binding of

labelled lectins (fluorescent, radiolabelled, ferritin-

conjugated, etc.) to isolated nuclei. Binding specificity

was assessed by competition of labelling with hapten

inhibitors. Occasionally, these studies were supplemented

by metabolic labelling experiments with radioactive fucose

which is primarily incorporated into fucoproteins with

little metabolism into other molecular species in eukaryotic

cells (Yurchenko et al., 1978).



Nuclear membranes. One early report suggested the

presence of fucose-containing structures on the cytoplasmic

face of isolated bovine nuclei (Nicolson et al., 1972).

Nicolson et al. (1972) found that purified nuclei were

agglutinated by the L-fucose-specific lectin UEA-I from Ulex

europaeus (Lis and Sharon, 1986), and agglutination was

inhibited by incubation in the presence of L-fucose,








5

suggesting that there were fucose-containing membrane-bound

oligosaccharides on the outer nuclear membrane. Similar

results were obtained with the mannose- and glucose-specific

lectin concanavalin A (Con A), which was also found to

agglutinate purified nuclei (Nicolson et al., 1972).

However, subsequent work by another group revealed that, as

evidenced by electron microscopic examination, ferritin-Con

A appeared to stain only damaged nuclei (Virtanen and

Wartiovaara, 1976) raising some concerns about the studies

by Nicolson et al. (1972). Likewise, in the aforementioned

studies (Nicolson et al., 1972), integrity of the isolated

nuclei was not determined, allowing for the possibility that

the lectin was interacting with lumenal fucoproteins that

escaped organelles during nuclei isolation or with

contaminating fucoprotein-containing membranes, such as

plasma membrane.



Chromatin. The existence of chromatin associated

fucose-containing proteins has been suggested by several

laboratories. Early on, Stein et al. (1975) reported the

presence of [3H] labelled glycoconjugates in purified

chromatin from HeLa cells grown in the presence of

[3 H]fucose. Although the label was not examined to

corroborate its presence as fucose, in the case of

radiolabelling macromolecules with radioactive sugar

precursors, fucose is an excellent candidate because it has








6

been shown to be incorporated as such, with minimal

metabolizing of the label (Yurchenko et al., 1978). In

these studies the authors examined a very pure chromatin

preparation, with essentially no contamination from nuclear

or plasma membranes. The authors argued against plasma

membrane contamination based on mixing experiments, in which

radiolabeled cell-surface trypsinates were combined with

unlabeled nuclear preparations prior to chromatin isolation.

The fucosylated chromatin-associated macromolecules were

deemed to be fucoproteins based on their sensitivity to

pronase digestion. Unfortunately, these studies have not

been pursued further, and many questions remain unanswered

with respect to their structure and biosynthesis.

Chromatin-associated fucoproteins were also detected in

normal rat liver and Novikoff hepatoma ascites cells

(Goldberg et al., 1978) using the L-fucose-specific lectin

asparagus pea agglutinin (APA). The authors reported a

strongly basic fucoprotein, that was sensitive to pronase

digestion. Based on the reactivity with the lectin, they

calculated that the protein was three times more

concentrated in tumor chromatin than in normal liver cells.

However, the method for chromatin purification did not

eliminate adequately nuclear inner membrane contamination,

and since binding to the lectin was assayed by

affinoelectrophoresis a molecular weight for the protein was

not determined (Goldberg et al., 1978).








7

In a more recent study on duodenal columnar cells, Kan

and Pinto da Silva (1986) used UEA-I conjugated to colloidal

gold in freeze-fracture electron microscopy of cross-

fractured nuclei. They compared binding of the conjugated

lectin to euchromatin, heterochromatin, and nucleolus;

compartments which can be differentiated ultrastructurally.

Binding of UEA-I showed that colloidal gold particles were

almost exclusively confined to cross-fractured areas where

euchromatin was exposed. Labelling was abolished by

pretreatment and incubation in the presence of L-fucose, as

expected for a specific label. Pre-digestion of the

fractions with trypsin also abolished labelling, suggesting

the receptors for UEA-I binding were glycoproteins. The

preferential binding to euchromatin may be of importance,

because replication and transcription take place at

euchromatin regions. Although the results reported are

intriguing, the authors did not identify the type of fucose-

containing proteins detected. DNA-associated proteins can

be classified as either histone or non-histone proteins, and

as summarized below, both types of proteins appear to be

fucosylated.



Histones. The histones are the most abundant proteins

associated with DNA. Histones are very basic proteins and

are found in all nuclei (Darnell et al., 1986). Based on

the specific binding to UEA-I, Levy-Wilson (1983) presented








8

evidence that histones isolated from the macronucleus of

Tetrahymena thermophila appear to contain fucose. These

results were strengthened by metabolic incorporation of

[3H]fucose into histones, which showed that all five

Tetrahymena histones, HI, H2A, H2B, H3, and H4, appear to

contain fucose, with H2A incorporating the highest amount

(Levy-Wilson, 1983). In these studies macronuclei were

isolated to high purity, and highly pure histone

preparations were obtained after extensive washing in high

salt to remove nonhistone proteins. Using Con A, the author

showed specific binding to histones, which was interpreted

as evidence for the presence of mannose residues. However,

Con A has previously been shown to also recognize D-glucose

residues. Based on the extent of fucose incorporation and

its specific radioactivity, the author estimated, as the

lowest estimate, that one in a thousand nucleosomes

contained a fucosylated H2A molecule. To date, the

glycosylation pathway of histones and the oligosaccharide

structures) present in histones remain unknown.



Nonhistone proteins. High mobility group (HMG)

proteins are fairly abundant nonhistone chromosomal proteins

classified according to their relative electrophoretic

mobilities (Darnell et al., 1986). HMG proteins undergo a

variety of posttranslational modifications, including

glycosylation. Since they appear to be preferentially








9

associated with actively transcribed DNA, it has been

speculated that glycosylation may influence gene activity.

Highly purified HMG14 and HMG17 from mouse Friend

erythroleukemic cells were found to contain fucose, among

other sugars, by direct composition analysis (Reeves et al.,

1981). These proteins bound UEA-I specifically and could be

metabolically labelled with [3H]fucose. The

oligosaccharides were largely insensitive to B-elimination,

suggesting an N-linkage to protein (Reeves et al., 1981).

When purified HMG14 and HMG17 were digested with mixed

glycosidases, the binding of HMG14 and HMG17 to the nuclear

matrix was abolished. Even though they did not employ a

fucosidase, making it difficult to ascertain the role of

fucose in binding, it was evident that glycosylation

influenced binding to the nuclear matrix (Reeves and Chang,

1983). These studies presented convincing evidence that HMG

14 and 17 are fucosylated, although they did not explore the

site of modification or the composition of the

oligosaccharide(s).



Nuclear fucoproteins and transcriptional activity.

Since some glycoproteins have been found associated with DNA

there has been speculation about a possible correlation

between the state of nuclear glycosylation and

transcriptional activity (Hart et al., 1989a). There are

examples in the literature of glycosylated transcription








10

factors and, at least in one case, glycosylation may have

influenced transcriptional activity (Lichtsteiner and

Schibler, 1989; Jackson and Tjian, 1989). As mentioned

earlier, in the case of fucosylated macromolecules, it was

found that in Novikoff hepatoma cells chromatin had three

times the amount of a fucoprotein of normal liver cells

based on APA binding (Goldberg et al., 1978). However, the

identity, size, or number of such proteins were not well

documented. Putative fucoproteins were found preferentially

associated with euchromatin (Kan and Pinto da Silva, 1986).

Fucosylated histones were found in the macronucleus of

Tetrahymena, where transcriptionally active chromatin is

compartmentalized (Levy-Wilson, 1983). Nevertheless,

histones from non-active chromatin were not studied, so no

comparisons can be made with heterochromatin. Levy-Wilson

(1983) suggested that the reason why other investigators

have failed to detect fucosylation in mammalian histones is

due to the low proportion of transcriptionally active genome

which would imply a low concentration of fucosylated

histones.

Although the data gathered in these reports are

interesting, they are the result of isolated studies from

diverse organisms and it is difficult to draw generalized

conclusions. In addition, virtually nothing is known about

the structure or biosynthesis of the fucose-containing

moieties. Even though the fractions in all these reports








11

appeared to have almost no contamination, structural studies

showing a different glycoconjugate from those found in other

organelles would convincingly argue against contamination.

Nevertheless, these provocative studies are encouraging and

deserve to be pursued further.



Cytosolic Fucoconjugate

The existence of glycoproteins in the cytosol has been

documented in the past and reviewed recently (Hart et al.,

1989a). Studies suggesting the existence of glycoproteins

in the cytosol were based on determinations of lectin

binding sites or biochemical compositional analyses (Hart et

al., 1989a). Some of the negative results reported by those

studies, that relied on lectin binding as confirmation for

the presence of a sugar residue, may be misleading since

lack of binding may reflect a poor choice of lectins to

probe with and not necessarily the absence of a fucose

residue. While there are several fucose-binding lectins

available that serve as useful biochemical tools, they do

not recognize every possible fucose-containing structure.

Lectin binding is dependent on a specific array of sugar

residues, and does not depend solely on the presence of

fucose (Lis and Sharon, 1986). However, as will be

described below, there is one biochemical study that reports

the existence of a cytosolic fucoconjugate. In many cell

fractionation experiments, the cytosolic compartment is









12

defined by the lack of sedimentation during high-speed

centrifugation. However, this criterion alone is not

sufficient, since cytosolic glycans might arise from

contamination by other organelles. The most convincing

evidence describes fucoconjugates that appear to be

preferentially enriched in the cytosol in relation to other

compartments.

In studies in rat brain, a soluble proteoglycan that

contains novel 0-linked mannose-containing oligosaccharides

was recovered in the cytosolic fraction (Margolis et al.,

1976; Finne et al., 1979). The proteoglycan was

characterized as a soluble chondroitin sulfate proteoglycan,

that contained neutral oligosaccharides releasable by mild

alkaline borohydride treatment. The oligosaccharides

contained mannose at their proximal ends, and one

oligosaccharide was proposed to be composed of mannose,

GlcNAc, fucose, and galactose (Finne et al., 1979).

However, the possibility remains that the oligosaccharides

are not integral components of the proteoglycan, but were

associated with other co-purified material. Nevertheless,

the oligosaccharides appeared to be endogenous to the

cytosol and not the result of contamination from other

fractions, since they were present in only trace amounts in

the microsomal or synaptosomal membrane fractions (Finne et

al., 1979). Unfortunately, the function and biosynthetic

pathway of these oligosaccharides remain unknown.








13

Fucose-Bindinq Proteins

The presence of fucoproteins in the nucleus and cytosol

prompted the idea that there might be specific proteins

inside the cell that bind, and/or modify these fucoproteins,

as is the case with the fucoproteins in the secretary

pathway. This would include fucose-binding proteins,

fucosyltransferases responsible for fucosylation, and

fucosidases responsible for fucose removal. There are some

studies suggesting the existence of fucose-binding lectins

and fucosyltransferases. However, there is no evidence for

a nuclear or cytosolic fucosidase.



Endogenous Lectins

In light of evidence for glycoproteins in the cytosolic

and nuclear compartments, investigators sought the existence

of carbohydrate-binding proteins that would colocalize with

such glycoproteins. Endogenous lectins have been detected

in preparations of nucleoplasmic and/or cytosolic fractions

of a variety of cells (Hart et al., 1989a).

Aided by fluorescein-labeled neoglycoproteins, Sbve et

al. (1986) have postulated the presence of endogenous

fucose-specific lectins. Baby hamster kidney cell nuclei

were isolated by two different procedures, cell lysis and

enucleation in Ficoll, to argue against contamination by

cytoplasmic or membrane-derived components. Using

fluorescein-labelled BSA conjugated to fucose in the order








14

of 205 sugar units per molecule, fluorescence microscopy

experiments suggested that the majority of the binding

appeared to be associated with nucleoli and nucleoplasmic

ribonucleoprotein elements (Seve et al., 1986).

Interestingly, it was shown by quantitative flow

microfluorometry that nuclei from exponentially growing

cells bound one order of magnitude more fucose-BSA than

nuclei from contact-inhibited cells. In spite of these

results, the authors acknowledge that it is impossible based

on the data to ascribe biological roles to the nuclear

fucose-binding sites or to conclude that they influence

cellular physiology (Seve et al., 1986).

In Dictyostelium discoideum a family of lectins, the

discoidins, has been identified, and discoidin I has been

the most extensively studied isoform. In erythrocyte

agglutination assays, agglutination by discoidin may be

inhibited by galactose, modified galactose residues, L-

fucose, D-fucose, and other sugars, suggesting the existence

of cell-surface sugar-dependent epitopes on erythrocytes

recognized by Discoidin (Barondes and Haywood, 1979).

Although discoidin I was originally thought to be a cell

surface protein involved in cell-cell adhesion, it has since

been established that it is primarily present in the cytosol

(Erdos and Whitaker, 1983). Although galactose and modified

galactose residues are the ligands bound by discoidin I with

highest affinity, it is possible that a fucose-containing








15

macromolecule may serve as a ligand for it or that another

lectin, possibly from the discoidin family, will be present

in the cytosol with fucose-binding capability.



Fucosyltransferases

Intracellular fucosyltransferases identified to date,

are localized in the lumen of the Golgi or possibly,

endoplasmic reticulum. Thus the presence of fucoproteins in

the nucleus and cytosolic compartments poses questions

pertaining to the mode of synthesis and/or intracellular

transport of such glycoproteins. Early on, Kawasaki and

Yamashina (1972) theorized, based on metabolic labelling,

that nuclear membrane glycoproteins were not synthesized in

and transported from microsomes, but were made in the

nuclear membrane or its vicinity. A few reports have

suggested the presence of glycosyltransferases in the

nucleus, nuclear membranes, or cytosol that may be involved

in the modification of several glycoproteins; however, none

of these glycoproteins was reported to contain fucose

(Richard et al., 1975; Galland et al., 1988, Haltiwanger et

al., 1990).

Though there is evidence for nuclear and cytosolic

fucosylated macromolecules, to my knowledge there are no

reports of nuclear or cytosolic fucosyltransferases.

Louisot and collaborators reported the purification and

separation of two soluble fucosyltransferase activities from








16

rat small intestinal mucosa (Martin et al., 1987). They

report the isolation of al,2 and al,3/1,4fucosyltransferases

that could be candidates for cytosolic fucosyltransferases

based on their inability to sediment after homogenization of

cells in 0.25 M sucrose, followed by 90 min, 200k x g

centrifugation (Martin et al., 1987). They compare the

al,2fucosyltransferase with the a2,6sialyltransferase, which

is normally a Golgi enzyme converted to a soluble form by

cleavage of the amino terminal signal anchor to allow for

secretion (Weinstein et al., 1987). In the case of the

fucosyltransferases, the authors did not report using

protease inhibitors during isolation, nor did they document

the partitioning of cellular markers in the different

fractions (Martin et al., 1987). The reasons the enzymes

localize to the high speed supernatant, which would usually

be considered as the cytosolic fraction, may include the

breakage of the microsomal vesicles or the fact that they

were initially present extracellularly in the body fluids of

the mucosa. In another report these investigators tested

for the presence of glycosyltransferases in the nucleus. In

highly purified rat liver nuclei there was a total absence

of fucosyltransferase activity when endogenous

macromolecules or asialofetuin were used as acceptors

(Richard et al., 1975).

One explanation for the lack of evidence of nuclear

and/or cytosolic fucosyltransferases activity in any








17

organism may be that there indeed are no

fucosyltransferases. Alternatively, if there are no

cytosolic fucosyltransferases, the presence of

fucoconjugates in the cytosol or nucleus may be explained by

membrane-associated enzymes that face the cytosolic

compartment, as has been reported for other

glycosyltransferases (Haltiwanger et al., 1990). In

addition, the lack of adequate acceptors, unfavorable

conditions for in vitro activity, and the scarcity of

sustained interest in the field, could account for the lack

of evidence for such fucosyltransferase(s) that, just as is

the case with other glycosyltransferases, are not of lumenal

origin.

Evidence for fucosylation, and glycosylation in

general, has been accumulating in the past decades. Until

recently, only sporadic reports about fucoproteins appeared

in the literature. With the identification of newly

reported glycoconjugates (Hart et al, 1989b), interest has

been revived in this area of research and currently there

appears to be much interest in the field. However, a

careful examination and sustained interest will be necessary

in order to elucidate the structures, and modes of

biosynthesis and functions of nuclear and cytosolic

fucoconjugates.














CHAPTER II
CHARACTERIZATION OF A FUCOSYLATION MUTANT


Introduction

Since fucosylation comprises numerous steps which are

potential sites for mutations, many fucosylation mutants

have been obtained. Fucosylation consists of the synthesis

of the sugar nucleotide donor GDP-fucose and the transfer of

fucose from the GDP-fucose to an acceptor (Kornfeld and

Kornfeld, 1985). The main source of GDP-fucose is the

conversion pathway of GDP-mannose to GDP-fucose (Yurchenco

et al., 1978; Flowers, 1985). Alternatively, synthesis of

GDP-fucose by the fucose salvage pathway can occur in the

presence of extracellular L-fucose, allowing cells defective

in the conversion pathway to phenotypically revert (Ripka

and Stanley, 1986; Reitman et al., 1980). Lesions may

affect the formation of GDP-fucose, the transport of GDP-

fucose to the fucosylation compartment, the transferases

responsible for fucosylation and/or the biosynthesis or

transport of the acceptor species to the fucosylation

compartment. Two mutants deficient in protein-associated

fucose were shown to be defective in the formation of GDP-

fucose (Reitman et al., 1980; Ripka et al., 1986). A

chinese hamster ovary (CHO) cell line that showed a marked








19

reduction of incorporation of fucose into macromolecules was

unable to synthesize complex-type N-linked oligosaccharides,

resulting in a deficiency of acceptors (Hirschberg et al.,

1982). In another case, two glycosylation mutants were

shown to express a fucosyltransferase activity absent in the

parental cell line with the concomitant expression of a

novel linkage (Campbell and Stanley, 1984).

There are several putative glycosylation mutants. One

of these mutants, HL250, was selected after mutagenesis of

the parental normal strain Ax3 for the inability to bind

anti-SP96 antiserum (Loomis, 1987). This mutant also failed

to express a fucose-dependent epitope recognized by the

monoclonal antibody 83.5 (West et al., 1986). The focus of

my initial investigation was to characterize the mutation in

HL250 with the help of biochemical and morphological

techniques. We determined that HL250 is a fucosylation

mutant that lacked cellular fucose when grown in the absence

of fucose and that the defect is probably a result of a

lesion detectable in vitro in the conversion pathway that

forms GDP-fucose from GDP-mannose.



Materials and Methods

Materials

GDP-[l-3Hjmannose (9.1 Ci/mmol) was purchased from New

England Nuclear and L-[(5,6)-3H]-fucose (60 Ci/mmol) from

American Radiochemical Corporation. KC1 and MgCl2 were








20

obtained from Mallinckrodt; formic acid from Fisher; ATP

(disodium salt, catalog number A-5394), niacinamide, NAD4,

NADPH, Trizma base, phenylmethylsulfonyl fluoride, Dowex-1

(1x8, minus 400, chloride form), hexokinase, GDP-mannose,

and Amberlite MB-3 were obtained from Sigma. ATP was stored

frozen at a concentration of 500 mM in 1 mM Tris-HCl (pH

7.4), resulting in a final pH of approximately 5.5. Dowex-1

format form was made as follows: 1) The column was washed

with 1 M HC1 until pH of eluate was below 2 (as determined

by pH paper). 2) The column was washed with water until pH

was higher than 4.5. 3) Subsequently, the column was washed

with 1 M NaOH until pH was higher than 13. 4) The column

was washed with water as described in step 2. 5) The column

was washed with 3 volumes of 1 M formic acid until pH of

eluate was below 2. 6) Lastly, the column was washed with

water as described in step 2. This procedure was also

followed for regeneration of column.



Strains and Conditions of Growth and Development

Dictyostelium discoideum amoebae were grown on HL-5, a

medium that contains glucose, yeast extract, and proteose

peptone (Loomis, 1971). The axenic strains Ax3 (from F.

Rothman) and HL250 (from W.F. Loomis) were maintained by

passage during logarithmic growth phase. Ax3 is the normal

strain and HL250 is a mutant obtained from Ax3 by N-methyl-

N'-nitro-N-nitrosoguanidine mutagenesis. For metabolic








21

labelling experiments, cells were grown for 4-6 doublings in

8-20 pCi/ml (0.10-0.26 MM) of [3H]fucose in FM medium, a

minimal defined medium that lacks fucose (Franke and Kessin,

1977). When appropriate, the medium was supplemented with

L-fucose (Sigma). For development, cells were plated in PDF

buffer (20 mM KC1, 45 mM sodium phosphate, 6 mM MgSO4, pH

5.8) on filters as previously described (West and Erdos,

1988).



Cell Lysis and Fractionation

Logarithmically growing amoebas were harvested and

washed in 50 mM Tris-HCl (pH 7.5) and resuspended to a

concentration of 2x108 cells/ml in the lysis buffer

consisting of 0.25 M sucrose, 50 mM Tris-HCl buffer (pH 7.5)

supplemented with 1 mM PMSF. All operations were carried

out at 0-4C. Cells were immediately lysed by forced

passage through a 5 Mm nuclepore polycarbonate filter, with

pore diameter slightly smaller than the diameter of the

cells (Das and Henderson, 1986). This method routinely

yields more than 99% cell breakage. The lysate was

centrifuged at 100k xg for 1 hour. The supernatant (S100)

was made 2.5% (v/v) in glycerol by addition of 100% glycerol

and either used immediately or saved at -80 without

significant loss of activity for four weeks.








22

Localization by Immunofluorescence

Prespore and spore cells were examined as described

previously (West and Loomis, 1985; GonzAlez-Yanes et al.,

1989). The monoclonal antibodies utilized have been

described previously elsewhere (West et al., 1986; GonzAlez-

Yanes et al., 1989).



Determination of Fucose Content and Specific Activity

The method is a modification of the protocol developed

by Yurchenco and Atkinson (1975) for HeLa cells. Spores

were harvested from sori not more than a day old and

resuspended in water without washing, since it has been

determined that a significant amount of spore coat protein

may be lost after washing spores in water (West and Erdos,

1990). Vegetative cells were grown in FM (in the presence

or absence of extracellular L-fucose) for determination of

sugar content. For determination of specific activity,

cells were grown in FM media supplemented with [3H]fucose.

After harvesting, cells were washed twice in PDF followed by

EtOH precipitation. The ethanol supernatant was reextracted

with EtOH and the resulting pellet pooled with the previous

precipitate. EtOH was evaporated under a stream of air.

Alternate methods such as TCA precipitation of the ethanol

supernatant, did not significantly increase the yield.

Samples were then hydrolyzed in a reacti-vial (Pierce) in

0.1 N HCl for 45 min at 100 on a heating block.








23

Macromolecules were EtOH precipitated and the remaining

supernatant dried down, resuspended in water, desalted on an

Amberlite MB-3 column and dried by vacuum centrifugation.

The samples were redissolved in water and analyzed using a

Dionex Bio-LC ion chromatograph by the method of Hardy et

al. (1988). Standards and modifications to the

chromatography procedure have been published elsewhere

(Gonzdlez-Yanes et al., 1989). Specific activity, when

applicable, was determined by counting elution fractions and

was expressed as radioactivity present in the fucose peak

divided by the amount of fucose detected by the pulsed

amperometric detector coupled to the HPLC, with reference to

previously established calibration curves for L-fucose

(Hardy et al., 1988). Greater than 95% of the eluted

radioactivity was recovered from the column eluate at the

elution position of fucose.



Assay for Conversion of GDP-mannose to GDP-fucose

The conversion assay was carried out essentially as in

Ripka et al. (1986). In short, the standard assay mixture

contained in a final volume of 1 ml 600-750 pg S100 protein,

10 mM niacinamide, 5 mM ATP, 0.2 mM NAD, 0.2 mM NADPH, 7.5

pM GDP-[ H]mannose (approximately 105 cpm) in 50 mM Tris-

HC1, pH 7.5. After incubation at 37 the reaction was

stopped with 50 pi of 2 N HC1, the reaction mixture boiled

for 20 min and subsequently neutralized with 55 pl of 2 N








24

NaOH. Quantitation of the conversion of GDP-mannose to GDP-

fucose was achieved by determining the amount of fucose

present after acid hydrolysis. Free mannose, released from

GDP-mannose, was phosphorylated by adding 4 units of

hexokinase in the presence of 5 mM ATP and 5 mM MgCl2 at 37

for 1 hour. Hexokinase catalyzes the transfer of one

phosphate to C-6 of any acceptor hexose. Since fucose lacks

a hydroxyl group at position C-6, it cannot be

phosphorylated. Parallel controls with no enzyme were used

to correct for losses. The mixture was passed over a Dowex-

1 formatt) column (0.6 x 5 cm) and eluted with water. 1 ml

fractions were collected an aliquot of 100 p1 from the

eluate was counted using ScintiVerse LC (Fisher); fucose-

associated radioactivity usually eluted by the first 2 ml.

Calculation of K and V was done by the Lineweaver-Burk
m max

double reciprocal plot (1/v vs. 1/[S]) as discussed by

Henderson (1985).



Results

Phenotypic Description of a Fucosylation Mutant

The normal strain Ax3 was mutagenized with

nitrosoguanidine and surviving clones were screened with

anti-SP96 antiserum (Loomis, 1987). One of the clones,

HL250, was selected due to its inability to bind anti-SP96

antiserum which recognizes carbohydrate and peptide

epitopes. HL250 has been found to have a more permeable








25

spore coat, lower germination efficiency in older spores,

and a longer doubling time when compared to the parental

strain Ax3 (Gonzalez-Yanes et al., 1989; West et al.,

manuscript in preparation).



Absence of a fucose-dependent epitope. The failure of

HL250 to react with anti-SP96 antiserum suggested that the

mutant might be deficient in a form of protein

glycosylation. This hypothesis was confirmed by finding

that HL250 did not react at appreciable levels with the

fucose-dependent monoclonal antibody (mAb) 83.5 (West et

al., 1986; Gonzalez-Yanes et al., 1989). Spores from mutant

and the parental normal strains were subjected to SDS-PAGE

and Western blotting and probed with mAbs 83.5 and A6.2

(West et al., 1986; Gonzalez-Yanes et al., 1989). The

latter monoclonal is specific for SP96. Consistent with the

supposition that glycosylation is affected in HL250, SP96

was reduced in apparent molecular weight compared to SP96

from Ax3 spores (West et al., 1986). The absence of the

fucose-dependent epitope was further examined in developing

cells and spores. Cells were plated for development and

slugs dissociated by shearing in the presence of EDTA.

Cells and spores were then processed for indirect

immunofluorescence. Figure 2-1 shows the localization of

SP96 in spores and prespore cells. Ax3 exhibits peripheral

labelling of spores and a punctate labelling from prespore

































Figure 2-1. Localization of SP96 in prespore and spore cells.

Slugs cells were dissociated, placed onto polylysine-coated
glass slides, fixed, permeabilized, and processed for
indirect immunofluorescence using mAbs 83.5 or A6.2.





















PRESPORE CELLS


83.5


A6-2


SPORES


83.5


A6-2









28

vesicles using both mAb. However, HL250 shows the same

pattern only when A6.2 is used, in agreement with the

results observed by Western blotting. The fact that

labelling of spores with A6.2 is similar in both strains

indicates that spore coat localization of SP96 is not

affected by the mutation.



Fucose content of normal and mutant strains. Epitope

recognition by mAb 83.5 was inhibited by L-fucose (West et

al., 1986), so the possibility of a defect in fucosylation

in strain HL250 was investigated. Initially, the

macromolecular fucose content of vegetative cells grown in

fucose-free media, and of spores was investigated. Ethanol

insoluble macromolecules were acid hydrolyzed, ethanol

precipitated, and the supernatant deionized and

chromatographed on an alkaline anion-exchange column

equipped with a pulsed amperometric detector. When

authentic [3H]fucose was fractionated in this manner, more

than 95% of the radioactivity eluted at the fucose position

(not shown). Fucose content was found to be negligible in

the mutant, HL250, both in spores and vegetative cells when

compared to the normal strain, Ax3 (table 2-1). However,

when HL250 cells were grown in FM supplemented with 1 mM L-

fucose, they possessed detectable amounts of fucose.

Previous investigators have found by autoradiography that

[3H]fucose incorporation is highest in prespore cells
















0,-I

or-i
+1 +1


:0
co
'-1
+1 ,w

C0 *
in o


1-I*'-
+1 +1
-I-I "1-1
iL n
m qw
ao o
rv).a.


00
o.* *r()
In -< +1 .-4
+1 +1 ( +1










*C4 0
+1 0 Lnl 0
S+1 r. +1

V-4 w C 4
.-i 1o oc4


0 0) a)
Sd 02 U)
0 0 0
U U
4-4
0 4- 4.4

02
C1 r- -1
o

*-4

+ +

o a s a s.

0)


4 ) Q) .U a
r-4 $4 $4 a) a()
'-4 00 0 0
U wO m (a (a




-.4 0 0
to Ln Ln
$-4 MC%4 mY CN
+4- x 9 X 9
ml < = 4


p 2: v0 0
a)rX4 Z (1)
(a 4- )

(0 i+
s*s


r.4 0
a ,



(0~ flU
S 4-4


4-1 W.4




ds D




a4 ) (~D
u 0 0






Sa) a,$
5 444







4n c .i
$4~4




4-4 q)a,*

a) a)
*-i6 n
'o ( 1 a







0 a) oU
4-) C 0L:












0 44J W4.3
0 a) a)

p> O' to
0 4-)M




^ u x

a) 4: n "




U wU.



Ml C: a) va r
a) -H *-4 0 w

04 p a)s a)
(n oil aa z 1o








30

compared to prestalk and vegetative cells, so higher levels

of fucose would be expected in spores (Lam and Siu, 1981;

Gregg and Karp, 1978). The levels of glucose and mannose

were measured for comparison to determine if there was a

difference in the amounts of other sugars in the mutant.

None of the contents of the other sugars were reduced in the

mutant when grown in FM. Thus it seems that the lesion in

HL250 is selectively affecting fucose metabolism, relative

to that of other sugars. This is consistent with the fact

that fucosylation is usually a terminal modification of

oligosaccharides, so its addition is not a prerequisite for

the addition of other sugars (Kornfeld and Kornfeld, 1985).



Characterization of the Mutant Lesion

Earlier investigations have described mutants with a

fucose minus phenotype that are the result of a defect in

GDP-fucose biosynthesis. The main source of GDP-fucose is

the conversion pathway of GDP-mannose to GDP-fucose

(Yurchenco et al., 1978; Flowers, 1985). It consists of the

reactions presented in figure 2-2. Alternatively, synthesis

of GDP-fucose by the fucose salvage pathway can occur in the

presence of extracellular L-fucose (figure 2-2), allowing

cells defective in the conversion pathway to phenotypically

revert (Ripka and Stanley, 1986; Reitman et al., 1980).

HL250 cells grown and developed in the presence of 1 mM L-

fucose reexpressed the fucose epitope (Gonzdlez-Yanes et

































Figure 2-2. Biosynthesis of GDP-fucose.

Diagrams of the conversion pathway elucidated in bacteria
and mammalian cells (adapted from Flowers, 1981) and the
salvage pathway as described in HeLa cells (adapted from
Yurchenko et al., 1978).




























CH20H CH3 H
0 'H O-0 lH C 0 4 .c~ H 0---0\
4,6-ehydrete C ,5-.pm.r 4- renuctm

O H. O' 0O-DP 0- N 11 1 L O-GDP

H H H H ON M OH H


A CONVERSION PATHWAY



H 0H
H N H

/ CH3 3 J fucm / CHI pyroplsphoryIloe |/CH \
1\ ON k~, nose N OMN ---- > H M y
I\LH OH"
OO N ON 0- p OH O-GDP
OH H OH H OH H


B. SALVAGE PATHWAY








33

al., 1989), and vegetative cells grown in 1 mM L-fucose had

detectable amounts of macromolecular-associated fucose

(table 2-2). These results suggested that HL250 had a

normal salvage pathway, but a defect in the GDP-mannose to

GDP-fucose conversion pathway.



In vitro conversion of GDP-mannose to GDP-fucose.

Conversion of GDP-mannose to GDP-fucose has been reported in

bacteria, in higher plant cells, and mammalian cells

(Kornfeld and Ginsburg, 1966; Liao and Barber, 1971; Ripka

et al., 1986). It was assumed that Dictyostelium would

share this ability with other species, so I assayed if cell

extracts in vitro were able to convert GDP-mannose to GDP-

fucose. High speed supernatants from Ax3 and HL250 were

assayed in vitro for their ability to convert GDP-

[ 3H]mannose into GDP-[3H]fucose. Cells were homogenized and

a 100k x g soluble fraction (Si00) assayed as described by

Ripka et al. (1986) and the effects of time, varying protein

and GDP-mannose concentrations were examined. Table 2-2

shows that the conversion activity in Ax3 was proportional

to the time of incubation. In contrast, mutant extracts

showed negligible activity. Ax3 and HL250 cytosols were

mixed to determine if a soluble inhibitor of GDP-mannose to

GDP-fucose conversion activity was present. When equal

amounts of extracts were mixed, activity was commensurate

with the Ax3 contribution (table 2-2) indicating HL250 does




























Table 2-2.
GDP-fucose.


Effect of time on conversion of GDP-mannose to


nmol fucose/mg protein


time (min)


7.5
15


Ax3

0.57
2.5
2.9
9.2


HL250

0
0
0.15
0


Ax3+HL250 (0.5:0.5)


0.21
1.3
2.2
4.3


Protein (600 pg total) from a 100,000 xg supernatant of a
vegetative cell-free extract was assayed for ability to
convert GDP-[ 14C]mannose (7.5 pM initial concentration) to GDP-
[14C]fucose; data are the result of the average of two
determinations








35

not contain an inhibitor for the activity. Conversion was

linear with respect to protein through 800 Mg (figure 2-3,

panel A). In agreement with the results from the time

dependence experiment, HL250 expressed less than 1% of the

activity possessed by Ax3 at all concentrations of protein

assayed (figure 2-3, panel A). Conversion by Ax3 S100 was

also dependent on the amount of GDP-mannose present while

the activity present in HL250 was insignificant (figure 2-3,

panel B). The Ax3 GDP-mannose to GDP-fucose conversion

activity showed an apparent KM of 14.1 yM and V of 18.3

nmol fucose/mg protein/30 min (fig. 2-3, panel C). Previous

reported values for the apparent KM of the conversion

pathway range from 2 MM in CHO cells (Ripka et al., 1986) to

160 pM in the higher plant Phaseolus vulgaris (Liao and

Barber, 1971).

The conversion activity is absent in mutant extracts in

vitro, if absent in vivo, this would explain the lack of

fucose in living cells and the correction by exogenous

fucose. Taken together, all these results point to the

conversion pathway as the site of the lesion in HL250.

Furthermore, the defect seems to be in the first step of the

conversion of GDP-mannose to GDP-fucose because no

radioactivity was recovered after hexokinase treatment. If

the 4,6-dehydratase was active, the product would have

eluted from the ion-exchange column after hydrolysis and

phosphorylation by hexokinase (see figure 2-2).




























Figure 2-3. Conversion of GDP-mannose to GDP-fucose by normal
and mutant strains. Closed circles, Ax3; open circles,
HL250. Results are the average of two determinations.

Panel A. Effect of protein concentration on conversion. 30
min. assay, 7.5 pM GDP-[3H]mannose.

Panel B. Effect of GDP-mannose concentration on conversion.
30 min assay, 750 pg protein for each strain.

Panel C. Apparent Michaelis constant for GDP-[3H]mannose,
determined for Ax3 conversion. Apparent K was determined
to be 14.1 pM and apparent V 18.3 nmol/mg protein/30 min
by the Lineweaver-Burk double reciprocal plot method.


























0 200 400 600 800
ig protein


0 5 10 15 20 25
GDP-mannose (uiM)


0.3--
c



0.2



0.1 -



0.0-"^
-0.09


-0.00 0.09 0.18 0.27
1/[S]








38

Specific activity of fucose pools. A lesion in the

GDP-mannose to GDP-fucose conversion pathway would render

the mutant defective in macromolecular fucosylation when

grown in the absence of fucose, as was shown earlier. Such

a scenario would require that the GDP-fucose inside the cell

be derived from the salvage pathway fed exclusively by

fucose from the extracellular media. In the case of the

Ax3, however, there would be a contribution of GDP-fucose

derived from the conversion pathway. If the mutation in

HL250 was indeed in the GDP-mannose to GDP-fucose

conversion, macromolecular fucose of cells grown in the

presence of L-[3H]fucose would have the same specific

activity as the fucose present in the media. To test this

hypothesis, mutant and normal cells were grown in fucose-

free defined media supplemented with [3H]fucose. Whole cell

preparations were then assayed for fucose content as

described above and the specific activity expressed as the

radioactivity present in the fucose peak divided by the

amount of fucose detected. As expected, the macromolecular

pool of the mutant had essentially the same specific

activity as the fucose in the medium. In contrast, the

specific activity in Ax3 was diluted approximately 400-fold

compared to the medium (table 2-3). These results confirmed

that HL250 derived its intracellular fucose from the salvage

pathway. Meanwhile it appears that in Ax3 the contribution






























Table 2-3. Specific activities of fucose.


fucose concentration


50 *M
0.1 PM


cpm/nmol fucose
medium macromolecular


1. 1x105
1.8x107


9.5x10'
4.7xlO4


Cells were grown for 3 days in FM media in the presence of
6x106 dpm/ml of [3H]fucose supplemented with non-radioactive
fucose to yield the noted concentration of fucose in the
media. Specific activity was determined as described in
Materials and Methods.


strain

HL250
Ax3









40

of fucose by the salvage pathway is minimal, 1/400 of the

total content. These studies agree with earlier reports on

fucose metabolism, where the majority of fucose in mammalian

cells is derived from the conversion of GDP-mannose to GDP-

fucose (Yurchenko et al., 1978).



Discussion

HL250 failed to express a fucose-dependent epitope

recognized by the mAb 83.5. However, it expressed SP96, one

of the polypeptides that bears the carbohydrate epitope

recognized by 83.5. These results were reproduced by

Western blot (GonzAlez-Yanes et al., 1989) and indirect

immunofluorescence. Interestingly, the immunofluorescence

microscopy studies showed that the mutant is not defective

in its ability to package SP96 in vesicles or in the

targeting of the glycoprotein to the spore coat. Similar

results were reported for other proteins in another

Dictyostelium glycosylation mutant (Aparicio et al., 1990;

West and Loomis, 1985). Measurements of fucose content of

cells and spores demonstrated that the mutant contained

almost undetectable amounts of fucose, in contrast to Ax3

which contained macromolecular-associated fucose in both

cell types.

Once HL250 was identified as having a mutation that

resulted in decreased macromolecular fucose, I tried to

identify the nature of the lesion. It was speculated that








41

the mutant HL250 may have a defect in (1) fucosyltransferase

activities, (2) endogenous acceptors for

fucosyltransferases, (3) transport of GDP-fucose into

microsomal vesicles, and/or (4) synthesis of GDP-fucose. In

vitro microsomal extracts of normal and mutant cells were

active in the transfer of [14C]fucose from GDP-[14C]fucose to

endogenous acceptors and the activity was latent (see

Chapter IV), so it was reasoned that fucosyltransferases may

be normal and probably the uptake of GDP-fucose by vesicles

was not impaired, so the lesion might be at another point in

the fucosylation pathway. Since fucose is normally added as

a terminal modification, the fucose minus phenotype could be

the result of a lack of formation of acceptors for the

fucosyltransferases (Stanley, 1984; Hirschberg et al.,

1982). HL250 expresses levels comparable with Ax3 of other

carbohydrate epitopes and has normal neutral monosaccharide

composition (Gonzdlez-Yanes et al., 1989; West et al.,

1986), for these reasons it was speculated that the defect

was not in an earlier step of glycosylation but it involved

the fucosylation pathway directly.

The glycosylation defect in HL250 appears to result

from an inability to produce GDP-fucose. I have found that

the GDP-mannose to GDP-fucose conversion activity in vitro

is reduced to undetectable levels in mutant cell extracts.

The fact that there is partial rescue when the cells are

grown in the presence of fucose, suggests that the cells are








42

producing GDP-fucose via the salvage pathway and that the

rest of the fucosylation machinery is probably normal.

Earlier studies have reported the phenotypic reversion of

mammalian mutants with a defective GDP-mannose to GDP-fucose

conversion pathway when the cells were supplied with

extracellular fucose (Ripka and Stanley, 1986; Reitman et

al., 1980). However, Ripka and Stanley (1986) used the

recovery of lectin sensitivity as a marker for phenotypic

reversion, but did not report measuring the fucose content

of the cells or show the data for lectin binding compared to

the parental strain. Reitman et al. (1980) showed that a

mouse lymphoma cell line which has a defect in the

conversion of GDP-mannose to GDP-fucose was defective in pea

lectin binding compared to the parental cell line. The

ability to bind pea lectin was restored to wild type

parental cell line levels after culturing in 10 mM fucose.

Fucose is an important determinant in the carbohydrate-

binding specificity of pea lectin (Kornfeld et al., 1981).

It is important to note that the mutant mouse lymphoma cell

line had approximately one fifth the amount of fucose and

one third the number of high affinity lectin-binding sites

as the parental line, indicating that mutant cells were

salvaging fucose from the medium, or that the mutation was

only partial. The evaluation for phenotypic reversion of

HL250 is more rigorous since it demands the expression of a








43

carbohydrate epitope and measures total levels of fucose

from a cell that was grown in fucose-free media.

The specific activity of the medium was compared to the

specific activity of the intracellular macromolecular

fucose. Consistent with a lesion in the GDP-mannose to GDP-

fucose conversion pathway, the mutant cells relied on

extracellular fucose as their only fucose source. In

contrast, extracellular fucose only contributed to a small

fraction of the total Ax3 fucose pool. This is useful

because it means that radioactivity from cells grown in

[3H]fucose can be used as a direct measure of fucosylation

in HL250.

In conclusion, even though HL250 has a severe

glycosylation lesion which renders it unable to carry out

fucosylation when grown in the absence of fucose, the strain

is able to grow, develop, and form spores. To my knowledge,

there are no previous reports in the literature of

eukaryotic cells defective in the GDP-mannose to GDP-fucose

conversion pathway that can survive in fucose-free media.

The fucosylation mutants reported are cell lines that have a

functional salvage pathway maintained in culture in the

presence of animal serum, so they can synthesize GDP-fucose

from the fucose present in the cell culture media (Ripka and

Stanley, 1986; Reitman et al., 1980). For this reason,

these investigators were unable to totally deprive the

mutants of fucose, as I am able to do with Dictyostelium.








44

Other lower eukaryotes, such as yeast, do not carry out

fucosylation (Kukuruzinska et al., 1987) so the existence of

a Dictyostelium fucosylation mutant could be very important

to study fucosylation. In any event, HL250 has already

served as a very useful tool in which to study fucosylation

events, as will be evident in the following chapters.















CHAPTER III
IDENTIFICATION OF A CYTOSOLIC FUCOPROTEIN


Introduction

There is some evidence that fucoconjugates are present

in the nucleus and cytosol (Hart et al., 1989a; Chapter I).

However, virtually nothing is known about the structure or

biosynthesis of these fucosylated macromolecules. Most

studies have limited themselves to reporting the existence

of evidence for nuclear or cytoplasmic glycoproteins, but

have not gone further to characterize the sugar-peptide

linkage or compare it with material derived from the

secretary pathway. As discussed in Chapter II, there is a

conditional fucosylation mutant, HL250, that can be readily

labelled when grown in radioactive fucose. Using this

strain, I have identified a fucoprotein that fractionated

with the cytosol and appeared to be the major fucosylated

species in the cytosol. The oligosaccharide-peptide linkage

was characterized and the fucoprotein was compared with

fucosylated material derived from vesicles, and

differentiated based upon several criteria.








46

Materials and Methods

Materials

L-[(5,6)-3H]-fucose (60 Ci/mmol) was obtained from

American Radiochemical Corporation and D-[(2,3)-3H]mannose

(24 Ci/mmol) from New England Nuclear. TS-1 was purchased

from Research Product International; POPOP was from

Mallinckrodt; SDS, leupeptin, aprotinin, phenyl

methylsulfonyl fluoride, Triton X-100,

dimethyldichlorosilane, MES, all nitro-phenyl substrates,

mannose-6-phosphate, bovine serum albumin (fraction V), blue

dextran, bromo phenol blue, and trypsin were from Sigma;

glycine, benzene, ammonium acetate, PPO, and toluene were

from Fisher. The concentrations of Triton X-100 and NP-40

are expressed as v/v, all others are expressed as w/v,

unless specified otherwise.



Strains and Conditions of Growth

Dictyostelium discoideum strains Ax3 (from S. Free) and

HL250 (from W.F. Loomis) were grown on HL-5, a complete

medium that contains glucose, yeast extract, and proteose

peptone (Loomis, 1971). Ax3 is the normal strain and HL250

is a mutant obtained from Ax3 by N-methyl-N'-nitro-N-

nitrosoguanidine mutagenesis (Loomis, 1987). HL250 lacks

the enzyme activity that converts GDP-mannose into GDP-

fucose which results in a lack of cell fucose (Gonzalez-

Yanes et al., 1989; also see Chapter II). In all








47

experiments cells were collected at the logarithmic growth

phase, with a cell density of 1 to 9 x 106 cells/ml. For

metabolic labelling experiments, cells were grown for 4-6

doublings in 2-20 pCi/ml (0.03-0.26 pM) of L-[3H]-fucose in

FM medium, a minimal defined medium that lacks fucose

(Franke and Kessin, 1977).



Cell Lysis and Fractionation

Logarithmically growing amoebae were harvested and

washed in 50 mM MES (pH 7.4) and resuspended to a

concentration of 2x108 cells/ml in the lysis buffer

consisting of 0.25 M sucrose, 50 mM MES buffer (pH 7.4)

supplemented with the protease inhibitors leupeptin (10

pg/ml), aprotinin (10 pg/ml), and PMSF (1 mM). When

specified, cells were fractionated in the presence of a

comprehensive cocktail of protease inhibitors which have

been developed for the isolation of various proteolitically

sensitive proteins in Dictyostelium (Goodloe-Holland & Luna,

1987; Stone, et al., 1987). All steps were carried out at

0-4C. At once, the cells were gently lysed by forced

passage through a 5 pm nuclepore polycarbonate filter, with

a pore diameter slightly smaller than the diameter of the

cells (Das and Henderson, 1986). This method routinely

yielded more than 99% cell breakage as assessed by contrast

phase microscopy. The lysate was clarified from unbroken

cells and nuclei by a 2k xg centrifugation for 5 min, then








48

it was centrifuged at 100k xg for 1 hour, unless otherwise

specified. The pellet (P100) was resuspended in lysis

buffer by pipetting to the same volume as the supernatant

(S100). For lysing vesicles, the P100 was sonicated using a

Branson Sonifier Cell Disrupter 185.

Slug cells were plated for development as described

earlier (West and Erdos, 1988), and harvested in buffer of

Berger and Clark (as described in West and Brownstein, 1987)

supplemented with 20 mM EDTA. Cells were dissociated in

this buffer by passing 20 times through a long, 9 inches,

pasteur pipette, followed by passing 20 times through a 23-

gauge needle. EDTA was washed by resuspending cells in 50

mM MES, pH 7.4, titrated with NaOH. Cells were resuspended

in lysis buffer and immediately lysed by passage through a 3

pm nuclepore polycarbonate filter (Das and Henderson, 1986).

Cell lysates were then treated as described above for

vegetative cells.



Gel Electrophoresis and Western Blotting

SDS-PAGE was carried out under reducing conditions

essentially as described in West & Loomis (1985). Samples

were resolved by 7-20% acrylamide linear gradient gels or

15% acrylamide gels and, initially, molecular weight

assigned using low MW markers kit (Sigma). In later

experiments, trypsin was used as a molecular weight marker.

Following electrophoresis, the gels were cut immediately








49

and/or stained and destined and then cut into either 2.2 mm

or 0.5 cm slices. Gel pieces were shaken and swollen

overnight in a scintillation cocktail composed of 111.1 ml

of tissue solubilizer (TS-1), 6.0 g PPO, 0.15 g POPOP, and

20 ml dH2O to 1 1 of toluene. Gel slices were counted and

recounted until dpm were determined to be stable, usually 1-

2 days later. For gel-purified material, the sample was run

in a 7-20% linear gradient gel and the 21 kD area

(approximately 1 cm below trypsin) cut out and electroeluted

overnight using a Bio-Rad electroeluter following

manufacturer's directions, except the Laemmli

electrophoresis buffer used for SDS-PAGE (West and Loomis,

1985) was used instead of the recommended volatile buffer to

avoid alkaline hydrolysis. Western blotting was carried out

as previously described (West and Loomis, 1985).


Partial Purification of FP21 by Anion Exchange
Chromatography

In preliminary studies, FP21 was partially purified by

fractionation of metabolically labelled S100 fraction on a

TSK DEAE-5PW 8 x 75-mm HPLC anion-exchange column (LKB)

preequilibrated with 10 mM NH4Ac, pH 7.0. Sample was

dialyzed against 2 1 of 10 mM NH4Ac, pH 7.0, for several

hours, and clarified by centrifugation at 10k x g for 10 min

prior to injection. Protein was eluted using an increasing

linear gradient (10 mM to 1 M NH4Ac, pH 7.0) for 40 min at a

rate of 0.75 ml per min, and the majority of FP21 was found









50

to elute at 0.5 M input buffer concentration. Fractions

were analyzed by SDS-PAGE and counting of the gel slices.



Protein Concentration Assay

The Bio-Rad protein assay was used for determination of

protein concentration, and bovine serum albumin used as

standard.



Enzyme Assays

a-glucosidase-2 assays contained 100-300 yg protein,

8.6 mM p-nitrophenyl-a-D-glucoside, 0.1% Triton X-100, in 21

mM citrate-phosphate buffer (pH 7.5) at 37 (Borts and

Dimond, 1981). Reaction was stopped after 1 hr by addition

of Na2CO3 to a concentration of 0.5 M and the absorbance

read at 420 nm. Glucose-6-phosphatase and mannose-6-

phosphatase were measured by release of phosphate from

mannose-6-phosphate, which has been previously shown to be a

suitable substrate for both enzymes (Arion et al., 1976).

Reaction mixtures contained 100-300 pg protein, 1 mM MgCl2,

2 mM mannose-6-phosphate, 0.1% Triton X-100 in 10 mM MES

titratedd with NaOH to a pH of 7.4) in a volume of 200 p1.

After 20 min incubations at 30 reactions were stopped by

adding 200 M1 20% ice-cold TCA (Snider et al., 1980). Tubes

were centrifuged at 14k rpm, for 10 min in an Eppendorf

table top microfuge and aliquots of the supernatants were

assayed for Pi by the method of Chen et al. (1956). Acid









51

phosphatase was assayed as described in McMahon et al.

(1977) except that Triton X-100 was included at a

concentration of 0.1% and absorbance was measured at 420 nm,

instead of 400 nm.



PNGase F Digestion

Gel-purified FP21 (3000-8000 dpm) was boiled in 0.5%

SDS in water for 3 min. For digestion, the protocol of

Tarentino et al. (1985) was followed. The sample was

incubated in 104 mM sodium phosphate, pH 8.6, 10 mM EDTA, 10

mM 1,10-phenanthroline (stock solution of 100 mM in

methanol), 2% NP-40, 0.21% SDS, and 20 U/ml PNGase F

(Boehringer-Mannheim) in 300 pi for 22 h at 37. Fetuin (B-

grade, Calbiochem) and RNAse B (Sigma) were treated

identically as controls and digestion was quantitative, or

nearly quantitative, as determined by shifts in molecular

weight in SDS-PAGE. In one trial FP21 was digested with

trypsin prior to PNGase F digestion by incubating gel

purified FP21 in 0.8 mg/ml trypsin, 1 mM CaCI2 for 1 hr at

37. The reaction was stopped by boiling in the presence of

1 mM PMSF and 10 pg/ml aprotinin for 3 min.



Pronase Digestion

Glycopeptides were prepared by exhaustive pronase

digestion as described (Ivatt et al, 1984). In short, after

gel purification and electroelution, the samples were dried








52

down and resuspended in water to a volume of 200 il

containing at least 5x103 dpm. In other experiments, 200 il

of the entire P100 fraction were left intact or made 0.1%

Triton X-100. 200 il of freshly dissolved 1% pronase

(CalBiochem) in 0.1 M Tris-HCl (pH 8.0), 1 mM CaCI2, were

added at 0, 24, and 48 hours. Incubation was at 50 and a

few drops of toluene were added to prevent microbial growth.

At 72 hours the reaction was stopped by incubating in a

boiling water bath for 3 min.



Oligosaccharide Release by Alkaline-Borohydride Treatment

The oligosaccharide in FP21 was released by mild

alkaline hydrolysis under reducing conditions, also known as

B-elimination. To approximately 5x103 dpm of gel purified

protein, freshly made NaOH and NaBH4 concentrated solutions

were added in that order to yield a final concentration of

0.1 M and 1 M, respectively. Samples were incubated for 15

hours in a water bath at 45. The reaction was stopped by

the addition of acetic acid to a final concentration of 1 M.

The samples were dried down by vacuum centrifugation,

resuspended once in 1 ml 100 mM HAc, dried again, and

resuspended twice in methanol and stored dry at -80 until

ready to use.

The oligosaccharide was also released by strong

alkaline-borohydride treatment. The method is that of Zinn

et al. (1978) and very similar to the procedure for 3-









53

elimination, with some exceptions: NaOH and NaBH4 were

present at a final concentration of 1 and 4 M, respectively,

and samples were incubated at 80 for 24 h. Reaction was

terminated by diluting the sample twofold with water and

adding acetic acid to a final concentration of 4 M. Borate

salts were removed by methanol evaporation as described

above.



P-4 Column Fractionation

Dry samples from B-elimination and strong alkaline-

borohydride treatment were resuspended in 800 4I of 50 mM

pyridinium acetate (pH 5.5). Pyridinium acetate was made in

the hood by mixing in water, to a final volume of 2 liters,

8.06 ml pyridine, 2.17 ml glacial acetic acid. After the

previous reagents were dissolved, 0.4 g of sodium azide was

added. The solution had a final pH of approximately 5.5.

The buffer was degassed and stored under chloroform

atmosphere. Samples from pronase digestion were centrifuged

for 5 min at 14k rpm on a Eppendorf table top microfuge and

the supernatant taken for P-4 chromatography.

Oligosaccharides and glycopeptides were fractionated in a

0.9 cm x 1 m BioGel P-4 column (-400 mesh) equilibrated with

50 mM pyridinium acetate (pH 5.5) as the mobile phase.

Prior to pouring, the column was acid washed overnight and

siliconized by coating with 1% (v/v) dimethyldichlorosilane

in benzene for 10 min. The column was calibrated with








54

glucose oligomers used as standards (Yamashita et al., 1982)

that were derived from a dextran hydrolysate which was

reduced with NaB3 H4 (kindly provided by J. Baezinger).

Twelve drop fractions (approximately 250 pi) were collected

and counted using ScintiVerse LC (Fisher Scientific). All

runs were performed at 37. Recovery varied somewhat

between runs, but it was between 30-80% of the loaded

radioactivity. The void volume (Vo) was determined with

blue dextran (at a concentration of 0.2%), and the inclusion

volume (Vi) with either bromo phenol blue (0.2%) or

[3H]mannose (approximately 2,000 dpm). The relative elution

coefficient (Rev) for each component was determined from the

elution volume (Ve): Rev = (Ve-Vo)/(Vi-Vo).



Results

Analysis of Cellular Fucoproteins by SDS-PAGE

HL250 amoebae were grown in minimal defined media

supplemented with L-[ 3H]fucose. During logarithmic growth

phase, cells were harvested, washed, and resuspended in 0.25

M sucrose buffer supplemented with protease inhibitors.

Immediately, the cells were lysed and the lysate was

clarified from unbroken cells and nuclei by centrifugation

at 2k x g for 5 min. The resulting supernatant from this

spin was centrifuged at 100k x g for 1 hour. Both fractions

were analyzed by SDS-PAGE and the gels stained, cut into 2.2

mm slices, and counted. While total protein distributed in








55

a ratio of almost 1:1 (P100:S100), the specific activity

(expressed as dpm/mg of protein) distributed roughly in a

6:1 ratio (P100:S100) (table 3-1). The majority of the

radioactivity fractionated with the P100. However, the

amounts of radioactivity recovered in the S100 were

unexpected; 36% of the radioactivity in the S100 (this value

varied in individual experiments from 30-68%) migrated as

one peak at the 21 kD position, slightly ahead of trypsin,

which was used as a molecular weight marker. On the other

hand, the P100 showed two main broad peaks, one at 10-25 kD

and another at 58-84 kD with 10% of the total radioactivity

in the P100 migrating at the 21 kD level (figure 3-1). The

proteinaceous nature of the S100 fucoprotein was confirmed

by digestion with the proteases trypsin and pronase with

quantitative recovery of radioactivity at lower MW positions

in unfixed gels (results from pronase digestion shown in

figure 3-2). The S100 fucosylated protein has been called

FP21 for fucoprotein of 21 kD molecular weight.

The abundance of FP21 was estimated based on the

specific activity of fucose and determined to be 103

molecules in FP21, assuming one fucose molecule per molecule

of FP21. Based on the dilution of fucose specific activity

in Ax3 (from table 2-3, Chapter II), there would be 4 x 105

molecules in Ax3. If there is one fucose per copy of FP21,

and all FP21 molecules are fucosylated (evidence for

quantitative FP21 fucosylation in Ax3 will be presented in



























Table 3-1. Distribution of protein and radioactivity in S100
and P100 fractions.


S100


total protein
(equivalents)


P100


0.95


specific activity
3 days
5 days


14 dpm/pg
13 dpm/pg


88 dpm/Mg
225 dpm/pg


HL250 amoebae were grown in the presence of 2 pCi/ml of
[13H]fucose in FM media for the indicated period of time. Data
are from one representative experiment. Cells were harvested,
filter lysed, and fractionated into an S100 and P100.





























Figure 3-1. Incorporation of [ 3H]fucose into macromolecular
species of the S100 and P100.

HL250 amoebae were metabolically labelled with 2 pCi/ml of
[3Hjfucose, lysed, fractionated into an S100 and P100, and
subjected to 7-20% linear gradient SDS-PAGE; the gel was
sliced into 2.2 mm pieces and counted. 100 pg of protein
were electrophoresed for the S100 and P100, respectively.
Open circles, P100; closed circles, S100; arrow, migration
of trypsin.























2000 600


P~- 500
1600- 500

] '-400
0n 1200- n
o 0
S-300

800n 2
n 1I200


400 1 0-10


0 -
0 10 20 30 40 50

gel slice
































Figure 3-2. Proteinaceous nature of FP21.

Approximately 1,500 dpm of metabolically [3H]fucose-labelled
FP21 from HL250 was gel purified, as described in Materials
and Methods, and subjected to either a mock or pronase
digestion. Resulting digests were electrophoresed on a 15%
SDS polyacrylamide gel, which was then sliced into 0.5 cm
pieces and counted. Open circles, FP21; closed circles,
FP21 digested with pronase.








60











160
trypsin dye front



120




o_ 80




40-


0 I I I I I I I I
0 2 4 6 8 10 12 14 16

gel slice









61

Chapter IV), and the number of copies of FP21 per cell is

not affected by the mutation, then there may be a maximum of

4 x 105 copies of FP21 per cell.



Evidence that FP21 is Endogenous to the Cytosol

Fucosylation has been shown to occur in the Golgi

apparatus in other organisms (Hirschberg and Snider, 1987;

Kornfeld and Kornfeld, 1985), so I assessed the possibility

that FP21 was a lumenal microsomal protein that leaked

during the P100 and S100 isolation procedure. To guard the

P100 from chemical lysis, the vesicles were prepared in a

cocktail of protease inhibitors that contained additional

inhibitors from those utilized in the standard fractionation

protocol, and which have been developed for the isolation of

various proteolitically sensitive proteins in Dictyostelium

(Goodloe-Holland & Luna, 1987; Stone, et al., 1987). To

address the possibility of mechanical disruption of the

vesicles, the P100 was disrupted by sonication and

recentrifuged at 100k x g for 1 h. Approximately 11% of the

radioactivity was released and the supernatant of this

centrifugation was analyzed by SDS-PAGE as described above

(figure 3-3). Although several radioactive peaks were

present in the supernatant of the second centrifugation, the

radioactivity profile was different from the S100 suggesting

that FP21 is not a protein released by disruption of the

vesicle fraction. Although radioactivity which comigrated





























Figure 3-3. Comparison of S100 and releasable P100 components.

HL250 amoebae were metabolically labelled with 2 pCi/ml of
[13H]fucose, lysed, fractionated into an S100 and P100. The
P100 was sonicated and recentrifuged. After centrifugation,
the resulting supernatant was examined by slicing 7-20%
linear gradient SDS-PAGE into 2.2 mm and counting the gel
pieces (panel B). Included for comparison, is the profile
from the S100 radiolabelled species from the same
preparation run on the same gel (panel A). 50 Mg of protein
were electrophoresed for the S100 and P100, respectively.
Arrow, migration of trypsin.































0 10 20 30 40 50
gel slice



200
B

150 -

0.
0 100


50


0"
0 10 20 30 40 50
gel slice








64

with FP21 was observed in the P100-derived supernatant, this

material was a minority of the radioactivity released, and

probably reflected the general heterogeneity of the P100

vesicle contents. Less than 1% of the total cell

radioactivity that migrated at the 21 kD molecular weight

position was released from the P100 by sonication,

indicating that the remaining FP21 is recovered in the S100.

In a different approach, P100 from in vivo [3H]fucose

labelled cells was mixed with unlabelled post-nuclear

supernatant and recentrifuged (table 3-2). More than 98% of

the radioactivity sedimented with the P100, suggesting that

once associated with the P100, radioactivity is not lost,

unless vesicles are purposely disrupted, as in sonication.

The enzymes a-glucosidase-2, glucose-6-phosphatase, and

acid phosphatase have been used as markers of the

endoplasmic reticulum, Golgi apparatus and endoplasmic

reticulum, and lysosomes, respectively (Borts and Dimond,

1981; McMahon et al., 1977). To examine the distribution of

these enzyme markers in the high speed fractions, HL250

amoebae were harvested, homogenized, fractionated into an

S100 and P100 and assayed for activity of the different

marker enzymes. As seen in table 3-3, the majority of the

activity was recovered in the P100, suggesting minimal

contamination of the S100 by vesicles containing these

enzymes. Less than 7% of the total a-glucosidase-2 activity



























Table 3-2. Radioactivity recovered in the second S100 after
different P100 treatments.

condition % radioactivity recovered

untreated 2.2%
sonicated 11.3%
mixed* 1.4%


Cells were labelled in vivo by growing in 2 MCi/ml of
[3H]fucose in FM media for 3 days, lysed, and fractionated
into an S100 and P100. The P100 was then subjected to
different treatments and recentrifuged at 100k x g for 1 hr.
The S100 from this second centrifugation was analysed for
radioactivity. *P100 from metabolically labelled cells was
mixed with unlabelled post-nuclear supernatant from cells
grown in FM media and recentrifuged at 100k x g for 1 hr.






























Table 3-3.
fractions.


Distribution of markers among S100 and P100


a-glucosidase-2
acid phosphatase
glucose-6-phosphatase


S100

6.5%
12.8%
n.d.


P100

93.5%
87.2%
100%


HL250 amoebae grown in HL-5 were fractionated into S100 and
P100, and assayed for activity as described in Materials and
Methods. Activity expressed as percentage of total activity
detected in both fractions, n.d., not detectable.








67

was found in the S100; mannose-6-phosphatase was only

detectable in the P100. More than 87% of the activity of

the lysosomal enzyme acid phosphatase was detected in the

P100. In Dictyostelium, this enzyme has been shown to be a

soluble lumenal lysosomal protein (Dimond, et.el., 1981).

Further evidence that P100 vesicles are stable, closed

structures comes from earlier studies from the laboratory.

Prespore proteins SP75 and SP96 sediment at 100k x g unless

cells are sonicated (West and Erdos, 1988). These spore

coat proteins are contained in secretary vesicles (Erdos and

West, 1989; West and Erdos, 1988), which appear to be intact

since they are resistant to proteolysis by Proteinase K

unless they are treated with 0.1% Triton X-100 (West et al.,

1986; Q.H. Yang and C.M. West, unpublished data). In other

studies, N-acetyl glucosaminyltransferase activity was

inhibited by EDTA in the P100 fraction only when assayed in

the presence of detergent, suggesting that Golgi-like

vesicles in the P100 were closed (R.B. Mandell and C.M.

West, unpublished observations)



FP21 is Unrelated to Other Known Cytoplasmic Proteins

The possible relationship of FP21 with other known

Dictyostelium proteins was investigated. I examined the

reactivity of antisera raised against discoidin I and II

(Erdos and Whitaker, 1983) and against gp24 (Knecht et al.,

1987) for FP21. On Western blots, antiserum against gp24








68

recognized a band that migrated with a slower mobility than

metabolically labelled FP21 (not shown). The possibility of

FP21 being the lectin discoidin was examined, since it is

primarily present in the cytosol and exhibits weak affinity

for L-fucose (Erdos and Whitaker, 1983; Bartles and Frazier,

1980). Metabolically labelled FP21 from HL250 was

electrophoresed in a 15% SDS polyacrylamide gel, while a

replicate lane was blotted onto nitrocellulose paper and

immunoprobed using an anti-discoidin antiserum. The

antiserum recognized a band of higher MW than FP21 with

mobility slower than trypsin, reproducing results reported

by others where discoidin I was shown to have a slower

mobility than trypsin in 15% SDS polyacrylamide gels

(Kohnken and Berger, 1987). Migration of purified discoidin

in SDS-PAGE differed upon boiling of the sample, migrating

as a tetramer (ca. 100 kD) when samples were not boiled,

whereas metabolically labelled FP21 was found to migrate as

a discrete peak of radioactivity ahead of trypsin regardless

of boiling (Q.H. Yang and C.M. West, unpublished data).

FP21 could be partially purified by HPLC DEAE

chromatography. Metabolically labelled S100 was

fractionated on an anion exchange column and FP21 recovery

monitored by counting of SDS-PAGE slices. FP21 eluted at

0.5 M NH4Ac with an increase of 24-fold the specific

activity relative to the starting sample (data not shown).

Discoidin eluted earlier than FP21 from the HPLC DEAE column








69

and no radioactivity was found associated with discoidin,

suggesting it does not bind to discoidin during

purification. I conclude that FP21 is not related to

discoidin or any discoidin isoforms, or gp24, and does not

bind to any discoidin isoforms.



Oligosaccharide Studies
FP21 from Ax3 and HL250 yield similar size
glycopeptides after pronase digestion. Pronase digested

gel-purified FP21 that had been metabolically labelled from

Ax3 and HL250 were compared (figure 3-4). More than 50% of

the radioactivity from both sources eluted as a major peak

with a relative elution coefficient (Rev) of 0.42 and 0.43,

respectively (ca. 5.5 glucose units). In the case of FP21

derived from Ax3, the rest of the radioactivity eluted

earlier in the void volume and distributed into minor peaks.

The digestion products from HL250 FP21 yielded a major peak,

with the rest of the radioactivity eluting earlier, and less

than 10% eluting after the major peak, although some

variability was observed in different runs regarding the

minor peaks. Radioactivity eluting at an earlier position

than the major peak may be explained by incomplete

digestion. The minor amount of radioactivity that eluted at

a later position may be due to breakdown of the

oligosaccharide. These results indicate that the main

glycopeptides derived from Ax3 and HL250 have the same




























Figure 3-4. Gel filtration chromatography of FP21 glycopeptides.

Ax3 and HL250 vegetative cells were metabolically labelled
with 2 pCi/ml of [3H]fucose. Cells were lysed, fractionated
into an S100 and P100, and FP21 isolated by SDS-PAGE and
electroelution from the S100. The samples were exhaustively
digested with pronase and analyzed by BioGel P-4 gel
filtration. Data obtained from one representative
experiment.

Panel A. Glycopeptides derived from Ax3 FP21, arrow
identifies major peak with a Rev of 0.42. Vo, 32; Vi, 142.

Panel B. Glycopeptides derived from HL250 FP21, arrow
identifies major peak with a Rev of 0.43. Vo, 26; Vi, 166.



















600-O


450


n 300 -
Q Vo jVi

150 V



0 20 40 60 80 100 120 140
fraction



240 -
B

180-


120

Vo Vi

60-


0'1
0 20 40 60 80 100 120 140 160
fraction








72

sizes, and suggest that both strains form in vivo the same

oligosaccharide when grown in fucose-containing media.



Oligosaccharide in FP21 is 0-linked. An enzymatic and

a chemical approach were used to determine whether the

fucose-containing oligosaccharide in FP21 is N-linked or 0-

linked. I first tried the enzyme PNGase F. The minimum

requirement of PNGase F is a di-N-acetylchitobiose core

(Chu, 1986; Tarentino et al., 1985). This enzyme has a

broad specificity and can cleave most asparagine linked N-

glycans (including high mannose and complex multibranched

oligosaccharides) provided they are not located at the amino

or carboxy termini (Chu, 1986). Gel-purified FP21 from

metabolically labelled Ax3 was digested with PNGase F. As

shown in figure 3-5, the fucose label eluted in the void

volume, while control substrates were quantitatively

digested as determined by SDS-PAGE (not shown). FP21

trypsinized prior to digestion with PNGase F also eluted in

the void volume (not shown). The inability of PNGase F to

release radioactivity from FP21 suggested that either the

fucose-containing oligosaccharide was not N-linked or the

oligosaccharide was insensitive to the enzyme.

I then considered the possibility that the

oligosaccharide in FP21 was 0-linked. Metabolically

labelled gel purified FP21 was subjected to mild alkaline,

reducing conditions, to release intact 0-linked































Figure 3-5. Gel filtration chromatography of PNGase F digests.

Ax3 vegetative cells were metabolically labelled with 2
PCi/ml of [3H]fucose, an S100 was prepared, and FP21 was
isolated by SDS-PAGE and electroelution. FP21 digested with
PNGase F as described in Materials and Methods, and analyzed
by BioGel P-4 gel filtration. Nine drops fraction were
collected, instead of the usual 12 drops; [3H]mannose was
used to determine Vi.





















1200

Vo

900




600- i




300-



0-
0 20 40 60 80 100 120 140 160 180

fraction








75

oligosaccharides (B-elimination). Under the conditions

employed, N-linked oligosaccharides are insensitive to

chemical release (Biermann, 1988). More than 95% of the

radioactivity was released from Ax3 (Rev 0.48) and HL250

(Rev 0.49) in vivo labelled FP21 and it was resolved as one

peak with a size of 4.8 glucose unit (figure 3-6, panels A

and B). The oligosaccharide appeared to have been released

by B-elimination and thus is concluded to be 0-linked.

Consistent with the pronase digestion studies reported

above, Ax3 and HL250 yielded a similar size oligosaccharide,

supporting the idea that both strains produced the same

oligosaccharide. The slightly smaller size of the

oligosaccharide compared to the glycopeptide is consistent

with the notion that the glycopeptide consisted of one or

more amino acids and the oligosaccharide chain. If the

glycopeptide resulting from pronase digestion consisted of

the oligosaccharide attached to two or more amino acids, it

was possible that mild alkaline hydrolysis under reducing

conditions resulted in further hydrolysis of the remaining

polypeptide backbone yielding one amino acid and the

oligosaccharide. This would yield a smaller radioactive

species, with concomitant increase in Rev. To investigate

this possibility, I employed harsher chemical conditions.

In vivo labelled FP21 from Ax3 was gel purified and

subjected to strong alkaline hydrolysis in the presence of

sodium borohydride, which cleaves N- and 0-linked



























Figure 3-6. Gel filtration chromatography of FP21
oligosaccharides.

Vegetative cells were metabolically labelled with 2 yCi/ml
of [3H]fucose, an S100 was prepared, and FP21 was isolated
by SDS-PAGE and electroelution. Gel purified FP21 from Ax3
and HL250 were subjected to B-elimination or strong alkaline
hydrolysis followed by fractionation by gel filtration
chromatography. Data obtained from one representative
experiment.

Panel A. B-elimination of Ax3 FP21. Vo, 36; Vi, 141.

Panel B. B-elimination of HL250 FP21. Vo, 30; Vi, 140.

Panel C. Alkaline hydrolysis of Ax3 FP21. Vo, 38; Vi, 134.








6000
A


4000

0
Vo I Vi






0 20 40 60 80 100 120 140
fraction

800-
B

600-


o 400

Vo vi





0 20 40 60 80 100 120 140
fraction

2500
C
2000


1500
0-
SVo Vi
1000


500 -

0*
0 20 40 60 80 100 120 140
fraction











oligosaccharides (Zinn et al., 1978; Biermann, 1988). The

results of this reaction are seen in figure 3-6, panel C.

More than 75% of the radioactivity was released, eluting

with a Rev of 0.49. Most of the remainder of the

radioactivity eluted in the void volume. I expected to see

a change in Rev by strong alkaline hydrolysis compared to

mild alkaline hydrolysis if mild alkaline hydrolysis did not

release the oligosaccharide, but hydrolysed the protein.

The fact that similar results are obtained by mild and

strong conditions suggested that B-elimination occurred to

release the oligosaccharide. Since mild alkaline hydrolysis

had been shown earlier not to cleave N-linked sugars

(Biermann, 1988), I conclude that the oligosaccharide in

FP21 is linked via an 0-linkage. The released

oligosaccharide eluted as an asymmetrical peak, both by mild

and strong alkaline hydrolysis. These results suggest that

there is more than one type of oligosaccharide in FP21 that

differ slightly in size. These results also suggest that

the slight heterogeneity seen in the glycopeptide size may

reflect amino acid heterogeneity.


Comparison of glycopeptides derived from vesicular and
cytosolic material. To further address the possibility of

FP21 arising by contamination of the S100 from the P100

fraction, glycopeptides derived from the Ax3 P100 were

compared to gel purified pronase digested FP21 from Ax3.

P100 derived glycopeptides were obtained in three different








79

manners. Metabolically labelled P100 was subjected to SDS-

PAGE and the material that comigrated with FP21 was gel

purified, pronase digested, and fractionated by gel

filtration chromatography. Entire P100 from metabolically

labelled cells was digested in the presence or absence of

detergent, and analyzed by BioGel P-4 gel filtration. As

seen earlier in figure 3-4, panel A, the S100 digest eluted

mainly as a single peak with a Rev of 0.42; on the other

hand, the digest from the FP21-comigrating P100 material,

fractionated as a major peak of 0.50 Rev (figure 3-7, panel

A). An analysis of glycopeptides from the entire P100

digestion showed a different elution profile from the FP21

digestion (figure 3-7, panel B). Note than in this

chromatograph there is a minor peak of 0.51 Rev, consistent

with the idea that the major peak seen in panel A is a minor

component of the entire P100 glycopeptide repertoire. Also

approximately 40% of the radioactivity eluted with the void

volume, suggesting it may be resistant to pronase digestion.

Digestion of the entire P100 fraction was carried out in the

presence of Triton X-100, to determine if solubilization of

the sample yielded a different digestion profile, by

facilitating accessibility of the enzyme to the substrates

(figure 3-7, panel C). The overall profile is similar, with

more than 30% of the radioactivity eluting at the void

volume, and a peak with a Rev of 0.51 is still a minor

component of the glycopeptides released. West et al. (1986)


























Figure 3-7. Gel filtration chromatography of P100 glycopeptides.

Vegetative Ax3 cells were metabolically labelled with 2
pCi/ml of [3H]fucose. A P100 was prepared and the 21 kD MW
material that comigrated with FP21 on SDS-PAGE was
electroeluted and pronase digested. In another assay,
entire P100 was pronase digested in the absence or presence
of 0.1% Triton X-100. After pronase digestion, samples were
subjected to BioGel P-4 fractionation. Arrow identifies the
position with a Rev of 0.51.

Panel A. Glycopeptides from 21 kD MW P100-derived material.
Vo, 34; Vi, 162.

Panel B. Glycopeptides from entire pronase-digested P100.
Vo, 36; Vi, 142.

Panel C. Glycopeptides from entire P100 digested with
pronase in the presence of Triton X-100. Vo, 32; Vi, 146.











2000
A.


1500


S1000
Q Vo Vi

500 -



0 20 40 60 80 100 120 140 160
fraction


2500


2000


1500
2
a-
S1000


500


0


0 20 40 60 80 100 120 140 160
fraction


C. v




Vvi







I I I R^


I I I I a a 1
20 40 60 80 100 120 140 160
fraction








82

have reported the existence of pronase-resistant material in

the particulate fraction of vegetative cells, so it seems

possible that the pronase-resistant material eluting in the

void volume is, or is related to, the smear previously

described, although SDS-PAGE analysis will be needed to

confirm this supposition.



Fucose is Covalently Bound to FP21

The fact that the radioactivity in FP21 was not

released by boiling in SDS/ B-mercaptoethanol, nor after

boiling in SDS under reducing conditions followed by SDS-

PAGE, suggested that 3H was covalently bound to in vivo

labelled FP21. Nevertheless, there have been reports of

covalent-bound enzymatic intermediates in the literature

(Scrimgeour, 1977). Although ES (enzyme-substrate)

intermediates are very reactive and usually cannot be

isolated without some sort of stabilization or chemical

trapping technique, the possibility that FP21 is really a

cytosolic fucosyltransferase that binds GDP-fucose or other

fucose metabolites covalently was considered. The Rev of

the radioactivity released by mild alkaline hydrolysis

suggests that it is not related to GDP-fucose or fucose

because it elutes with a different Rev than GDP-fucose or

fucose. Additional evidence that 3H is present as fucose

was presented in Chapter II, where it was shown that more








83

than 95% of the macromolecular-associated radioactivity from

metabolically labelled cells migrated as authentic fucose.



FP21 is Present in Migrating Slug Stage Cells

To investigate whether FP21 was present also in

developing cells, HL250 amoebae were plated for development

on nuclepore filters and 9 hours after plating the filters

were lifted and cells were metabolically labelled by

placement on 100 pCi of [3H]fucose. After 5 hours of

exposure to [3H]fucose, filters were lifted again, carefully

washed from H label, and cells were allowed to continue

development for two more hours. Cells were then harvested,

disaggregated, and fractionated into an S100 and P100

fractions, run in an SDS-PAGE, and the gel cut and the

pieces counted. As shown in figure 3-8, a fucosylated

macromolecule is preferentially fucosylated in the cytosol

and it has the same mobility in SDS-PAGE as FP21. Thus, it

seems that FP21 fucosylation is not restricted to the growth

phase of Dictyostelium.



Discussion

In this chapter I have presented evidence for the

existence of a fucosylated single molecular weight species,

which has been termed FP21 based on its mobility as

determined by SDS-PAGE. This protein cofractionates with





























Figure 3-8. Incorporation of [ 3H]fucose into macromolecular
species of slug stage cells.

HL250 cells were harvested, plated on filters for
development, and exposed to [3H]fucose for 5 h as described
in the text. Cells were then collected, disaggregated,
filter-lysed, and S100 and P100 fractions were prepared and
analyzed by 7-20% linear gradient SDS-PAGE. 70 pg and 24.6
pg of protein were electrophoresed for the S100 and P100,
respectively. Open circles, P100; closed circles, S100;
arrow, migration position of trypsin.




Full Text
CHAPTER IV
EVIDENCE FOR A CYTOSOLIC FUCOSYLTRANSFERASE
Introduction
In the preceding chapter, I identified a fucosylated
protein in the cytosol, FP21. The presence of FP21 in the
cytosol challenges the prevailing belief that fucoproteins
are restricted to the cell surface and lumenal compartments
of the cell. Even though in the past three decades evidence
has been accumulating on the presence of glycoproteins and
fucoproteins in non-lumenal locations (Hart et al., 1989a;
Hart et al., 1989b), to my knowledge, no one has shown the
existence of a fucosyltransferase in the cytosol. One
possibility is that fucosylation is restricted to the
microsomes, and cytosolic fucoproteins are
posttranslationally transported back across the membrane to
the cytosol.
However, even though there is no previous evidence for
cytosolic fucosylation there is precedent for glycosylation
in the cytosol. Studies on the biosynthesis of nuclear pore
proteins bearing O-GlcNAc suggested that the sugar was added
to the proteins within 5 min of their synthesis and before
they became associated with membranes (Davis and Blobel,
1987). These data suggested that the activity responsible
90


49
and/or stained and destained and then cut into either 2.2 mm
or 0.5 cm slices. Gel pieces were shaken and swollen
overnight in a scintillation cocktail composed of 111.1 ml
of tissue solubilizer (TS-1), 6.0 g PPO, 0.15 g POPOP, and
20 ml dH20 to 1 1 of toluene. Gel slices were counted and
recounted until dpm were determined to be stable, usually 1-
2 days later. For gel-purified material, the sample was run
in a 7-20% linear gradient gel and the 21 kD area
(approximately 1 cm below trypsin) cut out and electroeluted
overnight using a Bio-Rad electroeluter following
manufacturer's directions, except the Laemmli
electrophoresis buffer used for SDS-PAGE (West and Loomis,
1985) was used instead of the recommended volatile buffer to
avoid alkaline hydrolysis. Western blotting was carried out
as previously described (West and Loomis, 1985).
Partial Purification of FP21 by Anion Exchange
Chromatography
In preliminary studies, FP21 was partially purified by
fractionation of metabolically labelled S100 fraction on a
TSK DEAE-5PW 8 x 75-mm HPLC anion-exchange column (LKB)
preequilibrated with 10 mM NH^Ac, pH 7.0. Sample was
dialyzed against 2 1 of 10 mM NH4Ac, pH 7.0, for several
hours, and clarified by centrifugation at 10k x g for 10 min
prior to injection. Protein was eluted using an increasing
linear gradient (10 mM to 1 M NH^Ac, pH 7.0) for 40 min at a
rate of 0.75 ml per min, and the majority of FP21 was found


124
proportional to the amount of [14C] radioactivity added.
Since this study did not analyze the in vitro fucosylated
species by SDS-PAGE, it cannot be concluded that Ax3 was
able to fucosylate FP21. To investigate which MW species
served as acceptor for the Ax3 cytosolic fucosyltransferase,
the experiment was repeated using a new batch of partially
purified [14C]FP21. In vitro labelled FP21 eluted in
consecutive fractions 20 and 21 during HPLC gel filtration
chromatography, as confirmed by SDS-PAGE. Ax3 S100 was
added to an aliquot of fraction 21 that had previously been
dried on the bottom of the assay tube and assayed for
fucosyltransferase activity in the presence of GDP-
[3H]fucose and Mg++. The reaction was stopped by boiling in
sample electrophoresis buffer, resolved by SDS-PAGE, and the
gel sliced and counted. Figure 4-7 shows the comigration on
SDS-PAGE of the in vitro [3H] label resulting from
fucosylation by Ax3 with the trace [14C] labelled FP21 from
fraction 21. There was no [3H] radioactivity incorporation
into any MW species when the Ax3 S100 fraction was incubated
in the absence of purified HL250 FP21. In conclusion, Ax3
S100 has a cytosolic fucosyltransferase activity that
utilizes the same acceptor as the mutant cytosolic
fucosyltransferase.


Table 2-2.
GDP-fucose.
Effect of time on conversion of GDP-mannose to
nmol fucose/mq protein
time (min)
Ax3
HL250
Ax3+HL250 (0.5:0.5)
7.5
0.57
0
0.21
15
2.5
0
1.3
30
2.9
0.15
2.2
90
9.2
0
4.3
Protein (600
yq total)
from a
100,000 xg supernatant of
vegetative cell-free extract was assayed for ability to
convert GDP-[ 14C]mannose (7.5 initial concentration) to GDP-
[ 4C]f ucose; data are the result of the average of two
determinations


142
et al., 1990). In any event, I interpret the data presented
as evidence for a fucosyltransferase in the cytosol of
Dictyostelium discoideum.
The fact that Ax3 produced fucosylated FP21 suggested
that, as it occurred in the mutant, the normal strain may
have a cytosolic fucosyltransferase responsible for FP21
fucosylation. However, while activity was not detectable in
Ax3 S100 fraction, it could be reconstituted by addition of
mutant FP21, indicating that Ax3 possessed a cytosolic
fucosyltransferase equivalent to the mutant
fucosyltransferase.
In order to characterize the fucosyl linkage catalyzed
by the cytosolic fucosyltransferase, several acceptors were
used. Activity with synthetic acceptors was about an order
of magnitude higher for the Ax3 extract, which may be
attributed to competitive inhibition by the unfucosylated
FP21 in the mutant. Of those tested, the only suitable
acceptor was found to be a type I analog, 8-
methoxycarbonyloctyl galfil,3GlcNAcfi. Since the type II
analog [8-methoxycarbonyloctyl galfil,4GlcNAcH] did not work
as acceptor, it appears that the cytosolic
fucosyltransferase may be an al,4fucosyltransferase.
The cytosolic fucosyltransferase preferentially
recognized a type I analog, suggesting it was an
al,4fucosyltransferase that lacked al,3 activity. This
activity would differ from other al,4fucosyltransferase


Figure 3-4. Gel filtration chromatography of FP21 glycopeptides.
Ax3 and HL250 vegetative cells were metabolically labelled
with 2 pCi/ml of [3H]fucose. Cells were lysed, fractionated
into an S100 and P100, and FP21 isolated by SDS-PAGE and
electroelution from the S100. The samples were exhaustively
digested with pronase and analyzed by BioGel P-4 gel
filtration. Data obtained from one representative
experiment.
Panel A. Glycopeptides derived from Ax3 FP21, arrow
identifies major peak with a Rev of 0.42. Vo, 32; Vi, 142.
Panel B. Glycopeptides derived from HL250 FP21, arrow
identifies major peak with a Rev of 0.43. Vo, 26; Vi, 166.


44
Other lower eukaryotes, such as yeast, do not carry out
fucosylation (Kukuruzinska et al., 1987) so the existence of
a Dictyostelium fucosylation mutant could be very important
to study fucosylation. In any event, HL250 has already
served as a very useful tool in which to study fucosylation
events, as will be evident in the following chapters.


123
either because activity was not detected in Ax3 after
desalting through a P-2 column, while the HL250 S100
retained activity (table 4-4). Another explanation for the
lack of activity in the Ax3 S100 is that FP21 from Ax3 was
quantitatively fucosylated in vivo, leaving no acceptor
sites for the reaction in vitro. Evidence using purified
FP21 from HL250 supports this conclusion (see next section).
Fucosyltransferase Activity Is Detected in Ax3 S100 Fraction
Upon Addition of Mutant FP21
The absence of cytosolic fucosyltransferase activity in
Ax3 S100 extracts could be explained as a result of
guantitative fucosylation of FP21 in the living cell. It
was reasoned that if this model was correct, then addition
of mutant FP21 to Ax3 extracts would lead to incorporation
into FP21. FP21 was trace-labelled in vitro using GDP-
[14C]fucose, and partially purified by ammonium sulfate
precipitation, QAE-ion exchange chromatography, and HPLC gel
filtration. Fractions from the gel filtration step were
counted and examined by SDS-PAGE and those that contained
FP21 were pooled, brought to dryness, dissolved in water,
and added to Ax3 S100 extract. In vitro fucosylation was
determined as [3H] incorporated into TCA-insoluble material
in the presence of GDP-[3H]fucose. Since FP21 was trace-
labeled with [ 14C] fucose, the relative amount of the
acceptor added was estimated from [14C] dpm. West et al.
(unpublished results) showed that incorporation was


51
phosphatase was assayed as described in McMahon et al.
(1977) except that Triton X-100 was included at a
concentration of 0.1% and absorbance was measured at 420 nm,
instead of 400 nm.
PNGase F Digestion
Gel-purified FP21 (3000-8000 dpm) was boiled in 0.5%
SDS in water for 3 min. For digestion, the protocol of
Tarentino et al. (1985) was followed. The sample was
incubated in 104 mM sodium phosphate, pH 8.6, 10 mM EDTA, 10
mM 1,10-phenanthroline (stock solution of 100 mM in
methanol), 2% NP-40, 0.21% SDS, and 20 U/ml PNGase F
(Boehringer-Mannheim) in 300 jil for 22 h at 37. Fetuin (B-
grade, Calbiochem) and RNAse B (Sigma) were treated
identically as controls and digestion was quantitative, or
nearly quantitative, as determined by shifts in molecular
weight in SDS-PAGE. In one trial FP21 was digested with
trypsin prior to PNGase F digestion by incubating gel
purified FP21 in 0.8 mg/ml trypsin, 1 mM CaCl2 for 1 hr at
37. The reaction was stopped by boiling in the presence of
1 mM PMSF and 10 pg/ml aprotinin for 3 min.
Pronase Digestion
Glycopeptides were prepared by exhaustive pronase
digestion as described (Ivatt et al, 1984). In short, after
gel purification and electroelution, the samples were dried


98
fi-elimination of In Vitro Labelled Acceptor
S100 extracts from HL250 were fucosylated in vitro. To
corroborate that I obtained 21 kD MW fucosylated product
from the in vitro reaction, l/25th of the reaction was
terminated by 3 min boiling in sample buffer and analyzed by
SDS-PAGE. The remainder of the sample (containing
approximately 104 dpm) was stored at -80 until ready to
use. The reaction mixture was centrifuged for approximately
2 h in a centricon filter to reduce unused GDP-[14C]fucose.
After concentrating the volume to 200 pi, fi-elimination was
carried out as described in Materials and Methods, Chapter
III.
PNGase F Digestion of In Vitro Labelled FP21
FP21 was fucosylated in vitro as described above for fi-
elimination. After the volume was concentrated, PNGase F
digestion was carried out as described in Materials and
Methods, Chapter III.
Results
Cytosolic Fucosyltransferase Activity
The presence of FP21 in the cytosol suggested that a
fucosyltransferase might also be located there. To
investigate this possibility, HL250 cells were fractionated
to yield cytosolic supernatant (S100) and organelle (P100)
fractions. The fractions were analyzed for their ability to


97
[14C]fucose. Incorporation into FP21 was confirmed by
electrophoresing an aliquot and counting of SDS-PAGE slices.
The radiolabelled aliquot was mixed with the rest of the
unlabelled preparation and subjected to (NH4)2S04
fractionation. The 70-80% cut was dissolved in and dialyzed
against 100 mM NH4Ac, applied to a 14 ml bed of the strong
anion exchanger A25-QAE-Sephadex, and eluted with an
ascending gradient up to 1.5 M NH4Ac. [UC]FP21 eluted at
input buffer concentration of 0.49 M. This preparation was
then concentrated and desalted on Centricon and/or
Centriprep cartridges with nominal 10 kD MW cutoffs, and
then applied to an HPLC gel filtration column (8 x 300 mm
Toya Soda TSK GW-300) equilibrated in 100 mM NH4Ac, with a
flow rate of 0.5 ml/min. Sample was clarified by
centrifugation at 10k x g for 10 min prior to injection.
Radioactivity from the concentrated QAE-Sephadex eluate
eluted between the 14 kD and 29 kD MW standards. Fractions
were analyzed by SDS-PAGE using 15% polyacrylamide gels and
counting of the gel slices. For addition of purified FP21
to cell extracts, HPLC gel filtration fractions were brought
to dryness in a vacuum centrifuge, redissolved in dH20, and
brought to dryness again, in the 1.5 ml microcentrifuge tube
that was going to be used for the assay.


75
oligosaccharides (Jl-elimination). Under the conditions
employed, N-linked oligosaccharides are insensitive to
chemical release (Biermann, 1988). More than 95% of the
radioactivity was released from Ax3 (Rev 0.48) and HL250
(Rev 0.49) in vivo labelled FP21 and it was resolved as one
peak with a size of 4.8 glucose unit (figure 3-6, panels A
and B). The oligosaccharide appeared to have been released
by fl-elimination and thus is concluded to be O-linked.
Consistent with the pronase digestion studies reported
above, Ax3 and HL250 yielded a similar size oligosaccharide,
supporting the idea that both strains produced the same
oligosaccharide. The slightly smaller size of the
oligosaccharide compared to the glycopeptide is consistent
with the notion that the glycopeptide consisted of one or
more amino acids and the oligosaccharide chain. If the
glycopeptide resulting from pronase digestion consisted of
the oligosaccharide attached to two or more amino acids, it
was possible that mild alkaline hydrolysis under reducing
conditions resulted in further hydrolysis of the remaining
polypeptide backbone yielding one amino acid and the
oligosaccharide. This would yield a smaller radioactive
species, with concomitant increase in Rev. To investigate
this possibility, I employed harsher chemical conditions.
In vivo labelled FP21 from Ax3 was gel purified and
subjected to strong alkaline hydrolysis in the presence of
sodium borohydride, which cleaves N- and O-linked


137
The endogenous acceptor utilized by the S100
fucosyltransferase was a protein which comigrated with FP21
by SDS-PAGE. I compared the acceptor for the in vitro
fucosyltransferase reaction with metabolically labelled FP21
from Ax3 cells by SDS-PAGE. There was one main radioactive
peak, revealing in vivo and in vitro fucosylated acceptors
with the same mobility on polyacrylamide gels. These
results suggested that a cytosolic fucosyltransferase
existed that utilized FP21 as its primary acceptor species
in vitro, and may be responsible for fucosylation of FP21.
To investigate the origin of the S100
fucosyltransferase, I compared it to the bulk P100
fucosyltransferase activity, since the cytosol is the
default location of lumenal enzymes released by rupture of
vesicles. If both activities were indeed different, I
expected to detect enzymatic differences. Initially, I
examined the SDS-PAGE profiles of in vitro fucosylated
acceptors and found they were very similar to those obtained
from metabolic labelling. Incorporation by endogenous
acceptors was at the 21 kD MW position for the S100, and in
the P100 radioactivity migrated as two separate, broad
peaks.
In order to compare directly the soluble and the
sedimentable activities, I assayed the S100 and P100 in the
presence of detergent to overcome any differences in
accessibility for GDP-fucose by the fucosyltransferases. I


FP21 is synthesized and fucosylated in the cytosol by an
al,4fucosyltransferase.
xvi


110


Figure 3-5. Gel filtration chromatography of PNGase F digests.
Ax3 vegetative cells were metabolically labelled with 2
^/Ci/ml of [3H]fucose, an S100 was prepared, and FP21 was
isolated by SDS-PAGE and electroelution. FP21 digested with
PNGase F as described in Materials and Methods, and analyzed
by BioGel P-4 gel filtration. Nine drops fraction were
collected, instead of the usual 12 drops; [3H]mannose was
used to determine Vi.


46
Materials and Methods
Materials
L-[ (5,6)-3H]-fucose (60 Ci/mmol) was obtained from
American Radiochemical Corporation and D- [ (2,3)-3H]mannose
(24 Ci/mmol) from New England Nuclear. TS-1 was purchased
from Research Product International; POPOP was from
Mallinckrodt; SDS, leupeptin, aprotinin, phenyl
methylsulfonyl fluoride, Triton X-100,
dimethyldichlorosilane, MES, all nitro-phenyl substrates,
mannose-6-phosphate, bovine serum albumin (fraction V), blue
dextran, bromo phenol blue, and trypsin were from Sigma;
glycine, benzene, ammonium acetate, PPO, and toluene were
from Fisher. The concentrations of Triton X-100 and NP-40
are expressed as v/v, all others are expressed as w/v,
unless specified otherwise.
Strains and Conditions of Growth
Dictyostelium discoideum strains Ax3 (from S. Free) and
HL250 (from W.F. Loomis) were grown on HL-5, a complete
medium that contains glucose, yeast extract, and proteose
peptone (Loomis, 1971). Ax3 is the normal strain and HL250
is a mutant obtained from Ax3 by N-methyl-N'-nitro-N-
nitrosoguanidine mutagenesis (Loomis, 1987). HL250 lacks
the enzyme activity that converts GDP-mannose into GDP-
fucose which results in a lack of cell fucose (Gonzlez-
Yanes et al., 1989; also see Chapter II). In all


138
determined that the S100 fraction was dependent on divalent
cations, while the P100 was active in the absence of cations
and in the presence of the chelator EDTA. The activities in
both fractions were maximal at a similar pH range, but the
cytosolic fucosyltransferase was more sensitive to higher pH
than the P100 fucosyltransferase activity.
Glycosyltransferase activities have commonly been found to
be dependent on the presence of divalent cations. In the
case of fucosyltransferases, however, there are precedents
for al,2, al,3, and al,3/1,4 fucosyltransferases which are
active in the absence of cations, and are either stimulated
or inhibited by different cations (Beyer and Hill, 1980;
Campbell and Stanley, 1984; Foster et al.,1991; Stroup et
al., 1990; Zatz and Barondes, 1971).
The apparent affinity for GDP-fucose differed greatly
for S100 and P100 activities. The S100 fucosyltransferase
activity had a higher affinity for GDP-fucose than the P100
activity when both were assayed in the presence of Tween-20.
The lower apparent Km for the cytosolic fucosyltransferase
explained why activity is higher in the S100 at the low
concentration of GDP-fucose used in most assays, 0.36 pM.
At 0.36 pM the concentration of GDP-fucose was near its
apparent Km for the S100 fucosyltransferase (1.7 pM), but
well below the apparent Km for the P100 enzyme (38.2 pM) .
The dependence on GDP-fucose concentration was also examined
in the intact fractions to gain some insight into the


Figure 2-3. Conversion of GDP-mannose to GDP-fucose by normal
and mutant strains. Closed circles, Ax3; open circles,
HL250. Results are the average of two determinations.
Panel A. Effect of protein concentration on conversion. 30
min. assay, 7.5 /vM GDP-[3H]mannose.
Panel B. Effect of GDP-mannose concentration on conversion.
30 min assay, 750 pg protein for each strain.
Panel C. Apparent Michaelis constant for GDP-[3H]mannose,
determined for Ax3 conversion. Apparent Km was determined
to be 14.1 pM and apparent Vmax 18.3 nmol/mg protein/30 min
by the Lineweaver-Burk double reciprocal plot method.


141
activity in the P100 will require characterization of the
purified fucosyltransferases from the S100 and P100.
The results obtained from my investigation are based on
biochemical evidence in which a soluble fucosyltransferase
partitioned with the cytosol. Other investigators have
identified glycosylated proteins in the cytosol and/or
nucleus and have searched for an enzyme responsible for the
addition of the sugar (Haltiwanger et al., 1990). Their
biochemical studies showed that an activity capable of
adding GlcNAc to protein was recovered in both the soluble
and membrane fractions (Haltiwanger et al., 1990). However,
they showed that the membrane-associated activity was
releasable by high salt treatment and was oriented towards
the cytosol, not the lumen of the vesicles. Thus, it is
possible that a fraction of this newly discovered cytosolic
fucosyltransferase stayed associated with vesicles but since
it was in a minority, remained masked by other P100
fucosyltransferases. As more synthetic acceptors become
available, latency experiments in the presence and absence
of detergent can be done to address this question.
Alternatively, it is possible that a fucosyltransferase with
enzymatic properties similar to the cytosolic
fucosyltransferase is present in the lumen of P100 vesicles.
Still this will not contradict my findings and will imply
that there are two similar enzymes that reside in distinct
compartments, as has been reported for another enzyme (Lewin


50
to elute at 0.5 M input buffer concentration. Fractions
were analyzed by SDS-PAGE and counting of the gel slices.
Protein Concentration Assay
The Bio-Rad protein assay was used for determination of
protein concentration, and bovine serum albumin used as
standard.
Enzyme Assays
a-glucosidase-2 assays contained 100-300 pg protein,
8.6 mM p-nitrophenyl-a-D-glucoside, 0.1% Triton X-100, in 21
mM citrate-phosphate buffer (pH 7.5) at 37 (Borts and
Dimond, 1981). Reaction was stopped after 1 hr by addition
of Na2C03 to a concentration of 0.5 M and the absorbance
read at 420 nm. Glucose-6-phosphatase and mannose-e-
phosphatase were measured by release of phosphate from
mannose-6-phosphate, which has been previously shown to be a
suitable substrate for both enzymes (Arion et al., 1976).
Reaction mixtures contained 100-300 pg protein, 1 mM MgCl2,
2 mM mannose-6-phosphate, 0.1% Triton X-100 in 10 mM MES
(titrated with NaOH to a pH of 7.4) in a volume of 200 pi.
After 20 min incubations at 30 reactions were stopped by
adding 200 pi 20% ice-cold TCA (Snider et al., 1980). Tubes
were centrifuged at 14k rpm, for 10 min in an Eppendorf
table top microfuge and aliquots of the supernatants were
assayed for P1 by the method of Chen et al. ( 1956). Acid


LIST OF ABBREVIATIONS
APA
Asparagus Pea Agglutinin
ATP
Adenosine-5'-triphosphate
BSA
Bovine Serum Albumin
14c
Radioactive Carbon
cDNA
Complementary Deoxyribonucleic Acid
cm
Centimeter(s)
Con A
Concanavalin Agglutinin
cpm
Counts per Minute
CHO
Chinese Hamster Ovary
dH2
Distilled Water
dpm
Disintegrations per Minute
EDTA
Ethylenediaminetetraacetic Acid
EtOH
Ethanol
GlcNAc
N-acetyl Glucosamine
GDP
Guanosine-5'-diphosphate
h
Hour(s)
3h
Tritiated
HMG
High Mobility Group
HPLC
High Performance Liquid Chromatography
kD
Kilodaltons
K
m
Michaelis Constant
l
Liter(s)
xi


16
rat small intestinal mucosa (Martin et al., 1987). They
report the isolation of al,2 and ccl,3/l,4fucosyltransferases
that could be candidates for cytosolic fucosyltransferases
based on their inability to sediment after homogenization of
cells in 0.25 M sucrose, followed by 90 min, 200k x g
centrifugation (Martin et al., 1987). They compare the
al,2fucosyltransferase with the cc2,6sialyltransferase, which
is normally a Golgi enzyme converted to a soluble form by
cleavage of the amino terminal signal anchor to allow for
secretion (Weinstein et al., 1987). In the case of the
fucosyltransferases, the authors did not report using
protease inhibitors during isolation, nor did they document
the partitioning of cellular markers in the different
fractions (Martin et al., 1987). The reasons the enzymes
localize to the high speed supernatant, which would usually
be considered as the cytosolic fraction, may include the
breakage of the microsomal vesicles or the fact that they
were initially present extracellularly in the body fluids of
the mucosa. In another report these investigators tested
for the presence of glycosyltransferases in the nucleus. In
highly purified rat liver nuclei there was a total absence
of fucosyltransferase activity when endogenous
macromolecules or asialofetuin were used as acceptors
(Richard et al., 1975).
One explanation for the lack of evidence of nuclear
and/or cytosolic fucosyltransferases activity in any


REFERENCES 156
BIOGRAPHICAL SKETCH 166
vii


revealed that fucose was incorporated into an O-linked
oligosaccharide with an average size of 4.8 glucose units.
FP21 appeared to be endogenous to the cytosol, based on the
failure to release FP21 from P100 vesicles by sonication,
and the absence of FP21-like glycopeptides derived by
pronase digestion of the P100.
To determine if FP21 was fucosylated in the cytosol,
S100 and P100 fractions from HL250 were assayed for
fucosyltransferase activity, measured as ability to transfer
[14C] from GDP-[ 14C] fucose to endogenous acceptors.
Fucosylation of FP21 by the S100 was time- and protein
concentration-dependent. The cytosolic activity was
distinguished from the bulk P100 activity by its absolute
divalent cation dependence, alkaline pH sensitivity, and
very low apparent Km for GDP-fucose. Fucosyltransferase
activity was not detectable in Ax3 cytosol. However,
activity was reconstituted by addition of purified mutant
FP21, suggesting FP21 was already fucosylated in living
cells, and Ax3 possessed a cytosolic fucosyltransferase
equivalent to the HL250 enzyme. S100 fractions from Ax3 and
HL250 were able to fucosylate an al,4fucosyltransferase-
specific acceptor glycolipid analog, but not analogs capable
of being modified by al,2 and al,3 fucosyltransferases.
Since fucosylation of the analog was reduced by addition of
purified FP21, the same enzyme appears to be responsible for
fucosylation of both molecules. Thus it is proposed that
XV


A NOVEL FUCOSYLATION PATHWAY IN THE CYTOSOL
OF DICTYOSTELIUM DISCOIDEUM
By
BEATRIZ GONZALEZ-YANES
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1991


CHAPTER III
IDENTIFICATION OF A CYTOSOLIC FUCOPROTEIN
Introduction
There is some evidence that fucoconjugates are present
in the nucleus and cytosol (Hart et al., 1989a; Chapter I).
However, virtually nothing is known about the structure or
biosynthesis of these fucosylated macromolecules. Most
studies have limited themselves to reporting the existence
of evidence for nuclear or cytoplasmic glycoproteins, but
have not gone further to characterize the sugar-peptide
linkage or compare it with material derived from the
secretory pathway. As discussed in Chapter II, there is a
conditional fucosylation mutant, HL250, that can be readily
labelled when grown in radioactive fucose. Using this
strain, I have identified a fucoprotein that fractionated
with the cytosol and appeared to be the major fucosylated
species in the cytosol. The oligosaccharide-peptide linkage
was characterized and the fucoprotein was compared with
fucosylated material derived from vesicles, and
differentiated based upon several criteria.
45


CHAPTER I
HISTORICAL REVIEW AND BACKGROUND
Introduction
Glycoprotein synthesis and localization have been
subjects of intense study in the last few decades. Much has
been learned about glycosylation and some excellent reviews
are available (Kornfeld and Kornfeld, 1985; Hirschberg and
Snider, 1987). Carbohydrate moieties are usually
categorized as N and/or O-linked depending on whether the
carbohydrate moiety is attached to an asparagine by an amide
glycosidic linkage or to a serine or threonine,
respectively. Glycosylation is generally considered to be a
modification restricted to macromolecules that pass through
the secretory pathway, starting in the rough endoplasmic
reticulum where N-linked glycosylation is initiated
(Kornfeld and Kornfeld, 1985), but in some instances 0-
linked glycosylation also takes occurs (Spielman et al.,
1988). Glycosylation then continues in the Golgi apparatus
where further processing of N-linked oligosaccharides may
occur and O-linked glycosylation takes place (Abeijon and
Hirschberg, 1987). In the case of fucosylation, L-fucose
has been shown to be added as a terminal modification to
either N-linked or O-linked oligosaccharides in the Golgi
1


38
Specific activity of fucose pools. A lesion in the
GDP-mannose to GDP-fucose conversion pathway would render
the mutant defective in macromolecular fucosylation when
grown in the absence of fucose, as was shown earlier. Such
a scenario would require that the GDP-fucose inside the cell
be derived from the salvage pathway fed exclusively by
fucose from the extracellular media. In the case of the
Ax3, however, there would be a contribution of GDP-fucose
derived from the conversion pathway. If the mutation in
HL250 was indeed in the GDP-mannose to GDP-fucose
conversion, macromolecular fucose of cells grown in the
presence of L-[3H] fucose would have the same specific
activity as the fucose present in the media. To test this
hypothesis, mutant and normal cells were grown in fucose-
free defined media supplemented with [3H]fucose. Whole cell
preparations were then assayed for fucose content as
described above and the specific activity expressed as the
radioactivity present in the fucose peak divided by the
amount of fucose detected. As expected, the macromolecular
pool of the mutant had essentially the same specific
activity as the fucose in the medium. In contrast, the
specific activity in Ax3 was diluted approximately 400-fold
compared to the medium (table 2-3). These results confirmed
that HL250 derived its intracellular fucose from the salvage
pathway. Meanwhile it appears that in Ax3 the contribution


66
Table 3-3. Distribution of markers among S100 and P100
fractions.
a-glucosidase-2
acid phosphatase
glucose-6-phosphatase
S100
P100
6.5%
93.5%
12.8%
87.2%
n.d.
100%
HL250 amoebae grown in HL-5 were fractionated into S100 and
P100, and assayed for activity as described in Materials and
Methods. Activity expressed as percentage of total activity
detected in both fractions, n.d., not detectable.


parents, Beatriz and Germn, for everything. I firmly
believe that the education received during my first twenty-
one years of life in their company is ultimately responsible
for every achievement in my life.
v


HL250 AX3
27
PRESPORE CELLS
83.5 A6-2
SPORES
83.5
A6-2


104
Table 4-2. Failure to sediment S100 fucosyltransferase
activity.
S100 P100
100k
x g,
lh
943
444
170k
x g,
2.5h
1131
484
Cells were lysed, centrifuged, and fractionated into S100 and
P100. In this experiment lysis buffer (described in Materials
and Methods, pH 8.0) was supplemented with 1 mM chymostatin,
5 jjg/ml pepstatin, and 2 mM NBZ-phenylalanine. Fractions were
assayed immediately for C] incorporation from GDP-[ CJfucose
and expressed as total dpm incorporated in 30 min into TCA
insoluble endogenous acceptors. Reaction mixtures contained
0.36 nM GDP-fucose, 5 mM MgCl2, 240 pg protein, and were
incubated for 30 min. Results are expressed as average of two
determinations.


103
that were not sedimented before (if existent) and that could
have contained fucosyltransferase activity. To preserve the
intactness of the P100 vesicles, additional protease
inhibitors (Goodloe-Holland and Luna, 1987) from those
routinely used, were utilized during cell fractionation.
After these measures activity was still recovered in the
cytosol at similar levels (table 4-2). Additional evidence
supporting the notion that the vesicles in the P100 are not
damaged during filter lysis and centrifugation is presented
in Chapter III in the section of the origin of FP21. Taken
together, these results suggest that vesicles were not
measurably damaged during the isolation procedure;
therefore, the fucosyltransferase activity is probably
endogenous to the cytosol.
Comparison Between the P100 and S100 Fucosyltransferase
Activities
In order to compare the S100 and P100 activities, I
examined the identity of endogenous acceptors and the
effects of divalent cations, pH, and varying GDP-fucose
concentration on both fucosyltransferase activities. The
criteria of differential behavior has previously been used
to differentiate glycosyltransferases, since it is assumed
that under similar conditions, enzymes should behave in a
similar fashion (Campbell and Stanley, 1984; Galland et al.,
1988). Activities were measured in the presence of
detergent to circumvent any potential problem in substrate


Figure 3-1. Incorporation of [3H]fucose into macromolecular
species of the S100 and P100.
HL250 amoebae were metabolically labelled with 2 pCi/ml of
[3H]fucose, lysed, fractionated into an S100 and P100, and
subjected to 7-20% linear gradient SDS-PAGE; the gel was
sliced into 2.2 mm pieces and counted. 100 pg of protein
were electrophoresed for the S100 and P100, respectively.
Open circles, P100; closed circles, S100; arrow, migration
of trypsin.


159
Hallgreen, P., A. Lundblad, and S. Svensson. 1975. A new
type of carbohydrate-protein linkage in a glycopeptide
from normal human urine. J. Biol. Chem. 250:5312-5314.
Haltiwanger, R.S., G.D. Holt, and G.W. Hart. 1990. Enzymatic
addition of 0-GlcNAc to nuclear and cytoplasmic
proteins. J. Biol. Chem. 265:2563-2568.
Hardy, M.R., R.R. Townsend, and Y.C. Lee. 1988.
Monosaccharide analysis of glycoconjugates by anion
exchange chromatography with pulsed amperometric
detection. Anal. Biochem. 170:54-62.
Hart, G.W., R.S. Haltiwanger, G.D. Holt, and W.G. Kelly.
1989a. Glycosylation in the nucleus and cytoplasm.
Annu. Rev. Biochem. 58:841-874.
Hart, G.W., R.S. Haltiwanger, G.D. Holt, and W.G. Kelly.
1989b. Nucleoplasmic and cytoplasmic glycoproteins.
Ciha Found. Symp. 145:102-118.
Henderson, P.J.F. 1985. Statistical analysis of enzyme
kinetic data. In: Techniques in the Life Sciences, vol.
Bl/II, Protein and Enzyme Biochemistry, BS114:l-48,
Elsevier Scientific Publishers, Ltd., Ireland.
Hirschberg, C.B., M. Perez, M.D. Snider, W.L. Hanneman, J.
Esko, and C.R.H. Raetz. 1982. Autoradiographic
detection and characterization of a Chinese hamster
ovary cell mutant deficient in fucoproteins. J. Cell.
Physiol. 111:255-263.
Hirschberg, C.B. and M.D. Snider. 1987. Topography of
glycosylation in the rough endoplasmic reticulum and
golgi apparatus. Ann. Rev. Biochem. 56:63-87.
Ivatt, R.L., O.P. Das, E.J. Henderson, and P.W. Robbins.
1984. Glycoprotein biosynthesis in Dictyostelium
discoideum: developmental regulation of the protein-
linked glycans. Cell 38:561-567.
Jackson, S.P. and R. Tjian. 1988. O-Glycosylation of
eukaryotic transcription factors: implications for
mechanisms of transcriptional regulation. Cell 55:125-
133.
Kan, F.W.K. and P. Pinto da Silva. 1986. Preferential
association of glycoproteins to the euchromatin regions
of cross-fractured nuclei is revealed by fracture-
label. J. Cell Biol. 102:576-586.


92
macromolecular acceptors. In this chapter I present
evidence for a fucosyltransferase that partitions with FP21
in the cytosol. The fucosyltransferase was distinguished
from vesicular fucosyltransferase activity by several
criteria, and was characterized using hydrophobic synthetic
analogs.
Materials and Methods
Materials
GDP-[U-3H] fucose (6.6 Ci/mmol) and GDP-[U-14C ] fucose
(250 mCi/mmol) were from New England Nuclear (more than 90%
of the radiolabel was in the form of the fi anomer, as
indicated by the manufacturer). Reagent grade KC1, MnCl2,
CaCl2, BaCl2 and MgCl2 were from Mallinckrodt; GDP-fl-fucose
from Biocarb (stored frozen as a concentrated stock); GDP-a-
fucose (stored frozen as a concentrated stock), Tween-20,
CoCl2, Dowex-2 (2x8, minus 400, chloride form), Triton X-
100, chymostatin, pepstatin, NBZ-phenylalanine, bovine serum
albumin (BSA), and all p-nitro-phenyl acceptors were from
Sigma; Na2EDTA, FeCl3, formic acid, and trichloroacetic acid
were from Fisher. Cations and Na2EDTA were stored as
concentrated 500 mM solutions at 4. Hydrophobic synthetic
acceptors were generously provided by Monica Palcic. Making
and regeneration of Dowex-2 formate form column was as
described for Dowex-1 in Materials and Methods, Chapter II.
The concentrations of Triton X-100 and NP-40 are expressed


69
and no radioactivity was found associated with discoidin,
suggesting it does not bind to discoidin during
purification. I conclude that FP21 is not related to
discoidin or any discoidin isoforms, or gp24, and does not
bind to any discoidin isoforms.
Oligosaccharide Studies
FP21 from Ax3 and HL250 yield similar size
glycopeptides after pronase digestion. Pronase digested
gel-purified FP21 that had been metabolically labelled from
Ax3 and HL250 were compared (figure 3-4). More than 50% of
the radioactivity from both sources eluted as a major peak
with a relative elution coefficient (Rev) of 0.42 and 0.43,
respectively (ca. 5.5 glucose units). In the case of FP21
derived from Ax3, the rest of the radioactivity eluted
earlier in the void volume and distributed into minor peaks.
The digestion products from HL250 FP21 yielded a major peak,
with the rest of the radioactivity eluting earlier, and less
than 10% eluting after the major peak, although some
variability was observed in different runs regarding the
minor peaks. Radioactivity eluting at an earlier position
than the major peak may be explained by incomplete
digestion. The minor amount of radioactivity that eluted at
a later position may be due to breakdown of the
oligosaccharide. These results indicate that the main
glycopeptides derived from Ax3 and HL250 have the same


79
manners. Metabolically labelled P100 was subjected to SDS
PAGE and the material that comigrated with FP21 was gel
purified, pronase digested, and fractionated by gel
filtration chromatography. Entire P100 from metabolically
labelled cells was digested in the presence or absence of
detergent, and analyzed by BioGel P-4 gel filtration. As
seen earlier in figure 3-4, panel A, the S100 digest eluted
mainly as a single peak with a Rev of 0.42; on the other
hand, the digest from the FP21-comigrating P100 material,
fractionated as a major peak of 0.50 Rev (figure 3-7, panel
A). An analysis of glycopeptides from the entire P100
digestion showed a different elution profile from the FP21
digestion (figure 3-7, panel B). Note than in this
chromatograph there is a minor peak of 0.51 Rev, consistent
with the idea that the major peak seen in panel A is a minor
component of the entire P100 glycopeptide repertoire. Also
approximately 40% of the radioactivity eluted with the void
volume, suggesting it may be resistant to pronase digestion.
Digestion of the entire P100 fraction was carried out in the
presence of Triton X-100, to determine if solubilization of
the sample yielded a different digestion profile, by
facilitating accessibility of the enzyme to the substrates
(figure 3-7, panel C). The overall profile is similar, with
more than 30% of the radioactivity eluting at the void
volume, and a peak with a Rev of 0.51 is still a minor
component of the glycopeptides released. West et al. (1986)


19
reduction of incorporation of fucose into macromolecules was
unable to synthesize complex-type N-linked oligosaccharides,
resulting in a deficiency of acceptors (Hirschberg et al.,
1982). In another case, two glycosylation mutants were
shown to express a fucosyltransferase activity absent in the
parental cell line with the concomitant expression of a
novel linkage (Campbell and Stanley, 1984).
There are several putative glycosylation mutants. One
of these mutants, HL250, was selected after mutagenesis of
the parental normal strain Ax3 for the inability to bind
anti-SP96 antiserum (Loomis, 1987). This mutant also failed
to express a fucose-dependent epitope recognized by the
monoclonal antibody 83.5 (West et al., 1986). The focus of
my initial investigation was to characterize the mutation in
HL250 with the help of biochemical and morphological
techniques. We determined that HL250 is a fucosylation
mutant that lacked cellular fucose when grown in the absence
of fucose and that the defect is probably a result of a
lesion detectable in vitro in the conversion pathway that
forms GDP-fucose from GDP-mannose.
Materials and Methods
Materials
GDP-[ l-3H]mannose (9.1 Ci/mmol) was purchased from New
England Nuclear and L-[(5,6)-3H]-fucose (60 Ci/mmol) from
American Radiochemical Corporation. KC1 and MgCl2 were


Vi
Inclusion Volume
V
max
Maximal Velocity
Vo
Void volume
v/v
Volume per Volume
WGA
Wheat Germ Agglutinin
w/v
Weight per Volume
xiii


130
CH.,r¡H
1 D-gal-R
CH-.OH
2 gall3(I,3)G1cNAcI3-R
CH.-.CH
R= (CH2)8C00CH3


88
species (in accordance with Tsurchin, et.al. 1989) was
obtained with sizes unlike that of FP21 glycopeptides.
The designation of compartmentalization of a protein as
cytosolic is difficult since the cytosol is the site of
localization after disruption of organellar vesicles. This
task is complicated in the case of glycoproteins and
glycosylation enzymes, which are usually described as
components of the secretory pathway. However, some
glycoproteins have been identified as cytosolic, and
generally accepted as such (Hart et al., 1989a; Hart et al.,
1989b). My studies report the existence of a fucosylated
cytosolic protein. However, these results do not exclude
the possibility of FP21 being present in other locations
topologically continuous with the cytosol, such as the
nucleus, or being synthesized elsewhere and transported.
Due to its small size, FP21 could, in theory, be able to
diffuse freely into the nucleus.
Since glycoprotein fucosylation has been shown to take
place in enclosed organelles of the secretory pathway
(Hirschberg and Snider, 1987), the identification of FP21 in
the cytosol raises the question of where in the cell is
fucosylation of FP21 taking place. One scenario would have
FP21 being fucosylated in organellar vesicles (presumably
the Golgi apparatus) and subsequently transported to the
cytosol, while another would postulate the presence of a
fucosyltransferase that localized in the cytosol with FP21.


23
Macromolecules were EtOH precipitated and the remaining
supernatant dried down, resuspended in water, desalted on an
Amberlite MB-3 column and dried by vacuum centrifugation.
The samples were redissolved in water and analyzed using a
Dionex Bio-LC ion chromatograph by the method of Hardy et
al. (1988). Standards and modifications to the
chromatography procedure have been published elsewhere
(Gonzlez-Yanes et al., 1989). Specific activity, when
applicable, was determined by counting elution fractions and
was expressed as radioactivity present in the fucose peak
divided by the amount of fucose detected by the pulsed
amperometric detector coupled to the HPLC, with reference to
previously established calibration curves for L-fucose
(Hardy et al., 1988). Greater than 95% of the eluted
radioactivity was recovered from the column eluate at the
elution position of fucose.
Assay for Conversion of GDP-mannose to GDP-fucose
The conversion assay was carried out essentially as in
Ripka et al. (1986). In short, the standard assay mixture
contained in a final volume of 1 ml 600-750 pg S100 protein,
10 mM niacinamide, 5 mM ATP, 0.2 mM NAD+, 0.2 mM NADPH, 7.5
pM GDP-[3H]mannose (approximately 105 cpm) in 50 mM Tris-
HC1, pH 7.5. After incubation at 37 the reaction was
stopped with 50 pi of 2 N HC1, the reaction mixture boiled
for 20 min and subsequently neutralized with 55 pi of 2 N


BIOGRAPHICAL SKETCH
Beatriz Yadira Gonzlez-Yanes was born June 9, 1964, in
Fajardo, Puerto Rico. She graduated from Nuestra Seora del
Pilar High School in Ro Piedras, Puerto Rico, in 1981.
Following high school, she attended the University of Puerto
Rico in Ro Piedras, and earned a Bachelor of Science
degree, Magna Cum Laude, in biology in 1985. In August 1985
she entered graduate school at the University of Florida,
and joined the Department of Anatomy and Cell Biology in
February 1987. She completed the requirements for the
degree of Doctor of Philosophy in December 1991. She has
accepted a postdoctoral research position in the Animal
Science Department at the University of Florida in
Gainesville, Florida.
166


93
as volume/volume (v/v), all others are expressed as
weight/volume (w/v) unless specified otherwise.
Strains and Conditions of Growth
Dictyostelium discoideum strains Ax3 (from S. Free) and
HL250 (from W.F. Loomis) were grown on HL-5, a complete
medium that contains glucose, yeast extract, and proteose
peptone (Loomis, 1971). Ax3 is the normal strain and HL250
is a mutant obtained from Ax3 by N-methyl-N'-nitro-N-
nitrosoguanidine mutagenesis (Loomis, 1987). HL250 lacks
the enzyme activity that converts GDP-mannose into GDP-
fucose which results in a lack of cell fucose (Gonzlez-
Yanes et al., 1989; Chapter II). In all experiments cells
were collected at the logarithmic growth phase, with a cell
density of 1 to 9 x 106 cells/ml.
Cell Lysis and Fractionation
Vegetative and slug stage cells were fractionated into
S100 and P100 fractions as described in Materials and
Methods, Chapter III.
Fucosyltransferase Assay
Fucosyltransferase activity was assayed immediately
after obtaining the S100 and P100 fractions. Fractions were
found to be sensitive to freezing and thawing, and up to 70%
of the fucosyltransferase activity could be lost. The


DPM (P100)
107
DPM (S100)


Ill
To compare the S100 and P100 fucosyltransferase
activities, HL250 vegetative cells were harvested,
homogenized, and fractionated into an S100 and P100.
Fractions were assayed for fucosyltransferase activity under
different conditions in the presence of Tween-20 (table 4-
3). The fractions differed in that the S100
fucosyltransferase activity was approximately threefold more
efficient (on a per protein basis) than the P100 under
standard conditions (which contained 0.36 pM GDP-fucose and
5 mM MgCl2, see Materials and Methods). A major difference
between the bulk activities was their sensitivity to the
presence of divalent cations. In the absence of any added
cation, the P100 retains more than one fourth the activity
exhibited in the presence of Mg++, while the activity in the
S100 was almost negligible. The presence of EDTA does not
inhibit further the activity in the P100. One
interpretation is that the fucosyltransferase activity in
the S100 is dependent on added Mg++, while the activity in
the P100 is present in the absence and presence of Mg++,
being stimulated by the cation. An alternative explanation,
is that there are multiple enzymes in the P100, which differ
in their requirements for divalent cations.
Sensitivity to pH is a feature exhibited by enzymes,
including fucosyltransferases (Foster, et al. 1991; Kumazaki
and Yoshida, 1984). HL250 vegetative cells were harvested,
lysed, and fractionated into S100 and P100 fractions. Both


72
sizes, and suggest that both strains form in vivo the same
oligosaccharide when grown in fucose-containing media.
Oligosaccharide in FP21 is Q-linked. An enzymatic and
a chemical approach were used to determine whether the
fucose-containing oligosaccharide in FP21 is N-linked or 0-
linked. I first tried the enzyme PNGase F. The minimum
requirement of PNGase F is a di-N-acetylchitobiose core
(Chu, 1986; Tarentino et al., 1985). This enzyme has a
broad specificity and can cleave most asparagine linked N-
glycans (including high mannose and complex multibranched
oligosaccharides) provided they are not located at the amino
or carboxy termini (Chu, 1986). Gel-purified FP21 from
metabolically labelled Ax3 was digested with PNGase F. As
shown in figure 3-5, the fucose label eluted in the void
volume, while control substrates were quantitatively
digested as determined by SDS-PAGE (not shown). FP21
trypsinized prior to digestion with PNGase F also eluted in
the void volume (not shown). The inability of PNGase F to
release radioactivity from FP21 suggested that either the
fucose-containing oligosaccharide was not N-linked or the
oligosaccharide was insensitive to the enzyme.
I then considered the possibility that the
oligosaccharide in FP21 was 0-linked. Metabolically
labelled gel purified FP21 was subjected to mild alkaline,
reducing conditions, to release intact 0-linked


REFERENCES
Abeijon, C. and C.B. Hirschberg. 1990. Topography of
initiation of N-glycosylation reactions. J Biol. Chem.
265:14691-14695.
Aparicio, J., G.W. Erdos, and C.M. West. 1990. Spore coat is
altered in mod B glycosylation mutants of Dictyostelium
discoideum. J. Cell. Biochem. 42:255-266.
Arion, W.J., L.M. Balias, A.J. Lange, and B.K. Wallin. 1976.
Microsomal membrane permeability and the hepatic
glucose-6-phosphatase system. Interactions of the
system with D-mannose- 6-phosphate and D-mannose. J.
Biol. Chem. 251:4901-4907.
Barondes, S.H. and P.L. Haywood. 1979. Comparison of
developmentally regulated lectins from three species of
cellular slime mold. Biochim. Biophys. Acta 550:297-
308.
125
Bartles, J.R. and W.A. Frazier. 1980. Preparation of I-
discoidin I and the properties of its binding to
Dictyostelium discoideum cells. J. Biol. Chem. 255:30-
38.
Beniak, B., J. Orr, I. Brockhausen, G. Vella, and C. Phoebe.
1988. Separation of neutral reducing oligosaccharides
derived from glycoproteins by HPLC on a hydroxylated
polymeric support. Anal. Biochem. 175:96-105.
Bennett, G., C.P. Leblond, and A. Haddad. 1974. Migration of
glycoprotein from the Golgi apparatus to the surface of
various cell types as shown by radioautography after
labeled fucose injection into rats. J. Cell Biol.
60:259-284.
Beyer, T.A. and R.L. Hill. 1980. Enzymatic properties of the
B-galactoside al-2 fucosyltransferase from porcine
submaxillary gland. J. Biol. Chem. 255:5373-5379.
Beyer, T.A., J.E. Sadler, and R.L. Hill. 1980. Purification
to homogeneity of the H blood group fi-galactoside al-2
fucosyltransferase from porcine submaxillary gland. J.
Biol. Chem. 255:5364-5372.
156


108
interpret these results as an indication that the activity
that fucosylates FP21 in vivo is being assayed in vitro.
FP21 was fucosylated in vitro by incubating HL250 S100
fractions in the presence of GDP-[ 14C ] fucose, and the
oligosaccharide fucosylation in vitro was examined as before
for in vivo fucosylated FP21. In vitro fucosylated FP21 was
digested with PNGase F (not shown) or subjected to mild
alkaline hydrolysis and analyzed by gel filtration (figure
4-3, panel A). As was the case with metabolically labelled
FP21, PNGase F failed to release radioactivity and in vitro
labelled FP21 digested with PNGase F eluted in the void
volume. The digestion of the control substrates, fetuin and
ribonuclease B, was confirmed by SDS-PAGE. On the other
hand, approximately 20% of the radioactivity eluted with a
Rev of 0.50, the elution position of the metabolically
fucosylated oligosaccharide produced by Ax3 (reproduced for
comparison in figure 4-3, panel B, from figure 3-6, panel
A). The remainder of the radioactivity fractionated as
material of larger size. These results suggested that the
in vitro fucosylated oligosaccharide in FP21 was also O-
linked. The reasons for the discrepancies in size between
in vivo and in vitro fucosylated oligosaccharides are not
known, but may be due to incomplete release of the
oligosaccharide, accompanied by partial hydrolysis of the
polypeptide. Alternatively, it may indicate the presence of
oligosaccharides of various sizes.


9
associated with actively transcribed DNA, it has been
speculated that glycosylation may influence gene activity.
Highly purified HMG14 and HMG17 from mouse Friend
erythroleukemic cells were found to contain fucose, among
other sugars, by direct composition analysis (Reeves et al.,
1981). These proteins bound UEA-I specifically and could be
metabolically labelled with [3H]fucose. The
oligosaccharides were largely insensitive to fi-elimination,
suggesting an N-linkage to protein (Reeves et al., 1981).
When purified HMG14 and HMG17 were digested with mixed
glycosidases, the binding of HMG14 and HMG17 to the nuclear
matrix was abolished. Even though they did not employ a
fucosidase, making it difficult to ascertain the role of
fucose in binding, it was evident that glycosylation
influenced binding to the nuclear matrix (Reeves and Chang,
1983). These studies presented convincing evidence that HMG
14 and 17 are fucosylated, although they did not explore the
site of modification or the composition of the
oligosaccharide(s).
Nuclear fucoproteins and transcriptional activity.
Since some glycoproteins have been found associated with DNA
there has been speculation about a possible correlation
between the state of nuclear glycosylation and
transcriptional activity (Hart et al., 1989a). There are
examples in the literature of glycosylated transcription


22
Localization by Immunofluorescence
Prespore and spore cells were examined as described
previously (West and Loomis, 1985; Gonzlez-Yanes et al.,
1989). The monoclonal antibodies utilized have been
described previously elsewhere (West et al., 1986; Gonzlez-
Yanes et al., 1989).
Determination of Fucose Content and Specific Activity
The method is a modification of the protocol developed
by Yurchenco and Atkinson (1975) for HeLa cells. Spores
were harvested from sori not more than a day old and
resuspended in water without washing, since it has been
determined that a significant amount of spore coat protein
may be lost after washing spores in water (West and Erdos,
1990). Vegetative cells were grown in FM (in the presence
or absence of extracellular L-fucose) for determination of
sugar content. For determination of specific activity,
cells were grown in FM media supplemented with [3H]fucose.
After harvesting, cells were washed twice in PDF followed by
EtOH precipitation. The ethanol supernatant was reextracted
with EtOH and the resulting pellet pooled with the previous
precipitate. EtOH was evaporated under a stream of air.
Alternate methods such as TCA precipitation of the ethanol
supernatant, did not significantly increase the yield.
Samples were then hydrolyzed in a reacti-vial (Pierce) in
0.1 N HC1 for 45 min at 100 on a heating block.


TABLE OF CONTENTS
ACKNOWLEDGEMENTS iii
LIST OF TABLES viii
LIST OF FIGURES ix
LIST OF ABBREVIATIONS xi
ABSTRACT xiv
CHAPTERS
I HISTORICAL REVIEW AND BACKGROUND 1
Introduction 1
Fucosylated Macromolecules 3
Fucose-Binding Proteins 13
II CHARACTERIZATION OF A FUCOSYLATION MUTANT .... 18
Introduction 18
Materials and Methods 19
Results 24
Discussion 40
III IDENTIFICATION OF A CYTOSOLIC FUCOPROTEIN .... 45
Introduction 45
Materials and Methods 46
Results 54
Discussion 83
IV EVIDENCE FOR A CYTOSOLIC FUCOSYLTRANSFERASE ... 90
Introduction 90
Materials and Methods 92
Results 98
Discussion 135
V SUMMARY AND CONCLUSIONS 144
Summary of Results 144
Future Studies 149
vi


56
Table 3-1. Distribution of protein and radioactivity in S100
and P100 fractions.
S100 P100
total protein
(equivalents) 1
specific activity
3 days
5 days
14 dpm/pg
13 dpm/^g
0.95
88 dpm/pg
225 dpm/ng
HL250 amoebae were grown in the presence of 2 pCi/ml of
[3H]fucose in FM media for the indicated period of time. Data
are from one representative experiment. Cells were harvested,
filter lysed, and fractionated into an S100 and P100.


47
experiments cells were collected at the logarithmic growth
phase, with a cell density of 1 to 9 x 106 cells/ml. For
metabolic labelling experiments, cells were grown for 4-6
doublings in 2-20 pCi/ml (0.03-0.26 pM) of L-[3H]-fucose in
FM medium, a minimal defined medium that lacks fucose
(Franke and Kessin, 1977).
Cell Lysis and Fractionation
Logarithmically growing amoebae were harvested and
washed in 50 mM MES (pH 7.4) and resuspended to a
concentration of 2xl08 cells/ml in the lysis buffer
consisting of 0.25 M sucrose, 50 mM MES buffer (pH 7.4)
supplemented with the protease inhibitors leupeptin (10
pg/ml), aprotinin (10 pg/ml), and PMSF (1 mM). When
specified, cells were fractionated in the presence of a
comprehensive cocktail of protease inhibitors which have
been developed for the isolation of various proteolitically
sensitive proteins in Dictyostelium (Goodloe-Holland & Luna,
1987; Stone, et al., 1987). All steps were carried out at
0-4C. At once, the cells were gently lysed by forced
passage through a 5 pm nuclepore polycarbonate filter, with
a pore diameter slightly smaller than the diameter of the
cells (Das and Henderson, 1986). This method routinely
yielded more than 99% cell breakage as assessed by contrast
phase microscopy. The lysate was clarified from unbroken
cells and nuclei by a 2k xg centrifugation for 5 min, then


17
organism may be that there indeed are no
fucosyltransferases. Alternatively, if there are no
cytosolic fucosyltransferases, the presence of
fucoconjugates in the cytosol or nucleus may be explained by
membrane-associated enzymes that face the cytosolic
compartment, as has been reported for other
glycosyltransferases (Haltiwanger et al., 1990). In
addition, the lack of adequate acceptors, unfavorable
conditions for in vitro activity, and the scarcity of
sustained interest in the field, could account for the lack
of evidence for such fucosyltransferase(s) that, just as is
the case with other glycosyltransferases, are not of lumenal
origin.
Evidence for fucosylation, and glycosylation in
general, has been accumulating in the past decades. Until
recently, only sporadic reports about fucoproteins appeared
in the literature. With the identification of newly
reported glycoconjugates (Hart et al, 1989b), interest has
been revived in this area of research and currently there
appears to be much interest in the field. However, a
careful examination and sustained interest will be necessary
in order to elucidate the structures, and modes of
biosynthesis and functions of nuclear and cytosolic
fucoconjugates.


25
spore coat, lower germination efficiency in older spores,
and a longer doubling time when compared to the parental
strain Ax3 (Gonzlez-Yanes et al., 1989; West et al.,
manuscript in preparation).
Absence of a fucose-dependent epitope. The failure of
HL250 to react with anti-SP96 antiserum suggested that the
mutant might be deficient in a form of protein
glycosylation. This hypothesis was confirmed by finding
that HL250 did not react at appreciable levels with the
fucose-dependent monoclonal antibody (mAb) 83.5 (West et
al., 1986; Gonzlez-Yanes et al., 1989). Spores from mutant
and the parental normal strains were subjected to SDS-PAGE
and Western blotting and probed with mAbs 83.5 and A6.2
(West et al., 1986; Gonzlez-Yanes et al., 1989). The
latter monoclonal is specific for SP96. Consistent with the
supposition that glycosylation is affected in HL250, SP96
was reduced in apparent molecular weight compared to SP96
from Ax3 spores (West et al., 1986). The absence of the
fucose-dependent epitope was further examined in developing
cells and spores. Cells were plated for development and
slugs dissociated by shearing in the presence of EDTA.
Cells and spores were then processed for indirect
immunofluorescence. Figure 2-1 shows the localization of
SP96 in spores and prespore cells. Ax3 exhibits peripheral
labelling of spores and a punctate labelling from prespore


pmol fucose/mg protein/30 min
pmol fucoae/mg protein/30 min
O
115


89
I consider these possible alternatives in the following
chapter and present evidence for the presence of a
fucosyltransferase in the cytosol responsible for FP21
fucosylation.


96
Incorporation into hydrophobic synthetic acceptors was
determined as described by Palcic et al. (1988). The assay
was carried out as for endogenous acceptors, but GDP-
[3H]fucose was used instead of GDP-[ 14C ] f ucose and the
reaction was terminated by the addition of 1 ml ice-cold
water. The reaction mixture was loaded onto a C18 SepPak
column (Waters) under vacuum, and eluted with 6 successive 5
ml aliquots of water, and four 5 ml aliquots of methanol.
Eluates were counted by addition of 15 ml of ScintiVerse LC
(Fisher). In initial trials I determined that the
radioactivity eluted in the first methanol fraction, so in
subsequent experiments only the first methanol fraction was
used for determination of radioactivity incorporated. All
extracts were assayed in the absence of exogenous acceptor
and this value (usually about 20% of the dpm incorporated)
subtracted from experimental value to determine substrate-
dependent incorporation.
Purification of FP21
FP21 was purified in the following manner for
preparations which were to be added back to cytosolic
fractions to measure fucosylation acceptor activity.
Starting with 8 x 1010 cells, an S100 cytosolic fraction was
prepared from the mutant HL250. An aliquot of the fraction
(approximately 0.7% of the total volume) was incubated with
GDP-[ 14C]fucose and allowed to fucosylate FP21 with


134
Table 4-7
analog by
fraction
. Reduction of
purified FP21.
relative amount
fucosylation of acceptor type
r3Hlfucose incorporated
added
(dpm/mg protein/h)
0
499
20
lx
387
20
4x
318
21
lx
436
21
4x
<20
Transfer of [3H] from GDP-[3H]fucose into 4 pg (0.145 mM) of
type I acceptor analog by Ax3 S100 was measured using the C18
Sep-Pak method (see Materials and Methods). Data are the
results of one determination. 349 nq protein of Ax3 S100;
GDP-[3H]fucose concentration, 0.15 pM; 60 min assay.


64
with FP21 was observed in the PlOO-derived supernatant, this
material was a minority of the radioactivity released, and
probably reflected the general heterogeneity of the P100
vesicle contents. Less than 1% of the total cell
radioactivity that migrated at the 21 kD molecular weight
position was released from the P100 by sonication,
indicating that the remaining FP21 is recovered in the S100.
In a different approach, P100 from in vivo [3H]fucose
labelled cells was mixed with unlabelled post-nuclear
supernatant and recentrifuged (table 3-2). More than 98% of
the radioactivity sedimented with the P100, suggesting that
once associated with the P100, radioactivity is not lost,
unless vesicles are purposely disrupted, as in sonication.
The enzymes a-glucosidase-2, glucose-6-phosphatase, and
acid phosphatase have been used as markers of the
endoplasmic reticulum, Golgi apparatus and endoplasmic
reticulum, and lysosomes, respectively (Borts and Dimond,
1981; McMahon et al., 1977). To examine the distribution of
these enzyme markers in the high speed fractions, HL250
amoebae were harvested, homogenized, fractionated into an
S100 and P100 and assayed for activity of the different
marker enzymes. As seen in table 3-3, the majority of the
activity was recovered in the P100, suggesting minimal
contamination of the S100 by vesicles containing these
enzymes. Less than 7% of the total a-glucosidase-2 activity


Figure 4-5. Effect of GDP-fucose concentration on S100 and P100 fucosyltransferase
activities in the presence of Tween-20.
HL250 amoebae were harvested, homogenized, fractionated into an S100 and P100 and
fucosylated in vitro at varying concentrations of GDP-[ 14C]fucose. Reactions were
carried out for 30 min in the presence of 5 mM MgCl2, 0.1% Tween-20, 240 pg of
protein for the S100, and 432 pg of protein for the P100. Fucose incorporation was
calculated from the amount of TCA-precipitable [14C]radioactivity. Results expressed
as the mean of three determinations + s.e.m.
Panel A. Effect of GDP-fucose concentration on S100 fucosyltransferase activity in
the presence of Tween-20.
Panel B. Hanes single reciprocal plot ([S]/v vs. [S]) for the S100
fucosyltransferase activity in the presence of Tween-20; apparent Km=1.7 pM, apparent
Vmax=42.7 pmol/mg protein/30 min.
Panel C. Effect of GDP-fucose concentration on P100 fucosyltransferase activity in
the presence of Tween-20.
Panel D. Hanes single reciprocal plot ([S]/v vs. [S]) for the S100
fucosyltransferase activity in the presence of Tween-20; apparent Km=38.2 pM,
apparent Vmax=122 pmol/mg protein/30 min.


30
compared to prestalk and vegetative cells, so higher levels
of fucose would be expected in spores (Lam and Siu, 1981;
Gregg and Karp, 1978). The levels of glucose and mannose
were measured for comparison to determine if there was a
difference in the amounts of other sugars in the mutant.
None of the contents of the other sugars were reduced in the
mutant when grown in FM. Thus it seems that the lesion in
HL250 is selectively affecting fucose metabolism, relative
to that of other sugars. This is consistent with the fact
that fucosylation is usually a terminal modification of
oligosaccharides, so its addition is not a prerequisite for
the addition of other sugars (Kornfeld and Kornfeld, 1985).
Characterization of the Mutant Lesion
Earlier investigations have described mutants with a
fucose minus phenotype that are the result of a defect in
GDP-fucose biosynthesis. The main source of GDP-fucose is
the conversion pathway of GDP-mannose to GDP-fucose
(Yurchenco et al., 1978; Flowers, 1985). It consists of the
reactions presented in figure 2-2. Alternatively, synthesis
of GDP-fucose by the fucose salvage pathway can occur in the
presence of extracellular L-fucose (figure 2-2), allowing
cells defective in the conversion pathway to phenotypically
revert (Ripka and Stanley, 1986; Reitman et al., 1980).
HL250 cells grown and developed in the presence of 1 mM L-
fucose reexpressed the fucose epitope (Gonzlez-Yanes et


Figure 4-3. BioGel P-4 gel filtration chromatography of in vitro
labelled FP21 oligosaccharide.
HL250 S100 extracts were incubated in vitro in the presence
of GDP-[ 14C]fucose, desalted, subjected to fi-elimination (as
described in Materials and Methods), and analyzed by gel
filtration. For comparison, the profile resulting from fi-
elimination of in vivo labelled Ax3 gel purified FP21 is
presented in panel B (was panel A in figure 3-6, Chapter
III) .
Panel A. B-elimination of in vitro fucosylated FP21; arrow
identifies peak with Rev of 0.50. Vo, 38; Vi, 136.
Panel B. fi-elimination of Ax3 FP21. Vo, 36; Vi, 141.


Table 4-4. Fucosyltransferase activity in Ax3 S100 fraction.
Contribution
Intact3
Desalted13
Ax 3
HL250
DPM
rel. activity0
DPM
rel. activity
th. value*
1
0
3
<0.01
8
<0.01
0
0
1
1842
1.00
2715
1.00
1
1
1
1030
0.56
963
0.35
0.50
4
1
435
0.24
218
0.08
0.20
1
4
1425
0.77
2098
0.77
0.80
Ax3
S100 was
mixed with
mutant (HL250) S100
and assayed for fucosyltransferase activity
as
described
for the
standard assay for
30 min
. Results are the
average of two
determinations. The
samples were either
intact
or desalted prior
to assay, rel.
activity, relative activity; th. value, theoretical value. A total of 30 pi of extract
was used in each assay and mixing was done relative to volume contributed by each S100
fraction. 1:1 ratio was 15 pl:15 pi; 1:4, 6 pl:24 pi; 4:1, 24 pi: 6 pi. relative
activity calculated using HL250 activity as a unit. *theoretical value based solely on
contribution from HL250 fraction. aAx3 S100 was at a concentration of 5 pg of
protein/pl, and HL250 at 6 pg/pl. bAx3 protein concentration at 8.5 pg/pl and HL250 at
6 pg/pl.
122


ACKNOWLEDGEMENTS
First of all I would like to acknowledge my advisor,
Dr. Christopher M. West, for his time, training, and example
throughout these studies. I would also like to thank the
members of my committee, Drs. Robert Cohen, William Dunn,
and Carl Feldherr, for their time and guidance regarding
this project.
I would like to acknowledge and thank Dr. Michael Ross
for his interest and support of my graduate education.
Part of the work presented in Chapter II is reproduced
from the journal of Developmental Biology, 1989, volume 133,
pages 576-587 by copyright permission of Academic Press,
Inc.
I am very grateful to the following researchers and the
members of their laboratories for sharing equipment and/or
reagents with us: Drs. Gudrun Bennett, Ross Brown and Lina
Gritzali, William Dunn, Carl Feldherr, Michael Humphreys-
Beher, K.J. Kao, and Gillian Small.
In general, I would like to express my gratitude to the
faculty and staff in the Department of Anatomy and Cell
Biology for their support and interest in my advancement as
a scientific investigator. I would like to acknowledge Kari
Eissinger for her help in photographing and printing some of
iii


A NOVEL FUCOSYLATION PATHWAY IN THE CYTOSOL
OF DICTYOSTELIUM DISCOIDEUM
By
BEATRIZ GONZALEZ-YANES
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1991

A mis padres
(To my parents)

ACKNOWLEDGEMENTS
First of all I would like to acknowledge my advisor,
Dr. Christopher M. West, for his time, training, and example
throughout these studies. I would also like to thank the
members of my committee, Drs. Robert Cohen, William Dunn,
and Carl Feldherr, for their time and guidance regarding
this project.
I would like to acknowledge and thank Dr. Michael Ross
for his interest and support of my graduate education.
Part of the work presented in Chapter II is reproduced
from the journal of Developmental Biology, 1989, volume 133,
pages 576-587 by copyright permission of Academic Press,
Inc.
I am very grateful to the following researchers and the
members of their laboratories for sharing equipment and/or
reagents with us: Drs. Gudrun Bennett, Ross Brown and Lina
Gritzali, William Dunn, Carl Feldherr, Michael Humphreys-
Beher, K.J. Kao, and Gillian Small.
In general, I would like to express my gratitude to the
faculty and staff in the Department of Anatomy and Cell
Biology for their support and interest in my advancement as
a scientific investigator. I would like to acknowledge Kari
Eissinger for her help in photographing and printing some of
iii

the figures contained herein. Particularly, I would like to
thank Drs. Kelly Selman and Gillian Small for the advice and
friendship they have provided during these years. I am also
grateful to have been in contact with other students, past
and present, which have made the time spent in the
department more enjoyable. I am grateful to Dan Tuttle for
reviewing my dissertation. I would also like to acknowledge
all the people that have worked in the laboratory of Dr.
West, including Scherwin Henry, for making the time spent in
the laboratory more rewarding and Mel Fields for his
assistance in some experiments.
I would like to acknowledge my professors at the
University of Puerto Rico, who encouraged me to pursue a
career in biological research.
I appreciate all the friends I have made in Gainesville
in the past years and would like to acknowledge their
importance in my education. Especially, I thank Hans van
Oostrom for his moral support, encouragement, and for
teaching me about the wonders of the electronic world and
those rare creatures called computers.
Finally, and most importantly, I would like to thank my
family. I thank my grandparents, especially Abuela Emma and
Abuela Yuya, for the faith they have showed in me and for
their prayers, which have helped me through my graduate
education. My brothers Germn, Omar, and Carlos have been
great sources of happiness and pride. Lastly, I thank my
iv

parents, Beatriz and Germn, for everything. I firmly
believe that the education received during my first twenty-
one years of life in their company is ultimately responsible
for every achievement in my life.
v

TABLE OF CONTENTS
ACKNOWLEDGEMENTS iii
LIST OF TABLES viii
LIST OF FIGURES ix
LIST OF ABBREVIATIONS xi
ABSTRACT xiv
CHAPTERS
I HISTORICAL REVIEW AND BACKGROUND 1
Introduction 1
Fucosylated Macromolecules 3
Fucose-Binding Proteins 13
II CHARACTERIZATION OF A FUCOSYLATION MUTANT .... 18
Introduction 18
Materials and Methods 19
Results 24
Discussion 40
III IDENTIFICATION OF A CYTOSOLIC FUCOPROTEIN .... 45
Introduction 45
Materials and Methods 46
Results 54
Discussion 83
IV EVIDENCE FOR A CYTOSOLIC FUCOSYLTRANSFERASE ... 90
Introduction 90
Materials and Methods 92
Results 98
Discussion 135
V SUMMARY AND CONCLUSIONS 144
Summary of Results 144
Future Studies 149
vi

REFERENCES 156
BIOGRAPHICAL SKETCH 166
vii

LIST OF TABLES
Table Page
2-1. Fucose content of normal and mutant spores and
vegetative cells 29
2-2. Effect of time on conversion of GDP-mannose to
GDP-fucose 34
2-3. Specific activities of fucose 39
3-1. Distribution of protein and radioactivity in S100
and P100 fractions 56
3-2. Radioactivity recovered in the second S100 after
different P100 treatments 65
3-3. Distribution of markers among S100 and P100
fractions 66
4-1. Effect of different treatments on the cytosolic
fucosyltransferase activity 102
4-2. Failure to sediment S100 fucosyltransferase
activity 104
4-3. Comparison between the S100 and P100
fucosyltransferase activities 112
4-4. Fucosyltransferase activity in Ax3 S100 fraction 122
4-5. Utilization of 8-methoxycarbonyloctyl synthetic
acceptors by cytosolic fucosyltransferase activity
from Ax3 and HL250 128
4-6. Evaluation of the suitability of p-nitro-phenyl
glycosides as acceptors for cytosolic fucosyl
transferase activities in HL250 and Ax3 S100 . 132
4-7. Reduction of fucosylation of acceptor type I analog
by purified FP21 134
4-8. In vitro fucosyltransferase activity of HL250 and
Ax3 slug extracts 136
viii

LIST OF FIGURES
Figure Page
2-1. Localization of SP96 in prespore and spore cells 27
2-2. Biosynthesis of GDP-L-fucose 32
2-3. Conversion of GDP-mannose to GDP-fucose by normal
and mutant strains 37
3-1. Incorporation of [3H]fucose into macromolecular
species of the S100 and P100 58
3-2. Proteinaceous nature of FP21 60
3-3. Comparison of S100 and releasable P100 components 63
3-4. Gel filtration chromatography of FP21
glycopeptides 71
3-5. Gel filtration chromatography of PNGase F digests 74
3-6. Gel filtration chromatography of FP21
oligosaccharides 77
3-7. Gel filtration chromatography of P100
glycopeptides 81
3-8. Incorporation of [3H]fucose into macromolecular
species of slug stage cells 85
4-1. Fucosylation of endogenous acceptors by S100
fraction 101
4-2. SDS-PAGE profile of endogenous acceptors
fucosylated in vitro 107
4-3. BioGel P-4 gel filtration chromatography of in
vitro labelled FP21 oligosaccharide 110
4-4. Effect of pH on S100 and P100 fucosyltransferase
activities in the presence of Tween-20 115
ix

4-5. Effect of GDP-fucose concentration on S100 and P100
fucosyltransferase activities in the presence of
Tween-20 117
4-6. Effect of GDP-fucose concentration on intact S100
and P100 fucosyltransferase activities 120
4-7. Fucosylation of mutant FP21 by Ax3 S100 fraction 126
4-8. Haworth projections of the structures of the synthetic
glycolipid acceptors 130
x

LIST OF ABBREVIATIONS
APA
Asparagus Pea Agglutinin
ATP
Adenosine-5'-triphosphate
BSA
Bovine Serum Albumin
14c
Radioactive Carbon
cDNA
Complementary Deoxyribonucleic Acid
cm
Centimeter(s)
Con A
Concanavalin Agglutinin
cpm
Counts per Minute
CHO
Chinese Hamster Ovary
dH2
Distilled Water
dpm
Disintegrations per Minute
EDTA
Ethylenediaminetetraacetic Acid
EtOH
Ethanol
GlcNAc
N-acetyl Glucosamine
GDP
Guanosine-5'-diphosphate
h
Hour(s)
3h
Tritiated
HMG
High Mobility Group
HPLC
High Performance Liquid Chromatography
kD
Kilodaltons
K
m
Michaelis Constant
l
Liter(s)
xi

M
Molar Concentration
mCi
Millicurie(s)
mAb
Monoclonal Antibody
MES
2-(N-Morpholino)ethanesulfonic Acid
mg
Milligram(s)
ml
Milliliter(s)
min
Minute(s)
mm
Millimeter(s)
mM
Millimolar
MW
Molecular Weight
pCi
Microcurie(s)
V M
Micromolar
nm
Nanometers
P
probability
PAGE
Polyacrylamide Gel Electrophoresis
pmol
Picomole(s)
PMSF
Phenylmethylsulfonyl Fluoride
PNGase F
Peptide N-glycosidase F
Rev
Relative Elution Coefficient
RNAse B
Ribonuclease B
SDS
Sodium Dodecyl Sulphate
TCA
Trichloroacetic Acid
Tris
Tris(hydroxymethyl)aminomethane
U
Unit(s)
UEA-I
Ulex europaeus Agglutinin
Ve
Elution Volume
xii

Vi
Inclusion Volume
V
max
Maximal Velocity
Vo
Void volume
v/v
Volume per Volume
WGA
Wheat Germ Agglutinin
w/v
Weight per Volume
xiii

Abstract of Dissertation Presented to the Graduate School
of the University Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
A NOVEL FUCOSYLATION PATHWAY IN THE CYTOSOL
OF DICTYOSTELIUM DISCOIDEUM
By
Beatriz Gonzlez-Yanes
December 1991
Chairman: Dr. Christopher M. West
Major Department: Anatomy and Cell Biology
The existence of a fucosylation pathway in the cytosol
of Dictyostelium discoideum was investigated in the
glycosylation mutant strain, HL250, and the normal parental
strain, Ax3. HL250 was characterized as a conditional
mutant that cannot convert GDP(guanosine-5'-diphosphate)-
mannose to GDP-fucose, resulting in a lack of macromolecular
fucosylation unless grown in the presence of extracellular
fucose. HL250 or Ax3 cells were metabolically labelled with
[3H]fucose, filter-lysed, and fractionated by high speed
centrifugation into sedimentable (P100) and soluble (S100)
fractions. The fractions exhibited unique profiles of
fucoconjugates as analyzed by gel electrophoresis. The
major acceptor in the S100 was a 21 kilodalton molecular
weight protein, FP21. Analysis of FP21 oligosaccharide by
mild alkaline hydrolysis and gel filtration chromatography
xiv

revealed that fucose was incorporated into an O-linked
oligosaccharide with an average size of 4.8 glucose units.
FP21 appeared to be endogenous to the cytosol, based on the
failure to release FP21 from P100 vesicles by sonication,
and the absence of FP21-like glycopeptides derived by
pronase digestion of the P100.
To determine if FP21 was fucosylated in the cytosol,
S100 and P100 fractions from HL250 were assayed for
fucosyltransferase activity, measured as ability to transfer
[14C] from GDP-[ 14C] fucose to endogenous acceptors.
Fucosylation of FP21 by the S100 was time- and protein
concentration-dependent. The cytosolic activity was
distinguished from the bulk P100 activity by its absolute
divalent cation dependence, alkaline pH sensitivity, and
very low apparent Km for GDP-fucose. Fucosyltransferase
activity was not detectable in Ax3 cytosol. However,
activity was reconstituted by addition of purified mutant
FP21, suggesting FP21 was already fucosylated in living
cells, and Ax3 possessed a cytosolic fucosyltransferase
equivalent to the HL250 enzyme. S100 fractions from Ax3 and
HL250 were able to fucosylate an al,4fucosyltransferase-
specific acceptor glycolipid analog, but not analogs capable
of being modified by al,2 and al,3 fucosyltransferases.
Since fucosylation of the analog was reduced by addition of
purified FP21, the same enzyme appears to be responsible for
fucosylation of both molecules. Thus it is proposed that
XV

FP21 is synthesized and fucosylated in the cytosol by an
al,4fucosyltransferase.
xvi

CHAPTER I
HISTORICAL REVIEW AND BACKGROUND
Introduction
Glycoprotein synthesis and localization have been
subjects of intense study in the last few decades. Much has
been learned about glycosylation and some excellent reviews
are available (Kornfeld and Kornfeld, 1985; Hirschberg and
Snider, 1987). Carbohydrate moieties are usually
categorized as N and/or O-linked depending on whether the
carbohydrate moiety is attached to an asparagine by an amide
glycosidic linkage or to a serine or threonine,
respectively. Glycosylation is generally considered to be a
modification restricted to macromolecules that pass through
the secretory pathway, starting in the rough endoplasmic
reticulum where N-linked glycosylation is initiated
(Kornfeld and Kornfeld, 1985), but in some instances 0-
linked glycosylation also takes occurs (Spielman et al.,
1988). Glycosylation then continues in the Golgi apparatus
where further processing of N-linked oligosaccharides may
occur and O-linked glycosylation takes place (Abeijon and
Hirschberg, 1987). In the case of fucosylation, L-fucose
has been shown to be added as a terminal modification to
either N-linked or O-linked oligosaccharides in the Golgi
1

2
apparatus utilizing GDP-fucose as the sugar nucleotide donor
(Bennett et al., 1974; Hirschberg and Snider, 1987).
However, there are a few exceptions to these remarks, since
fucose has been found to be attached directly to serine and
threonine, although the site for this modification has not
been identified (Hallgreen et al., 1975), and to be present
in homopolymers, as in fucoidans (Flowers, 1981). It has
been suggested that fucosylation also occurs in the
endoplasmic reticulum as well as in the Golgi apparatus in
thyrotrophs under different physiological conditions (Magner
et al., 1986). There are two pathways for the biosynthesis
of GDP-fucose. The main source of GDP-fucose is the
conversion pathway of GDP-mannose to GDP-fucose (Yurchenco
et al., 1978; Flowers, 1985). Alternatively, synthesis of
GDP-fucose may occur by the fucose salvage pathway in the
presence of extracellular L-fucose (Yurchenco et al., 1978;
Flowers, 1985; Ripka and Stanley, 1986; Reitman et al.,
1980).
Glycosylation is carried out by diverse specific
glycosyltransferases which have been located in the
endoplasmic reticulum and Golgi apparatus. One underlying
assumption is that the acceptor macromolecule must
colocalize with the transferase enzyme in order to be
glycosylated. However, there have been occasional reports
of glycoproteins in compartments topologically discontinuous
with the lumen of the rough endoplasmic reticulum and Golgi

3
apparatus. These findings contradict the dogma that
glycosylation is strictly an endoplasmic reticulum- and
Golgi apparatus-dependent event. Two models could account
for the existence of glycoproteins outside the realms of the
secretory pathway; one postulates that lumenal glycosylated
proteins translocate across the membrane back to the
cytosolic space, the other that glycosyltransferases are
localized outside of the lumen of the endoplasmic reticulum
or Golgi apparatus. Evidence has been accumulating in the
past decade that supports the latter model. In this chapter
I shall be concerned with the presence of fucosylated
proteins in compartments topologically discontinuous with
the lumen of the endoplasmic reticulum and Golgi apparatus.
Subsequently, I shall analyze some of the studies that
suggest the presence of fucosyltransferases that would co
compartmentalize with such fucoproteins. In light of the
results presented in this dissertation, a review focussing
on the presence of fucoproteins and fucosyltransferases
outside the secretory pathway will be useful. An excellent
review is available that examines nuclear and cytosolic
glycosylation in general (Hart et al., 1989a).
Fucosylated Macromolecules
The presence of fucosylated proteins in compartments
discontinuous with the secretory pathway has been documented
since the 1970's. Various techniques have been employed in

4
these reports, including the direct analysis of carbohydrate
content, use of lectins, use of radiolabelled fucose and/or
a combination of biochemical and morphological techniques.
Nuclear Fucosylated Macromolecules
In contrast to the prevailing dogma, fucosylated
macromolecules have been detected in the nucleus. Such
studies demonstrated lectin binding to nuclear membranes,
chromatin, nuclear proteins, and the nuclear matrix. These
studies generally involved localization by binding of
labelled lectins (fluorescent, radiolabelled, ferritin-
conjugated, etc.) to isolated nuclei. Binding specificity
was assessed by competition of labelling with hapten
inhibitors. Occasionally, these studies were supplemented
by metabolic labelling experiments with radioactive fucose
which is primarily incorporated into fucoproteins with
little metabolism into other molecular species in eukaryotic
cells (Yurchenko et al., 1978).
Nuclear membranes. One early report suggested the
presence of fucose-containing structures on the cytoplasmic
face of isolated bovine nuclei (Nicolson et al., 1972).
Nicolson et al. (1972) found that purified nuclei were
agglutinated by the L-fucose-specific lectin UEA-I from Ulex
europaeus (Lis and Sharon, 1986), and agglutination was
inhibited by incubation in the presence of L-fucose,

5
suggesting that there were fucose-containing membrane-bound
oligosaccharides on the outer nuclear membrane. Similar
results were obtained with the mannose- and glucose-specific
lectin concanavalin A (Con A), which was also found to
agglutinate purified nuclei (Nicolson et al., 1972).
However, subsequent work by another group revealed that, as
evidenced by electron microscopic examination, ferritin-Con
A appeared to stain only damaged nuclei (Virtanen and
Wartiovaara, 1976) raising some concerns about the studies
by Nicolson et al. (1972). Likewise, in the aforementioned
studies (Nicolson et al., 1972), integrity of the isolated
nuclei was not determined, allowing for the possibility that
the lectin was interacting with lumenal fucoproteins that
escaped organelles during nuclei isolation or with
contaminating fucoprotein-containing membranes, such as
plasma membrane.
Chromatin. The existence of chromatin associated
fucose-containing proteins has been suggested by several
laboratories. Early on, Stein et al. (1975) reported the
presence of [3H] labelled glycoconjugates in purified
chromatin from HeLa cells grown in the presence of
[3H]fucose. Although the label was not examined to
corroborate its presence as fucose, in the case of
radiolabelling macromolecules with radioactive sugar
precursors, fucose is an excellent candidate because it has

6
been shown to be incorporated as such, with minimal
metabolizing of the label (Yurchenko et al., 1978). In
these studies the authors examined a very pure chromatin
preparation, with essentially no contamination from nuclear
or plasma membranes. The authors argued against plasma
membrane contamination based on mixing experiments, in which
radiolabeled cell-surface trypsinates were combined with
unlabeled nuclear preparations prior to chromatin isolation.
The fucosylated chromatin-associated macromolecules were
deemed to be fucoproteins based on their sensitivity to
pronase digestion. Unfortunately, these studies have not
been pursued further, and many questions remain unanswered
with respect to their structure and biosynthesis.
Chromatin-associated fucoproteins were also detected in
normal rat liver and Novikoff hepatoma ascites cells
(Goldberg et al., 1978) using the L-fucose-specific lectin
asparagus pea agglutinin (APA). The authors reported a
strongly basic fucoprotein, that was sensitive to pronase
digestion. Based on the reactivity with the lectin, they
calculated that the protein was three times more
concentrated in tumor chromatin than in normal liver cells.
However, the method for chromatin purification did not
eliminate adequately nuclear inner membrane contamination,
and since binding to the lectin was assayed by
affinoelectrophoresis a molecular weight for the protein was
not determined (Goldberg et al., 1978).

7
In a more recent study on duodenal columnar cells, Kan
and Pinto da Silva (1986) used UEA-I conjugated to colloidal
gold in freeze-fracture electron microscopy of cross-
fractured nuclei. They compared binding of the conjugated
lectin to euchromatin, heterochromatin, and nucleolus;
compartments which can be differentiated ultrastructurally.
Binding of UEA-I showed that colloidal gold particles were
almost exclusively confined to cross-fractured areas where
euchromatin was exposed. Labelling was abolished by
pretreatment and incubation in the presence of L-fucose, as
expected for a specific label. Pre-digestion of the
fractions with trypsin also abolished labelling, suggesting
the receptors for UEA-I binding were glycoproteins. The
preferential binding to euchromatin may be of importance,
because replication and transcription take place at
euchromatin regions. Although the results reported are
intriguing, the authors did not identify the type of fucose-
containing proteins detected. DNA-associated proteins can
be classified as either histone or non-histone proteins, and
as summarized below, both types of proteins appear to be
fucosylated.
Histones. The histones are the most abundant proteins
associated with DNA. Histones are very basic proteins and
are found in all nuclei (Darnell et al., 1986). Based on
the specific binding to UEA-I, Levy-Wilson (1983) presented

8
evidence that histones isolated from the macronucleus of
Tetrahymena thermophila appear to contain fucose. These
results were strengthened by metabolic incorporation of
[3H]fucose into histones, which showed that all five
Tetrahymena histones, HI, H2A, H2B, H3, and H4, appear to
contain fucose, with H2A incorporating the highest amount
(Levy-Wilson, 1983). In these studies macronuclei were
isolated to high purity, and highly pure histone
preparations were obtained after extensive washing in high
salt to remove nonhistone proteins. Using Con A, the author
showed specific binding to histones, which was interpreted
as evidence for the presence of mannose residues. However,
Con A has previously been shown to also recognize D-glucose
residues. Based on the extent of fucose incorporation and
its specific radioactivity, the author estimated, as the
lowest estimate, that one in a thousand nucleosomes
contained a fucosylated H2A molecule. To date, the
glycosylation pathway of histones and the oligosaccharide
structure(s) present in histones remain unknown.
Nonhistone proteins. High mobility group (HMG)
proteins are fairly abundant nonhistone chromosomal proteins
classified according to their relative electrophoretic
mobilities (Darnell et al., 1986). HMG proteins undergo a
variety of posttranslational modifications, including
glycosylation. Since they appear to be preferentially

9
associated with actively transcribed DNA, it has been
speculated that glycosylation may influence gene activity.
Highly purified HMG14 and HMG17 from mouse Friend
erythroleukemic cells were found to contain fucose, among
other sugars, by direct composition analysis (Reeves et al.,
1981). These proteins bound UEA-I specifically and could be
metabolically labelled with [3H]fucose. The
oligosaccharides were largely insensitive to fi-elimination,
suggesting an N-linkage to protein (Reeves et al., 1981).
When purified HMG14 and HMG17 were digested with mixed
glycosidases, the binding of HMG14 and HMG17 to the nuclear
matrix was abolished. Even though they did not employ a
fucosidase, making it difficult to ascertain the role of
fucose in binding, it was evident that glycosylation
influenced binding to the nuclear matrix (Reeves and Chang,
1983). These studies presented convincing evidence that HMG
14 and 17 are fucosylated, although they did not explore the
site of modification or the composition of the
oligosaccharide(s).
Nuclear fucoproteins and transcriptional activity.
Since some glycoproteins have been found associated with DNA
there has been speculation about a possible correlation
between the state of nuclear glycosylation and
transcriptional activity (Hart et al., 1989a). There are
examples in the literature of glycosylated transcription

10
factors and, at least in one case, glycosylation may have
influenced transcriptional activity (Lichtsteiner and
Schibler, 1989; Jackson and Tjian, 1989). As mentioned
earlier, in the case of fucosylated macromolecules, it was
found that in Novikoff hepatoma cells chromatin had three
times the amount of a fucoprotein of normal liver cells
based on APA binding (Goldberg et al., 1978). However, the
identity, size, or number of such proteins were not well
documented. Putative fucoproteins were found preferentially
associated with euchromatin (Kan and Pinto da Silva, 1986).
Fucosylated histones were found in the macronucleus of
Tetrahymena, where transcriptionally active chromatin is
compartmentalized (Levy-Wilson, 1983). Nevertheless,
histones from non-active chromatin were not studied, so no
comparisons can be made with heterochromatin. Levy-Wilson
(1983) suggested that the reason why other investigators
have failed to detect fucosylation in mammalian histones is
due to the low proportion of transcriptionally active genome
which would imply a low concentration of fucosylated
histones.
Although the data gathered in these reports are
interesting, they are the result of isolated studies from
diverse organisms and it is difficult to draw generalized
conclusions. In addition, virtually nothing is known about
the structure or biosynthesis of the fucose-containing
moieties. Even though the fractions in all these reports

11
appeared to have almost no contamination, structural studies
showing a different glycoconjugate from those found in other
organelles would convincingly argue against contamination.
Nevertheless, these provocative studies are encouraging and
deserve to be pursued further.
Cytosolic Fucoconjugate
The existence of glycoproteins in the cytosol has been
documented in the past and reviewed recently (Hart et al.,
1989a). Studies suggesting the existence of glycoproteins
in the cytosol were based on determinations of lectin
binding sites or biochemical compositional analyses (Hart et
al., 1989a). Some of the negative results reported by those
studies, that relied on lectin binding as confirmation for
the presence of a sugar residue, may be misleading since
lack of binding may reflect a poor choice of lectins to
probe with and not necessarily the absence of a fucose
residue. While there are several fucose-binding lectins
available that serve as useful biochemical tools, they do
not recognize every possible fucose-containing structure.
Lectin binding is dependent on a specific array of sugar
residues, and does not depend solely on the presence of
fucose (Lis and Sharon, 1986). However, as will be
described below, there is one biochemical study that reports
the existence of a cytosolic fucoconjugate. In many cell
fractionation experiments, the cytosolic compartment is

12
defined by the lack of sedimentation during high-speed
centrifugation. However, this criterion alone is not
sufficient, since cytosolic glycans might arise from
contamination by other organelles. The most convincing
evidence describes fucoconjugates that appear to be
preferentially enriched in the cytosol in relation to other
compartments.
In studies in rat brain, a soluble proteoglycan that
contains novel O-linked mannose-containing oligosaccharides
was recovered in the cytosolic fraction (Margolis et al.,
1976; Finne et al., 1979). The proteoglycan was
characterized as a soluble chondroitin sulfate proteoglycan,
that contained neutral oligosaccharides releasable by mild
alkaline borohydride treatment. The oligosaccharides
contained mannose at their proximal ends, and one
oligosaccharide was proposed to be composed of mannose,
GlcNAc, fucose, and galactose (Finne et al., 1979).
However, the possibility remains that the oligosaccharides
are not integral components of the proteoglycan, but were
associated with other co-purified material. Nevertheless,
the oligosaccharides appeared to be endogenous to the
cytosol and not the result of contamination from other
fractions, since they were present in only trace amounts in
the microsomal or synaptosomal membrane fractions (Finne et
al., 1979). Unfortunately, the function and biosynthetic
pathway of these oligosaccharides remain unknown.

13
Fucose-Bindinq Proteins
The presence of fucoproteins in the nucleus and cytosol
prompted the idea that there might be specific proteins
inside the cell that bind, and/or modify these fucoproteins,
as is the case with the fucoproteins in the secretory
pathway. This would include fucose-binding proteins,
fucosyltransferases responsible for fucosylation, and
fucosidases responsible for fucose removal. There are some
studies suggesting the existence of fucose-binding lectins
and fucosyltransferases. However, there is no evidence for
a nuclear or cytosolic fucosidase.
Endogenous Lectins
In light of evidence for glycoproteins in the cytosolic
and nuclear compartments, investigators sought the existence
of carbohydrate-binding proteins that would colocalize with
such glycoproteins. Endogenous lectins have been detected
in preparations of nucleoplasmic and/or cytosolic fractions
of a variety of cells (Hart et al., 1989a).
Aided by fluorescein-labeled neoglycoproteins, Sve et
al. (1986) have postulated the presence of endogenous
fucose-specific lectins. Baby hamster kidney cell nuclei
were isolated by two different procedures, cell lysis and
enucleation in Ficoll, to argue against contamination by
cytoplasmic or membrane-derived components. Using
fluorescein-labelled BSA conjugated to fucose in the order

14
of 20+5 sugar units per molecule, fluorescence microscopy
experiments suggested that the majority of the binding
appeared to be associated with nucleoli and nucleoplasmic
ribonucleoprotein elements (Sve et al., 1986).
Interestingly, it was shown by quantitative flow
microfluorometry that nuclei from exponentially growing
cells bound one order of magnitude more fucose-BSA than
nuclei from contact-inhibited cells. In spite of these
results, the authors acknowledge that it is impossible based
on the data to ascribe biological roles to the nuclear
fucose-binding sites or to conclude that they influence
cellular physiology (Sve et al., 1986).
In Dictyostelium discoideum a family of lectins, the
discoidins, has been identified, and discoidin I has been
the most extensively studied isoform. In erythrocyte
agglutination assays, agglutination by discoidin may be
inhibited by galactose, modified galactose residues, L-
fucose, D-fucose, and other sugars, suggesting the existence
of cell-surface sugar-dependent epitopes on erythrocytes
recognized by Discoidin (Barondes and Haywood, 1979).
Although discoidin I was originally thought to be a cell
surface protein involved in cell-cell adhesion, it has since
been established that it is primarily present in the cytosol
(Erdos and Whitaker, 1983). Although galactose and modified
galactose residues are the ligands bound by discoidin I with
highest affinity, it is possible that a fucose-containing

15
macromolecule may serve as a ligand for it or that another
lectin, possibly from the discoidin family, will be present
in the cytosol with fucose-binding capability.
Fucosyltransferases
Intracellular fucosyltransferases identified to date,
are localized in the lumen of the Golgi or possibly,
endoplasmic reticulum. Thus the presence of fucoproteins in
the nucleus and cytosolic compartments poses questions
pertaining to the mode of synthesis and/or intracellular
transport of such glycoproteins. Early on, Kawasaki and
Yamashina (1972) theorized, based on metabolic labelling,
that nuclear membrane glycoproteins were not synthesized in
and transported from microsomes, but were made in the
nuclear membrane or its vicinity. A few reports have
suggested the presence of glycosyltransferases in the
nucleus, nuclear membranes, or cytosol that may be involved
in the modification of several glycoproteins; however, none
of these glycoproteins was reported to contain fucose
(Richard et al., 1975; Galland et al., 1988, Haltiwanger et
al., 1990).
Though there is evidence for nuclear and cytosolic
fucosylated macromolecules, to my knowledge there are no
reports of nuclear or cytosolic fucosyltransferases.
Louisot and collaborators reported the purification and
separation of two soluble fucosyltransferase activities from

16
rat small intestinal mucosa (Martin et al., 1987). They
report the isolation of al,2 and ccl,3/l,4fucosyltransferases
that could be candidates for cytosolic fucosyltransferases
based on their inability to sediment after homogenization of
cells in 0.25 M sucrose, followed by 90 min, 200k x g
centrifugation (Martin et al., 1987). They compare the
al,2fucosyltransferase with the cc2,6sialyltransferase, which
is normally a Golgi enzyme converted to a soluble form by
cleavage of the amino terminal signal anchor to allow for
secretion (Weinstein et al., 1987). In the case of the
fucosyltransferases, the authors did not report using
protease inhibitors during isolation, nor did they document
the partitioning of cellular markers in the different
fractions (Martin et al., 1987). The reasons the enzymes
localize to the high speed supernatant, which would usually
be considered as the cytosolic fraction, may include the
breakage of the microsomal vesicles or the fact that they
were initially present extracellularly in the body fluids of
the mucosa. In another report these investigators tested
for the presence of glycosyltransferases in the nucleus. In
highly purified rat liver nuclei there was a total absence
of fucosyltransferase activity when endogenous
macromolecules or asialofetuin were used as acceptors
(Richard et al., 1975).
One explanation for the lack of evidence of nuclear
and/or cytosolic fucosyltransferases activity in any

17
organism may be that there indeed are no
fucosyltransferases. Alternatively, if there are no
cytosolic fucosyltransferases, the presence of
fucoconjugates in the cytosol or nucleus may be explained by
membrane-associated enzymes that face the cytosolic
compartment, as has been reported for other
glycosyltransferases (Haltiwanger et al., 1990). In
addition, the lack of adequate acceptors, unfavorable
conditions for in vitro activity, and the scarcity of
sustained interest in the field, could account for the lack
of evidence for such fucosyltransferase(s) that, just as is
the case with other glycosyltransferases, are not of lumenal
origin.
Evidence for fucosylation, and glycosylation in
general, has been accumulating in the past decades. Until
recently, only sporadic reports about fucoproteins appeared
in the literature. With the identification of newly
reported glycoconjugates (Hart et al, 1989b), interest has
been revived in this area of research and currently there
appears to be much interest in the field. However, a
careful examination and sustained interest will be necessary
in order to elucidate the structures, and modes of
biosynthesis and functions of nuclear and cytosolic
fucoconjugates.

CHAPTER II
CHARACTERIZATION OF A FUCOSYLATION MUTANT
Introduction
Since fucosylation comprises numerous steps which are
potential sites for mutations, many fucosylation mutants
have been obtained. Fucosylation consists of the synthesis
of the sugar nucleotide donor GDP-fucose and the transfer of
fucose from the GDP-fucose to an acceptor (Kornfeld and
Kornfeld, 1985). The main source of GDP-fucose is the
conversion pathway of GDP-mannose to GDP-fucose (Yurchenco
et al., 1978; Flowers, 1985). Alternatively, synthesis of
GDP-fucose by the fucose salvage pathway can occur in the
presence of extracellular L-fucose, allowing cells defective
in the conversion pathway to phenotypically revert (Ripka
and Stanley, 1986; Reitman et al., 1980). Lesions may
affect the formation of GDP-fucose, the transport of GDP-
fucose to the fucosylation compartment, the transferases
responsible for fucosylation and/or the biosynthesis or
transport of the acceptor species to the fucosylation
compartment. Two mutants deficient in protein-associated
fucose were shown to be defective in the formation of GDP-
fucose (Reitman et al., 1980; Ripka et al., 1986). A
Chinese hamster ovary (CHO) cell line that showed a marked
18

19
reduction of incorporation of fucose into macromolecules was
unable to synthesize complex-type N-linked oligosaccharides,
resulting in a deficiency of acceptors (Hirschberg et al.,
1982). In another case, two glycosylation mutants were
shown to express a fucosyltransferase activity absent in the
parental cell line with the concomitant expression of a
novel linkage (Campbell and Stanley, 1984).
There are several putative glycosylation mutants. One
of these mutants, HL250, was selected after mutagenesis of
the parental normal strain Ax3 for the inability to bind
anti-SP96 antiserum (Loomis, 1987). This mutant also failed
to express a fucose-dependent epitope recognized by the
monoclonal antibody 83.5 (West et al., 1986). The focus of
my initial investigation was to characterize the mutation in
HL250 with the help of biochemical and morphological
techniques. We determined that HL250 is a fucosylation
mutant that lacked cellular fucose when grown in the absence
of fucose and that the defect is probably a result of a
lesion detectable in vitro in the conversion pathway that
forms GDP-fucose from GDP-mannose.
Materials and Methods
Materials
GDP-[ l-3H]mannose (9.1 Ci/mmol) was purchased from New
England Nuclear and L-[(5,6)-3H]-fucose (60 Ci/mmol) from
American Radiochemical Corporation. KC1 and MgCl2 were

20
obtained from Mallinckrodt; formic acid from Fisher; ATP
(disodium salt, catalog number A-5394), niacinamide, NAD+,
NADPH, Trizma base, phenylmethylsulfonyl fluoride, Dowex-1
(1x8, minus 400, chloride form), hexokinase, GDP-mannose,
and Amberlite MB-3 were obtained from Sigma. ATP was stored
frozen at a concentration of 500 mM in 1 mM Tris-HCl (pH
7.4), resulting in a final pH of approximately 5.5. Dowex-1
formate form was made as follows: 1) The column was washed
with 1 M HC1 until pH of eluate was below 2 (as determined
by pH paper). 2) The column was washed with water until pH
was higher than 4.5. 3) Subsequently, the column was washed
with 1 M NaOH until pH was higher than 13. 4) The column
was washed with water as described in step 2. 5) The column
was washed with 3 volumes of 1 M formic acid until pH of
eluate was below 2. 6) Lastly, the column was washed with
water as described in step 2. This procedure was also
followed for regeneration of column.
Strains and Conditions of Growth and Development
Dictyostelium discoideum amoebae were grown on HL-5, a
medium that contains glucose, yeast extract, and proteose
peptone (Loomis, 1971). The axenic strains Ax3 (from F.
Rothman) and HL250 (from W.F. Loomis) were maintained by
passage during logarithmic growth phase. Ax3 is the normal
strain and HL250 is a mutant obtained from Ax3 by N-methyl-
N'-nitro-N-nitrosoguanidine mutagenesis. For metabolic

21
labelling experiments, cells were grown for 4-6 doublings in
8-20 pCi/ml (0.10-0.26 pM) of [3H]fucose in FM medium, a
minimal defined medium that lacks fucose (Franke and Kessin,
1977). When appropriate, the medium was supplemented with
L-fucose (Sigma). For development, cells were plated in PDF
buffer (20 mM KC1, 45 mM sodium phosphate, 6 mM MgS04, pH
5.8) on filters as previously described (West and Erdos,
1988) .
Cell Lysis and Fractionation
Logarithmically growing amoebas were harvested and
washed in 50 mM Tris-HCl (pH 7.5) and resuspended to a
concentration of 2xl08 cells/ml in the lysis buffer
consisting of 0.25 M sucrose, 50 mM Tris-HCl buffer (pH 7.5)
supplemented with 1 mM PMSF. All operations were carried
out at 0-4C. Cells were immediately lysed by forced
passage through a 5 pm nuclepore polycarbonate filter, with
pore diameter slightly smaller than the diameter of the
cells (Das and Henderson, 1986). This method routinely
yields more than 99% cell breakage. The lysate was
centrifuged at 100k xg for 1 hour. The supernatant (S100)
was made 2.5% (v/v) in glycerol by addition of 100% glycerol
and either used immediately or saved at -80 without
significant loss of activity for four weeks.

22
Localization by Immunofluorescence
Prespore and spore cells were examined as described
previously (West and Loomis, 1985; Gonzlez-Yanes et al.,
1989). The monoclonal antibodies utilized have been
described previously elsewhere (West et al., 1986; Gonzlez-
Yanes et al., 1989).
Determination of Fucose Content and Specific Activity
The method is a modification of the protocol developed
by Yurchenco and Atkinson (1975) for HeLa cells. Spores
were harvested from sori not more than a day old and
resuspended in water without washing, since it has been
determined that a significant amount of spore coat protein
may be lost after washing spores in water (West and Erdos,
1990). Vegetative cells were grown in FM (in the presence
or absence of extracellular L-fucose) for determination of
sugar content. For determination of specific activity,
cells were grown in FM media supplemented with [3H]fucose.
After harvesting, cells were washed twice in PDF followed by
EtOH precipitation. The ethanol supernatant was reextracted
with EtOH and the resulting pellet pooled with the previous
precipitate. EtOH was evaporated under a stream of air.
Alternate methods such as TCA precipitation of the ethanol
supernatant, did not significantly increase the yield.
Samples were then hydrolyzed in a reacti-vial (Pierce) in
0.1 N HC1 for 45 min at 100 on a heating block.

23
Macromolecules were EtOH precipitated and the remaining
supernatant dried down, resuspended in water, desalted on an
Amberlite MB-3 column and dried by vacuum centrifugation.
The samples were redissolved in water and analyzed using a
Dionex Bio-LC ion chromatograph by the method of Hardy et
al. (1988). Standards and modifications to the
chromatography procedure have been published elsewhere
(Gonzlez-Yanes et al., 1989). Specific activity, when
applicable, was determined by counting elution fractions and
was expressed as radioactivity present in the fucose peak
divided by the amount of fucose detected by the pulsed
amperometric detector coupled to the HPLC, with reference to
previously established calibration curves for L-fucose
(Hardy et al., 1988). Greater than 95% of the eluted
radioactivity was recovered from the column eluate at the
elution position of fucose.
Assay for Conversion of GDP-mannose to GDP-fucose
The conversion assay was carried out essentially as in
Ripka et al. (1986). In short, the standard assay mixture
contained in a final volume of 1 ml 600-750 pg S100 protein,
10 mM niacinamide, 5 mM ATP, 0.2 mM NAD+, 0.2 mM NADPH, 7.5
pM GDP-[3H]mannose (approximately 105 cpm) in 50 mM Tris-
HC1, pH 7.5. After incubation at 37 the reaction was
stopped with 50 pi of 2 N HC1, the reaction mixture boiled
for 20 min and subsequently neutralized with 55 pi of 2 N

24
NaOH. Quantitation of the conversion of GDP-mannose to GDP-
fucose was achieved by determining the amount of fucose
present after acid hydrolysis. Free mannose, released from
GDP-mannose, was phosphorylated by adding 4 units of
hexokinase in the presence of 5 mM ATP and 5 mM MgCl2 at 37
for 1 hour. Hexokinase catalyzes the transfer of one
phosphate to C-6 of any acceptor hexose. Since fucose lacks
a hydroxyl group at position C-6, it cannot be
phosphorylated. Parallel controls with no enzyme were used
to correct for losses. The mixture was passed over a Dowex-
1 (formate) column (0.6 x 5 cm) and eluted with water. 1 ml
fractions were collected an aliquot of 100 pi from the
eluate was counted using ScintiVerse LC (Fisher); fucose-
associated radioactivity usually eluted by the first 2 ml.
Calculation of K and V was done by the Lineweaver-Burk
double reciprocal plot (1/v vs. 1/[S]) as discussed by
Henderson (1985).
Results
Phenotypic Description of a Fucosylation Mutant
The normal strain Ax3 was mutagenized with
nitrosoguanidine and surviving clones were screened with
anti-SP96 antiserum (Loomis, 1987). One of the clones,
HL250, was selected due to its inability to bind anti-SP96
antiserum which recognizes carbohydrate and peptide
epitopes. HL250 has been found to have a more permeable

25
spore coat, lower germination efficiency in older spores,
and a longer doubling time when compared to the parental
strain Ax3 (Gonzlez-Yanes et al., 1989; West et al.,
manuscript in preparation).
Absence of a fucose-dependent epitope. The failure of
HL250 to react with anti-SP96 antiserum suggested that the
mutant might be deficient in a form of protein
glycosylation. This hypothesis was confirmed by finding
that HL250 did not react at appreciable levels with the
fucose-dependent monoclonal antibody (mAb) 83.5 (West et
al., 1986; Gonzlez-Yanes et al., 1989). Spores from mutant
and the parental normal strains were subjected to SDS-PAGE
and Western blotting and probed with mAbs 83.5 and A6.2
(West et al., 1986; Gonzlez-Yanes et al., 1989). The
latter monoclonal is specific for SP96. Consistent with the
supposition that glycosylation is affected in HL250, SP96
was reduced in apparent molecular weight compared to SP96
from Ax3 spores (West et al., 1986). The absence of the
fucose-dependent epitope was further examined in developing
cells and spores. Cells were plated for development and
slugs dissociated by shearing in the presence of EDTA.
Cells and spores were then processed for indirect
immunofluorescence. Figure 2-1 shows the localization of
SP96 in spores and prespore cells. Ax3 exhibits peripheral
labelling of spores and a punctate labelling from prespore

Figure 2-1. Localization of SP96 in prespore and spore cells.
Slugs cells were dissociated, placed onto polylysine-coated
glass slides, fixed, permeabilized, and processed for
indirect immunofluorescence using mAbs 83.5 or A6.2.

HL250 AX3
27
PRESPORE CELLS
83.5 A6-2
SPORES
83.5
A6-2

28
vesicles using both mAb. However, HL250 shows the same
pattern only when A6.2 is used, in agreement with the
results observed by Western blotting. The fact that
labelling of spores with A6.2 is similar in both strains
indicates that spore coat localization of SP96 is not
affected by the mutation.
Fucose content of normal and mutant strains. Epitope
recognition by mAb 83.5 was inhibited by L-fucose (West et
al., 1986), so the possibility of a defect in fucosylation
in strain HL250 was investigated. Initially, the
macromolecular fucose content of vegetative cells grown in
fucose-free media, and of spores was investigated. Ethanol
insoluble macromolecules were acid hydrolyzed, ethanol
precipitated, and the supernatant deionized and
chromatographed on an alkaline anion-exchange column
equipped with a pulsed amperometric detector. When
authentic [3H]fucose was fractionated in this manner, more
than 95% of the radioactivity eluted at the fucose position
(not shown). Fucose content was found to be negligible in
the mutant, HL250, both in spores and vegetative cells when
compared to the normal strain, Ax3 (table 2-1). However,
when HL250 cells were grown in FM supplemented with 1 mM L-
fucose, they possessed detectable amounts of fucose.
Previous investigators have found by autoradiography that
[3H]fucose incorporation is highest in prespore cells

Fucose content of normal and mutant spores and vegetative cells.
Table 2-1.
strain
cell type
conditions of qrowth
fucose
qlucose
mannose
Ax 3
spores
581+18
n.d.
n.d.
HL250
spores
0.44
n.d.
n.d.
Ax 3
amoebae
FM
11.0+1.3
83.6+5.0
30.0+0.7
FM + ImM L-fucose
6.5+0.2
86.8+1.4
17.1+1.4
HL250
amoebae
FM
0.075
116.3+3.8
39.5+1.3
FM + ImM L-fucose
2.5 + 0.0
69.0+1.3
40.3+1.4
Spores were collected from fresh fruiting bodies from cells that were plated after
growing in HL-5. Vegetative cells were grown in fucose-free media (FM) or FM
supplemented with fucose. Fucose content was determined as described in Materials and
Methods. Results expressed as nmoles of sugar/ mg of protein as the mean of three
determinations s.e.m.; n.d., not determined.

30
compared to prestalk and vegetative cells, so higher levels
of fucose would be expected in spores (Lam and Siu, 1981;
Gregg and Karp, 1978). The levels of glucose and mannose
were measured for comparison to determine if there was a
difference in the amounts of other sugars in the mutant.
None of the contents of the other sugars were reduced in the
mutant when grown in FM. Thus it seems that the lesion in
HL250 is selectively affecting fucose metabolism, relative
to that of other sugars. This is consistent with the fact
that fucosylation is usually a terminal modification of
oligosaccharides, so its addition is not a prerequisite for
the addition of other sugars (Kornfeld and Kornfeld, 1985).
Characterization of the Mutant Lesion
Earlier investigations have described mutants with a
fucose minus phenotype that are the result of a defect in
GDP-fucose biosynthesis. The main source of GDP-fucose is
the conversion pathway of GDP-mannose to GDP-fucose
(Yurchenco et al., 1978; Flowers, 1985). It consists of the
reactions presented in figure 2-2. Alternatively, synthesis
of GDP-fucose by the fucose salvage pathway can occur in the
presence of extracellular L-fucose (figure 2-2), allowing
cells defective in the conversion pathway to phenotypically
revert (Ripka and Stanley, 1986; Reitman et al., 1980).
HL250 cells grown and developed in the presence of 1 mM L-
fucose reexpressed the fucose epitope (Gonzlez-Yanes et

Figure 2-2. Biosynthesis of GDP-fucose.
Diagrams of the conversion pathway elucidated in bacteria
and mammalian cells (adapted from Flowers, 1981) and the
salvage pathway as described in HeLa cells (adapted from
Yurchenko et al., 1978).

32
4,6-dehy O-GDP
A CONVERSION PATHWAY
O-GDP
B. SALVAGE PATHWAY

33
al., 1989), and vegetative cells grown in 1 mM L-fucose had
detectable amounts of macromolecular-associated fucose
(table 2-2). These results suggested that HL250 had a
normal salvage pathway, but a defect in the GDP-mannose to
GDP-fucose conversion pathway.
In vitro conversion of GDP-mannose to GDP-fucose.
Conversion of GDP-mannose to GDP-fucose has been reported in
bacteria, in higher plant cells, and mammalian cells
(Kornfeld and Ginsburg, 1966; Liao and Barber, 1971; Ripka
et al., 1986). It was assumed that Dictyostelium would
share this ability with other species, so I assayed if cell
extracts in vitro were able to convert GDP-mannose to GDP-
fucose. High speed supernatants from Ax3 and HL250 were
assayed in vitro for their ability to convert GDP-
[3H]mannose into GDP-[ 3H] fucose. Cells were homogenized and
a 100k x g soluble fraction (S100) assayed as described by
Ripka et al. (1986) and the effects of time, varying protein
and GDP-mannose concentrations were examined. Table 2-2
shows that the conversion activity in Ax3 was proportional
to the time of incubation. In contrast, mutant extracts
showed negligible activity. Ax3 and HL250 cytosols were
mixed to determine if a soluble inhibitor of GDP-mannose to
GDP-fucose conversion activity was present. When equal
amounts of extracts were mixed, activity was commensurate
with the Ax3 contribution (table 2-2) indicating HL250 does

Table 2-2.
GDP-fucose.
Effect of time on conversion of GDP-mannose to
nmol fucose/mq protein
time (min)
Ax3
HL250
Ax3+HL250 (0.5:0.5)
7.5
0.57
0
0.21
15
2.5
0
1.3
30
2.9
0.15
2.2
90
9.2
0
4.3
Protein (600
yq total)
from a
100,000 xg supernatant of
vegetative cell-free extract was assayed for ability to
convert GDP-[ 14C]mannose (7.5 initial concentration) to GDP-
[ 4C]f ucose; data are the result of the average of two
determinations

35
not contain an inhibitor for the activity. Conversion was
linear with respect to protein through 800 pg (figure 2-3,
panel A). In agreement with the results from the time
dependence experiment, HL250 expressed less than 1% of the
activity possessed by Ax3 at all concentrations of protein
assayed (figure 2-3, panel A). Conversion by Ax3 S100 was
also dependent on the amount of GDP-mannose present while
the activity present in HL250 was insignificant (figure 2-3,
panel B). The Ax3 GDP-mannose to GDP-fucose conversion
activity showed an apparent Km of 14.1 pM and Vnwy of 18.3
nmol fucose/mg protein/30 min (fig. 2-3, panel C). Previous
reported values for the apparent Km of the conversion
pathway range from 2 pM in CHO cells (Ripka et al., 1986) to
160 pM in the higher plant Phaseolus vulgaris (Liao and
Barber, 1971).
The conversion activity is absent in mutant extracts in
vitro, if absent in vivo, this would explain the lack of
fucose in living cells and the correction by exogenous
fucose. Taken together, all these results point to the
conversion pathway as the site of the lesion in HL250.
Furthermore, the defect seems to be in the first step of the
conversion of GDP-mannose to GDP-fucose because no
radioactivity was recovered after hexokinase treatment. If
the 4,6-dehydratase was active, the product would have
eluted from the ion-exchange column after hydrolysis and
phosphorylation by hexokinase (see figure 2-2).

Figure 2-3. Conversion of GDP-mannose to GDP-fucose by normal
and mutant strains. Closed circles, Ax3; open circles,
HL250. Results are the average of two determinations.
Panel A. Effect of protein concentration on conversion. 30
min. assay, 7.5 /vM GDP-[3H]mannose.
Panel B. Effect of GDP-mannose concentration on conversion.
30 min assay, 750 pg protein for each strain.
Panel C. Apparent Michaelis constant for GDP-[3H]mannose,
determined for Ax3 conversion. Apparent Km was determined
to be 14.1 pM and apparent Vmax 18.3 nmol/mg protein/30 min
by the Lineweaver-Burk double reciprocal plot method.

-0.09 -0.00 0.09 0.18 0.27
1/[S]
1 /V
nmol fucose/mg protein/30 min
nmol fucose/30 min

38
Specific activity of fucose pools. A lesion in the
GDP-mannose to GDP-fucose conversion pathway would render
the mutant defective in macromolecular fucosylation when
grown in the absence of fucose, as was shown earlier. Such
a scenario would require that the GDP-fucose inside the cell
be derived from the salvage pathway fed exclusively by
fucose from the extracellular media. In the case of the
Ax3, however, there would be a contribution of GDP-fucose
derived from the conversion pathway. If the mutation in
HL250 was indeed in the GDP-mannose to GDP-fucose
conversion, macromolecular fucose of cells grown in the
presence of L-[3H] fucose would have the same specific
activity as the fucose present in the media. To test this
hypothesis, mutant and normal cells were grown in fucose-
free defined media supplemented with [3H]fucose. Whole cell
preparations were then assayed for fucose content as
described above and the specific activity expressed as the
radioactivity present in the fucose peak divided by the
amount of fucose detected. As expected, the macromolecular
pool of the mutant had essentially the same specific
activity as the fucose in the medium. In contrast, the
specific activity in Ax3 was diluted approximately 400-fold
compared to the medium (table 2-3). These results confirmed
that HL250 derived its intracellular fucose from the salvage
pathway. Meanwhile it appears that in Ax3 the contribution

39
Table 2-3. Specific activities of fucose.
cpm/nmol
fucose
strain
fucose concentration
medium
macromolecular
HL250
50 pM
1.lxlO5
9.5x10*
Ax3
0.1 pM
1.8xl07
4.7x10*
Cells were grown for 3 days in FM media in the presence of
6xl06 dpm/ml of [3H]fucose supplemented with non-radioactive
fucose to yield the noted concentration of fucose in the
media. Specific activity was determined as described in
Materials and Methods.

40
of fucose by the salvage pathway is minimal, 1/400 of the
total content. These studies agree with earlier reports on
fucose metabolism, where the majority of fucose in mammalian
cells is derived from the conversion of GDP-mannose to GDP-
fucose (Yurchenko et al., 1978).
Discussion
HL250 failed to express a fucose-dependent epitope
recognized by the mAb 83.5. However, it expressed SP96, one
of the polypeptides that bears the carbohydrate epitope
recognized by 83.5. These results were reproduced by
Western blot (Gonzlez-Yanes et al., 1989) and indirect
immunofluorescence. Interestingly, the immunofluorescence
microscopy studies showed that the mutant is not defective
in its ability to package SP96 in vesicles or in the
targeting of the glycoprotein to the spore coat. Similar
results were reported for other proteins in another
Dictyostelium glycosylation mutant (Aparicio et al., 1990;
West and Loomis, 1985). Measurements of fucose content of
cells and spores demonstrated that the mutant contained
almost undetectable amounts of fucose, in contrast to Ax3
which contained macromolecular-associated fucose in both
cell types.
Once HL250 was identified as having a mutation that
resulted in decreased macromolecular fucose, I tried to
identify the nature of the lesion. It was speculated that

41
the mutant HL250 may have a defect in (1) fucosyltransferase
activities, (2) endogenous acceptors for
fucosyltransferases, (3) transport of GDP-fucose into
microsomal vesicles, and/or (4) synthesis of GDP-fucose. In
vitro microsomal extracts of normal and mutant cells were
active in the transfer of [14C]fucose from GDP-[ 14C] fucose to
endogenous acceptors and the activity was latent (see
Chapter IV), so it was reasoned that fucosyltransferases may
be normal and probably the uptake of GDP-fucose by vesicles
was not impaired, so the lesion might be at another point in
the fucosylation pathway. Since fucose is normally added as
a terminal modification, the fucose minus phenotype could be
the result of a lack of formation of acceptors for the
fucosyltransferases (Stanley, 1984; Hirschberg et al.,
1982). HL250 expresses levels comparable with Ax3 of other
carbohydrate epitopes and has normal neutral monosaccharide
composition (Gonzlez-Yanes et al., 1989; West et al.,
1986), for these reasons it was speculated that the defect
was not in an earlier step of glycosylation but it involved
the fucosylation pathway directly.
The glycosylation defect in HL250 appears to result
from an inability to produce GDP-fucose. I have found that
the GDP-mannose to GDP-fucose conversion activity in vitro
is reduced to undetectable levels in mutant cell extracts.
The fact that there is partial rescue when the cells are
grown in the presence of fucose, suggests that the cells are

42
producing GDP-fucose via the salvage pathway and that the
rest of the fucosylation machinery is probably normal.
Earlier studies have reported the phenotypic reversion of
mammalian mutants with a defective GDP-mannose to GDP-fucose
conversion pathway when the cells were supplied with
extracellular fucose (Ripka and Stanley, 1986; Reitman et
al., 1980). However, Ripka and Stanley (1986) used the
recovery of lectin sensitivity as a marker for phenotypic
reversion, but did not report measuring the fucose content
of the cells or show the data for lectin binding compared to
the parental strain. Reitman et al. (1980) showed that a
mouse lymphoma cell line which has a defect in the
conversion of GDP-mannose to GDP-fucose was defective in pea
lectin binding compared to the parental cell line. The
ability to bind pea lectin was restored to wild type
parental cell line levels after culturing in 10 mM fucose.
Fucose is an important determinant in the carbohydrate
binding specificity of pea lectin (Kornfeld et al., 1981).
It is important to note that the mutant mouse lymphoma cell
line had approximately one fifth the amount of fucose and
one third the number of high affinity lectin-binding sites
as the parental line, indicating that mutant cells were
salvaging fucose from the medium, or that the mutation was
only partial. The evaluation for phenotypic reversion of
HL250 is more rigorous since it demands the expression of a

43
carbohydrate epitope and measures total levels of fucose
from a cell that was grown in fucose-free media.
The specific activity of the medium was compared to the
specific activity of the intracellular macromolecular
fucose. Consistent with a lesion in the GDP-mannose to GDP-
fucose conversion pathway, the mutant cells relied on
extracellular fucose as their only fucose source. In
contrast, extracellular fucose only contributed to a small
fraction of the total Ax3 fucose pool. This is useful
because it means that radioactivity from cells grown in
[3H]fucose can be used as a direct measure of fucosylation
in HL250.
In conclusion, even though HL250 has a severe
glycosylation lesion which renders it unable to carry out
fucosylation when grown in the absence of fucose, the strain
is able to grow, develop, and form spores. To my knowledge,
there are no previous reports in the literature of
eukaryotic cells defective in the GDP-mannose to GDP-fucose
conversion pathway that can survive in fucose-free media.
The fucosylation mutants reported are cell lines that have a
functional salvage pathway maintained in culture in the
presence of animal serum, so they can synthesize GDP-fucose
from the fucose present in the cell culture media (Ripka and
Stanley, 1986; Reitman et al., 1980). For this reason,
these investigators were unable to totally deprive the
mutants of fucose, as I am able to do with Dictyostelium.

44
Other lower eukaryotes, such as yeast, do not carry out
fucosylation (Kukuruzinska et al., 1987) so the existence of
a Dictyostelium fucosylation mutant could be very important
to study fucosylation. In any event, HL250 has already
served as a very useful tool in which to study fucosylation
events, as will be evident in the following chapters.

CHAPTER III
IDENTIFICATION OF A CYTOSOLIC FUCOPROTEIN
Introduction
There is some evidence that fucoconjugates are present
in the nucleus and cytosol (Hart et al., 1989a; Chapter I).
However, virtually nothing is known about the structure or
biosynthesis of these fucosylated macromolecules. Most
studies have limited themselves to reporting the existence
of evidence for nuclear or cytoplasmic glycoproteins, but
have not gone further to characterize the sugar-peptide
linkage or compare it with material derived from the
secretory pathway. As discussed in Chapter II, there is a
conditional fucosylation mutant, HL250, that can be readily
labelled when grown in radioactive fucose. Using this
strain, I have identified a fucoprotein that fractionated
with the cytosol and appeared to be the major fucosylated
species in the cytosol. The oligosaccharide-peptide linkage
was characterized and the fucoprotein was compared with
fucosylated material derived from vesicles, and
differentiated based upon several criteria.
45

46
Materials and Methods
Materials
L-[ (5,6)-3H]-fucose (60 Ci/mmol) was obtained from
American Radiochemical Corporation and D- [ (2,3)-3H]mannose
(24 Ci/mmol) from New England Nuclear. TS-1 was purchased
from Research Product International; POPOP was from
Mallinckrodt; SDS, leupeptin, aprotinin, phenyl
methylsulfonyl fluoride, Triton X-100,
dimethyldichlorosilane, MES, all nitro-phenyl substrates,
mannose-6-phosphate, bovine serum albumin (fraction V), blue
dextran, bromo phenol blue, and trypsin were from Sigma;
glycine, benzene, ammonium acetate, PPO, and toluene were
from Fisher. The concentrations of Triton X-100 and NP-40
are expressed as v/v, all others are expressed as w/v,
unless specified otherwise.
Strains and Conditions of Growth
Dictyostelium discoideum strains Ax3 (from S. Free) and
HL250 (from W.F. Loomis) were grown on HL-5, a complete
medium that contains glucose, yeast extract, and proteose
peptone (Loomis, 1971). Ax3 is the normal strain and HL250
is a mutant obtained from Ax3 by N-methyl-N'-nitro-N-
nitrosoguanidine mutagenesis (Loomis, 1987). HL250 lacks
the enzyme activity that converts GDP-mannose into GDP-
fucose which results in a lack of cell fucose (Gonzlez-
Yanes et al., 1989; also see Chapter II). In all

47
experiments cells were collected at the logarithmic growth
phase, with a cell density of 1 to 9 x 106 cells/ml. For
metabolic labelling experiments, cells were grown for 4-6
doublings in 2-20 pCi/ml (0.03-0.26 pM) of L-[3H]-fucose in
FM medium, a minimal defined medium that lacks fucose
(Franke and Kessin, 1977).
Cell Lysis and Fractionation
Logarithmically growing amoebae were harvested and
washed in 50 mM MES (pH 7.4) and resuspended to a
concentration of 2xl08 cells/ml in the lysis buffer
consisting of 0.25 M sucrose, 50 mM MES buffer (pH 7.4)
supplemented with the protease inhibitors leupeptin (10
pg/ml), aprotinin (10 pg/ml), and PMSF (1 mM). When
specified, cells were fractionated in the presence of a
comprehensive cocktail of protease inhibitors which have
been developed for the isolation of various proteolitically
sensitive proteins in Dictyostelium (Goodloe-Holland & Luna,
1987; Stone, et al., 1987). All steps were carried out at
0-4C. At once, the cells were gently lysed by forced
passage through a 5 pm nuclepore polycarbonate filter, with
a pore diameter slightly smaller than the diameter of the
cells (Das and Henderson, 1986). This method routinely
yielded more than 99% cell breakage as assessed by contrast
phase microscopy. The lysate was clarified from unbroken
cells and nuclei by a 2k xg centrifugation for 5 min, then

48
it was centrifuged at 100k xg for 1 hour, unless otherwise
specified. The pellet (P100) was resuspended in lysis
buffer by pipetting to the same volume as the supernatant
(S100). For lysing vesicles, the P100 was sonicated using a
Branson Sonifier Cell Disrupter 185.
Slug cells were plated for development as described
earlier (West and Erdos, 1988), and harvested in buffer of
Berger and Clark (as described in West and Brownstein, 1987)
supplemented with 20 mM EDTA. Cells were dissociated in
this buffer by passing 20 times through a long, 9 inches,
pasteur pipette, followed by passing 20 times through a 23-
gauge needle. EDTA was washed by resuspending cells in 50
mM MES, pH 7.4, titrated with NaOH. Cells were resuspended
in lysis buffer and immediately lysed by passage through a 3
pm nuclepore polycarbonate filter (Das and Henderson, 1986).
Cell lysates were then treated as described above for
vegetative cells.
Gel Electrophoresis and Western Blotting
SDS-PAGE was carried out under reducing conditions
essentially as described in West & Loomis (1985). Samples
were resolved by 7-20% acrylamide linear gradient gels or
15% acrylamide gels and, initially, molecular weight
assigned using low MW markers kit (Sigma). In later
experiments, trypsin was used as a molecular weight marker.
Following electrophoresis, the gels were cut immediately

49
and/or stained and destained and then cut into either 2.2 mm
or 0.5 cm slices. Gel pieces were shaken and swollen
overnight in a scintillation cocktail composed of 111.1 ml
of tissue solubilizer (TS-1), 6.0 g PPO, 0.15 g POPOP, and
20 ml dH20 to 1 1 of toluene. Gel slices were counted and
recounted until dpm were determined to be stable, usually 1-
2 days later. For gel-purified material, the sample was run
in a 7-20% linear gradient gel and the 21 kD area
(approximately 1 cm below trypsin) cut out and electroeluted
overnight using a Bio-Rad electroeluter following
manufacturer's directions, except the Laemmli
electrophoresis buffer used for SDS-PAGE (West and Loomis,
1985) was used instead of the recommended volatile buffer to
avoid alkaline hydrolysis. Western blotting was carried out
as previously described (West and Loomis, 1985).
Partial Purification of FP21 by Anion Exchange
Chromatography
In preliminary studies, FP21 was partially purified by
fractionation of metabolically labelled S100 fraction on a
TSK DEAE-5PW 8 x 75-mm HPLC anion-exchange column (LKB)
preequilibrated with 10 mM NH^Ac, pH 7.0. Sample was
dialyzed against 2 1 of 10 mM NH4Ac, pH 7.0, for several
hours, and clarified by centrifugation at 10k x g for 10 min
prior to injection. Protein was eluted using an increasing
linear gradient (10 mM to 1 M NH^Ac, pH 7.0) for 40 min at a
rate of 0.75 ml per min, and the majority of FP21 was found

50
to elute at 0.5 M input buffer concentration. Fractions
were analyzed by SDS-PAGE and counting of the gel slices.
Protein Concentration Assay
The Bio-Rad protein assay was used for determination of
protein concentration, and bovine serum albumin used as
standard.
Enzyme Assays
a-glucosidase-2 assays contained 100-300 pg protein,
8.6 mM p-nitrophenyl-a-D-glucoside, 0.1% Triton X-100, in 21
mM citrate-phosphate buffer (pH 7.5) at 37 (Borts and
Dimond, 1981). Reaction was stopped after 1 hr by addition
of Na2C03 to a concentration of 0.5 M and the absorbance
read at 420 nm. Glucose-6-phosphatase and mannose-e-
phosphatase were measured by release of phosphate from
mannose-6-phosphate, which has been previously shown to be a
suitable substrate for both enzymes (Arion et al., 1976).
Reaction mixtures contained 100-300 pg protein, 1 mM MgCl2,
2 mM mannose-6-phosphate, 0.1% Triton X-100 in 10 mM MES
(titrated with NaOH to a pH of 7.4) in a volume of 200 pi.
After 20 min incubations at 30 reactions were stopped by
adding 200 pi 20% ice-cold TCA (Snider et al., 1980). Tubes
were centrifuged at 14k rpm, for 10 min in an Eppendorf
table top microfuge and aliquots of the supernatants were
assayed for P1 by the method of Chen et al. ( 1956). Acid

51
phosphatase was assayed as described in McMahon et al.
(1977) except that Triton X-100 was included at a
concentration of 0.1% and absorbance was measured at 420 nm,
instead of 400 nm.
PNGase F Digestion
Gel-purified FP21 (3000-8000 dpm) was boiled in 0.5%
SDS in water for 3 min. For digestion, the protocol of
Tarentino et al. (1985) was followed. The sample was
incubated in 104 mM sodium phosphate, pH 8.6, 10 mM EDTA, 10
mM 1,10-phenanthroline (stock solution of 100 mM in
methanol), 2% NP-40, 0.21% SDS, and 20 U/ml PNGase F
(Boehringer-Mannheim) in 300 jil for 22 h at 37. Fetuin (B-
grade, Calbiochem) and RNAse B (Sigma) were treated
identically as controls and digestion was quantitative, or
nearly quantitative, as determined by shifts in molecular
weight in SDS-PAGE. In one trial FP21 was digested with
trypsin prior to PNGase F digestion by incubating gel
purified FP21 in 0.8 mg/ml trypsin, 1 mM CaCl2 for 1 hr at
37. The reaction was stopped by boiling in the presence of
1 mM PMSF and 10 pg/ml aprotinin for 3 min.
Pronase Digestion
Glycopeptides were prepared by exhaustive pronase
digestion as described (Ivatt et al, 1984). In short, after
gel purification and electroelution, the samples were dried

52
down and resuspended in water to a volume of 200 pi
containing at least 5xl03 dpm. In other experiments, 200 pi
of the entire P100 fraction were left intact or made 0.1%
Triton X-100. 200 pi of freshly dissolved 1% pronase
(CalBiochem) in 0.1 M Tris-HCl (pH 8.0), 1 mM CaCl2, were
added at 0, 24, and 48 hours. Incubation was at 50 and a
few drops of toluene were added to prevent microbial growth.
At 72 hours the reaction was stopped by incubating in a
boiling water bath for 3 min.
Oligosaccharide Release by Alkaline-Borohydride Treatment
The oligosaccharide in FP21 was released by mild
alkaline hydrolysis under reducing conditions, also known as
fl-elimination. To approximately 5xl03 dpm of gel purified
protein, freshly made NaOH and NaBH4 concentrated solutions
were added in that order to yield a final concentration of
0.1 M and 1 M, respectively. Samples were incubated for 15
hours in a water bath at 45. The reaction was stopped by
the addition of acetic acid to a final concentration of 1 M.
The samples were dried down by vacuum centrifugation,
resuspended once in 1 ml 100 mM HAc, dried again, and
resuspended twice in methanol and stored dry at -80 until
ready to use.
The oligosaccharide was also released by strong
alkaline-borohydride treatment. The method is that of Zinn
et al. (1978) and very similar to the procedure for B-

53
elimination, with some exceptions: NaOH and NaBH4 were
present at a final concentration of 1 and 4 M, respectively,
and samples were incubated at 80 for 24 h. Reaction was
terminated by diluting the sample twofold with water and
adding acetic acid to a final concentration of 4 M. Borate
salts were removed by methanol evaporation as described
above.
P-4 Column Fractionation
Dry samples from fl-elimination and strong alkaline-
borohydride treatment were resuspended in 800 pi of 50 mM
pyridinium acetate (pH 5.5). Pyridinium acetate was made in
the hood by mixing in water, to a final volume of 2 liters,
8.06 ml pyridine, 2.17 ml glacial acetic acid. After the
previous reagents were dissolved, 0.4 g of sodium azide was
added. The solution had a final pH of approximately 5.5.
The buffer was degassed and stored under chloroform
atmosphere. Samples from pronase digestion were centrifuged
for 5 min at 14k rpm on a Eppendorf table top microfuge and
the supernatant taken for P-4 chromatography.
Oligosaccharides and glycopeptides were fractionated in a
0.9 cm x 1 m BioGel P-4 column (-400 mesh) equilibrated with
50 mM pyridinium acetate (pH 5.5) as the mobile phase.
Prior to pouring, the column was acid washed overnight and
siliconized by coating with 1% (v/v) dimethyldichlorosilane
in benzene for 10 min. The column was calibrated with

54
glucose oligomers used as standards (Yamashita et al., 1982)
that were derived from a dextran hydrolysate which was
reduced with NaB3H4 (kindly provided by J. Baezinger).
Twelve drop fractions (approximately 250 pi) were collected
and counted using ScintiVerse LC (Fisher Scientific). All
runs were performed at 37. Recovery varied somewhat
between runs, but it was between 30-80% of the loaded
radioactivity. The void volume (Vo) was determined with
blue dextran (at a concentration of 0.2%), and the inclusion
volume (Vi) with either bromo phenol blue (0.2%) or
[3H]mannose (approximately 2,000 dpm). The relative elution
coefficient (Rev) for each component was determined from the
elution volume (Ve): Rev = (Ve-Vo)/(Vi-Vo).
Results
Analysis of Cellular Fucoproteins by SDS-PAGE
HL250 amoebae were grown in minimal defined media
supplemented with L-[ 3H]fucose. During logarithmic growth
phase, cells were harvested, washed, and resuspended in 0.25
M sucrose buffer supplemented with protease inhibitors.
Immediately, the cells were lysed and the lysate was
clarified from unbroken cells and nuclei by centrifugation
at 2k x g for 5 min. The resulting supernatant from this
spin was centrifuged at 100k x g for 1 hour. Both fractions
were analyzed by SDS-PAGE and the gels stained, cut into 2.2
mm slices, and counted. While total protein distributed in

55
a ratio of almost 1:1 (P100:S100), the specific activity
(expressed as dpm/mg of protein) distributed roughly in a
6:1 ratio (P100:S100) (table 3-1). The majority of the
radioactivity fractionated with the P100. However, the
amounts of radioactivity recovered in the S100 were
unexpected; 36% of the radioactivity in the S100 (this value
varied in individual experiments from 30-68%) migrated as
one peak at the 21 kD position, slightly ahead of trypsin,
which was used as a molecular weight marker. On the other
hand, the P100 showed two main broad peaks, one at 10-25 kD
and another at 58-84 kD with 10% of the total radioactivity
in the P100 migrating at the 21 kD level (figure 3-1). The
proteinaceous nature of the S100 fucoprotein was confirmed
by digestion with the proteases trypsin and pronase with
quantitative recovery of radioactivity at lower MW positions
in unfixed gels (results from pronase digestion shown in
figure 3-2). The S100 fucosylated protein has been called
FP21 for fucoprotein of 21 kD molecular weight.
The abundance of FP21 was estimated based on the
specific activity of fucose and determined to be 103
molecules in FP21, assuming one fucose molecule per molecule
of FP21. Based on the dilution of fucose specific activity
in Ax3 (from table 2-3, Chapter II), there would be 4 x 105
molecules in Ax3. If there is one fucose per copy of FP21,
and all FP21 molecules are fucosylated (evidence for
quantitative FP21 fucosylation in Ax3 will be presented in

56
Table 3-1. Distribution of protein and radioactivity in S100
and P100 fractions.
S100 P100
total protein
(equivalents) 1
specific activity
3 days
5 days
14 dpm/pg
13 dpm/^g
0.95
88 dpm/pg
225 dpm/ng
HL250 amoebae were grown in the presence of 2 pCi/ml of
[3H]fucose in FM media for the indicated period of time. Data
are from one representative experiment. Cells were harvested,
filter lysed, and fractionated into an S100 and P100.

Figure 3-1. Incorporation of [3H]fucose into macromolecular
species of the S100 and P100.
HL250 amoebae were metabolically labelled with 2 pCi/ml of
[3H]fucose, lysed, fractionated into an S100 and P100, and
subjected to 7-20% linear gradient SDS-PAGE; the gel was
sliced into 2.2 mm pieces and counted. 100 pg of protein
were electrophoresed for the S100 and P100, respectively.
Open circles, P100; closed circles, S100; arrow, migration
of trypsin.

S100 (DPM)
58
600
500
400
300
200
100
0
P100 (DPM)

Figure 3-2. Proteinaceous nature of FP21.
Approximately 1,500 dpm of metabolically [3H]fucose-labelled
FP21 from HL250 was gel purified, as described in Materials
and Methods, and subjected to either a mock or pronase
digestion. Resulting digests were electrophoresed on a 15%
SDS polyacrylamide gel, which was then sliced into 0.5 cm
pieces and counted. Open circles, FP21; closed circles,
FP21 digested with pronase.

60
gel slice

61
Chapter IV), and the number of copies of FP21 per cell is
not affected by the mutation, then there may be a maximum of
4 x 105 copies of FP21 per cell.
Evidence that FP21 is Endogenous to the Cytosol
Fucosylation has been shown to occur in the Golgi
apparatus in other organisms (Hirschberg and Snider, 1987;
Kornfeld and Kornfeld, 1985), so I assessed the possibility
that FP21 was a lumenal microsomal protein that leaked
during the P100 and S100 isolation procedure. To guard the
P100 from chemical lysis, the vesicles were prepared in a
cocktail of protease inhibitors that contained additional
inhibitors from those utilized in the standard fractionation
protocol, and which have been developed for the isolation of
various proteolitically sensitive proteins in Dictyostelium
(Goodloe-Holland & Luna, 1987; Stone, et al., 1987). To
address the possibility of mechanical disruption of the
vesicles, the P100 was disrupted by sonication and
recentrifuged at 100k x g for 1 h. Approximately 11% of the
radioactivity was released and the supernatant of this
centrifugation was analyzed by SDS-PAGE as described above
(figure 3-3). Although several radioactive peaks were
present in the supernatant of the second centrifugation, the
radioactivity profile was different from the S100 suggesting
that FP21 is not a protein released by disruption of the
vesicle fraction. Although radioactivity which comigrated

Figure 3-3. Comparison of S100 and releasable P100 components.
HL250 amoebae were metabolically labelled with 2 pCi/ml of
[3H]fucose, lysed, fractionated into an S100 and P100. The
P100 was sonicated and recentrifuged. After centrifugation,
the resulting supernatant was examined by slicing 7-20%
linear gradient SDS-PAGE into 2.2 mm and counting the gel
pieces (panel B). Included for comparison, is the profile
from the S100 radiolabelled species from the same
preparation run on the same gel (panel A). 50 pg of protein
were electrophoresed for the S100 and P100, respectively.
Arrow, migration of trypsin.

DPM DPM
63

64
with FP21 was observed in the PlOO-derived supernatant, this
material was a minority of the radioactivity released, and
probably reflected the general heterogeneity of the P100
vesicle contents. Less than 1% of the total cell
radioactivity that migrated at the 21 kD molecular weight
position was released from the P100 by sonication,
indicating that the remaining FP21 is recovered in the S100.
In a different approach, P100 from in vivo [3H]fucose
labelled cells was mixed with unlabelled post-nuclear
supernatant and recentrifuged (table 3-2). More than 98% of
the radioactivity sedimented with the P100, suggesting that
once associated with the P100, radioactivity is not lost,
unless vesicles are purposely disrupted, as in sonication.
The enzymes a-glucosidase-2, glucose-6-phosphatase, and
acid phosphatase have been used as markers of the
endoplasmic reticulum, Golgi apparatus and endoplasmic
reticulum, and lysosomes, respectively (Borts and Dimond,
1981; McMahon et al., 1977). To examine the distribution of
these enzyme markers in the high speed fractions, HL250
amoebae were harvested, homogenized, fractionated into an
S100 and P100 and assayed for activity of the different
marker enzymes. As seen in table 3-3, the majority of the
activity was recovered in the P100, suggesting minimal
contamination of the S100 by vesicles containing these
enzymes. Less than 7% of the total a-glucosidase-2 activity

65
Table 3-2. Radioactivity recovered in the second S100 after
different P100 treatments.
condition
% radioactivity recovered
untreated
sonicated
mixed*
2.2%
11.3%
1.4%
Cells were labelled in vivo by growing in 2 pCi/ml of
[3H]fucose in FM media for 3 days, lysed, and fractionated
into an S100 and P100. The P100 was then subjected to
different treatments and recentrifuged at 100k x g for 1 hr.
The S100 from this second centrifugation was analysed for
radioactivity. *P100 from metabolically labelled cells was
mixed with unlabelled post-nuclear supernatant from cells
grown in FM media and recentrifuged at 100k x g for 1 hr.

66
Table 3-3. Distribution of markers among S100 and P100
fractions.
a-glucosidase-2
acid phosphatase
glucose-6-phosphatase
S100
P100
6.5%
93.5%
12.8%
87.2%
n.d.
100%
HL250 amoebae grown in HL-5 were fractionated into S100 and
P100, and assayed for activity as described in Materials and
Methods. Activity expressed as percentage of total activity
detected in both fractions, n.d., not detectable.

67
was found in the S100; mannose-6-phosphatase was only
detectable in the P100. More than 87% of the activity of
the lysosomal enzyme acid phosphatase was detected in the
P100. In Dictyostelium, this enzyme has been shown to be a
soluble lumenal lysosomal protein (Dimond, et.el., 1981).
Further evidence that P100 vesicles are stable, closed
structures comes from earlier studies from the laboratory.
Prespore proteins SP75 and SP96 sediment at 100k x g unless
cells are sonicated (West and Erdos, 1988). These spore
coat proteins are contained in secretory vesicles (Erdos and
West, 1989; West and Erdos, 1988), which appear to be intact
since they are resistant to proteolysis by Proteinase K
unless they are treated with 0.1% Triton X-100 (West et al.,
1986; Q.H. Yang and C.M. West, unpublished data). In other
studies, N-acetyl glucosaminyltransferase activity was
inhibited by EDTA in the P100 fraction only when assayed in
the presence of detergent, suggesting that Golgi-like
vesicles in the P100 were closed (R.B. Mandell and C.M.
West, unpublished observations)
FP21 is Unrelated to Other Known Cytoplasmic Proteins
The possible relationship of FP21 with other known
Dictyostelium proteins was investigated. I examined the
reactivity of antisera raised against discoidin I and II
(Erdos and Whitaker, 1983) and against gp24 (Knecht et al.,
1987) for FP21. On Western blots, antiserum against gp24

68
recognized a band that migrated with a slower mobility than
metabolically labelled FP21 (not shown). The possibility of
FP21 being the lectin discoidin was examined, since it is
primarily present in the cytosol and exhibits weak affinity
for L-fucose (Erdos and Whitaker, 1983; Bartles and Frazier,
1980). Metabolically labelled FP21 from HL250 was
electrophoresed in a 15% SDS polyacrylamide gel, while a
replicate lane was blotted onto nitrocellulose paper and
immunoprobed using an anti-discoidin antiserum. The
antiserum recognized a band of higher MW than FP21 with
mobility slower than trypsin, reproducing results reported
by others where discoidin I was shown to have a slower
mobility than trypsin in 15% SDS polyacrylamide gels
(Kohnken and Berger, 1987). Migration of purified discoidin
in SDS-PAGE differed upon boiling of the sample, migrating
as a tetramer (ca. 100 kD) when samples were not boiled,
whereas metabolically labelled FP21 was found to migrate as
a discrete peak of radioactivity ahead of trypsin regardless
of boiling (Q.H. Yang and C.M. West, unpublished data).
FP21 could be partially purified by HPLC DEAE
chromatography. Metabolically labelled S100 was
fractionated on an anion exchange column and FP21 recovery
monitored by counting of SDS-PAGE slices. FP21 eluted at
0.5 M NH4Ac with an increase of 24-fold the specific
activity relative to the starting sample (data not shown).
Discoidin eluted earlier than FP21 from the HPLC DEAE column

69
and no radioactivity was found associated with discoidin,
suggesting it does not bind to discoidin during
purification. I conclude that FP21 is not related to
discoidin or any discoidin isoforms, or gp24, and does not
bind to any discoidin isoforms.
Oligosaccharide Studies
FP21 from Ax3 and HL250 yield similar size
glycopeptides after pronase digestion. Pronase digested
gel-purified FP21 that had been metabolically labelled from
Ax3 and HL250 were compared (figure 3-4). More than 50% of
the radioactivity from both sources eluted as a major peak
with a relative elution coefficient (Rev) of 0.42 and 0.43,
respectively (ca. 5.5 glucose units). In the case of FP21
derived from Ax3, the rest of the radioactivity eluted
earlier in the void volume and distributed into minor peaks.
The digestion products from HL250 FP21 yielded a major peak,
with the rest of the radioactivity eluting earlier, and less
than 10% eluting after the major peak, although some
variability was observed in different runs regarding the
minor peaks. Radioactivity eluting at an earlier position
than the major peak may be explained by incomplete
digestion. The minor amount of radioactivity that eluted at
a later position may be due to breakdown of the
oligosaccharide. These results indicate that the main
glycopeptides derived from Ax3 and HL250 have the same

Figure 3-4. Gel filtration chromatography of FP21 glycopeptides.
Ax3 and HL250 vegetative cells were metabolically labelled
with 2 pCi/ml of [3H]fucose. Cells were lysed, fractionated
into an S100 and P100, and FP21 isolated by SDS-PAGE and
electroelution from the S100. The samples were exhaustively
digested with pronase and analyzed by BioGel P-4 gel
filtration. Data obtained from one representative
experiment.
Panel A. Glycopeptides derived from Ax3 FP21, arrow
identifies major peak with a Rev of 0.42. Vo, 32; Vi, 142.
Panel B. Glycopeptides derived from HL250 FP21, arrow
identifies major peak with a Rev of 0.43. Vo, 26; Vi, 166.

71

72
sizes, and suggest that both strains form in vivo the same
oligosaccharide when grown in fucose-containing media.
Oligosaccharide in FP21 is Q-linked. An enzymatic and
a chemical approach were used to determine whether the
fucose-containing oligosaccharide in FP21 is N-linked or 0-
linked. I first tried the enzyme PNGase F. The minimum
requirement of PNGase F is a di-N-acetylchitobiose core
(Chu, 1986; Tarentino et al., 1985). This enzyme has a
broad specificity and can cleave most asparagine linked N-
glycans (including high mannose and complex multibranched
oligosaccharides) provided they are not located at the amino
or carboxy termini (Chu, 1986). Gel-purified FP21 from
metabolically labelled Ax3 was digested with PNGase F. As
shown in figure 3-5, the fucose label eluted in the void
volume, while control substrates were quantitatively
digested as determined by SDS-PAGE (not shown). FP21
trypsinized prior to digestion with PNGase F also eluted in
the void volume (not shown). The inability of PNGase F to
release radioactivity from FP21 suggested that either the
fucose-containing oligosaccharide was not N-linked or the
oligosaccharide was insensitive to the enzyme.
I then considered the possibility that the
oligosaccharide in FP21 was 0-linked. Metabolically
labelled gel purified FP21 was subjected to mild alkaline,
reducing conditions, to release intact 0-linked

Figure 3-5. Gel filtration chromatography of PNGase F digests.
Ax3 vegetative cells were metabolically labelled with 2
^/Ci/ml of [3H]fucose, an S100 was prepared, and FP21 was
isolated by SDS-PAGE and electroelution. FP21 digested with
PNGase F as described in Materials and Methods, and analyzed
by BioGel P-4 gel filtration. Nine drops fraction were
collected, instead of the usual 12 drops; [3H]mannose was
used to determine Vi.

DPM
74

75
oligosaccharides (Jl-elimination). Under the conditions
employed, N-linked oligosaccharides are insensitive to
chemical release (Biermann, 1988). More than 95% of the
radioactivity was released from Ax3 (Rev 0.48) and HL250
(Rev 0.49) in vivo labelled FP21 and it was resolved as one
peak with a size of 4.8 glucose unit (figure 3-6, panels A
and B). The oligosaccharide appeared to have been released
by fl-elimination and thus is concluded to be O-linked.
Consistent with the pronase digestion studies reported
above, Ax3 and HL250 yielded a similar size oligosaccharide,
supporting the idea that both strains produced the same
oligosaccharide. The slightly smaller size of the
oligosaccharide compared to the glycopeptide is consistent
with the notion that the glycopeptide consisted of one or
more amino acids and the oligosaccharide chain. If the
glycopeptide resulting from pronase digestion consisted of
the oligosaccharide attached to two or more amino acids, it
was possible that mild alkaline hydrolysis under reducing
conditions resulted in further hydrolysis of the remaining
polypeptide backbone yielding one amino acid and the
oligosaccharide. This would yield a smaller radioactive
species, with concomitant increase in Rev. To investigate
this possibility, I employed harsher chemical conditions.
In vivo labelled FP21 from Ax3 was gel purified and
subjected to strong alkaline hydrolysis in the presence of
sodium borohydride, which cleaves N- and O-linked

Figure 3-6. Gel filtration chromatography of FP21
oligosaccharides.
Vegetative cells were metabolically labelled with 2 ^Ci/ml
of [3H]fucose, an S100 was prepared, and FP21 was isolated
by SDS-PAGE and electroelution. Gel purified FP21 from Ax3
and HL250 were subjected to fi-elimination or strong alkaline
hydrolysis followed by fractionation by gel filtration
chromatography. Data obtained from one representative
experiment.
Panel A. ft-elimination of Ax3 FP21. Vo, 36; Vi, 141.
Panel
B.
fi-elimination of HL250
FP21. Vo,
30;
<
p.
140.
Panel
C.
Alkaline hydrolysis of
Ax3 FP21.
o
>
38;
Vi,

DPM DPM DPM
77
fraction

78
oligosaccharides (Zinn et al., 1978; Biermann, 1988). The
results of this reaction are seen in figure 3-6, panel C.
More than 75% of the radioactivity was released, eluting
with a Rev of 0.49. Most of the remainder of the
radioactivity eluted in the void volume. I expected to see
a change in Rev by strong alkaline hydrolysis compared to
mild alkaline hydrolysis if mild alkaline hydrolysis did not
release the oligosaccharide, but hydrolysed the protein.
The fact that similar results are obtained by mild and
strong conditions suggested that fi-elimination occurred to
release the oligosaccharide. Since mild alkaline hydrolysis
had been shown earlier not to cleave N-linked sugars
(Biermann, 1988), I conclude that the oligosaccharide in
FP21 is linked via an O-linkage. The released
oligosaccharide eluted as an asymmetrical peak, both by mild
and strong alkaline hydrolysis. These results suggest that
there is more than one type of oligosaccharide in FP21 that
differ slightly in size. These results also suggest that
the slight heterogeneity seen in the glycopeptide size may
reflect amino acid heterogeneity.
Comparison of qlycopeptides derived from vesicular and
cytosolic material. To further address the possibility of
FP21 arising by contamination of the S100 from the P100
fraction, glycopeptides derived from the Ax3 P100 were
compared to gel purified pronase digested FP21 from Ax3.
P100 derived glycopeptides were obtained in three different

79
manners. Metabolically labelled P100 was subjected to SDS
PAGE and the material that comigrated with FP21 was gel
purified, pronase digested, and fractionated by gel
filtration chromatography. Entire P100 from metabolically
labelled cells was digested in the presence or absence of
detergent, and analyzed by BioGel P-4 gel filtration. As
seen earlier in figure 3-4, panel A, the S100 digest eluted
mainly as a single peak with a Rev of 0.42; on the other
hand, the digest from the FP21-comigrating P100 material,
fractionated as a major peak of 0.50 Rev (figure 3-7, panel
A). An analysis of glycopeptides from the entire P100
digestion showed a different elution profile from the FP21
digestion (figure 3-7, panel B). Note than in this
chromatograph there is a minor peak of 0.51 Rev, consistent
with the idea that the major peak seen in panel A is a minor
component of the entire P100 glycopeptide repertoire. Also
approximately 40% of the radioactivity eluted with the void
volume, suggesting it may be resistant to pronase digestion.
Digestion of the entire P100 fraction was carried out in the
presence of Triton X-100, to determine if solubilization of
the sample yielded a different digestion profile, by
facilitating accessibility of the enzyme to the substrates
(figure 3-7, panel C). The overall profile is similar, with
more than 30% of the radioactivity eluting at the void
volume, and a peak with a Rev of 0.51 is still a minor
component of the glycopeptides released. West et al. (1986)

Figure 3-7. Gel filtration chromatography of P100 glycopeptides.
Vegetative Ax3 cells were metabolically labelled with 2
/Ci/ml of [3H]fucose. A P100 was prepared and the 21 kD MW
material that comigrated with FP21 on SDS-PAGE was
electroeluted and pronase digested. In another assay,
entire P100 was pronase digested in the absence or presence
of 0.1% Triton X-100. After pronase digestion, samples were
subjected to BioGel P-4 fractionation. Arrow identifies the
position with a Rev of 0.51.
Panel A. Glycopeptides from 21 kD MW P100-derived material.
Vo, 34; Vi, 162.
Panel B. Glycopeptides from entire pronase-digested P100.
Vo, 36; Vi, 142.
Panel C. Glycopeptides from entire P100 digested with
pronase in the presence of Triton X-100. Vo, 32; Vi, 146.

81
fraction
fraction
fraction

82
have reported the existence of pronase-resistant material in
the particulate fraction of vegetative cells, so it seems
possible that the pronase-resistant material eluting in the
void volume is, or is related to, the smear previously
described, although SDS-PAGE analysis will be needed to
confirm this supposition.
Fucose is Covalently Bound to FP21
The fact that the radioactivity in FP21 was not
released by boiling in SDS/ fl-mercaptoethanol, nor after
boiling in SDS under reducing conditions followed by SDS-
PAGE, suggested that 3H was covalently bound to in vivo
labelled FP21. Nevertheless, there have been reports of
covalent-bound enzymatic intermediates in the literature
(Scrimgeour, 1977). Although ES (enzyme-substrate)
intermediates are very reactive and usually cannot be
isolated without some sort of stabilization or chemical
trapping technique, the possibility that FP21 is really a
cytosolic fucosyltransferase that binds GDP-fucose or other
fucose metabolites covalently was considered. The Rev of
the radioactivity released by mild alkaline hydrolysis
suggests that it is not related to GDP-fucose or fucose
because it elutes with a different Rev than GDP-fucose or
fucose. Additional evidence that 3H is present as fucose
was presented in Chapter II, where it was shown that more

83
than 95% of the macromolecular-associated radioactivity from
metabolically labelled cells migrated as authentic fucose.
FP21 is Present in Migrating Slug Stage Cells
To investigate whether FP21 was present also in
developing cells, HL250 amoebae were plated for development
on nuclepore filters and 9 hours after plating the filters
were lifted and cells were metabolically labelled by
placement on 100 pCi of [3H]fucose. After 5 hours of
exposure to [3H]fucose, filters were lifted again, carefully
washed from 3H label, and cells were allowed to continue
development for two more hours. Cells were then harvested,
disaggregated, and fractionated into an S100 and P100
fractions, run in an SDS-PAGE, and the gel cut and the
pieces counted. As shown in figure 3-8, a fucosylated
macromolecule is preferentially fucosylated in the cytosol
and it has the same mobility in SDS-PAGE as FP21. Thus, it
seems that FP21 fucosylation is not restricted to the growth
phase of Dictyostelium.
Discussion
In this chapter I have presented evidence for the
existence of a fucosylated single molecular weight species,
which has been termed FP21 based on its mobility as
determined by SDS-PAGE. This protein cofractionates with

Figure 3-8. Incorporation of [3H]fucose into macromolecular
species of slug stage cells.
HL250 cells were harvested, plated on filters for
development, and exposed to [3H]fucose for 5 h as described
in the text. Cells were then collected, disaggregated,
filter-lysed, and S100 and P100 fractions were prepared and
analyzed by 7-20% linear gradient SDS-PAGE. 70 jug and 24.6
fjq of protein were electrophoresed for the S100 and P100,
respectively. Open circles, P100; closed circles, S100;
arrow, migration position of trypsin.

85
P100 (DPM)

86
the high speed supernatant, S100, being the major
fucosylated species in this fraction. The recoverability of
FP21 after TCA precipitation, HPLC anion exchange
chromatography, boiling in SDS/fi-mercaptoethanol, and
methanol/acetic acid fixation of the gels suggested a
covalent nature for the association of radioactivity. It is
unlikely that FP21 is a cytosolic fucosyltransferase that
binds GDP-[3H]fucose or another fucose metabolite
covalently, since FP21 does not copurify with the cytosolic
fucosyltransferase (C.M. West, unpublished results).
Employing enzymatic and chemical analysis, the
oligosaccharide in FP21 was examined and characterized as a
small (4.8 glucose unit) oligosaccharide. The
oligosaccharide appeared to be O-linked based on its
insensitivity to PNGase F and the releasability from FP21
under alkaline, reducing conditions. Ax3 and HL250 produced
FP21-derived glycopeptides and oligosaccharides of similar
size, suggesting both strains carry out similar
modifications in vivo, despite the starvation for fucose in
HL250. Thus, it appears there are no competing reactions
for the unfucosylated oligosaccharide, unlike the case for
outer fucose in N-linked glycans from mammals (Paulson et
al., 1978). Based on antiserum specificity, molecular
weight, and/or fractionation by HPLC anion exchange
chromatography, FP21 was shown to be a protein unrelated to
discoidin or gp24. Finally, a 21 kD fucoprotein was present

87
in the cytosol of developing cells, suggesting FP21 was not
limited to the vegetative stage in Dictyostelium, and was
fucosylated during development.
I believe that FP21 is recovered in the S100 because it
resides in the cytosol in living cells, and not as default
location from ruptured vesicles, for several reasons. The
S100 fraction was shown to be equivalent to the cytosol and
essentially devoid of organellar markers. S100 and P100
fractions from [3H]fucose metabolically labelled cells
exhibited a different radioactive profile by SDS-PAGE.
Sonication of the P100 fraction failed to release FP21 into
the supernatant. FP21 appeared to be endogenous to the
cytosol and not derived from organellar vesicles, because
control experiments suggested there was no generalized
breakage of vesicles during the preparation of the cytosolic
fraction. In addition, as a control for contamination from
P100 material, glycopeptides derived from comigrating 21 kD
MW species from the P100 fraction were compared with FP21
glycopeptides. The fucopeptide in FP21 (ca. 5.5 glucose
unit) does not seem to be a product of vesicular
fucosylation, since it is not shared by macromolecules of 21
kD MW in the P100 fraction which yielded a major peak with a
different Rev (ca. 4.3 glucose unit) than the one derived
from FP21. When the entire P100 fraction was subjected to
Pronase digestion, a heterogeneous mixture of fucosylated

88
species (in accordance with Tsurchin, et.al. 1989) was
obtained with sizes unlike that of FP21 glycopeptides.
The designation of compartmentalization of a protein as
cytosolic is difficult since the cytosol is the site of
localization after disruption of organellar vesicles. This
task is complicated in the case of glycoproteins and
glycosylation enzymes, which are usually described as
components of the secretory pathway. However, some
glycoproteins have been identified as cytosolic, and
generally accepted as such (Hart et al., 1989a; Hart et al.,
1989b). My studies report the existence of a fucosylated
cytosolic protein. However, these results do not exclude
the possibility of FP21 being present in other locations
topologically continuous with the cytosol, such as the
nucleus, or being synthesized elsewhere and transported.
Due to its small size, FP21 could, in theory, be able to
diffuse freely into the nucleus.
Since glycoprotein fucosylation has been shown to take
place in enclosed organelles of the secretory pathway
(Hirschberg and Snider, 1987), the identification of FP21 in
the cytosol raises the question of where in the cell is
fucosylation of FP21 taking place. One scenario would have
FP21 being fucosylated in organellar vesicles (presumably
the Golgi apparatus) and subsequently transported to the
cytosol, while another would postulate the presence of a
fucosyltransferase that localized in the cytosol with FP21.

89
I consider these possible alternatives in the following
chapter and present evidence for the presence of a
fucosyltransferase in the cytosol responsible for FP21
fucosylation.

CHAPTER IV
EVIDENCE FOR A CYTOSOLIC FUCOSYLTRANSFERASE
Introduction
In the preceding chapter, I identified a fucosylated
protein in the cytosol, FP21. The presence of FP21 in the
cytosol challenges the prevailing belief that fucoproteins
are restricted to the cell surface and lumenal compartments
of the cell. Even though in the past three decades evidence
has been accumulating on the presence of glycoproteins and
fucoproteins in non-lumenal locations (Hart et al., 1989a;
Hart et al., 1989b), to my knowledge, no one has shown the
existence of a fucosyltransferase in the cytosol. One
possibility is that fucosylation is restricted to the
microsomes, and cytosolic fucoproteins are
posttranslationally transported back across the membrane to
the cytosol.
However, even though there is no previous evidence for
cytosolic fucosylation there is precedent for glycosylation
in the cytosol. Studies on the biosynthesis of nuclear pore
proteins bearing O-GlcNAc suggested that the sugar was added
to the proteins within 5 min of their synthesis and before
they became associated with membranes (Davis and Blobel,
1987). These data suggested that the activity responsible
90

91
for the addition of O-GlcNAc was in the same topological
compartment where translation takes place, the cytosol.
Thus it is possible that fucosylation may take place in the
cytosol, but it has escaped detection by previous
investigators for a variety of reasons. One of the
difficulties in assaying cytosolic enzymes is that the
endogenous acceptors for the enzymes may be present in low
quantities in the cell, complicating purification of large
amounts for use as substrates. If acceptors are already
fucosylated, in vitro assays that utilize endogenous
acceptors would not detect enzymatic activity. Hart and
coworkers have circumvented this problem with the use of
synthetic peptides with a sequence based on O-GlcNAc
glycosylation sites (Hart et al., 1989b). They have
identified an enzymatic activity capable of O-linked GlcNAc
transfer in rat hepatocytes that was recovered in both the
soluble and membrane fractions (Haltiwanger et al., 1990).
The membrane-associated activity was releasable by high salt
treatment and oriented towards the cytosol, not the lumen of
the vesicles (Haltiwanger et al., 1990).
With the help of the conditional fucosylation mutant
HL250, I have addressed the existence of a fucosylation
pathway in the cytosol. Total protein in HL250 is
underfucosylated relative to the normal strain, so it was
reasoned that it would be a useful strain to assay
fucosylation in vitro due to the availability of

92
macromolecular acceptors. In this chapter I present
evidence for a fucosyltransferase that partitions with FP21
in the cytosol. The fucosyltransferase was distinguished
from vesicular fucosyltransferase activity by several
criteria, and was characterized using hydrophobic synthetic
analogs.
Materials and Methods
Materials
GDP-[U-3H] fucose (6.6 Ci/mmol) and GDP-[U-14C ] fucose
(250 mCi/mmol) were from New England Nuclear (more than 90%
of the radiolabel was in the form of the fi anomer, as
indicated by the manufacturer). Reagent grade KC1, MnCl2,
CaCl2, BaCl2 and MgCl2 were from Mallinckrodt; GDP-fl-fucose
from Biocarb (stored frozen as a concentrated stock); GDP-a-
fucose (stored frozen as a concentrated stock), Tween-20,
CoCl2, Dowex-2 (2x8, minus 400, chloride form), Triton X-
100, chymostatin, pepstatin, NBZ-phenylalanine, bovine serum
albumin (BSA), and all p-nitro-phenyl acceptors were from
Sigma; Na2EDTA, FeCl3, formic acid, and trichloroacetic acid
were from Fisher. Cations and Na2EDTA were stored as
concentrated 500 mM solutions at 4. Hydrophobic synthetic
acceptors were generously provided by Monica Palcic. Making
and regeneration of Dowex-2 formate form column was as
described for Dowex-1 in Materials and Methods, Chapter II.
The concentrations of Triton X-100 and NP-40 are expressed

93
as volume/volume (v/v), all others are expressed as
weight/volume (w/v) unless specified otherwise.
Strains and Conditions of Growth
Dictyostelium discoideum strains Ax3 (from S. Free) and
HL250 (from W.F. Loomis) were grown on HL-5, a complete
medium that contains glucose, yeast extract, and proteose
peptone (Loomis, 1971). Ax3 is the normal strain and HL250
is a mutant obtained from Ax3 by N-methyl-N'-nitro-N-
nitrosoguanidine mutagenesis (Loomis, 1987). HL250 lacks
the enzyme activity that converts GDP-mannose into GDP-
fucose which results in a lack of cell fucose (Gonzlez-
Yanes et al., 1989; Chapter II). In all experiments cells
were collected at the logarithmic growth phase, with a cell
density of 1 to 9 x 106 cells/ml.
Cell Lysis and Fractionation
Vegetative and slug stage cells were fractionated into
S100 and P100 fractions as described in Materials and
Methods, Chapter III.
Fucosyltransferase Assay
Fucosyltransferase activity was assayed immediately
after obtaining the S100 and P100 fractions. Fractions were
found to be sensitive to freezing and thawing, and up to 70%
of the fucosyltransferase activity could be lost. The

94
standard fucosyltransferase assay contains 30 pi of extract
(100-350 pg of protein), 0.35 pM GDP-fi-[ 14C]-fucose, 5 mM
MgCl2, 0.25 mM NaF, and 5 mM ATP in 50 mM MES (titrated with
NaOH to a pH of 7.4) in a 50 pi volume incubated at 30 for
the specified amount of time. Endogenous macromolecules
were used as acceptors. To terminate the assay, 1 ml of
ice-cold 15% TCA was added to each sample along with 50 pg
BSA (to serve as carrier protein) and the precipitate
collected on 2.4 cm GF/C glass filters by vacuum filtration,
washed with 10 ml 10% ice-cold TCA, 10 ml acetone, and
counted after air-drying inside the vials for approximately
30 min using 10 ml of Bio-HP LC scintillation fluid
(Fisher). Background was subtracted from experimental
values, and was determined as the amount of TCA-precipitable
radioactivity at time zero; it was usually between 20-40
dpm. When indicated, the disodium EDTA salt was used. In
preliminary trials, 5 mM Mg++ was found to support maximal
activity, and this concentration was used in all the assays
unless indicated. GDP-fucose had been shown previously to
be only slightly decomposed under similar conditions (Nez
and Barker, 1976). Nevertheless, the extent of GDP-
[14C]fucose hydrolysis during the fucosyltransferase assay
was examined by carrying out the reaction for two hours and
separating products on a Dowex-2 formate column by
sequentially eluting with 5 ml of water, 3 M formic acid and
15 M formic acid as described (Sommers and Hirschberg,

95
1982). After 2 hours of incubation less than 20% of the
initial GDP-[ 14C ] fucose had been hydrolyzed. Incorporation
of 14C did not exceed 30% of the initial radioactivity on
any given experiment. For analysis by gel electrophoresis
of the endogenous acceptors of the in vitro
fucosyltransferase activity, the reaction was terminated by
3 min boiling in sample buffer. For the determination of pH
optima experiments, assays were buffered using concentrated
solutions of MES previously adjusted to different pH values
with either HC1 or NaOH. The final pH value of each
reaction was determined on an equivalent reaction mixture
lOOx the volume, without GDP-[ 14C]fucose. Where indicated,
the S100 was desalted on a 0.85 x 13 cm on a BioRad BioGel
P-2 column (200-400 mesh) equilibrated with 50 mM MES, pH
7.4, titrated with NaOH, at 4 with a flow rate of 0.5
ml/min. If supplied, synthetic acceptors and/or FP21 were
previously dried down onto the bottom of the assay tubes in
a vacuum centrifuge. Acceptors were subsequently
resuspended in water or in the reaction mixture. Results
are expressed as average of two determinations (variations
in the duplicates did not exceed 15% of the average value)
or average of three measurements + standard error of the
mean (s.e.m.). Calculations of K and V were done by the
Hanes single reciprocal plot ([S]/v vs. [S]) as discussed by
Henderson (1985).

96
Incorporation into hydrophobic synthetic acceptors was
determined as described by Palcic et al. (1988). The assay
was carried out as for endogenous acceptors, but GDP-
[3H]fucose was used instead of GDP-[ 14C ] f ucose and the
reaction was terminated by the addition of 1 ml ice-cold
water. The reaction mixture was loaded onto a C18 SepPak
column (Waters) under vacuum, and eluted with 6 successive 5
ml aliquots of water, and four 5 ml aliquots of methanol.
Eluates were counted by addition of 15 ml of ScintiVerse LC
(Fisher). In initial trials I determined that the
radioactivity eluted in the first methanol fraction, so in
subsequent experiments only the first methanol fraction was
used for determination of radioactivity incorporated. All
extracts were assayed in the absence of exogenous acceptor
and this value (usually about 20% of the dpm incorporated)
subtracted from experimental value to determine substrate-
dependent incorporation.
Purification of FP21
FP21 was purified in the following manner for
preparations which were to be added back to cytosolic
fractions to measure fucosylation acceptor activity.
Starting with 8 x 1010 cells, an S100 cytosolic fraction was
prepared from the mutant HL250. An aliquot of the fraction
(approximately 0.7% of the total volume) was incubated with
GDP-[ 14C]fucose and allowed to fucosylate FP21 with

97
[14C]fucose. Incorporation into FP21 was confirmed by
electrophoresing an aliquot and counting of SDS-PAGE slices.
The radiolabelled aliquot was mixed with the rest of the
unlabelled preparation and subjected to (NH4)2S04
fractionation. The 70-80% cut was dissolved in and dialyzed
against 100 mM NH4Ac, applied to a 14 ml bed of the strong
anion exchanger A25-QAE-Sephadex, and eluted with an
ascending gradient up to 1.5 M NH4Ac. [UC]FP21 eluted at
input buffer concentration of 0.49 M. This preparation was
then concentrated and desalted on Centricon and/or
Centriprep cartridges with nominal 10 kD MW cutoffs, and
then applied to an HPLC gel filtration column (8 x 300 mm
Toya Soda TSK GW-300) equilibrated in 100 mM NH4Ac, with a
flow rate of 0.5 ml/min. Sample was clarified by
centrifugation at 10k x g for 10 min prior to injection.
Radioactivity from the concentrated QAE-Sephadex eluate
eluted between the 14 kD and 29 kD MW standards. Fractions
were analyzed by SDS-PAGE using 15% polyacrylamide gels and
counting of the gel slices. For addition of purified FP21
to cell extracts, HPLC gel filtration fractions were brought
to dryness in a vacuum centrifuge, redissolved in dH20, and
brought to dryness again, in the 1.5 ml microcentrifuge tube
that was going to be used for the assay.

98
fi-elimination of In Vitro Labelled Acceptor
S100 extracts from HL250 were fucosylated in vitro. To
corroborate that I obtained 21 kD MW fucosylated product
from the in vitro reaction, l/25th of the reaction was
terminated by 3 min boiling in sample buffer and analyzed by
SDS-PAGE. The remainder of the sample (containing
approximately 104 dpm) was stored at -80 until ready to
use. The reaction mixture was centrifuged for approximately
2 h in a centricon filter to reduce unused GDP-[14C]fucose.
After concentrating the volume to 200 pi, fi-elimination was
carried out as described in Materials and Methods, Chapter
III.
PNGase F Digestion of In Vitro Labelled FP21
FP21 was fucosylated in vitro as described above for fi-
elimination. After the volume was concentrated, PNGase F
digestion was carried out as described in Materials and
Methods, Chapter III.
Results
Cytosolic Fucosyltransferase Activity
The presence of FP21 in the cytosol suggested that a
fucosyltransferase might also be located there. To
investigate this possibility, HL250 cells were fractionated
to yield cytosolic supernatant (S100) and organelle (P100)
fractions. The fractions were analyzed for their ability to

99
transfer [ 14C ] from GDP-[ 14C ] fucose into TCA precipitable
endogenous material. As shown in figure 4-1, the cytosolic
activity was dependent on time and protein content. The
cytosolic fucosyltransferase activity had the properties of
being enzyme-mediated. Table 4-1 shows the effect of
boiling, denaturants and temperature on the cytosolic
fucosyltransferase activity. While the non-ionic detergent
Triton X-100 inhibited all activity, Tween-20 was only
slightly inhibitory. 30 was the optimal temperature of
those tested (22, 30, and 37). Consistent with an
enzyme-mediated process, only unlabelled GDP-fi-fucose was
able to inhibit incorporation of radioactivity proportionate
to its relative concentration (GDP-a-fucose was without
effect), demonstrating the stereospecificity of the enzyme
(lower section of table 4-1). It also implies a
fucosyltransferase that catalyses an alpha-fucosyl linkage
is being assayed.
Earlier I explored the possibility of FP21 arising by
contamination from the vesicular fraction. The same
question was asked about the fucosyltransferase activity in
the cytosol, since known fucosyltransferases are Golgi
enzymes and activity was detectable in the P100 (see next
section). Hence, I tried to deplete the S100 of
fucosyltransferase activity by centrifuging at 170k x g for
2.5 hours (instead of 1 hr at 100k x g). This step was used
to sediment any population of small or low density vesicles

Figure 4-1. Fucosylation of endogenous acceptors by S100
fraction.
Vegetative HL250 cells were harvested, homogenized, and an
S100 obtained as described in detail in Materials and
Methods. The indicated amount of S100 protein was incubated
in the presence of 0.36 /M GDP-[14C]fucose, 5 mM MgCl2, 5 mM
ATP, 0.25 mM NaF, in 50 mM MES, pH 7.4 for the indicated
amount of time. Fucose incorporation was calculated from
the amount of TCA-precipitable [14C]radioactivity. Results
expressed as the mean of three determinations s.e.m.
Panel A. Effect of time on fucosylation; 159 fug protein.
Panel B. Effect of protein concentration on fucosylation;
30 min assay.

pmol fucose/ 30 min
o no co cn
time (min)
pmol fucose/ mg protein
101

Table 4-1. Effect of different treatments on the cytosolic fucosyltransferase activity.
pmol fucose/mq/30 min
condition
GDP-r Clfucose (pM)
control
experimental
relative activity
no cation
0.36
17.3
0.97
0.06
5 mM EDTA
1.32
9.23
0.06
0.01
boiled
1.32
9.23
<0.01
<0.01
0.15% Triton X-100
0.36
18.5
0.22
0.02
0.1% Tween-20
0.36
16.6
10.5
0.63
10% ethanol
6
54.4
3.00
0.06
22
6
31.2
22.8
0.73
37
6
35.3
16.1
0.46
6 pM GDP-a-fucose
0.36
20.8
21.3
1.02
6 pM GDP-fi-fucose
0.36
20.8
1.09
0.05
HL250 vegetative cells were harvested, filter-lysed, and fractionated into an S100 and
P100 fractions. The S100 was assayed for fucosyltransferase activity in the presence
of 100-300 pg protein, the noted concentration of GDP-[ 14C ] fucose, 5 mM ATP, 0.25 mM NaF,
5 mM MgCl2, in 50 mM MES, pH 7.4, for 30 min at 30 (standard conditions), unless other
conditions are specified. Additives were added at the noted concentration; when EDTA
was present divalent cations were omitted from the reaction mixture; the experiments
performed at 22 contained 10 mM MgCl2 and 10 mM MnCl ; those at 37 were supplemented
with 10 mM MgCl2. The results are a compilation of different experiments carried out
at the given GDP-[ C]fucose concentrations. Relative activity refers to the activity
exhibited under experimental conditions compared to control conditions. Fucose
incorporation into endogenous acceptors was calculated from the amount of TCA-
precipitable [14C]radioactivity. Results are the average of two determinations.
102

103
that were not sedimented before (if existent) and that could
have contained fucosyltransferase activity. To preserve the
intactness of the P100 vesicles, additional protease
inhibitors (Goodloe-Holland and Luna, 1987) from those
routinely used, were utilized during cell fractionation.
After these measures activity was still recovered in the
cytosol at similar levels (table 4-2). Additional evidence
supporting the notion that the vesicles in the P100 are not
damaged during filter lysis and centrifugation is presented
in Chapter III in the section of the origin of FP21. Taken
together, these results suggest that vesicles were not
measurably damaged during the isolation procedure;
therefore, the fucosyltransferase activity is probably
endogenous to the cytosol.
Comparison Between the P100 and S100 Fucosyltransferase
Activities
In order to compare the S100 and P100 activities, I
examined the identity of endogenous acceptors and the
effects of divalent cations, pH, and varying GDP-fucose
concentration on both fucosyltransferase activities. The
criteria of differential behavior has previously been used
to differentiate glycosyltransferases, since it is assumed
that under similar conditions, enzymes should behave in a
similar fashion (Campbell and Stanley, 1984; Galland et al.,
1988). Activities were measured in the presence of
detergent to circumvent any potential problem in substrate

104
Table 4-2. Failure to sediment S100 fucosyltransferase
activity.
S100 P100
100k
x g,
lh
943
444
170k
x g,
2.5h
1131
484
Cells were lysed, centrifuged, and fractionated into S100 and
P100. In this experiment lysis buffer (described in Materials
and Methods, pH 8.0) was supplemented with 1 mM chymostatin,
5 jjg/ml pepstatin, and 2 mM NBZ-phenylalanine. Fractions were
assayed immediately for C] incorporation from GDP-[ CJfucose
and expressed as total dpm incorporated in 30 min into TCA
insoluble endogenous acceptors. Reaction mixtures contained
0.36 nM GDP-fucose, 5 mM MgCl2, 240 pg protein, and were
incubated for 30 min. Results are expressed as average of two
determinations.

105
or cation accessibility. Tween-20 at a concentration of
0.1% was chosen because it did not inhibit considerably the
activity in the S100 (table 4-1).
I first examined the profile of in vitro endogenous
acceptors by SDS-PAGE analysis. Standard S100 and P100
fractions from HL250 cells were isolated and added to
fucosyltransferase reaction mixtures. After 90 min of
incubation, reactions were boiled in sample electrophoresis
buffer, subjected to SDS-PAGE, and the gels cut and counted.
The profile of radiolabel incorporation in vitro by the S100
was similar to that observed in metabolic labelling
experiments, with more than 70% of the radioactivity
migrating as one discrete peak with a MW of 21 kD (compare
figure 4-2, panel A with figure 3-1). Incorporation into 21
kD MW material by the S100 fraction varied from 70-95% in
different experiments, with the remainder of the
radioactivity migrating near the dye front or top of the
gel. In the P100 fraction, radioactivity distributed in two
peaks, similar to what was seen in metabolically labelled
P100. To determine if in vitro fucosylated protein had the
same apparent MW as metabolically labelled FP21, Ax3 gel-
purified FP21 was mixed with in vitro fucosylated HL250
S100, as described in the figure legend (figure 4-2, panel
B). The migration of in vitro fucosylated material
coincided with that of metabolically labelled FP21. I

Figure 4-2. SDS-PAGE profile of endogenous acceptors fucosylated
in vitro.
Panel A. SDS-PAGE profile of S100 and P100 endogenous
acceptors fucosylated in vitro. HL250 cells in logarithmic
growth phase were harvested, homogenized, and an S100 and
P100 prepared. Both fractions were fucosylated in vitro in
the presence of 4.4 pM and 8.8 pM of GDP-[ 14C]fucose for the
S100 and the P100, respectively, 5 mM MgCl for 90 min.
Fucosyltransferase reactions were stopped by boiling in
SDS/fl-mercaptoethanol sample buffer and resolved by 7-20%
linear gradient SDS-PAGE; the gel was sliced into 2.2 mm
pieces and counted. Electrophoresis was from left to right.
138 pg and 150 pg of protein were electrophoresed for the
S100 and P100, respectively. Open circles, P100 endogenous
acceptor species; closed circles, S100 endogenous acceptor
species; arrow, migration of trypsin.
Panel B. Comigration of in vitro fucosylated FP21 with
metabolically labelled FP21 on SDS-PAGE. Ax3 vegetative
cells grown in [3H]fucose were harvested, lysed, and
fractionated into an S100 and P100. The S100 was subjected
to SDS-PAGE and Ax3 FP21 was gel purified and electroeluted.
Independently, HL250 amoebae were harvested, homogenized,
fractionated, and the S100 obtained from the fractionation
fucosylated in vitro in the presence of 0.36 pM GDP-
[14C] fucose, 5 mM MgCl, 140 pg protein, for 30 min.
Reaction was stopped by mixing with gel purified Ax3
[3H]FP21 followed by boiling in SDS/B-mercaptoethanol.
Samples were coelectrophoresed in a 15% SDS-polyacrylamide
gel, and the gel cut into 0.5 cm slices and counted. Open
circles, Ax3 [3H] metabolically labelled gel purified FP21;
closed circles, [14C] in vitro labelled mutant S100 extract.

DPM (P100)
107
DPM (S100)

108
interpret these results as an indication that the activity
that fucosylates FP21 in vivo is being assayed in vitro.
FP21 was fucosylated in vitro by incubating HL250 S100
fractions in the presence of GDP-[ 14C ] fucose, and the
oligosaccharide fucosylation in vitro was examined as before
for in vivo fucosylated FP21. In vitro fucosylated FP21 was
digested with PNGase F (not shown) or subjected to mild
alkaline hydrolysis and analyzed by gel filtration (figure
4-3, panel A). As was the case with metabolically labelled
FP21, PNGase F failed to release radioactivity and in vitro
labelled FP21 digested with PNGase F eluted in the void
volume. The digestion of the control substrates, fetuin and
ribonuclease B, was confirmed by SDS-PAGE. On the other
hand, approximately 20% of the radioactivity eluted with a
Rev of 0.50, the elution position of the metabolically
fucosylated oligosaccharide produced by Ax3 (reproduced for
comparison in figure 4-3, panel B, from figure 3-6, panel
A). The remainder of the radioactivity fractionated as
material of larger size. These results suggested that the
in vitro fucosylated oligosaccharide in FP21 was also O-
linked. The reasons for the discrepancies in size between
in vivo and in vitro fucosylated oligosaccharides are not
known, but may be due to incomplete release of the
oligosaccharide, accompanied by partial hydrolysis of the
polypeptide. Alternatively, it may indicate the presence of
oligosaccharides of various sizes.

Figure 4-3. BioGel P-4 gel filtration chromatography of in vitro
labelled FP21 oligosaccharide.
HL250 S100 extracts were incubated in vitro in the presence
of GDP-[ 14C]fucose, desalted, subjected to fi-elimination (as
described in Materials and Methods), and analyzed by gel
filtration. For comparison, the profile resulting from fi-
elimination of in vivo labelled Ax3 gel purified FP21 is
presented in panel B (was panel A in figure 3-6, Chapter
III) .
Panel A. B-elimination of in vitro fucosylated FP21; arrow
identifies peak with Rev of 0.50. Vo, 38; Vi, 136.
Panel B. fi-elimination of Ax3 FP21. Vo, 36; Vi, 141.

110

Ill
To compare the S100 and P100 fucosyltransferase
activities, HL250 vegetative cells were harvested,
homogenized, and fractionated into an S100 and P100.
Fractions were assayed for fucosyltransferase activity under
different conditions in the presence of Tween-20 (table 4-
3). The fractions differed in that the S100
fucosyltransferase activity was approximately threefold more
efficient (on a per protein basis) than the P100 under
standard conditions (which contained 0.36 pM GDP-fucose and
5 mM MgCl2, see Materials and Methods). A major difference
between the bulk activities was their sensitivity to the
presence of divalent cations. In the absence of any added
cation, the P100 retains more than one fourth the activity
exhibited in the presence of Mg++, while the activity in the
S100 was almost negligible. The presence of EDTA does not
inhibit further the activity in the P100. One
interpretation is that the fucosyltransferase activity in
the S100 is dependent on added Mg++, while the activity in
the P100 is present in the absence and presence of Mg++,
being stimulated by the cation. An alternative explanation,
is that there are multiple enzymes in the P100, which differ
in their requirements for divalent cations.
Sensitivity to pH is a feature exhibited by enzymes,
including fucosyltransferases (Foster, et al. 1991; Kumazaki
and Yoshida, 1984). HL250 vegetative cells were harvested,
lysed, and fractionated into S100 and P100 fractions. Both

112
Table 4-3. Comparison between the S100 and P100
fucosyltransferase activities.
condition
pmol fucose/mq protein/45 min
S100 P100
MgCl2*
no cation
EDTA
6.95+2.60
0.18+0.08
0.05+0.03
1.71+0.45
0.68+0.10
0.64+0.10
HL250 amoebae were harvested, filter-lysed, and centrifuged to
prepare S100 and P100 fractions. Fucosyltransferase reaction
mixtures contained 0.36 pM GDP-[14C]fucose, 0.1% Tween-20, and
135 pg or 138 pg of protein from the S100 and P100,
respectively, in the absence of added divalent cations. MgCl2
and EDTA were present at 5 mM. *this is the standard assay,
as described in Materials and Methods. Results are the
average of three measurements + s.e.m.

113
fractions were assayed for fucosyltransferase activity at
various pH values as described in Materials and Methods.
The pH profiles of the S100 and P100 fucosyltransferase
activities in the presence of Tween-20 are shown in figure
4-4. Activity was maximal for the S100 from pH 6.8 to 7.8
and for the P100 from pH 6.4 to 7.8. At pH 9.6 the activity
in the S100 was inhibited more than 20-fold compared to
maximum (p<0.05), whereas the activity in the P100 was only
inhibited threefold (p<0.05). The pH-dependent activity
profiles were not affected by the exclusion of Tween-20 (not
shown). Thus, while the general profile is similar for the
activities from the S100 and P100, the cytosolic activity
was more sensitive to alkaline pH than the P100 activity.
The dependence of S100 and P100 fucosyltransferase
activities on the GDP-fucose concentration in the presence
of Tween-20 was also studied. S100 and P100 fractions were
prepared from HL250 vegetative cells and assayed for
fucosyltransferase activity under standard conditions, at
increasing concentrations of GDP-fucose in the presence of
0.1% Tween 20 (figure 4-5). The apparent Km for GDP-fucose
was 1.7 pM and 38.2 pM for the S100 and P100, respectively.
The apparent Vmax for the S100 was 42.7 pmol fuc/ mg protein/
30 min and 122 pmol fuc/ mg protein/ 30 min for the P100.
As evidenced by the lower apparent Km (22-fold lower), the
affinity of the S100 fucosyltransferase for GDP-fucose was
higher than the one from the P100 activity. This accounts

Figure 4-4. Effect of pH on S100 and P100 fucosyltransferase
activities in the presence of Tween-20.
S100 and P100 fractions prepared from vegetative HL250 cells
were assayed for fucosyltransferase activities in the
presence of Tween-20 at different pH values (pH values
determined as described in materials and methods). GDP-
[14C]fucose concentration was 0.36 pM; MgCl2, 5 mM; Tween-20,
0.1%. The assay was carried out for 30 min, and 150 pg and
228 jug of protein were supplied from the S100 and P100,
respectively. Fucose incorporation was calculated from the
amount of TCA-precipitable [14C]radioactivity. Results
expressed as the mean of three determinations s.e.m.
Panel A. Effect of pH on S100 fucosyltransferase activity
in the presence of Tween-20.
Panel B. Effect of pH on P100 fucosyltransferase activity
in the presence of Tween-20.

pmol fucose/mg protein/30 min
pmol fucoae/mg protein/30 min
O
115

Figure 4-5. Effect of GDP-fucose concentration on S100 and P100 fucosyltransferase
activities in the presence of Tween-20.
HL250 amoebae were harvested, homogenized, fractionated into an S100 and P100 and
fucosylated in vitro at varying concentrations of GDP-[ 14C]fucose. Reactions were
carried out for 30 min in the presence of 5 mM MgCl2, 0.1% Tween-20, 240 pg of
protein for the S100, and 432 pg of protein for the P100. Fucose incorporation was
calculated from the amount of TCA-precipitable [14C]radioactivity. Results expressed
as the mean of three determinations + s.e.m.
Panel A. Effect of GDP-fucose concentration on S100 fucosyltransferase activity in
the presence of Tween-20.
Panel B. Hanes single reciprocal plot ([S]/v vs. [S]) for the S100
fucosyltransferase activity in the presence of Tween-20; apparent Km=1.7 pM, apparent
Vmax=42.7 pmol/mg protein/30 min.
Panel C. Effect of GDP-fucose concentration on P100 fucosyltransferase activity in
the presence of Tween-20.
Panel D. Hanes single reciprocal plot ([S]/v vs. [S]) for the S100
fucosyltransferase activity in the presence of Tween-20; apparent Km=38.2 pM,
apparent Vmax=122 pmol/mg protein/30 min.

40 60 80 1 00 -40 -20 0 20 40 60
GDP-fucose (pM) GDP-fucose (|iM)
pmol fuc/mg protein/30 min
pmol fuc/mg protein/30 min
II

118
for the higher activity of the S100 fraction in the presence
of detergent at the concentration of GDP-fucose used in most
assays (0.36 pM), despite the higher Vmay of the P100.
Since the above results were obtained in the presence
of detergent, I investigated the effect of GDP-fucose
concentration on the fucosyltransferase activities in the
intact fractions. It was reasoned that, in the case of the
P100, it would give us some insight into the overall
fucosylation process, including uptake of GDP-fucose into
the vesicles. S100 and P100 fractions were prepared from
HL250 vegetative cells and assayed for fucosyltransferase
activity under standard conditions at increasing
concentrations of GDP-fucose (figure 4-6). An apparent Km
of 0.44 pM and Vmax of 25.5 pmol fuc/ mg protein/ 30 min was
calculated for the S100. For the P100 an apparent Kn of
28.3 pM and Vmax of 233 pmol fuc/ mg protein/ 30 min was
determined.
The fucosyltransferase activities from the S100 and
P100 differed in the acceptor species that were fucosylated,
the sensitivity to high pH, divalent cation dependence, and
apparent affinity for GDP-fucose. These differences in
enzymatic behavior support a model for separate
compartmentalization of the two fucosyltransferase
activities; the S100 activity is free in the cytosol and the
P100 activity is in a membrane bound organelle.

Figure 4-6. Effect of GDP-fucose concentration on intact S100 and P100 fucosyltransferase
activities.
HL250 amoebae were harvested, homogenized, fractionated into an S100 and P100 and
fucosylated in vitro at varying concentrations of GDP-[1 C]fucose. Reactions were
carried out for 30 min in the presence of 5 mM MgCl2, 240 pg of protein for the S100,
and 432 pg of protein for the P100. Fucose incorporation was calculated from the
amount of TCA-precipitable [ 14C]radioactivity. Results expressed as the mean of
three determinations s.e.m.
Panel A. Effect of GDP-fucose concentration on intact S100 fucosyltransferase
activity.
Panel B. Hanes single reciprocal plot ([S]/v vs. [S]) for the S100
fucosyltransferase activity; apparent Km=0.44 pM, apparent V =25.5 pmol/mg
protein/30 min.
Panel C. Effect of GDP-fucose concentration on intact P100 fucosyltransferase
activity.
Panel D. Hanes single reciprocal plot ([S]/v vs. [S]) for the S100
fucosyltransferase activity; apparent Km=28.3 pM, apparent Vnwy=233 pmol/mg protein/30
min.

pmol fuc/mg protein/30 min
pmol fuc/mg protein/30 mil)20
S/v
S/v
031

121
Fucosyltransferase Activity Cannot be Detected In Vitro in
Ax3 S100 Extracts
Normal growing cells expressed [3H]fucose metabolically
labelled FP21 and the main glycopeptide and oligosaccharide
products released by pronase digestion and fl-elimination,
respectively, were indistinguishable from those derived from
the mutant HL250 (Chapter III). All the studies reported
above on the cytosolic fucosyltransferase activity were
carried out in HL250 because in vitro transfer of fucose
from GDP-fucose to endogenous acceptors cannot be detected
in Ax3 S100 fractions by TCA-precipitation, SDS-PAGE
analysis, or C18 SepPak fractionation (see Materials and
Methods for description of methods). A plausible
explanation for the lack of activity in the Ax3 would be
that the S100 from Ax3 contained an inhibitor for the
activity. To investigate this possibility, I performed
several mixing experiments. Vegetative Ax3 and HL250 cells
were harvested, filter-lysed and fractionated into an S100
and P100. The S100 fractions were assayed individually or
mixed in different ratios before assaying for
fucosyltransferase activity. The fractions were used intact
or desalted prior to assay (table 4-4). There was no
evidence for an inhibitor since experiments in which S100
fractions from Ax3 and HL250 were mixed in different ratios
showed activity commensurate to the HL250 contribution
(table 4-4). Dilution of labelled GDP-fucose with
endogenous unlabelled GDP-fucose is not an explanation

Table 4-4. Fucosyltransferase activity in Ax3 S100 fraction.
Contribution
Intact3
Desalted13
Ax 3
HL250
DPM
rel. activity0
DPM
rel. activity
th. value*
1
0
3
<0.01
8
<0.01
0
0
1
1842
1.00
2715
1.00
1
1
1
1030
0.56
963
0.35
0.50
4
1
435
0.24
218
0.08
0.20
1
4
1425
0.77
2098
0.77
0.80
Ax3
S100 was
mixed with
mutant (HL250) S100
and assayed for fucosyltransferase activity
as
described
for the
standard assay for
30 min
. Results are the
average of two
determinations. The
samples were either
intact
or desalted prior
to assay, rel.
activity, relative activity; th. value, theoretical value. A total of 30 pi of extract
was used in each assay and mixing was done relative to volume contributed by each S100
fraction. 1:1 ratio was 15 pl:15 pi; 1:4, 6 pl:24 pi; 4:1, 24 pi: 6 pi. relative
activity calculated using HL250 activity as a unit. *theoretical value based solely on
contribution from HL250 fraction. aAx3 S100 was at a concentration of 5 pg of
protein/pl, and HL250 at 6 pg/pl. bAx3 protein concentration at 8.5 pg/pl and HL250 at
6 pg/pl.
122

123
either because activity was not detected in Ax3 after
desalting through a P-2 column, while the HL250 S100
retained activity (table 4-4). Another explanation for the
lack of activity in the Ax3 S100 is that FP21 from Ax3 was
quantitatively fucosylated in vivo, leaving no acceptor
sites for the reaction in vitro. Evidence using purified
FP21 from HL250 supports this conclusion (see next section).
Fucosyltransferase Activity Is Detected in Ax3 S100 Fraction
Upon Addition of Mutant FP21
The absence of cytosolic fucosyltransferase activity in
Ax3 S100 extracts could be explained as a result of
guantitative fucosylation of FP21 in the living cell. It
was reasoned that if this model was correct, then addition
of mutant FP21 to Ax3 extracts would lead to incorporation
into FP21. FP21 was trace-labelled in vitro using GDP-
[14C]fucose, and partially purified by ammonium sulfate
precipitation, QAE-ion exchange chromatography, and HPLC gel
filtration. Fractions from the gel filtration step were
counted and examined by SDS-PAGE and those that contained
FP21 were pooled, brought to dryness, dissolved in water,
and added to Ax3 S100 extract. In vitro fucosylation was
determined as [3H] incorporated into TCA-insoluble material
in the presence of GDP-[3H]fucose. Since FP21 was trace-
labeled with [ 14C] fucose, the relative amount of the
acceptor added was estimated from [14C] dpm. West et al.
(unpublished results) showed that incorporation was

124
proportional to the amount of [14C] radioactivity added.
Since this study did not analyze the in vitro fucosylated
species by SDS-PAGE, it cannot be concluded that Ax3 was
able to fucosylate FP21. To investigate which MW species
served as acceptor for the Ax3 cytosolic fucosyltransferase,
the experiment was repeated using a new batch of partially
purified [14C]FP21. In vitro labelled FP21 eluted in
consecutive fractions 20 and 21 during HPLC gel filtration
chromatography, as confirmed by SDS-PAGE. Ax3 S100 was
added to an aliquot of fraction 21 that had previously been
dried on the bottom of the assay tube and assayed for
fucosyltransferase activity in the presence of GDP-
[3H]fucose and Mg++. The reaction was stopped by boiling in
sample electrophoresis buffer, resolved by SDS-PAGE, and the
gel sliced and counted. Figure 4-7 shows the comigration on
SDS-PAGE of the in vitro [3H] label resulting from
fucosylation by Ax3 with the trace [14C] labelled FP21 from
fraction 21. There was no [3H] radioactivity incorporation
into any MW species when the Ax3 S100 fraction was incubated
in the absence of purified HL250 FP21. In conclusion, Ax3
S100 has a cytosolic fucosyltransferase activity that
utilizes the same acceptor as the mutant cytosolic
fucosyltransferase.

Figure 4-7. Fucosylation of mutant FP21 by Ax3 S100 fraction.
HL250 amoebae were harvested, homogenized, and fractionated
into S100 and P100 fractions. Unlabeled HL250 S100 was
mixed with an aliquot of in vitro [ 14C ] fucosylated HL250
S100. FP21 was purified from the S100 fraction by (NH4)S02
precipitation, QAE-ion exchange chromatography, and HPLC gel
filtration (described in detail in Materials and Methods).
[14C]FP21 eluted in fractions 20 and 21 of the HPLC gel
filtration chromatography, as confirmed by SDS-PAGE. Ax3
S100 was added to an aliquot of fraction 21 that was
previously dried down in the bottom of an assay tube in a
vacuum centrifuge to serve as acceptor in the in vitro Ax3
S100 fucosyltransferase reaction. The reaction mixture
contained 0.15 fjM GDP-[3H] fucose, 5 mM MgCl2, 349 pg of Ax3
S100 protein, was incubated for 60 min and the reaction
stopped by boiling in SDS/fi-mercaptoethanol electrophoresis
buffer. Sample was resolved on a 15% SDS-polyacrylamide
gel, which was cut into 0.5 cm slices and counted. No [ H]
radioactivity was incorporated into any MW species when the
wild type Ax3 S100 fraction was incubated in the absence of
added FP21 from mutant source (not shown). Open circles,
[14C] label derived from in vitro labelled purified FP21
from HL250; closed circles, [3H] radioactivity from in vitro
fucosylation by the Ax3 S100 fraction.

126

127
Cytosolic Fucosyltransferase Preferentially Fucosylated a
Type I Acceptor
The size of the fucose-containing oligosaccharide in
FP21 (determined in Chapter III to be 4.8 glucose units)
implies there is more than one sugar residue, so the
acceptor site on FP21 may be another sugar. As a first step
to determine whether the cytosolic fucosyltransferase could
fucosylate model acceptor analogs utilized by known
fucosyltransferases, the S100 from Ax3 and mutant origin
were screened for activity towards hydrophobic model
acceptors. The incorporation of radioactivity into
synthetic sugar acceptors that contained 8-methoxy
carbonyloctyl, or methyl nonanoate, [CH3(CH2)7COOCH3,
referred to as R throughout the text] as the hydrophobic
tail by the S100 was determined by the C18 Sep-Pak method,
which employs a hydrophobic interaction column. Unreacted
GDP-fucose does not interact with the column, eluting in the
water wash while the glycolipid acceptor is eluted from the
column with methanol. As shown in table 4-5, only the type
I acceptor analog (known as lacto-N-biose I or
gal/11,3GlcNAc/l-R) sustained activity in the Ax3 S100. In
contrast, type II (known as N-acetyllactosamine or gal/11,4-
GlcNAc-R) and /1-gal-R were not suitable acceptors (see
figure 4-8 for structures). The mutant S100 was also active
with gal/11,3GlcNAc/l-R, but only one tenth as active (on a
per protein basis) as Ax3.

128
Table 4-5. Utilization of 8-methoxycarbonyloctyl synthetic
acceptors by cytosolic fucosyltransferase activity from Ax3
and HL250.
pmol fucose/mq/h
substrate concentration (mM) Ax3 HL250
fi-gal-R 1 und. n.d.
galfil,3GlcNAcfi-R 0.15 0.38 0.039
galfll,4GlcNAcfi-R 0.15 und. n.d.
Incorporation was determined by the C18 Sep-Pak assay (see
Materials and Methods). Experiment performed by C.M. West and
is the result of one determination, und, undetectable; n.d.,
not determined; -R is -(CH2)0COOCH3; GDP-[ 3H] fucose
concentration, 0.15 jiM.

Figure 4-8. Haworth projections of the structures of the
synthetic glycolipid acceptors.

130
CH.,r¡H
1 D-gal-R
CH-.OH
2 gall3(I,3)G1cNAcI3-R
CH.-.CH
R= (CH2)8C00CH3

131
8-methoxycarbonyloctyl synthetic acceptors are not
available commercially, so I investigated the possibility
that other hydrophobic glycosides, which can be readily
obtained from commercial sources, would serve as acceptors.
Some of these phenyl derivatives have been shown by others
to be suitable acceptors for a fucosyltransferase (Potvin,
et al., 1990; Palcic et al., 1988). S100 fractions from
HL250 and Ax3 were assayed for fucosyltransferase activity
in the presence of p-nitro-phenyl glycoside derivatives;
fucosylation of endogenous substrates was monitored by TCA-
precipitation, and fucosylation of p-nitro-phenyl glycosides
by the Sep-Pak method (table 4-6). Inhibition of
incorporation of radioactivity into FP21 in HL250 was
examined by TCA precipitation of endogenous acceptors.
Millimolar concentrations of these compounds failed to
inhibit significantly incorporation. Likewise, none of the
p-nitro-phenyl glycosides served as acceptor for the mutant
nor the Ax3 cytosolic fucosyltransferase activity when
assayed by the C Sep-Pak method. Thus it is concluded,
that from the acceptor candidates examined, only
galfil,3GlcNAcfi-R is a suitable acceptor under the conditions
used.
The activity responsible for fucosylation of the type I
acceptor analog in Ax3 was examined by varying the
concentration of acceptor or the concentration of nucleotide
sugar donor. The type I acceptor analog was fucosylated

Table 4-6. Evaluation of the suitability of p-nitro-phenyl glycosides as acceptors for
cytosolic fucosyltransferase activity in HL250 and Ax3 S100.
treatment concentration
control
p-nitro-0-cr-D-glucose
p-nitro-0-fi-D-glucose
p-nitro-0-a-D-GlcNAc
p-nitro-0-i-D-GlcNAc
p-nitro-0-fi-D-galactose
p-nitro-0-fi-D-glc
tetraacetate
DPM HL250
DPM HL250
DPM Ax3
(TCA)
(SepPak C18)
(SepPak i
1003
<3
<3
996
21
4
1020
19
n.d.
1066
26
10
1032
31
18
1059
38
3
1134
24
10
10
mM
10
mM
10
mM
10
mM
10
mM
10
mM
CO
ho
S100 fractions from Ax3 and HL250 were assayed for fucosyltransferase activity and
fucose incorporation determined by TCA-precipitation or SepPak C method (see Materials
and Methods fordetaiis). n=l; n.d., not determined; 0, phenyl; 0.89 pM GDP-[ 14C ] f ucose;
0.1% Tween-20; 619 pg protein; 60 min assay.

133
with an apparent of approximately 1 mM for the acceptor
and 1.6 fjM for GDP-fucose (C.M. West, unpublished results).
The similarity in the apparent Km for GDP-fucose for the
type I and FP21 fucosyltransferase suggested that the same
enzyme may be responsible for both reactions. The fact that
HL250 was able to fucosylate so poorly the analog when
compared to Ax3, suggested that the availability of
endogenous substrate (FP21) inhibited incorporation into the
synthetic acceptor, and supported the idea of the same
enzyme fucosylating both substrates. The notion was further
reinforced by reduction of type I analog fucosylation in Ax3
S100 extracts by purified FP21. Fractions 20 and 21 from
the HPLC gel filtration chromatograph (see section above on
reconstitution of Ax3 fucosyltransferase activity by
purified FP21) reduced incorporation of radioactivity into
type I acceptor (table 4-7). Even though reduction was not
strictly proportional, it was evident that it increased with
increasing amounts of FP21. Taken together the results of
this section and the preceding one, it appears that Ax3
possesses a fucosyltransferase activity in the cytosol
capable of fucosylating FP21 and the type I analog acceptor.
Cytosolic Fucosyltransferase Activity is Present in
Migrating Slug Stage Cells
Cytosolic FP21 was detectable by metabolic labelling in
slug stage HL250 cells (Chapter III). Reasoning that a
fucosyltransferase responsible for its modification would be

134
Table 4-7
analog by
fraction
. Reduction of
purified FP21.
relative amount
fucosylation of acceptor type
r3Hlfucose incorporated
added
(dpm/mg protein/h)
0
499
20
lx
387
20
4x
318
21
lx
436
21
4x
<20
Transfer of [3H] from GDP-[3H]fucose into 4 pg (0.145 mM) of
type I acceptor analog by Ax3 S100 was measured using the C18
Sep-Pak method (see Materials and Methods). Data are the
results of one determination. 349 nq protein of Ax3 S100;
GDP-[3H]fucose concentration, 0.15 pM; 60 min assay.

135
present at this stage of development, I assayed developing
cells for their ability to incorporate [14C] from GDP-
[14C]fucose into endogenous acceptors. HL250 cells were
plated for development and at slug stage harvested,
disaggregated, fractionated into an S100 and P100, and
assayed for fucosyltransferase activity as described in
Materials and Methods. For comparison, fractions from Ax3
slugs were assayed (table 4-8). As seen with amoebae cells,
HL250 S100 was active whereas the Ax3 S100 did not
incorporate radioactivity. On the other hand, the P100 was
active in both strains. Thus, it seems that the cytosolic
fucosyltransferase is not restricted to the growth phase and
is present in developing cells.
Discussion
The presence of FP21 in the cytosol suggested that a
fucosyltransferase might also be located in the cytosol. An
S100 fucosyltransferase activity was detected which was both
time- and protein concentration-dependent. The activity was
strictly divalent cation dependent. Incorporation of
radioactivity was sensitive to temperature, certain
detergents, and ethanol. A variety of sugars and sugar-
derivatives failed to inhibit activity, except GDP-ii-fucose,
which inhibited in a dose-responsive manner. The activity
could not be sedimented by higher centrifugation force in
the presence of an extensive list of protease inhibitors.

136
Table 4-8. In vitro fucosyltransferase activity of HL250 and
Ax3 slug extracts.
pmol fucose/mq protein/30 min
strain SIPO P100
HL250 40.6 36.3
Ax3 <0.1 30.7
Normal and mutant amoebae were allowed to develop, harvested,
disaggregated, filter-lysed, and fractionated into an S100 and
P100. Intact fractions were assayed for fucosyltransferase
activity in the presence of 0.36 pM GDP-[ 14C] fucose, 5 mM
MgCl2, 12-48 pg of protein, for 30 min. Results are the
average of two determinations.

137
The endogenous acceptor utilized by the S100
fucosyltransferase was a protein which comigrated with FP21
by SDS-PAGE. I compared the acceptor for the in vitro
fucosyltransferase reaction with metabolically labelled FP21
from Ax3 cells by SDS-PAGE. There was one main radioactive
peak, revealing in vivo and in vitro fucosylated acceptors
with the same mobility on polyacrylamide gels. These
results suggested that a cytosolic fucosyltransferase
existed that utilized FP21 as its primary acceptor species
in vitro, and may be responsible for fucosylation of FP21.
To investigate the origin of the S100
fucosyltransferase, I compared it to the bulk P100
fucosyltransferase activity, since the cytosol is the
default location of lumenal enzymes released by rupture of
vesicles. If both activities were indeed different, I
expected to detect enzymatic differences. Initially, I
examined the SDS-PAGE profiles of in vitro fucosylated
acceptors and found they were very similar to those obtained
from metabolic labelling. Incorporation by endogenous
acceptors was at the 21 kD MW position for the S100, and in
the P100 radioactivity migrated as two separate, broad
peaks.
In order to compare directly the soluble and the
sedimentable activities, I assayed the S100 and P100 in the
presence of detergent to overcome any differences in
accessibility for GDP-fucose by the fucosyltransferases. I

138
determined that the S100 fraction was dependent on divalent
cations, while the P100 was active in the absence of cations
and in the presence of the chelator EDTA. The activities in
both fractions were maximal at a similar pH range, but the
cytosolic fucosyltransferase was more sensitive to higher pH
than the P100 fucosyltransferase activity.
Glycosyltransferase activities have commonly been found to
be dependent on the presence of divalent cations. In the
case of fucosyltransferases, however, there are precedents
for al,2, al,3, and al,3/1,4 fucosyltransferases which are
active in the absence of cations, and are either stimulated
or inhibited by different cations (Beyer and Hill, 1980;
Campbell and Stanley, 1984; Foster et al.,1991; Stroup et
al., 1990; Zatz and Barondes, 1971).
The apparent affinity for GDP-fucose differed greatly
for S100 and P100 activities. The S100 fucosyltransferase
activity had a higher affinity for GDP-fucose than the P100
activity when both were assayed in the presence of Tween-20.
The lower apparent Km for the cytosolic fucosyltransferase
explained why activity is higher in the S100 at the low
concentration of GDP-fucose used in most assays, 0.36 pM.
At 0.36 pM the concentration of GDP-fucose was near its
apparent Km for the S100 fucosyltransferase (1.7 pM), but
well below the apparent Km for the P100 enzyme (38.2 pM) .
The dependence on GDP-fucose concentration was also examined
in the intact fractions to gain some insight into the

139
overall fucosylation process in the P100 fraction, including
transport into the intact vesicles. The apparent Michaelis
constants for the P100 activity in the presence of Tween-20
and in the intact fraction were 38.2 pM and 28.3 /jM,
respectively. The similarity of the apparent Km values
suggested that the GDP-fucose transporter in the P100
vesicles had a similar or lower K relative to that of the
m
bulk P100 fucosyltransferase activity, since if it had a
much higher apparent Km, GDP-fucose transport would have
been rate limiting. The GDP-fucose transporter from rat
liver Golgi-enriched vesicles has an apparent Km of 7.5 /jM
(Sommers and Hirschberg, 1982). The apparent Km for the
cytosolic fucosyltransferase is relatively low compared to
that of the bulk P100 activity. Though the relative
concentrations of GDP-fucose in the cytosol and vesicles are
not known, vesicles have the ability to concentrate GDP-
fucose relative to the outside (Perez and Hirschberg, 1986).
Thus it is not unreasonable to predict that a cytosolic
fucosyltransferase would have a higher affinity for GDP-
fucose since the concentration of GDP-fucose is probably
lower in the cytosol than in the vesicles.
The studies described in this chapter concerning the
P100, characterized the bulk activity in the fraction and
cannot differentiate among different fucosyltransferases
that may be present. The fucosyltransferase activity in the
P100 may be a product of different fucosyltransferases with

140
different specificities. This may be the case in
Dictyostelium because, even though fucosyltransferases have
not been well characterized in this organism, various
fucosyltransferases have been localized to microsomes in
other eukaryotes (Hirschberg and Snider, 1987; Kornfeld and
Kornfeld, 1985). The fact that I was able to differentiate
the bulk activity in the P100 from the cytosolic
fucosyltransferase supported the idea that the
fucosyltransferase in the S100 is unrelated to the P100
activity and thus endogenous to the cytosol.
However, my observations do not rule out other
possibilities. For example, the cytosolic
fucosyltransferase could have derived from vesicles but was
preferentially lost during isolation and the remaining
enzyme, though with distinct properties from the majority of
the P100 activity, is in the minority. The inability of
EDTA to inhibit activity further when compared to no
addition of divalent cations to the P100, may mean that the
enzyme does not need cations at all. Conversely, since the
bulk activity in the P100 is stimulated by cations, it is
possible that the activity retains tightly bound cations
which EDTA cannot remove. Another possibility is that the
acceptor, FP21, is not present in the P100, either due to a
cytosolic compartmentalization, or to leakage from the
vesicles. A definitive confirmation that the cytosolic
fucosyltransferase is different from any fucosyltransferase

141
activity in the P100 will require characterization of the
purified fucosyltransferases from the S100 and P100.
The results obtained from my investigation are based on
biochemical evidence in which a soluble fucosyltransferase
partitioned with the cytosol. Other investigators have
identified glycosylated proteins in the cytosol and/or
nucleus and have searched for an enzyme responsible for the
addition of the sugar (Haltiwanger et al., 1990). Their
biochemical studies showed that an activity capable of
adding GlcNAc to protein was recovered in both the soluble
and membrane fractions (Haltiwanger et al., 1990). However,
they showed that the membrane-associated activity was
releasable by high salt treatment and was oriented towards
the cytosol, not the lumen of the vesicles. Thus, it is
possible that a fraction of this newly discovered cytosolic
fucosyltransferase stayed associated with vesicles but since
it was in a minority, remained masked by other P100
fucosyltransferases. As more synthetic acceptors become
available, latency experiments in the presence and absence
of detergent can be done to address this question.
Alternatively, it is possible that a fucosyltransferase with
enzymatic properties similar to the cytosolic
fucosyltransferase is present in the lumen of P100 vesicles.
Still this will not contradict my findings and will imply
that there are two similar enzymes that reside in distinct
compartments, as has been reported for another enzyme (Lewin

142
et al., 1990). In any event, I interpret the data presented
as evidence for a fucosyltransferase in the cytosol of
Dictyostelium discoideum.
The fact that Ax3 produced fucosylated FP21 suggested
that, as it occurred in the mutant, the normal strain may
have a cytosolic fucosyltransferase responsible for FP21
fucosylation. However, while activity was not detectable in
Ax3 S100 fraction, it could be reconstituted by addition of
mutant FP21, indicating that Ax3 possessed a cytosolic
fucosyltransferase equivalent to the mutant
fucosyltransferase.
In order to characterize the fucosyl linkage catalyzed
by the cytosolic fucosyltransferase, several acceptors were
used. Activity with synthetic acceptors was about an order
of magnitude higher for the Ax3 extract, which may be
attributed to competitive inhibition by the unfucosylated
FP21 in the mutant. Of those tested, the only suitable
acceptor was found to be a type I analog, 8-
methoxycarbonyloctyl galfil,3GlcNAcfi. Since the type II
analog [8-methoxycarbonyloctyl galfil,4GlcNAcH] did not work
as acceptor, it appears that the cytosolic
fucosyltransferase may be an al,4fucosyltransferase.
The cytosolic fucosyltransferase preferentially
recognized a type I analog, suggesting it was an
al,4fucosyltransferase that lacked al,3 activity. This
activity would differ from other al,4fucosyltransferase

143
described, which exhibit al,3 activity as well (Kukowska-
Latallo et al, 1990; Stroup et al., 1990). However, there
are some limitations to the studies employing synthetic
acceptors. To conserve synthetic acceptors, which were not
commercially available, the concentration of the acceptors
was well below the Km (0.145 mM, while the apparent Kn was
determined to be approximately 1 mM). The possibility still
exists that the enzyme is able to use 8-methoxycarbonyloctyl
galiil,4GlcNAcfi as acceptor, but will only be evident at
higher concentrations. Tentatively, an al,4 specificity is
being assigned to the cytosolic fucosyltransferase, but
definitive proof will require characterization of the enzyme
purified to homogeneity.
Finally, slug stage extracts were examined for
fucosyltransferase activity, because it was found by
metabolic labelling experiments in Chapter III that a
fucosylated protein of 21 kD fractionated with the S100.
The S100 and P100 fractions from HL250 had considerable
activity, but from the Ax3 fractions only the P100 showed
activity, consistent with the results from vegetative cells.
The detection of a cytosolic fucosyltransferase in slug-
stage cell extracts is consistent with their ability to
fucosylate FP21 in vivo as determined by metabolic
labelling. The apparent absence of activity in Ax3 cells
indicated that, as found for vegetative stage cells, FP21
was quantitatively fucosylated.

CHAPTER V
SUMMARY AND CONCLUSIONS
Summary of Results
Fucosylation has generally been regarded as a
modification restricted to the secretory compartment,
however, there is evidence of fucosylated macromolecules in
the nucleus and cytosol (see Chapter I). In the present
study, I identified a novel fucosylation pathway in the
cytosol of Dictyostelium discoideum. In the next three
paragraphs a short summary is presented of the results
reported in this dissertation, followed by a proposed model
of fucosylation in the cytosol.
In chapter II the mutant HL250 was characterized as a
conditional fucosylation mutant. The results are summarized
as follows: 1) Spores and vegetative cells from the mutant
strain contained negligible amounts of macromolecular-
associated and total cell fucose when compared to the normal
strain, Ax3, as determined chemically in acid hydrolysates.
2) The phenotype was conditional to growth in the absence of
fucose. When vegetative cells were grown in fucose-
supplemented media, they expressed macromolecular fucose
conjugates. The fucose specific activity of the medium was
not diluted relative to the intracellular fucose. 3)
144

145
Mutant extracts were incapable of carrying out the
conversion of GDP-mannose to GDP-fucose in vitro. In other
organisms, this pathway is the sole pathway of GDP-fucose
synthesis in the absence of extracellular fucose. The low
fucose biochemical phenotype can be explained by the model
that the conversion pathway is defective. HL250 cells and
extracts in vitro can still fucosylate, showing that GDP-
fucose transport and fucosyltransferase(s) are still active.
Although the possibility remains that there are other
genetic defects in this mutagenized strain, there is no
reason to suspect that other genes of the fucosylation
pathway have been affected.
After determining that the source of macromolecular
fucose in HL250 grown in normal medium was derived from
extracellular fucose, I examined the compartmentalization of
fucosylation. The results of the experiments described in
Chapter III show the existence of a fucosylated protein in
the cytosol and are summarized as follows: 1) The major
fucosylated species in the S100 is FP21. It is present in
both Ax3 and HL250. 2) Analysis of FP21 revealed that the
oligosaccharide in FP21 was O-linked with a size of 4.8
glucose units. 3) FP21 appears to be endogenous to the
cytosol, and not derived from a sedimentable compartment
during preparation of the extracts. 4) Glycopeptides
released from FP21 by pronase digestion differ from 21 kD MW

146
PlOO-derived glycopeptides, which reinforced the notion that
contaminating P100 material was not the source of FP21.
The presence of a cytosolic fucosylated protein
suggested the existence of a cytosolic fucosylation pathway.
In vitro analysis of subcellular fractions led to the
detection of a fucosyltransferase activity in the cytosol.
The results of this investigation described in Chapter IV
are summarized as follows: 1) Using a fucosylation assay
dependent on endogenous acceptor substrates, I detected
fucosyltransferase activity in cytosolic and vesicular
fractions. 2) Activities from S100 and P100 fractions
differed in the acceptor species fucosylated, their
sensitivities to alkaline pH and divalent cations, and
affinities for GDP-fucose, as evidenced by differences in
apparent Km. I consider these results to be an indication
that the S100 fucosyltransferase did not arise from vesicles
by rupturing during cell fractionation. 3) The cytosolic
fucosyltransferase activity was absolutely dependent on
availability of a non-fucosylated acceptor. Accordingly, in
vitro cytosolic fucosylation could be detected in mutant
extracts, but not in Ax3 fractions. However, cytosolic
fucosyltransferase activity was reconstituted in Ax3
fractions by addition of purified mutant FP21. 4) A
fucosyltransferase activity was detected in the S100 with
the use of synthetic hydrophobic acceptors. Based on the
utilization of these acceptors, the activity was determined

147
to be an al,4fucosyltransferase lacking ctl,3 activity. 5)
Fucosylation of the type I acceptor analog (galfil,3GlcNAcfi-
8-methoxycarbonyloctyl) was inhibited by addition of
purified FP21, suggesting the same activity was responsible
for fucosylation of both molecules.
Based on the results obtained in my studies, I propose
a model for fucosylation in Dictyostelium, acknowledging the
existence of a fucosyltransferase in the cytosol that
fucosylates a cytosolic protein, FP21. There are
fucosyltransferases in vesicles and in the cytosol; the
preferential acceptor for the cytosolic fucosyltransferase
is FP21. This model is appealing because all of the
elements necessary for fucosylation, biosynthesis of GDP-
fucose, a fucosyltransferase, and the acceptor,
compartmentalize in the cytosol. The model also concurs
with emerging views of glycosylation in the cytosol (Hart et
al., 1989a; Hart et al., 1989b). Initially I showed that
Dictyostelium possesses a GDP-fucose conversion pathway
similar to that reported earlier for other organisms
(Kornfeld and Ginsburg, 1966; Liao and Barber, 1971; Ripka
et al., 1986). It was shown that Dictyostelium can convert
GDP-mannose into GDP-fucose, and that when this biosynthetic
pathway is defective, GDP-fucose is formed from fucose
supplied in the extracellular medium by the salvage pathway.
This is the first time evidence has been presented that
suggests Dictyostelium has GDP-fucose biosynthetic pathways

148
similar to those found in bacteria (Kornfeld and Ginsburg,
1966), a higher plant (Liao and Barber, 1971), and mammalian
cells (Ripka et al., 1986; Reitman et al., 1980).
However, this model is not the only one that could
account for the data obtained during the course of my
investigation. Alternatively, the presence of FP21 in the
cytosol could be explained by fucosylation in vesicles and
rapid posttranslational transport to the cytosol. The
absence of FP21 and FP21-like glycopeptides in the vesicular
fraction was interpreted earlier as evidence for the absence
of FP21 in the P100. However, it does not rule out the
possibility that FP21 was fucosylated in vesicles and soon
thereafter transported back into the cytosol, but was not
detected because it did not accumulate in the P100. The
presence of a fucosyltransferase in the cytosol would then
be accounted by leakage from the vesicular fraction.
Clearly, this model must then explain the export of FP21
into the cytosol by novel and unknown mechanisms. Another
model that would account for my results is that both FP21
and the fucosyltransferase detected in the S100 leaked into
the supernatant during fractionation. Since known vesicular
markers were shown to remain in the P100, this model would
require FP21, from all fucoconjugates in the P100, to be
released preferentially. It would also require the leakage
of a fucosyltransferase capable of fucosylating FP21.
Nevertheless, in order to distinguish between the model

149
proposed and the other possible models, additional studies
are needed.
Future Studies
The results presented in this dissertation lay the
ground work for future studies with immense possibilities.
With the help of a strain with a conditional fucosylation
mutation, I was able to recognize the presence of a
fucosylation pathway that otherwise may have gone
undetected. Future research should focus on FP21 or the
cytosolic fucosyltransferase.
Studies on FP21
The first question to be addressed will be the
compartmentalization of FP21 using an independent approach
from that followed in my studies. An initial step would be
to raise antibodies against FP21. A protocol to purify FP21
is being improved in the laboratory, and should prove useful
for this purpose. An antibody against FP21 will be useful
for immunolocalization of the acceptor in fixed cells.
Currently, FP21 is detected by SDS-PAGE as a fucose-labelled
21 kD MW species. In the P100, the presence of other
fucoconjugates of similar MW on SDS-PAGE could mask FP21,
although, as discussed earlier, it appears that FP21 is not
present in the P100 fraction. However, if FP21 was entirely
released into the S100 fraction during cell fractionation

150
that would explain its absence in the P100. Alternatively,
a higher MW precursor may exist in the P100.
Immunolocalization of FP21 will help clarify this point. An
antibody against FP21 will help in localizing FP21 in other
compartments of the cell, if present, such as in the nucleus
or nuclear membranes.
Another aspect of interest is the other sugar residues
present in the FP21 oligosaccharide. Due to its size (4.8
glucose units) I suspect the carbohydrate moiety is truly an
oligosaccharide, containing more than one sugar residue.
There is evidence for a peptide-GlcNAc transferase in the
cytosol of rat hepatocytes (Haltiwanger et al., 1990), so it
is possible that the oligosaccharide is O-linked to the
polypeptide backbone via a GlcNAc residue. The fact that
the FP21 radioactive peak released by alkaline hydrolysis
was not symmetric, suggested there is more than one type of
oligosaccharide. The first step would be to separate and
purify the oligosaccharides. For this purpose, a longer P-4
column could be used. Alternatively, the oligosaccharides
could be separated by other chromatographic methods
(Townsend et al., 1989; Beniak et al., 1988). Once
separated, the oligosaccharides can be examined by nuclear
magnetic resonance spectrometry.
Fucosylated FP21 was present in the cytosol of
vegetative and developing cells, but at this moment the
relative levels of expression at different developmental

151
times, nor if its preferentially expressed in any cell type
during development, have been determined. Ideally, it would
be useful to produce antibodies with specificity for the
glycosylated protein, and specificity for the peptide moiety
of FP21 (similar to other mAb produced in the laboratory;
see West et al., 1986). Using these antibodies,
fucosylation of FP21 during development could be followed by
immunoprecipitation of FP21.
Studies on the Cytosolic Fucosyltransferase
Another aspect of my project was the evidence presented
for a novel fucosyltransferase that appears to be cytosolic
and seems to differ from the bulk sedimentable
fucosyltransferase activity. The first question to be
addressed will be the compartmentalization of the enzyme.
The cytosolic enzyme could be purified by conventional
methods (Beyer et al., 1980; Foster et al., 1991; Martin et
al., 1987). Once purified, antibodies could be raised
against the enzyme and used for immunolocalization of the
fucosyltransferase. Currently, a purification protocol is
being developed in the laboratory. If localization of the
enzyme is done by immunofluorescence and the enzyme is a
soluble cytosolic protein, it should be possible to observe
a cytosolic distribution of the enzyme and an absence from
intracellular vesicles. However, if the immunofluorescence
pattern shows labelling of vesicles, the results will need

152
to be examined more carefully. It is possible that the
enzyme fucosylates the cytosolic acceptor, FP21, while being
membrane-associated, but facing the cytosol. There is a
membrane-associated glycosyltransferase that utilizes
cytosolic acceptors (Haltiwanger et al., 1990). If a
portion of the fucosyltransferase pool was to partition to
the outside of the vesicles, the activity on intact vesicles
should fucosylate the type I analog. Fucosylation of this
synthetic acceptor by intact P100 vesicles should be
dependent on added Mg++.
Another approach to study the cytosolic
fucosyltransferase is to clone and seguence the enzyme,
avoiding purification of the protein. To date, only two
fucosyltransferases have been sequenced, one encodes an
al,3/l,4fucosyltransferase and the other an
al,3fucosyltransferase (Kukowska-Latallo et al., 1990; Goelz
et al., 1990). There is 57% identity between the two
enzymes at the C-terminus, for a stretch of two-thirds the
length of the protein (Goelz et al., 1990). Both enzymes
appear to be type II transmembrane proteins, each composed
of a short amino-terminal cytoplasmic domain with no
discernible signal sequence, and a putative single
transmembrane signal/anchor domain (Kukowska-Latallo et al.,
1990; Goelz et al., 1990). The sequenced
fucosyltransferases possessed N-linked glycosylation sites,
and one of them was shown to be a glycoprotein (Kukowska-

153
Latallo et al., 1990). Since the fucosyltransferase
reported in my studies appears to be cytosolic, it would be
important to determine what is the relationship between
microsomal and cytosolic fucosyltransferases. All
fucosyltransferases utilize the same sugar nucleotide donor,
GDP-fucose, so it is likely that the GDP-fucose binding site
would be similar for all enzymes. In addition, comparisons
among the fucosyltransferases may reveal important
information regarding intracellular targeting and possible
evolutionary relationships. There is evidence that a
retaining sequence allows glycosyltransferases to remain in
the Golgi apparatus and endoplasmic reticulum (Paulson and
Colley, 1989). The fact that the fucosyltransferase
reported in these studies localizes to the cytosol raises
the possibility that the fucosyltransferase would lack the
targeting and retaining sequences.
In order to compare the cytosolic fucosyltransferase
with the sequenced fucosyltransferases (Kukowska-Latallo et
al., 1990; Goelz et al., 1990), it will be necessary to
sequence the cytosolic fucosyltransferase. The
aforementioned enzymes were cloned using a gene transfer
system in which cloned cDNAs determined the expression of
the enzyme in a recipient host that did not express such
activity. It could be possible to do the same for the
cytosolic fucosyltransferase, using the type I analog
synthetic acceptor to screen for activity of transfected

154
clones. A suitable host to express the cytosolic
fucosyltransferase cDNA would be a mutant Dictyostelium
strain, although there are no such mutants available at the
present. On the other hand, yeast could be used, since it
has been shown yeast cells do not carry out fucosylation
(Kukuruzinska et al., 1987). One of the complications that
may arise in trying to screen for clones expressing the
cytosolic fucosyltransferase is the transfection of
microsomal fucosyltransferases. It remains to be determined
whether the P100 fucosyltransferase activity is capable of
fucosylating the type I analog. If the activity in the P100
does not utilize the type I analog [galfl( 1,3) GlcNAc/1-8-
methoxycarbonyloctyl] as acceptor, then clones can be
screened using the synthetic acceptor. However, if there
are fucosyltransferases in the P100 that utilize the type I
analog as acceptor, it will be necessary to differentiate
the activity in vitro before the transfection experiments.
P100 extracts will be assayed for the ability to fucosylate
galfil,3GlcNAcfi-8-methoxycarbonyloctyl. If the fraction
fucosylates the acceptor, the sensitivity to cations will be
examined for possible differences with the cytosolic
fucosyltransferase activity. If the fraction is active in
the presence of EDTA, fucosyltransferase positive clones may
be screened in the presence and absence of Mg++. Those that
express activity only in the presence of Mg++ may represent
positive clones. In the event that the activity in the P100

155
is dependent on Mg++ in a fashion similar to the cytosolic
fucosyltransferase, other inhibitors should be tried,
including tunicamycin and N-ethylmaleimide (Galland et al.,
1988; Campbell and Stanley; 1984).

REFERENCES
Abeijon, C. and C.B. Hirschberg. 1990. Topography of
initiation of N-glycosylation reactions. J Biol. Chem.
265:14691-14695.
Aparicio, J., G.W. Erdos, and C.M. West. 1990. Spore coat is
altered in mod B glycosylation mutants of Dictyostelium
discoideum. J. Cell. Biochem. 42:255-266.
Arion, W.J., L.M. Balias, A.J. Lange, and B.K. Wallin. 1976.
Microsomal membrane permeability and the hepatic
glucose-6-phosphatase system. Interactions of the
system with D-mannose- 6-phosphate and D-mannose. J.
Biol. Chem. 251:4901-4907.
Barondes, S.H. and P.L. Haywood. 1979. Comparison of
developmentally regulated lectins from three species of
cellular slime mold. Biochim. Biophys. Acta 550:297-
308.
125
Bartles, J.R. and W.A. Frazier. 1980. Preparation of I-
discoidin I and the properties of its binding to
Dictyostelium discoideum cells. J. Biol. Chem. 255:30-
38.
Beniak, B., J. Orr, I. Brockhausen, G. Vella, and C. Phoebe.
1988. Separation of neutral reducing oligosaccharides
derived from glycoproteins by HPLC on a hydroxylated
polymeric support. Anal. Biochem. 175:96-105.
Bennett, G., C.P. Leblond, and A. Haddad. 1974. Migration of
glycoprotein from the Golgi apparatus to the surface of
various cell types as shown by radioautography after
labeled fucose injection into rats. J. Cell Biol.
60:259-284.
Beyer, T.A. and R.L. Hill. 1980. Enzymatic properties of the
B-galactoside al-2 fucosyltransferase from porcine
submaxillary gland. J. Biol. Chem. 255:5373-5379.
Beyer, T.A., J.E. Sadler, and R.L. Hill. 1980. Purification
to homogeneity of the H blood group fi-galactoside al-2
fucosyltransferase from porcine submaxillary gland. J.
Biol. Chem. 255:5364-5372.
156

157
Biermann, C.J. 1988. Hydrolysis and other cleavages of
glycosidic linkages in polysaccharides. Adv. Carbohydr.
Chem. Biochem. 46:251-271.
Borts, R.H. and R.L. Dimond. 1981. The a-glucosidases of
Dictyostelium discoideum. Develop. Biol. 87:176-184.
Campbell, C. and P. Stanley. 1984. The Chinese hamster ovary
glycosylation mutants LECH and LEC12 express two novel
GDP-fucose:N-acetylglucosaminide 3-a-L-
fucosyltransferase enzymes. J. Biol. Chem. 259:11208-
11214.
Chen, P.S., Jr., T.Y. Toribara, and H. Warner. 1956.
Microdetermination of phosphorus. Analyt. Chem.
28:1756-1758.
Chu, F.K. 1986. Requirements of cleavage of high mannose
oligosaccharides in glycoproteins by peptide N-
glycosidase F. J. Biol. Chem. 261:172-177.
Darnell, J., H. Lodish, and D. Baltimore. 1986. Molecular
Cell Biology. Scientific American Books, Inc., New
York.
Das, O.P. and E.J. Henderson. 1986. A novel technique for
gentle lysis of eukaryotic cells: isolation of plasma
membranes from Dictyostelium discoideum. Biochim.
Biophys. Acta 736:43-56.
Davis, L.I. and G. Blobel. 1987. Nuclear pore complex
contains a family of glycoproteins that includes p62:
glycosylation through a previously unidentified
cellular pathway. Proc. Natl. Acad. Sci. USA 84:7552-
7556 .
Dimond, R.L., R.A. Burns, and K.B. Jordan. 1981. Secretion
of lysosomal enzymes in the cellular slime mold
Dictyostelium discoideum. J. Biol. Chem. 256:6565-6572.
Erdos, G.W. and C.M. West. 1989. Formation and organization
of the spore coat of Dictyostelium discoideum. Exp.
Mycol. 13:169-182.
Erdos, G.W. and D. Whitaker. 1983. Failure to detect
immunocytochemically reactive endogenous lectin on the
cell surface of Dictyostelium discoideum. J. Cell Biol.
97:993-1000.

158
Finne, J., T. Krusius, R.K. Margolis, and R.U. Margolis.
1979. Novel mannitol-containing oligosaccharides
obtained by mild alkaline borohydride treatment of a
chondroitin sulfated proteoglycan from brain. J. Biol.
Chem. 254:10295-10300.
Flowers, H.M. 1981. Chemistry and biochemistry of D- and L-
fucose. Adv. Carbohydr. Chem. Biochem. 39:279-345.
Foster, C.S., D.R.B. Gillies, and M.C. Glick. 1991.
Purification and characterization of GDP-L-fuc-N-
acetyl-fi-D-glucosaminide al-3fucosyltransferase from
human neuroblastoma cells. Unusual substrate
specificities of the tumor enzyme. J. Biol. Chem.
266:3526-3531.
Franke, J. and R. Kessin. 1977. A defined minimal medium for
axenic strains of Dictyostelium discoideum. Proc. Natl.
Acad. Sci. USA 74:2157-2161.
Galland, S., A. Degiuli, J. Frot-Coutaz, and R. Got. 1988.
Transfer of N-acetylglucosamine to endogenous
glycoproteins in the nucleus and in non-nuclear
membranes of rat hepatocytes: electrophoretic analysis
of the endogenous acceptors. Biochem. Inti. 17:59-67.
Goelz, S.E., C. Hession, D. Goff, B. Griffiths, R. Tizard,
B. Newman, G. Chi-Rosso, and R. Lobb. 1990. ELFT: a
gene that directs the expression of an ELAM-1 ligand.
Cell. 63:1349-1356.
Goldberg, A.H., L.C. Yeoman, and H. Busch. 1978. Chromatin-
associated glycoproteins of normal rat liver and
Novikoff hepatoma ascites cells. Cancer Res. 38:1052-
1056.
Gonzlez-Yanes, B., R.B. Mandell, M. Girard, S. Henry, 0.
Aparicio, M. Gritzalli, R.D. Brown, Jr., G.W. Erdos,
and C.M. West. 1989. The spore coat of a fucosylation
mutant in Dictyostelium discoideum. Develop. Biol.
133:576-587.
Goodloe-Holland, C.M. and E.J. Luna. 1987. Purification and
characterization of Dictyostelium discoideum plasma
membranes. Meth. Cell Biol. 28:215-229.
Gregg, J.H. and G.C. Karp. 1978. Patterns of cell
differentiation revealed by L-[3H]fucose incorporation
in Dictyostelium. Exptl. Cell Res. 112:31-46.

159
Hallgreen, P., A. Lundblad, and S. Svensson. 1975. A new
type of carbohydrate-protein linkage in a glycopeptide
from normal human urine. J. Biol. Chem. 250:5312-5314.
Haltiwanger, R.S., G.D. Holt, and G.W. Hart. 1990. Enzymatic
addition of 0-GlcNAc to nuclear and cytoplasmic
proteins. J. Biol. Chem. 265:2563-2568.
Hardy, M.R., R.R. Townsend, and Y.C. Lee. 1988.
Monosaccharide analysis of glycoconjugates by anion
exchange chromatography with pulsed amperometric
detection. Anal. Biochem. 170:54-62.
Hart, G.W., R.S. Haltiwanger, G.D. Holt, and W.G. Kelly.
1989a. Glycosylation in the nucleus and cytoplasm.
Annu. Rev. Biochem. 58:841-874.
Hart, G.W., R.S. Haltiwanger, G.D. Holt, and W.G. Kelly.
1989b. Nucleoplasmic and cytoplasmic glycoproteins.
Ciha Found. Symp. 145:102-118.
Henderson, P.J.F. 1985. Statistical analysis of enzyme
kinetic data. In: Techniques in the Life Sciences, vol.
Bl/II, Protein and Enzyme Biochemistry, BS114:l-48,
Elsevier Scientific Publishers, Ltd., Ireland.
Hirschberg, C.B., M. Perez, M.D. Snider, W.L. Hanneman, J.
Esko, and C.R.H. Raetz. 1982. Autoradiographic
detection and characterization of a Chinese hamster
ovary cell mutant deficient in fucoproteins. J. Cell.
Physiol. 111:255-263.
Hirschberg, C.B. and M.D. Snider. 1987. Topography of
glycosylation in the rough endoplasmic reticulum and
golgi apparatus. Ann. Rev. Biochem. 56:63-87.
Ivatt, R.L., O.P. Das, E.J. Henderson, and P.W. Robbins.
1984. Glycoprotein biosynthesis in Dictyostelium
discoideum: developmental regulation of the protein-
linked glycans. Cell 38:561-567.
Jackson, S.P. and R. Tjian. 1988. O-Glycosylation of
eukaryotic transcription factors: implications for
mechanisms of transcriptional regulation. Cell 55:125-
133.
Kan, F.W.K. and P. Pinto da Silva. 1986. Preferential
association of glycoproteins to the euchromatin regions
of cross-fractured nuclei is revealed by fracture-
label. J. Cell Biol. 102:576-586.

160
Kawasaki, T. and I. Yamashina. 1972. Isolation and
characterization of glycopeptides from rat liver
nuclear membrane. J. Biochem. 72:1517-1525.
Knecht, D.A., D.L. Fuller, and W.F. Loomis. 1987. Surface
glycoprotein, gp24, involved in early adhesion.
Develop. Biol. 121:277-283.
Kohnken, R.E. and E.A. Berger. 1987. Assay and
characterization of carbohydrate binding by the lectin
discoidin I immobilized on nitrocellulose. Biochemistry
25:3949-3957.
Kornfeld, R.H. and V. Ginsburg. 1966. Control of synthesis
of guanosine 5'-diphosphate D-mannose and guanosine 5'-
diphosphate L-fucose in bacteria. Biochim. Biophys.
Acta 117:79-87.
Kornfeld, R. and S. Kornfeld. 1985. Assembly of asparagine-
linked oligosaccharides. Annu. Rev. Biochem. 54:631-
664 .
Kornfeld, K., M.L. Reitman, and R. Kornfeld. 1981. The
carbohydrate-binding specificity of pea and lentil
lectins. Fucose is an important determinant. J. Biol.
Chem. 256:6633-6640.
Kukowska-Latallo, J.F., R.D. Larsen, R.P. Nair, and J.B.
Lowe. 1990. A cloned human cDNA determines expression
of a mouse stage-specific embryonic antigen and the
Lewis blood group (al,3/al,4)fucosyltransferase. Genes
Dev. 4:1288-1303.
Kukuruzinska, M.A., M.L.E. Bergh, and B.J. Jackson. 1987.
Protein glycosylation in yeast. Ann. Rev. Biochem.
56:915-944.
Kumazaki, T. and A. Yoshida. 1984. Biochemical evidence that
secretor gene, Se, is a structural gene encoding a
specific fucosyltransferase. Proc. Natl. Acad. Sci. USA
81:4193-4197.
Lam, T.Y. and C.-H. Siu. 1981. Synthesis of stage-specific
glycoproteins in Dictyostelium discoideum during
development. Develop. Biol. 83:127-137.
Levy-Wilson, B. 1983. Glycosylation, ADP-ribosylation, and
methylation of Tetrahymena histones. Biochemistry
22:484-489.

161
Lewin, A.S., V. Hines, and G.M. Small. 1990. Citrate
synthase encoded by the CIT2 gene of Saccharomyces
cerevisiae is peroxisomal. Mol. Cell Biol. 10:1399-
1405.
Liao, T.-H. and G.A. Barber. 1971. The synthesis of
guanosine 5'-diphosphate L-fucose by enzymes of a
higher plant. Biochim. Biophys. Acta 230:64-71.
Lichtsteiner, S. and U. Schibler. 1989. A glycosylated
liver-specific transcription factor stimulated
transcription of the albumin gene. Cell 57:1179-1187.
Lis, H. and N. Sharon. 1986. Lectins as molecules and as
tools. Annu. Rev. Biochem. 55:35-67.
Loomis, W.F. 1971. Sensitivity of Dictyostelium discoideum
to nucleic acid analogues. Exptl. Cell Res. 64:484-486.
Loomis, W.F. 1987. Genetic tools for Dictyostelium
discoideum. Methods Cell Biol. 28:31-65.
Magner, J.A., W. Novak, and E. Papagiannes. 1986.
Subcellular localization of fucose incorporation into
mouse thyrotropin and free a-subunits: studies
employing subcellular fractionation and inhibitors of
the intracellular translocation of proteins.
Endocrinology 119:1315-1328.
Margolis, R.U., K. Lalley. W.-L. Kiang, C. Crockett, and
R.K. Margolis. 1976. Isolation and properties of a
soluble chondroitin sulfate proteoglycan from brain.
Biochem. Biophys. Res. Comm. 73:1018-1024.
Martin, A., M.-C. Biol, M. Richard, and P. Louisot. 1987.
Purification and separation of two soluble
fucosyltransferase activities of small intestinal
mucosa. Comp. Biochem. Physiol. 87B:725-731.
McMahon, D., M. Miller, and S. Long. 1977. The involvement
of the plasma membrane in the development of
Dictyostelium discoideum. Biochim. Biophys. Acta
465:224-241.
Nicolson, G., M. Lacorbire, and P. Delmonte. 1972. Outer
membrane terminal saccharides of bovine liver nuclei
and mitochondria. Exptl. Cell Res. 71:468-473.
Nez, H.A. and R. Barker. 1976. The metal ion catalyzed
decomposition of nucleoside diphosphate sugars.
Biochemistry 15:3843-3847.

162
Palcic, M.M., L.D. Heerze, M. Pierce, and 0. Hindsgaul.
1988. The use of hydrophobic synthetic glycosides as
acceptors in glycosyltransferase assays. Glycoconjugate
J. 5:49-63.
Paulson, J.C. and K.J. Colley. 1989. Glycosyltransferases.
Structure, localization, and control of cell type-
specific glycosylation. J. Biol. Chem. 264:17615-17618.
Paulson, J.C., J.-P. Prieels, L.R. Glasgow, and R.L. Hill.
1978. Sialyl- and fucosyltransferases in the
biosynthesis of asparaginyl-linked oligosaccharides in
glycoproteins. Mutually exclusive glycosylation by fi-
galactoside a2-6 sialyltransferase and N-
acetylglucosaminide al-3 fucosyltransferase. J. Biol.
Chem. 253:5617-5624.
Perez, M. and C.B. Hirschberg. 1986. Transport of sugar
nucleotides and adenosine 3'-phosphate 5'-
phosphosulfate into vesicles derived from the Golgi
apparatus. Biochim. Biophys. Acta 864:213-222.
Poole, S., R.A. Firtel, and E. Lamar. 1981. Sequence and
expression of the discoidin I gene family in
Dictyostelium discoideum. J. Mol. Biol. 153:273-289.
Potvin, B., R. Kumar, D.R. Howard, and P. Stanley. 1990.
Transfection of a human a-(1,3)fucosyltransferase gene
into Chinese hamster ovary cells. Complication arise
from activion of endogenous a-(1,3)fucosyltransferases.
J. Biol. Chem. 265:1615-1622.
Rajan, V.P., R.D. Larsen, S. Ajmera, L.K. Ernst, and J.B.
Lowe. 1989. A cloned human DNA restriction fragment
determines expression of a GDP-L-fucose:B-D-galactoside
2-a-L-fucosyltransferase in transfected cells. J. Biol.
Chem. 264:11158-11167.
Reeves, R. and D. Chang. 1983. Investigations of the
possible functions for glycosylation in the high
mobility group proteins. Evidence for a role in nuclear
matrix association. J. Biol. Chem. 258:679-687.
Reeves, R., D. Chang, and S.-C. Chung. 1981. Carbohydrate
modifications of the high mobility group proteins.
Proc. Natl. Acad. Sci. USA 78:6704-6708.
Reitman, M.L., I.S. Trowbridge, and S. Kornfeld. 1980. Mouse
lymphoma cell lines resistant to pea lectin are
defective in fucose metabolism. J. Biol. Chem.
255:9900-9906.

163
Richard, M., A. Martin, and P. Louisot. 1975. Evidence for
glycosyl-transferases in rat liver nuclei. Biochem.
Biophys. Res. Comm. 64:108-114.
Ripka, J., A. Adamany, and P. Stanley. 1986. Two Chinese
hamster ovary glycosylation mutants affected in the
conversion of GDP-mannose to GDP-fucose. Arch. Biochem.
Biophys. 249:533-545.
Ripka, J. and P. Stanley. 1986. Lectin-resistant CHO cells:
selection of four new pea lectin-resistant phenotypes.
Somatic Cell Mol. Genet. 12:51-62.
Satir, B.H., C. Srisomsap, M. Reichman, and R.B. Marchase.
1990. Parafusin, an exocytic-sensitive phosphoprotein,
is the primary acceptor for the
glucosylphosphotransferase in Paramecium tetraurelia
and rat liver. J. Cell Biol. 111:901-907.
Scrimgeour, K.G. 1977. Chemistry and Control of Enzyme
Reactions. Academic Press, London and New York.
Sve, A.-P., J. Hubert, D. Bouvier, C. Bourgeois, P. Midoux,
A.-C. Roche, and M. Monsigny. 1986. Analysis of sugar
binding sites in mammalian cell nuclei by quantitative
flow microfluorometry. Proc. Natl. Acad. USA 83:5997-
6001.
Snider, M.D., L.A. Sultzman, and P.W. Robbins. 1980.
Transmembrane location of oligosaccharide-lipid
synthesis in microsomal vesicles. Cell 21:385-392.
Sommers, L.W. and C.B. Hirschberg. 1982. Transport of sugar
nucleotides into rat liver Golgi. A new Golgi marker
activity. J. Biol. Chem. 257:10811-10817.
Spielman, J., S.R. Hull, Z.Q. Sheng, R. Kanterman, A.
Bright, K.L. Carraway. 1988. Biosynthesis of a tumor
cell surface sialomucin. Maturation and effects of
monensin. J. Biol. Chem. 263:9621-9629.
Stanley, P. 1984. Glycosylation mutants of animal cells.
Ann. Rev. Genet. 18:525-552.
Stein, G.S., R.M. Roberts, J.L. Davis, W.J. Head, J.L.
Stein, C.L. Thrall, J. van Veen, and D.W. Welch. 1975.
Are glycoproteins and glycosaminoglycans components of
the eukaryotic genome? Nature 258:639-641.
Stone, D.B., P.M.G. Curmi, and R.A. Mendelson. 1987.
Preparation of deuterated actin from Dictyostelium
discoideum. Meth. Cell Biol. 28:215-229.

164
Stroup, G.B., K.R. Anumula, T.F. Kline, and M.M. Caltabiano.
1990. Identification and characterization of two
distinct a-(1-3)-L-fucosyltransferase activities in
human colon carcinoma. Cancer Res. 50:6787-6792.
Tarentino, A.L., C.M. Gmez, and T.H. Plummer, Jr. 1985.
Deglycosylation of asparagine-linked glycans by
peptide:N-glycosidase F. Biochemistry 24:4665-4671.
Townsend, R.R., M.R. Hardy, D.A. Cumming, J.P. Carver, and
B. Bendiak. 1989. Separation of branched sialylated
oligosaccharides using high-pH anion-exchange
chromatography with pulsed amperometric detection.
Anal. Biochem. 182:1-8.
Tschursin, E., G.R. Riley, and E.J. Henderson. 1989.
Differential regulation of glycoprotein sulfation and
fucosylation during growth of Dictyostelium discoideum.
Differentiation 40:1-9.
Virtanen, I. and J. Wartiovaara. 1976. Lectin receptor sites
on rat liver cell nuclear membranes. J. Cell Sci.
22:335-344.
Weinstein, J., E.U. Lee, K. McEntee, P.-H. Lai, and J.C.
Paulson. 1987. Primary structure of B-galactoside a2,6-
sialyltransferase. Conversion of membrane-bound enzyme
to soluble forms by cleavage of the NH2-terminal signal
anchor. J. Biol. Chem. 262:17735-17743.
West, C.M. and S.A. Brownstein. 1988. EDTA treatment alters
protein glycosylation in the cellular slime mold
Dictyostelium discoideum. Exptl. Cell Res. 175:26-36.
West, C.M. and G.W. Erdos. 1988. The expression of
glycoproteins in the extracellular matrix of the
cellular slime mold Dictyostelium discoideum. Cell
Differ. 23:1-6.
West, C.M. and G.W. Erdos. 1990. Formation of the
Dictyostelium spore coat. Dev. Genet. 11:492-506.
West, C.M., G.W. Erdos, and R. Davis. 1986. Glycoantigen
expression is regulated both temporally and spatially
during development in the cellular slime molds
Dictyostelium discoideum and L. mucoroides. Mol. Cell.
Biochem. 72:121-140.

165
West, C.M. and W.F. Loomis. 1985. Absence of a carbohydrate
modification does not affect the level or subcellular
localization of three membrane glycoproteins in modB
mutants of Dictyostelium discoideum. J. Biol. Chem.
260:13803-13809.
Yamashita, K., T. Mizuochi, and A. Kobata. 1982. Analysis of
oligosaccharides by gel filtration. Meth. Enzymol.
83:105-126.
Yurchenko, P.D. and P.H. Atkinson. 1975. Fucosyl-
glycoprotein and precursor pools in HeLa cells.
Biochemistry 14:3107-3114.
Yurchenko, P.D., C. Ceccarini, and P.H. Atkinson. 1978.
Labeling complex carbohydrates of animal cells with
monosaccharides. Meth. Enzymol. 50:175-204.
Zatz, M. and S.H. Barondes. 1971. Particulate and
solubilized fucosyltransferases from mouse brain. J.
Neurochem. 18:1625-1637.
Zinn, A.B., J.S. Marshall, and D.M. Carlson. 1978.
Preparation of glycopeptides and oligosaccharides from
thyroxine-binding globulin. J. Biol. Chem. 253:6761-
6767.

BIOGRAPHICAL SKETCH
Beatriz Yadira Gonzlez-Yanes was born June 9, 1964, in
Fajardo, Puerto Rico. She graduated from Nuestra Seora del
Pilar High School in Ro Piedras, Puerto Rico, in 1981.
Following high school, she attended the University of Puerto
Rico in Ro Piedras, and earned a Bachelor of Science
degree, Magna Cum Laude, in biology in 1985. In August 1985
she entered graduate school at the University of Florida,
and joined the Department of Anatomy and Cell Biology in
February 1987. She completed the requirements for the
degree of Doctor of Philosophy in December 1991. She has
accepted a postdoctoral research position in the Animal
Science Department at the University of Florida in
Gainesville, Florida.
166

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
A A.;
Christopher M. West, Chair
Associate Professor of Anatomy and
Cell Biology
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Carl M. Feldherr
Professor of Anatomy and Cell
Biology
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree qf Doc
of Philosophy.
William A. Dunn, Jr.i
Assistant Professor of Anatomy and
Cell Biology
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Robert . Cohen
Associate Professor of Biochemistry
and Molecular Biology
This dissertation was submitted to the Graduate Faculty
of the College of Medicine and to the Graduate School and
was accepted as partial fulfillment of the requirements for
the degree of Doctor of Philosophy.
December 1991
Dean, College of Medicine
Dean, Graduate School



83
than 95% of the macromolecular-associated radioactivity from
metabolically labelled cells migrated as authentic fucose.
FP21 is Present in Migrating Slug Stage Cells
To investigate whether FP21 was present also in
developing cells, HL250 amoebae were plated for development
on nuclepore filters and 9 hours after plating the filters
were lifted and cells were metabolically labelled by
placement on 100 pCi of [3H]fucose. After 5 hours of
exposure to [3H]fucose, filters were lifted again, carefully
washed from 3H label, and cells were allowed to continue
development for two more hours. Cells were then harvested,
disaggregated, and fractionated into an S100 and P100
fractions, run in an SDS-PAGE, and the gel cut and the
pieces counted. As shown in figure 3-8, a fucosylated
macromolecule is preferentially fucosylated in the cytosol
and it has the same mobility in SDS-PAGE as FP21. Thus, it
seems that FP21 fucosylation is not restricted to the growth
phase of Dictyostelium.
Discussion
In this chapter I have presented evidence for the
existence of a fucosylated single molecular weight species,
which has been termed FP21 based on its mobility as
determined by SDS-PAGE. This protein cofractionates with


14
of 20+5 sugar units per molecule, fluorescence microscopy
experiments suggested that the majority of the binding
appeared to be associated with nucleoli and nucleoplasmic
ribonucleoprotein elements (Sve et al., 1986).
Interestingly, it was shown by quantitative flow
microfluorometry that nuclei from exponentially growing
cells bound one order of magnitude more fucose-BSA than
nuclei from contact-inhibited cells. In spite of these
results, the authors acknowledge that it is impossible based
on the data to ascribe biological roles to the nuclear
fucose-binding sites or to conclude that they influence
cellular physiology (Sve et al., 1986).
In Dictyostelium discoideum a family of lectins, the
discoidins, has been identified, and discoidin I has been
the most extensively studied isoform. In erythrocyte
agglutination assays, agglutination by discoidin may be
inhibited by galactose, modified galactose residues, L-
fucose, D-fucose, and other sugars, suggesting the existence
of cell-surface sugar-dependent epitopes on erythrocytes
recognized by Discoidin (Barondes and Haywood, 1979).
Although discoidin I was originally thought to be a cell
surface protein involved in cell-cell adhesion, it has since
been established that it is primarily present in the cytosol
(Erdos and Whitaker, 1983). Although galactose and modified
galactose residues are the ligands bound by discoidin I with
highest affinity, it is possible that a fucose-containing


164
Stroup, G.B., K.R. Anumula, T.F. Kline, and M.M. Caltabiano.
1990. Identification and characterization of two
distinct a-(1-3)-L-fucosyltransferase activities in
human colon carcinoma. Cancer Res. 50:6787-6792.
Tarentino, A.L., C.M. Gmez, and T.H. Plummer, Jr. 1985.
Deglycosylation of asparagine-linked glycans by
peptide:N-glycosidase F. Biochemistry 24:4665-4671.
Townsend, R.R., M.R. Hardy, D.A. Cumming, J.P. Carver, and
B. Bendiak. 1989. Separation of branched sialylated
oligosaccharides using high-pH anion-exchange
chromatography with pulsed amperometric detection.
Anal. Biochem. 182:1-8.
Tschursin, E., G.R. Riley, and E.J. Henderson. 1989.
Differential regulation of glycoprotein sulfation and
fucosylation during growth of Dictyostelium discoideum.
Differentiation 40:1-9.
Virtanen, I. and J. Wartiovaara. 1976. Lectin receptor sites
on rat liver cell nuclear membranes. J. Cell Sci.
22:335-344.
Weinstein, J., E.U. Lee, K. McEntee, P.-H. Lai, and J.C.
Paulson. 1987. Primary structure of B-galactoside a2,6-
sialyltransferase. Conversion of membrane-bound enzyme
to soluble forms by cleavage of the NH2-terminal signal
anchor. J. Biol. Chem. 262:17735-17743.
West, C.M. and S.A. Brownstein. 1988. EDTA treatment alters
protein glycosylation in the cellular slime mold
Dictyostelium discoideum. Exptl. Cell Res. 175:26-36.
West, C.M. and G.W. Erdos. 1988. The expression of
glycoproteins in the extracellular matrix of the
cellular slime mold Dictyostelium discoideum. Cell
Differ. 23:1-6.
West, C.M. and G.W. Erdos. 1990. Formation of the
Dictyostelium spore coat. Dev. Genet. 11:492-506.
West, C.M., G.W. Erdos, and R. Davis. 1986. Glycoantigen
expression is regulated both temporally and spatially
during development in the cellular slime molds
Dictyostelium discoideum and L. mucoroides. Mol. Cell.
Biochem. 72:121-140.


135
present at this stage of development, I assayed developing
cells for their ability to incorporate [14C] from GDP-
[14C]fucose into endogenous acceptors. HL250 cells were
plated for development and at slug stage harvested,
disaggregated, fractionated into an S100 and P100, and
assayed for fucosyltransferase activity as described in
Materials and Methods. For comparison, fractions from Ax3
slugs were assayed (table 4-8). As seen with amoebae cells,
HL250 S100 was active whereas the Ax3 S100 did not
incorporate radioactivity. On the other hand, the P100 was
active in both strains. Thus, it seems that the cytosolic
fucosyltransferase is not restricted to the growth phase and
is present in developing cells.
Discussion
The presence of FP21 in the cytosol suggested that a
fucosyltransferase might also be located in the cytosol. An
S100 fucosyltransferase activity was detected which was both
time- and protein concentration-dependent. The activity was
strictly divalent cation dependent. Incorporation of
radioactivity was sensitive to temperature, certain
detergents, and ethanol. A variety of sugars and sugar-
derivatives failed to inhibit activity, except GDP-ii-fucose,
which inhibited in a dose-responsive manner. The activity
could not be sedimented by higher centrifugation force in
the presence of an extensive list of protease inhibitors.


LIST OF FIGURES
Figure Page
2-1. Localization of SP96 in prespore and spore cells 27
2-2. Biosynthesis of GDP-L-fucose 32
2-3. Conversion of GDP-mannose to GDP-fucose by normal
and mutant strains 37
3-1. Incorporation of [3H]fucose into macromolecular
species of the S100 and P100 58
3-2. Proteinaceous nature of FP21 60
3-3. Comparison of S100 and releasable P100 components 63
3-4. Gel filtration chromatography of FP21
glycopeptides 71
3-5. Gel filtration chromatography of PNGase F digests 74
3-6. Gel filtration chromatography of FP21
oligosaccharides 77
3-7. Gel filtration chromatography of P100
glycopeptides 81
3-8. Incorporation of [3H]fucose into macromolecular
species of slug stage cells 85
4-1. Fucosylation of endogenous acceptors by S100
fraction 101
4-2. SDS-PAGE profile of endogenous acceptors
fucosylated in vitro 107
4-3. BioGel P-4 gel filtration chromatography of in
vitro labelled FP21 oligosaccharide 110
4-4. Effect of pH on S100 and P100 fucosyltransferase
activities in the presence of Tween-20 115
ix


28
vesicles using both mAb. However, HL250 shows the same
pattern only when A6.2 is used, in agreement with the
results observed by Western blotting. The fact that
labelling of spores with A6.2 is similar in both strains
indicates that spore coat localization of SP96 is not
affected by the mutation.
Fucose content of normal and mutant strains. Epitope
recognition by mAb 83.5 was inhibited by L-fucose (West et
al., 1986), so the possibility of a defect in fucosylation
in strain HL250 was investigated. Initially, the
macromolecular fucose content of vegetative cells grown in
fucose-free media, and of spores was investigated. Ethanol
insoluble macromolecules were acid hydrolyzed, ethanol
precipitated, and the supernatant deionized and
chromatographed on an alkaline anion-exchange column
equipped with a pulsed amperometric detector. When
authentic [3H]fucose was fractionated in this manner, more
than 95% of the radioactivity eluted at the fucose position
(not shown). Fucose content was found to be negligible in
the mutant, HL250, both in spores and vegetative cells when
compared to the normal strain, Ax3 (table 2-1). However,
when HL250 cells were grown in FM supplemented with 1 mM L-
fucose, they possessed detectable amounts of fucose.
Previous investigators have found by autoradiography that
[3H]fucose incorporation is highest in prespore cells


126


157
Biermann, C.J. 1988. Hydrolysis and other cleavages of
glycosidic linkages in polysaccharides. Adv. Carbohydr.
Chem. Biochem. 46:251-271.
Borts, R.H. and R.L. Dimond. 1981. The a-glucosidases of
Dictyostelium discoideum. Develop. Biol. 87:176-184.
Campbell, C. and P. Stanley. 1984. The Chinese hamster ovary
glycosylation mutants LECH and LEC12 express two novel
GDP-fucose:N-acetylglucosaminide 3-a-L-
fucosyltransferase enzymes. J. Biol. Chem. 259:11208-
11214.
Chen, P.S., Jr., T.Y. Toribara, and H. Warner. 1956.
Microdetermination of phosphorus. Analyt. Chem.
28:1756-1758.
Chu, F.K. 1986. Requirements of cleavage of high mannose
oligosaccharides in glycoproteins by peptide N-
glycosidase F. J. Biol. Chem. 261:172-177.
Darnell, J., H. Lodish, and D. Baltimore. 1986. Molecular
Cell Biology. Scientific American Books, Inc., New
York.
Das, O.P. and E.J. Henderson. 1986. A novel technique for
gentle lysis of eukaryotic cells: isolation of plasma
membranes from Dictyostelium discoideum. Biochim.
Biophys. Acta 736:43-56.
Davis, L.I. and G. Blobel. 1987. Nuclear pore complex
contains a family of glycoproteins that includes p62:
glycosylation through a previously unidentified
cellular pathway. Proc. Natl. Acad. Sci. USA 84:7552-
7556 .
Dimond, R.L., R.A. Burns, and K.B. Jordan. 1981. Secretion
of lysosomal enzymes in the cellular slime mold
Dictyostelium discoideum. J. Biol. Chem. 256:6565-6572.
Erdos, G.W. and C.M. West. 1989. Formation and organization
of the spore coat of Dictyostelium discoideum. Exp.
Mycol. 13:169-182.
Erdos, G.W. and D. Whitaker. 1983. Failure to detect
immunocytochemically reactive endogenous lectin on the
cell surface of Dictyostelium discoideum. J. Cell Biol.
97:993-1000.


149
proposed and the other possible models, additional studies
are needed.
Future Studies
The results presented in this dissertation lay the
ground work for future studies with immense possibilities.
With the help of a strain with a conditional fucosylation
mutation, I was able to recognize the presence of a
fucosylation pathway that otherwise may have gone
undetected. Future research should focus on FP21 or the
cytosolic fucosyltransferase.
Studies on FP21
The first question to be addressed will be the
compartmentalization of FP21 using an independent approach
from that followed in my studies. An initial step would be
to raise antibodies against FP21. A protocol to purify FP21
is being improved in the laboratory, and should prove useful
for this purpose. An antibody against FP21 will be useful
for immunolocalization of the acceptor in fixed cells.
Currently, FP21 is detected by SDS-PAGE as a fucose-labelled
21 kD MW species. In the P100, the presence of other
fucoconjugates of similar MW on SDS-PAGE could mask FP21,
although, as discussed earlier, it appears that FP21 is not
present in the P100 fraction. However, if FP21 was entirely
released into the S100 fraction during cell fractionation


105
or cation accessibility. Tween-20 at a concentration of
0.1% was chosen because it did not inhibit considerably the
activity in the S100 (table 4-1).
I first examined the profile of in vitro endogenous
acceptors by SDS-PAGE analysis. Standard S100 and P100
fractions from HL250 cells were isolated and added to
fucosyltransferase reaction mixtures. After 90 min of
incubation, reactions were boiled in sample electrophoresis
buffer, subjected to SDS-PAGE, and the gels cut and counted.
The profile of radiolabel incorporation in vitro by the S100
was similar to that observed in metabolic labelling
experiments, with more than 70% of the radioactivity
migrating as one discrete peak with a MW of 21 kD (compare
figure 4-2, panel A with figure 3-1). Incorporation into 21
kD MW material by the S100 fraction varied from 70-95% in
different experiments, with the remainder of the
radioactivity migrating near the dye front or top of the
gel. In the P100 fraction, radioactivity distributed in two
peaks, similar to what was seen in metabolically labelled
P100. To determine if in vitro fucosylated protein had the
same apparent MW as metabolically labelled FP21, Ax3 gel-
purified FP21 was mixed with in vitro fucosylated HL250
S100, as described in the figure legend (figure 4-2, panel
B). The migration of in vitro fucosylated material
coincided with that of metabolically labelled FP21. I


68
recognized a band that migrated with a slower mobility than
metabolically labelled FP21 (not shown). The possibility of
FP21 being the lectin discoidin was examined, since it is
primarily present in the cytosol and exhibits weak affinity
for L-fucose (Erdos and Whitaker, 1983; Bartles and Frazier,
1980). Metabolically labelled FP21 from HL250 was
electrophoresed in a 15% SDS polyacrylamide gel, while a
replicate lane was blotted onto nitrocellulose paper and
immunoprobed using an anti-discoidin antiserum. The
antiserum recognized a band of higher MW than FP21 with
mobility slower than trypsin, reproducing results reported
by others where discoidin I was shown to have a slower
mobility than trypsin in 15% SDS polyacrylamide gels
(Kohnken and Berger, 1987). Migration of purified discoidin
in SDS-PAGE differed upon boiling of the sample, migrating
as a tetramer (ca. 100 kD) when samples were not boiled,
whereas metabolically labelled FP21 was found to migrate as
a discrete peak of radioactivity ahead of trypsin regardless
of boiling (Q.H. Yang and C.M. West, unpublished data).
FP21 could be partially purified by HPLC DEAE
chromatography. Metabolically labelled S100 was
fractionated on an anion exchange column and FP21 recovery
monitored by counting of SDS-PAGE slices. FP21 eluted at
0.5 M NH4Ac with an increase of 24-fold the specific
activity relative to the starting sample (data not shown).
Discoidin eluted earlier than FP21 from the HPLC DEAE column


Figure 4-4. Effect of pH on S100 and P100 fucosyltransferase
activities in the presence of Tween-20.
S100 and P100 fractions prepared from vegetative HL250 cells
were assayed for fucosyltransferase activities in the
presence of Tween-20 at different pH values (pH values
determined as described in materials and methods). GDP-
[14C]fucose concentration was 0.36 pM; MgCl2, 5 mM; Tween-20,
0.1%. The assay was carried out for 30 min, and 150 pg and
228 jug of protein were supplied from the S100 and P100,
respectively. Fucose incorporation was calculated from the
amount of TCA-precipitable [14C]radioactivity. Results
expressed as the mean of three determinations s.e.m.
Panel A. Effect of pH on S100 fucosyltransferase activity
in the presence of Tween-20.
Panel B. Effect of pH on P100 fucosyltransferase activity
in the presence of Tween-20.


128
Table 4-5. Utilization of 8-methoxycarbonyloctyl synthetic
acceptors by cytosolic fucosyltransferase activity from Ax3
and HL250.
pmol fucose/mq/h
substrate concentration (mM) Ax3 HL250
fi-gal-R 1 und. n.d.
galfil,3GlcNAcfi-R 0.15 0.38 0.039
galfll,4GlcNAcfi-R 0.15 und. n.d.
Incorporation was determined by the C18 Sep-Pak assay (see
Materials and Methods). Experiment performed by C.M. West and
is the result of one determination, und, undetectable; n.d.,
not determined; -R is -(CH2)0COOCH3; GDP-[ 3H] fucose
concentration, 0.15 jiM.


94
standard fucosyltransferase assay contains 30 pi of extract
(100-350 pg of protein), 0.35 pM GDP-fi-[ 14C]-fucose, 5 mM
MgCl2, 0.25 mM NaF, and 5 mM ATP in 50 mM MES (titrated with
NaOH to a pH of 7.4) in a 50 pi volume incubated at 30 for
the specified amount of time. Endogenous macromolecules
were used as acceptors. To terminate the assay, 1 ml of
ice-cold 15% TCA was added to each sample along with 50 pg
BSA (to serve as carrier protein) and the precipitate
collected on 2.4 cm GF/C glass filters by vacuum filtration,
washed with 10 ml 10% ice-cold TCA, 10 ml acetone, and
counted after air-drying inside the vials for approximately
30 min using 10 ml of Bio-HP LC scintillation fluid
(Fisher). Background was subtracted from experimental
values, and was determined as the amount of TCA-precipitable
radioactivity at time zero; it was usually between 20-40
dpm. When indicated, the disodium EDTA salt was used. In
preliminary trials, 5 mM Mg++ was found to support maximal
activity, and this concentration was used in all the assays
unless indicated. GDP-fucose had been shown previously to
be only slightly decomposed under similar conditions (Nez
and Barker, 1976). Nevertheless, the extent of GDP-
[14C]fucose hydrolysis during the fucosyltransferase assay
was examined by carrying out the reaction for two hours and
separating products on a Dowex-2 formate column by
sequentially eluting with 5 ml of water, 3 M formic acid and
15 M formic acid as described (Sommers and Hirschberg,


Fucose content of normal and mutant spores and vegetative cells.
Table 2-1.
strain
cell type
conditions of qrowth
fucose
qlucose
mannose
Ax 3
spores
581+18
n.d.
n.d.
HL250
spores
0.44
n.d.
n.d.
Ax 3
amoebae
FM
11.0+1.3
83.6+5.0
30.0+0.7
FM + ImM L-fucose
6.5+0.2
86.8+1.4
17.1+1.4
HL250
amoebae
FM
0.075
116.3+3.8
39.5+1.3
FM + ImM L-fucose
2.5 + 0.0
69.0+1.3
40.3+1.4
Spores were collected from fresh fruiting bodies from cells that were plated after
growing in HL-5. Vegetative cells were grown in fucose-free media (FM) or FM
supplemented with fucose. Fucose content was determined as described in Materials and
Methods. Results expressed as nmoles of sugar/ mg of protein as the mean of three
determinations s.e.m.; n.d., not determined.


112
Table 4-3. Comparison between the S100 and P100
fucosyltransferase activities.
condition
pmol fucose/mq protein/45 min
S100 P100
MgCl2*
no cation
EDTA
6.95+2.60
0.18+0.08
0.05+0.03
1.71+0.45
0.68+0.10
0.64+0.10
HL250 amoebae were harvested, filter-lysed, and centrifuged to
prepare S100 and P100 fractions. Fucosyltransferase reaction
mixtures contained 0.36 pM GDP-[14C]fucose, 0.1% Tween-20, and
135 pg or 138 pg of protein from the S100 and P100,
respectively, in the absence of added divalent cations. MgCl2
and EDTA were present at 5 mM. *this is the standard assay,
as described in Materials and Methods. Results are the
average of three measurements + s.e.m.


Figure 4-2. SDS-PAGE profile of endogenous acceptors fucosylated
in vitro.
Panel A. SDS-PAGE profile of S100 and P100 endogenous
acceptors fucosylated in vitro. HL250 cells in logarithmic
growth phase were harvested, homogenized, and an S100 and
P100 prepared. Both fractions were fucosylated in vitro in
the presence of 4.4 pM and 8.8 pM of GDP-[ 14C]fucose for the
S100 and the P100, respectively, 5 mM MgCl for 90 min.
Fucosyltransferase reactions were stopped by boiling in
SDS/fl-mercaptoethanol sample buffer and resolved by 7-20%
linear gradient SDS-PAGE; the gel was sliced into 2.2 mm
pieces and counted. Electrophoresis was from left to right.
138 pg and 150 pg of protein were electrophoresed for the
S100 and P100, respectively. Open circles, P100 endogenous
acceptor species; closed circles, S100 endogenous acceptor
species; arrow, migration of trypsin.
Panel B. Comigration of in vitro fucosylated FP21 with
metabolically labelled FP21 on SDS-PAGE. Ax3 vegetative
cells grown in [3H]fucose were harvested, lysed, and
fractionated into an S100 and P100. The S100 was subjected
to SDS-PAGE and Ax3 FP21 was gel purified and electroeluted.
Independently, HL250 amoebae were harvested, homogenized,
fractionated, and the S100 obtained from the fractionation
fucosylated in vitro in the presence of 0.36 pM GDP-
[14C] fucose, 5 mM MgCl, 140 pg protein, for 30 min.
Reaction was stopped by mixing with gel purified Ax3
[3H]FP21 followed by boiling in SDS/B-mercaptoethanol.
Samples were coelectrophoresed in a 15% SDS-polyacrylamide
gel, and the gel cut into 0.5 cm slices and counted. Open
circles, Ax3 [3H] metabolically labelled gel purified FP21;
closed circles, [14C] in vitro labelled mutant S100 extract.


8
evidence that histones isolated from the macronucleus of
Tetrahymena thermophila appear to contain fucose. These
results were strengthened by metabolic incorporation of
[3H]fucose into histones, which showed that all five
Tetrahymena histones, HI, H2A, H2B, H3, and H4, appear to
contain fucose, with H2A incorporating the highest amount
(Levy-Wilson, 1983). In these studies macronuclei were
isolated to high purity, and highly pure histone
preparations were obtained after extensive washing in high
salt to remove nonhistone proteins. Using Con A, the author
showed specific binding to histones, which was interpreted
as evidence for the presence of mannose residues. However,
Con A has previously been shown to also recognize D-glucose
residues. Based on the extent of fucose incorporation and
its specific radioactivity, the author estimated, as the
lowest estimate, that one in a thousand nucleosomes
contained a fucosylated H2A molecule. To date, the
glycosylation pathway of histones and the oligosaccharide
structure(s) present in histones remain unknown.
Nonhistone proteins. High mobility group (HMG)
proteins are fairly abundant nonhistone chromosomal proteins
classified according to their relative electrophoretic
mobilities (Darnell et al., 1986). HMG proteins undergo a
variety of posttranslational modifications, including
glycosylation. Since they appear to be preferentially


LIST OF TABLES
Table Page
2-1. Fucose content of normal and mutant spores and
vegetative cells 29
2-2. Effect of time on conversion of GDP-mannose to
GDP-fucose 34
2-3. Specific activities of fucose 39
3-1. Distribution of protein and radioactivity in S100
and P100 fractions 56
3-2. Radioactivity recovered in the second S100 after
different P100 treatments 65
3-3. Distribution of markers among S100 and P100
fractions 66
4-1. Effect of different treatments on the cytosolic
fucosyltransferase activity 102
4-2. Failure to sediment S100 fucosyltransferase
activity 104
4-3. Comparison between the S100 and P100
fucosyltransferase activities 112
4-4. Fucosyltransferase activity in Ax3 S100 fraction 122
4-5. Utilization of 8-methoxycarbonyloctyl synthetic
acceptors by cytosolic fucosyltransferase activity
from Ax3 and HL250 128
4-6. Evaluation of the suitability of p-nitro-phenyl
glycosides as acceptors for cytosolic fucosyl
transferase activities in HL250 and Ax3 S100 . 132
4-7. Reduction of fucosylation of acceptor type I analog
by purified FP21 134
4-8. In vitro fucosyltransferase activity of HL250 and
Ax3 slug extracts 136
viii


A mis padres
(To my parents)


53
elimination, with some exceptions: NaOH and NaBH4 were
present at a final concentration of 1 and 4 M, respectively,
and samples were incubated at 80 for 24 h. Reaction was
terminated by diluting the sample twofold with water and
adding acetic acid to a final concentration of 4 M. Borate
salts were removed by methanol evaporation as described
above.
P-4 Column Fractionation
Dry samples from fl-elimination and strong alkaline-
borohydride treatment were resuspended in 800 pi of 50 mM
pyridinium acetate (pH 5.5). Pyridinium acetate was made in
the hood by mixing in water, to a final volume of 2 liters,
8.06 ml pyridine, 2.17 ml glacial acetic acid. After the
previous reagents were dissolved, 0.4 g of sodium azide was
added. The solution had a final pH of approximately 5.5.
The buffer was degassed and stored under chloroform
atmosphere. Samples from pronase digestion were centrifuged
for 5 min at 14k rpm on a Eppendorf table top microfuge and
the supernatant taken for P-4 chromatography.
Oligosaccharides and glycopeptides were fractionated in a
0.9 cm x 1 m BioGel P-4 column (-400 mesh) equilibrated with
50 mM pyridinium acetate (pH 5.5) as the mobile phase.
Prior to pouring, the column was acid washed overnight and
siliconized by coating with 1% (v/v) dimethyldichlorosilane
in benzene for 10 min. The column was calibrated with


136
Table 4-8. In vitro fucosyltransferase activity of HL250 and
Ax3 slug extracts.
pmol fucose/mq protein/30 min
strain SIPO P100
HL250 40.6 36.3
Ax3 <0.1 30.7
Normal and mutant amoebae were allowed to develop, harvested,
disaggregated, filter-lysed, and fractionated into an S100 and
P100. Intact fractions were assayed for fucosyltransferase
activity in the presence of 0.36 pM GDP-[ 14C] fucose, 5 mM
MgCl2, 12-48 pg of protein, for 30 min. Results are the
average of two determinations.


32
4,6-dehy O-GDP
A CONVERSION PATHWAY
O-GDP
B. SALVAGE PATHWAY


148
similar to those found in bacteria (Kornfeld and Ginsburg,
1966), a higher plant (Liao and Barber, 1971), and mammalian
cells (Ripka et al., 1986; Reitman et al., 1980).
However, this model is not the only one that could
account for the data obtained during the course of my
investigation. Alternatively, the presence of FP21 in the
cytosol could be explained by fucosylation in vesicles and
rapid posttranslational transport to the cytosol. The
absence of FP21 and FP21-like glycopeptides in the vesicular
fraction was interpreted earlier as evidence for the absence
of FP21 in the P100. However, it does not rule out the
possibility that FP21 was fucosylated in vesicles and soon
thereafter transported back into the cytosol, but was not
detected because it did not accumulate in the P100. The
presence of a fucosyltransferase in the cytosol would then
be accounted by leakage from the vesicular fraction.
Clearly, this model must then explain the export of FP21
into the cytosol by novel and unknown mechanisms. Another
model that would account for my results is that both FP21
and the fucosyltransferase detected in the S100 leaked into
the supernatant during fractionation. Since known vesicular
markers were shown to remain in the P100, this model would
require FP21, from all fucoconjugates in the P100, to be
released preferentially. It would also require the leakage
of a fucosyltransferase capable of fucosylating FP21.
Nevertheless, in order to distinguish between the model


160
Kawasaki, T. and I. Yamashina. 1972. Isolation and
characterization of glycopeptides from rat liver
nuclear membrane. J. Biochem. 72:1517-1525.
Knecht, D.A., D.L. Fuller, and W.F. Loomis. 1987. Surface
glycoprotein, gp24, involved in early adhesion.
Develop. Biol. 121:277-283.
Kohnken, R.E. and E.A. Berger. 1987. Assay and
characterization of carbohydrate binding by the lectin
discoidin I immobilized on nitrocellulose. Biochemistry
25:3949-3957.
Kornfeld, R.H. and V. Ginsburg. 1966. Control of synthesis
of guanosine 5'-diphosphate D-mannose and guanosine 5'-
diphosphate L-fucose in bacteria. Biochim. Biophys.
Acta 117:79-87.
Kornfeld, R. and S. Kornfeld. 1985. Assembly of asparagine-
linked oligosaccharides. Annu. Rev. Biochem. 54:631-
664 .
Kornfeld, K., M.L. Reitman, and R. Kornfeld. 1981. The
carbohydrate-binding specificity of pea and lentil
lectins. Fucose is an important determinant. J. Biol.
Chem. 256:6633-6640.
Kukowska-Latallo, J.F., R.D. Larsen, R.P. Nair, and J.B.
Lowe. 1990. A cloned human cDNA determines expression
of a mouse stage-specific embryonic antigen and the
Lewis blood group (al,3/al,4)fucosyltransferase. Genes
Dev. 4:1288-1303.
Kukuruzinska, M.A., M.L.E. Bergh, and B.J. Jackson. 1987.
Protein glycosylation in yeast. Ann. Rev. Biochem.
56:915-944.
Kumazaki, T. and A. Yoshida. 1984. Biochemical evidence that
secretor gene, Se, is a structural gene encoding a
specific fucosyltransferase. Proc. Natl. Acad. Sci. USA
81:4193-4197.
Lam, T.Y. and C.-H. Siu. 1981. Synthesis of stage-specific
glycoproteins in Dictyostelium discoideum during
development. Develop. Biol. 83:127-137.
Levy-Wilson, B. 1983. Glycosylation, ADP-ribosylation, and
methylation of Tetrahymena histones. Biochemistry
22:484-489.


67
was found in the S100; mannose-6-phosphatase was only
detectable in the P100. More than 87% of the activity of
the lysosomal enzyme acid phosphatase was detected in the
P100. In Dictyostelium, this enzyme has been shown to be a
soluble lumenal lysosomal protein (Dimond, et.el., 1981).
Further evidence that P100 vesicles are stable, closed
structures comes from earlier studies from the laboratory.
Prespore proteins SP75 and SP96 sediment at 100k x g unless
cells are sonicated (West and Erdos, 1988). These spore
coat proteins are contained in secretory vesicles (Erdos and
West, 1989; West and Erdos, 1988), which appear to be intact
since they are resistant to proteolysis by Proteinase K
unless they are treated with 0.1% Triton X-100 (West et al.,
1986; Q.H. Yang and C.M. West, unpublished data). In other
studies, N-acetyl glucosaminyltransferase activity was
inhibited by EDTA in the P100 fraction only when assayed in
the presence of detergent, suggesting that Golgi-like
vesicles in the P100 were closed (R.B. Mandell and C.M.
West, unpublished observations)
FP21 is Unrelated to Other Known Cytoplasmic Proteins
The possible relationship of FP21 with other known
Dictyostelium proteins was investigated. I examined the
reactivity of antisera raised against discoidin I and II
(Erdos and Whitaker, 1983) and against gp24 (Knecht et al.,
1987) for FP21. On Western blots, antiserum against gp24


146
PlOO-derived glycopeptides, which reinforced the notion that
contaminating P100 material was not the source of FP21.
The presence of a cytosolic fucosylated protein
suggested the existence of a cytosolic fucosylation pathway.
In vitro analysis of subcellular fractions led to the
detection of a fucosyltransferase activity in the cytosol.
The results of this investigation described in Chapter IV
are summarized as follows: 1) Using a fucosylation assay
dependent on endogenous acceptor substrates, I detected
fucosyltransferase activity in cytosolic and vesicular
fractions. 2) Activities from S100 and P100 fractions
differed in the acceptor species fucosylated, their
sensitivities to alkaline pH and divalent cations, and
affinities for GDP-fucose, as evidenced by differences in
apparent Km. I consider these results to be an indication
that the S100 fucosyltransferase did not arise from vesicles
by rupturing during cell fractionation. 3) The cytosolic
fucosyltransferase activity was absolutely dependent on
availability of a non-fucosylated acceptor. Accordingly, in
vitro cytosolic fucosylation could be detected in mutant
extracts, but not in Ax3 fractions. However, cytosolic
fucosyltransferase activity was reconstituted in Ax3
fractions by addition of purified mutant FP21. 4) A
fucosyltransferase activity was detected in the S100 with
the use of synthetic hydrophobic acceptors. Based on the
utilization of these acceptors, the activity was determined


81
fraction
fraction
fraction


150
that would explain its absence in the P100. Alternatively,
a higher MW precursor may exist in the P100.
Immunolocalization of FP21 will help clarify this point. An
antibody against FP21 will help in localizing FP21 in other
compartments of the cell, if present, such as in the nucleus
or nuclear membranes.
Another aspect of interest is the other sugar residues
present in the FP21 oligosaccharide. Due to its size (4.8
glucose units) I suspect the carbohydrate moiety is truly an
oligosaccharide, containing more than one sugar residue.
There is evidence for a peptide-GlcNAc transferase in the
cytosol of rat hepatocytes (Haltiwanger et al., 1990), so it
is possible that the oligosaccharide is O-linked to the
polypeptide backbone via a GlcNAc residue. The fact that
the FP21 radioactive peak released by alkaline hydrolysis
was not symmetric, suggested there is more than one type of
oligosaccharide. The first step would be to separate and
purify the oligosaccharides. For this purpose, a longer P-4
column could be used. Alternatively, the oligosaccharides
could be separated by other chromatographic methods
(Townsend et al., 1989; Beniak et al., 1988). Once
separated, the oligosaccharides can be examined by nuclear
magnetic resonance spectrometry.
Fucosylated FP21 was present in the cytosol of
vegetative and developing cells, but at this moment the
relative levels of expression at different developmental


pmol fuc/mg protein/30 min
pmol fuc/mg protein/30 mil)20
S/v
S/v
031


Figure 3-7. Gel filtration chromatography of P100 glycopeptides.
Vegetative Ax3 cells were metabolically labelled with 2
/Ci/ml of [3H]fucose. A P100 was prepared and the 21 kD MW
material that comigrated with FP21 on SDS-PAGE was
electroeluted and pronase digested. In another assay,
entire P100 was pronase digested in the absence or presence
of 0.1% Triton X-100. After pronase digestion, samples were
subjected to BioGel P-4 fractionation. Arrow identifies the
position with a Rev of 0.51.
Panel A. Glycopeptides from 21 kD MW P100-derived material.
Vo, 34; Vi, 162.
Panel B. Glycopeptides from entire pronase-digested P100.
Vo, 36; Vi, 142.
Panel C. Glycopeptides from entire P100 digested with
pronase in the presence of Triton X-100. Vo, 32; Vi, 146.


I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
A A.;
Christopher M. West, Chair
Associate Professor of Anatomy and
Cell Biology
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Carl M. Feldherr
Professor of Anatomy and Cell
Biology
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree qf Doc
of Philosophy.
William A. Dunn, Jr.i
Assistant Professor of Anatomy and
Cell Biology
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Robert . Cohen
Associate Professor of Biochemistry
and Molecular Biology
This dissertation was submitted to the Graduate Faculty
of the College of Medicine and to the Graduate School and
was accepted as partial fulfillment of the requirements for
the degree of Doctor of Philosophy.
December 1991
Dean, College of Medicine
Dean, Graduate School


3
apparatus. These findings contradict the dogma that
glycosylation is strictly an endoplasmic reticulum- and
Golgi apparatus-dependent event. Two models could account
for the existence of glycoproteins outside the realms of the
secretory pathway; one postulates that lumenal glycosylated
proteins translocate across the membrane back to the
cytosolic space, the other that glycosyltransferases are
localized outside of the lumen of the endoplasmic reticulum
or Golgi apparatus. Evidence has been accumulating in the
past decade that supports the latter model. In this chapter
I shall be concerned with the presence of fucosylated
proteins in compartments topologically discontinuous with
the lumen of the endoplasmic reticulum and Golgi apparatus.
Subsequently, I shall analyze some of the studies that
suggest the presence of fucosyltransferases that would co
compartmentalize with such fucoproteins. In light of the
results presented in this dissertation, a review focussing
on the presence of fucoproteins and fucosyltransferases
outside the secretory pathway will be useful. An excellent
review is available that examines nuclear and cytosolic
glycosylation in general (Hart et al., 1989a).
Fucosylated Macromolecules
The presence of fucosylated proteins in compartments
discontinuous with the secretory pathway has been documented
since the 1970's. Various techniques have been employed in


133
with an apparent of approximately 1 mM for the acceptor
and 1.6 fjM for GDP-fucose (C.M. West, unpublished results).
The similarity in the apparent Km for GDP-fucose for the
type I and FP21 fucosyltransferase suggested that the same
enzyme may be responsible for both reactions. The fact that
HL250 was able to fucosylate so poorly the analog when
compared to Ax3, suggested that the availability of
endogenous substrate (FP21) inhibited incorporation into the
synthetic acceptor, and supported the idea of the same
enzyme fucosylating both substrates. The notion was further
reinforced by reduction of type I analog fucosylation in Ax3
S100 extracts by purified FP21. Fractions 20 and 21 from
the HPLC gel filtration chromatograph (see section above on
reconstitution of Ax3 fucosyltransferase activity by
purified FP21) reduced incorporation of radioactivity into
type I acceptor (table 4-7). Even though reduction was not
strictly proportional, it was evident that it increased with
increasing amounts of FP21. Taken together the results of
this section and the preceding one, it appears that Ax3
possesses a fucosyltransferase activity in the cytosol
capable of fucosylating FP21 and the type I analog acceptor.
Cytosolic Fucosyltransferase Activity is Present in
Migrating Slug Stage Cells
Cytosolic FP21 was detectable by metabolic labelling in
slug stage HL250 cells (Chapter III). Reasoning that a
fucosyltransferase responsible for its modification would be


161
Lewin, A.S., V. Hines, and G.M. Small. 1990. Citrate
synthase encoded by the CIT2 gene of Saccharomyces
cerevisiae is peroxisomal. Mol. Cell Biol. 10:1399-
1405.
Liao, T.-H. and G.A. Barber. 1971. The synthesis of
guanosine 5'-diphosphate L-fucose by enzymes of a
higher plant. Biochim. Biophys. Acta 230:64-71.
Lichtsteiner, S. and U. Schibler. 1989. A glycosylated
liver-specific transcription factor stimulated
transcription of the albumin gene. Cell 57:1179-1187.
Lis, H. and N. Sharon. 1986. Lectins as molecules and as
tools. Annu. Rev. Biochem. 55:35-67.
Loomis, W.F. 1971. Sensitivity of Dictyostelium discoideum
to nucleic acid analogues. Exptl. Cell Res. 64:484-486.
Loomis, W.F. 1987. Genetic tools for Dictyostelium
discoideum. Methods Cell Biol. 28:31-65.
Magner, J.A., W. Novak, and E. Papagiannes. 1986.
Subcellular localization of fucose incorporation into
mouse thyrotropin and free a-subunits: studies
employing subcellular fractionation and inhibitors of
the intracellular translocation of proteins.
Endocrinology 119:1315-1328.
Margolis, R.U., K. Lalley. W.-L. Kiang, C. Crockett, and
R.K. Margolis. 1976. Isolation and properties of a
soluble chondroitin sulfate proteoglycan from brain.
Biochem. Biophys. Res. Comm. 73:1018-1024.
Martin, A., M.-C. Biol, M. Richard, and P. Louisot. 1987.
Purification and separation of two soluble
fucosyltransferase activities of small intestinal
mucosa. Comp. Biochem. Physiol. 87B:725-731.
McMahon, D., M. Miller, and S. Long. 1977. The involvement
of the plasma membrane in the development of
Dictyostelium discoideum. Biochim. Biophys. Acta
465:224-241.
Nicolson, G., M. Lacorbire, and P. Delmonte. 1972. Outer
membrane terminal saccharides of bovine liver nuclei
and mitochondria. Exptl. Cell Res. 71:468-473.
Nez, H.A. and R. Barker. 1976. The metal ion catalyzed
decomposition of nucleoside diphosphate sugars.
Biochemistry 15:3843-3847.


158
Finne, J., T. Krusius, R.K. Margolis, and R.U. Margolis.
1979. Novel mannitol-containing oligosaccharides
obtained by mild alkaline borohydride treatment of a
chondroitin sulfated proteoglycan from brain. J. Biol.
Chem. 254:10295-10300.
Flowers, H.M. 1981. Chemistry and biochemistry of D- and L-
fucose. Adv. Carbohydr. Chem. Biochem. 39:279-345.
Foster, C.S., D.R.B. Gillies, and M.C. Glick. 1991.
Purification and characterization of GDP-L-fuc-N-
acetyl-fi-D-glucosaminide al-3fucosyltransferase from
human neuroblastoma cells. Unusual substrate
specificities of the tumor enzyme. J. Biol. Chem.
266:3526-3531.
Franke, J. and R. Kessin. 1977. A defined minimal medium for
axenic strains of Dictyostelium discoideum. Proc. Natl.
Acad. Sci. USA 74:2157-2161.
Galland, S., A. Degiuli, J. Frot-Coutaz, and R. Got. 1988.
Transfer of N-acetylglucosamine to endogenous
glycoproteins in the nucleus and in non-nuclear
membranes of rat hepatocytes: electrophoretic analysis
of the endogenous acceptors. Biochem. Inti. 17:59-67.
Goelz, S.E., C. Hession, D. Goff, B. Griffiths, R. Tizard,
B. Newman, G. Chi-Rosso, and R. Lobb. 1990. ELFT: a
gene that directs the expression of an ELAM-1 ligand.
Cell. 63:1349-1356.
Goldberg, A.H., L.C. Yeoman, and H. Busch. 1978. Chromatin-
associated glycoproteins of normal rat liver and
Novikoff hepatoma ascites cells. Cancer Res. 38:1052-
1056.
Gonzlez-Yanes, B., R.B. Mandell, M. Girard, S. Henry, 0.
Aparicio, M. Gritzalli, R.D. Brown, Jr., G.W. Erdos,
and C.M. West. 1989. The spore coat of a fucosylation
mutant in Dictyostelium discoideum. Develop. Biol.
133:576-587.
Goodloe-Holland, C.M. and E.J. Luna. 1987. Purification and
characterization of Dictyostelium discoideum plasma
membranes. Meth. Cell Biol. 28:215-229.
Gregg, J.H. and G.C. Karp. 1978. Patterns of cell
differentiation revealed by L-[3H]fucose incorporation
in Dictyostelium. Exptl. Cell Res. 112:31-46.


Figure 4-1. Fucosylation of endogenous acceptors by S100
fraction.
Vegetative HL250 cells were harvested, homogenized, and an
S100 obtained as described in detail in Materials and
Methods. The indicated amount of S100 protein was incubated
in the presence of 0.36 /M GDP-[14C]fucose, 5 mM MgCl2, 5 mM
ATP, 0.25 mM NaF, in 50 mM MES, pH 7.4 for the indicated
amount of time. Fucose incorporation was calculated from
the amount of TCA-precipitable [14C]radioactivity. Results
expressed as the mean of three determinations s.e.m.
Panel A. Effect of time on fucosylation; 159 fug protein.
Panel B. Effect of protein concentration on fucosylation;
30 min assay.


21
labelling experiments, cells were grown for 4-6 doublings in
8-20 pCi/ml (0.10-0.26 pM) of [3H]fucose in FM medium, a
minimal defined medium that lacks fucose (Franke and Kessin,
1977). When appropriate, the medium was supplemented with
L-fucose (Sigma). For development, cells were plated in PDF
buffer (20 mM KC1, 45 mM sodium phosphate, 6 mM MgS04, pH
5.8) on filters as previously described (West and Erdos,
1988) .
Cell Lysis and Fractionation
Logarithmically growing amoebas were harvested and
washed in 50 mM Tris-HCl (pH 7.5) and resuspended to a
concentration of 2xl08 cells/ml in the lysis buffer
consisting of 0.25 M sucrose, 50 mM Tris-HCl buffer (pH 7.5)
supplemented with 1 mM PMSF. All operations were carried
out at 0-4C. Cells were immediately lysed by forced
passage through a 5 pm nuclepore polycarbonate filter, with
pore diameter slightly smaller than the diameter of the
cells (Das and Henderson, 1986). This method routinely
yields more than 99% cell breakage. The lysate was
centrifuged at 100k xg for 1 hour. The supernatant (S100)
was made 2.5% (v/v) in glycerol by addition of 100% glycerol
and either used immediately or saved at -80 without
significant loss of activity for four weeks.


M
Molar Concentration
mCi
Millicurie(s)
mAb
Monoclonal Antibody
MES
2-(N-Morpholino)ethanesulfonic Acid
mg
Milligram(s)
ml
Milliliter(s)
min
Minute(s)
mm
Millimeter(s)
mM
Millimolar
MW
Molecular Weight
pCi
Microcurie(s)
V M
Micromolar
nm
Nanometers
P
probability
PAGE
Polyacrylamide Gel Electrophoresis
pmol
Picomole(s)
PMSF
Phenylmethylsulfonyl Fluoride
PNGase F
Peptide N-glycosidase F
Rev
Relative Elution Coefficient
RNAse B
Ribonuclease B
SDS
Sodium Dodecyl Sulphate
TCA
Trichloroacetic Acid
Tris
Tris(hydroxymethyl)aminomethane
U
Unit(s)
UEA-I
Ulex europaeus Agglutinin
Ve
Elution Volume
xii


20
obtained from Mallinckrodt; formic acid from Fisher; ATP
(disodium salt, catalog number A-5394), niacinamide, NAD+,
NADPH, Trizma base, phenylmethylsulfonyl fluoride, Dowex-1
(1x8, minus 400, chloride form), hexokinase, GDP-mannose,
and Amberlite MB-3 were obtained from Sigma. ATP was stored
frozen at a concentration of 500 mM in 1 mM Tris-HCl (pH
7.4), resulting in a final pH of approximately 5.5. Dowex-1
formate form was made as follows: 1) The column was washed
with 1 M HC1 until pH of eluate was below 2 (as determined
by pH paper). 2) The column was washed with water until pH
was higher than 4.5. 3) Subsequently, the column was washed
with 1 M NaOH until pH was higher than 13. 4) The column
was washed with water as described in step 2. 5) The column
was washed with 3 volumes of 1 M formic acid until pH of
eluate was below 2. 6) Lastly, the column was washed with
water as described in step 2. This procedure was also
followed for regeneration of column.
Strains and Conditions of Growth and Development
Dictyostelium discoideum amoebae were grown on HL-5, a
medium that contains glucose, yeast extract, and proteose
peptone (Loomis, 1971). The axenic strains Ax3 (from F.
Rothman) and HL250 (from W.F. Loomis) were maintained by
passage during logarithmic growth phase. Ax3 is the normal
strain and HL250 is a mutant obtained from Ax3 by N-methyl-
N'-nitro-N-nitrosoguanidine mutagenesis. For metabolic


Figure 3-3. Comparison of S100 and releasable P100 components.
HL250 amoebae were metabolically labelled with 2 pCi/ml of
[3H]fucose, lysed, fractionated into an S100 and P100. The
P100 was sonicated and recentrifuged. After centrifugation,
the resulting supernatant was examined by slicing 7-20%
linear gradient SDS-PAGE into 2.2 mm and counting the gel
pieces (panel B). Included for comparison, is the profile
from the S100 radiolabelled species from the same
preparation run on the same gel (panel A). 50 pg of protein
were electrophoresed for the S100 and P100, respectively.
Arrow, migration of trypsin.


CHAPTER V
SUMMARY AND CONCLUSIONS
Summary of Results
Fucosylation has generally been regarded as a
modification restricted to the secretory compartment,
however, there is evidence of fucosylated macromolecules in
the nucleus and cytosol (see Chapter I). In the present
study, I identified a novel fucosylation pathway in the
cytosol of Dictyostelium discoideum. In the next three
paragraphs a short summary is presented of the results
reported in this dissertation, followed by a proposed model
of fucosylation in the cytosol.
In chapter II the mutant HL250 was characterized as a
conditional fucosylation mutant. The results are summarized
as follows: 1) Spores and vegetative cells from the mutant
strain contained negligible amounts of macromolecular-
associated and total cell fucose when compared to the normal
strain, Ax3, as determined chemically in acid hydrolysates.
2) The phenotype was conditional to growth in the absence of
fucose. When vegetative cells were grown in fucose-
supplemented media, they expressed macromolecular fucose
conjugates. The fucose specific activity of the medium was
not diluted relative to the intracellular fucose. 3)
144


4-5. Effect of GDP-fucose concentration on S100 and P100
fucosyltransferase activities in the presence of
Tween-20 117
4-6. Effect of GDP-fucose concentration on intact S100
and P100 fucosyltransferase activities 120
4-7. Fucosylation of mutant FP21 by Ax3 S100 fraction 126
4-8. Haworth projections of the structures of the synthetic
glycolipid acceptors 130
x


DPM DPM DPM
77
fraction


12
defined by the lack of sedimentation during high-speed
centrifugation. However, this criterion alone is not
sufficient, since cytosolic glycans might arise from
contamination by other organelles. The most convincing
evidence describes fucoconjugates that appear to be
preferentially enriched in the cytosol in relation to other
compartments.
In studies in rat brain, a soluble proteoglycan that
contains novel O-linked mannose-containing oligosaccharides
was recovered in the cytosolic fraction (Margolis et al.,
1976; Finne et al., 1979). The proteoglycan was
characterized as a soluble chondroitin sulfate proteoglycan,
that contained neutral oligosaccharides releasable by mild
alkaline borohydride treatment. The oligosaccharides
contained mannose at their proximal ends, and one
oligosaccharide was proposed to be composed of mannose,
GlcNAc, fucose, and galactose (Finne et al., 1979).
However, the possibility remains that the oligosaccharides
are not integral components of the proteoglycan, but were
associated with other co-purified material. Nevertheless,
the oligosaccharides appeared to be endogenous to the
cytosol and not the result of contamination from other
fractions, since they were present in only trace amounts in
the microsomal or synaptosomal membrane fractions (Finne et
al., 1979). Unfortunately, the function and biosynthetic
pathway of these oligosaccharides remain unknown.


155
is dependent on Mg++ in a fashion similar to the cytosolic
fucosyltransferase, other inhibitors should be tried,
including tunicamycin and N-ethylmaleimide (Galland et al.,
1988; Campbell and Stanley; 1984).


11
appeared to have almost no contamination, structural studies
showing a different glycoconjugate from those found in other
organelles would convincingly argue against contamination.
Nevertheless, these provocative studies are encouraging and
deserve to be pursued further.
Cytosolic Fucoconjugate
The existence of glycoproteins in the cytosol has been
documented in the past and reviewed recently (Hart et al.,
1989a). Studies suggesting the existence of glycoproteins
in the cytosol were based on determinations of lectin
binding sites or biochemical compositional analyses (Hart et
al., 1989a). Some of the negative results reported by those
studies, that relied on lectin binding as confirmation for
the presence of a sugar residue, may be misleading since
lack of binding may reflect a poor choice of lectins to
probe with and not necessarily the absence of a fucose
residue. While there are several fucose-binding lectins
available that serve as useful biochemical tools, they do
not recognize every possible fucose-containing structure.
Lectin binding is dependent on a specific array of sugar
residues, and does not depend solely on the presence of
fucose (Lis and Sharon, 1986). However, as will be
described below, there is one biochemical study that reports
the existence of a cytosolic fucoconjugate. In many cell
fractionation experiments, the cytosolic compartment is


163
Richard, M., A. Martin, and P. Louisot. 1975. Evidence for
glycosyl-transferases in rat liver nuclei. Biochem.
Biophys. Res. Comm. 64:108-114.
Ripka, J., A. Adamany, and P. Stanley. 1986. Two Chinese
hamster ovary glycosylation mutants affected in the
conversion of GDP-mannose to GDP-fucose. Arch. Biochem.
Biophys. 249:533-545.
Ripka, J. and P. Stanley. 1986. Lectin-resistant CHO cells:
selection of four new pea lectin-resistant phenotypes.
Somatic Cell Mol. Genet. 12:51-62.
Satir, B.H., C. Srisomsap, M. Reichman, and R.B. Marchase.
1990. Parafusin, an exocytic-sensitive phosphoprotein,
is the primary acceptor for the
glucosylphosphotransferase in Paramecium tetraurelia
and rat liver. J. Cell Biol. 111:901-907.
Scrimgeour, K.G. 1977. Chemistry and Control of Enzyme
Reactions. Academic Press, London and New York.
Sve, A.-P., J. Hubert, D. Bouvier, C. Bourgeois, P. Midoux,
A.-C. Roche, and M. Monsigny. 1986. Analysis of sugar
binding sites in mammalian cell nuclei by quantitative
flow microfluorometry. Proc. Natl. Acad. USA 83:5997-
6001.
Snider, M.D., L.A. Sultzman, and P.W. Robbins. 1980.
Transmembrane location of oligosaccharide-lipid
synthesis in microsomal vesicles. Cell 21:385-392.
Sommers, L.W. and C.B. Hirschberg. 1982. Transport of sugar
nucleotides into rat liver Golgi. A new Golgi marker
activity. J. Biol. Chem. 257:10811-10817.
Spielman, J., S.R. Hull, Z.Q. Sheng, R. Kanterman, A.
Bright, K.L. Carraway. 1988. Biosynthesis of a tumor
cell surface sialomucin. Maturation and effects of
monensin. J. Biol. Chem. 263:9621-9629.
Stanley, P. 1984. Glycosylation mutants of animal cells.
Ann. Rev. Genet. 18:525-552.
Stein, G.S., R.M. Roberts, J.L. Davis, W.J. Head, J.L.
Stein, C.L. Thrall, J. van Veen, and D.W. Welch. 1975.
Are glycoproteins and glycosaminoglycans components of
the eukaryotic genome? Nature 258:639-641.
Stone, D.B., P.M.G. Curmi, and R.A. Mendelson. 1987.
Preparation of deuterated actin from Dictyostelium
discoideum. Meth. Cell Biol. 28:215-229.


48
it was centrifuged at 100k xg for 1 hour, unless otherwise
specified. The pellet (P100) was resuspended in lysis
buffer by pipetting to the same volume as the supernatant
(S100). For lysing vesicles, the P100 was sonicated using a
Branson Sonifier Cell Disrupter 185.
Slug cells were plated for development as described
earlier (West and Erdos, 1988), and harvested in buffer of
Berger and Clark (as described in West and Brownstein, 1987)
supplemented with 20 mM EDTA. Cells were dissociated in
this buffer by passing 20 times through a long, 9 inches,
pasteur pipette, followed by passing 20 times through a 23-
gauge needle. EDTA was washed by resuspending cells in 50
mM MES, pH 7.4, titrated with NaOH. Cells were resuspended
in lysis buffer and immediately lysed by passage through a 3
pm nuclepore polycarbonate filter (Das and Henderson, 1986).
Cell lysates were then treated as described above for
vegetative cells.
Gel Electrophoresis and Western Blotting
SDS-PAGE was carried out under reducing conditions
essentially as described in West & Loomis (1985). Samples
were resolved by 7-20% acrylamide linear gradient gels or
15% acrylamide gels and, initially, molecular weight
assigned using low MW markers kit (Sigma). In later
experiments, trypsin was used as a molecular weight marker.
Following electrophoresis, the gels were cut immediately


4
these reports, including the direct analysis of carbohydrate
content, use of lectins, use of radiolabelled fucose and/or
a combination of biochemical and morphological techniques.
Nuclear Fucosylated Macromolecules
In contrast to the prevailing dogma, fucosylated
macromolecules have been detected in the nucleus. Such
studies demonstrated lectin binding to nuclear membranes,
chromatin, nuclear proteins, and the nuclear matrix. These
studies generally involved localization by binding of
labelled lectins (fluorescent, radiolabelled, ferritin-
conjugated, etc.) to isolated nuclei. Binding specificity
was assessed by competition of labelling with hapten
inhibitors. Occasionally, these studies were supplemented
by metabolic labelling experiments with radioactive fucose
which is primarily incorporated into fucoproteins with
little metabolism into other molecular species in eukaryotic
cells (Yurchenko et al., 1978).
Nuclear membranes. One early report suggested the
presence of fucose-containing structures on the cytoplasmic
face of isolated bovine nuclei (Nicolson et al., 1972).
Nicolson et al. (1972) found that purified nuclei were
agglutinated by the L-fucose-specific lectin UEA-I from Ulex
europaeus (Lis and Sharon, 1986), and agglutination was
inhibited by incubation in the presence of L-fucose,


61
Chapter IV), and the number of copies of FP21 per cell is
not affected by the mutation, then there may be a maximum of
4 x 105 copies of FP21 per cell.
Evidence that FP21 is Endogenous to the Cytosol
Fucosylation has been shown to occur in the Golgi
apparatus in other organisms (Hirschberg and Snider, 1987;
Kornfeld and Kornfeld, 1985), so I assessed the possibility
that FP21 was a lumenal microsomal protein that leaked
during the P100 and S100 isolation procedure. To guard the
P100 from chemical lysis, the vesicles were prepared in a
cocktail of protease inhibitors that contained additional
inhibitors from those utilized in the standard fractionation
protocol, and which have been developed for the isolation of
various proteolitically sensitive proteins in Dictyostelium
(Goodloe-Holland & Luna, 1987; Stone, et al., 1987). To
address the possibility of mechanical disruption of the
vesicles, the P100 was disrupted by sonication and
recentrifuged at 100k x g for 1 h. Approximately 11% of the
radioactivity was released and the supernatant of this
centrifugation was analyzed by SDS-PAGE as described above
(figure 3-3). Although several radioactive peaks were
present in the supernatant of the second centrifugation, the
radioactivity profile was different from the S100 suggesting
that FP21 is not a protein released by disruption of the
vesicle fraction. Although radioactivity which comigrated


40
of fucose by the salvage pathway is minimal, 1/400 of the
total content. These studies agree with earlier reports on
fucose metabolism, where the majority of fucose in mammalian
cells is derived from the conversion of GDP-mannose to GDP-
fucose (Yurchenko et al., 1978).
Discussion
HL250 failed to express a fucose-dependent epitope
recognized by the mAb 83.5. However, it expressed SP96, one
of the polypeptides that bears the carbohydrate epitope
recognized by 83.5. These results were reproduced by
Western blot (Gonzlez-Yanes et al., 1989) and indirect
immunofluorescence. Interestingly, the immunofluorescence
microscopy studies showed that the mutant is not defective
in its ability to package SP96 in vesicles or in the
targeting of the glycoprotein to the spore coat. Similar
results were reported for other proteins in another
Dictyostelium glycosylation mutant (Aparicio et al., 1990;
West and Loomis, 1985). Measurements of fucose content of
cells and spores demonstrated that the mutant contained
almost undetectable amounts of fucose, in contrast to Ax3
which contained macromolecular-associated fucose in both
cell types.
Once HL250 was identified as having a mutation that
resulted in decreased macromolecular fucose, I tried to
identify the nature of the lesion. It was speculated that


Figure 4-6. Effect of GDP-fucose concentration on intact S100 and P100 fucosyltransferase
activities.
HL250 amoebae were harvested, homogenized, fractionated into an S100 and P100 and
fucosylated in vitro at varying concentrations of GDP-[1 C]fucose. Reactions were
carried out for 30 min in the presence of 5 mM MgCl2, 240 pg of protein for the S100,
and 432 pg of protein for the P100. Fucose incorporation was calculated from the
amount of TCA-precipitable [ 14C]radioactivity. Results expressed as the mean of
three determinations s.e.m.
Panel A. Effect of GDP-fucose concentration on intact S100 fucosyltransferase
activity.
Panel B. Hanes single reciprocal plot ([S]/v vs. [S]) for the S100
fucosyltransferase activity; apparent Km=0.44 pM, apparent V =25.5 pmol/mg
protein/30 min.
Panel C. Effect of GDP-fucose concentration on intact P100 fucosyltransferase
activity.
Panel D. Hanes single reciprocal plot ([S]/v vs. [S]) for the S100
fucosyltransferase activity; apparent Km=28.3 pM, apparent Vnwy=233 pmol/mg protein/30
min.


54
glucose oligomers used as standards (Yamashita et al., 1982)
that were derived from a dextran hydrolysate which was
reduced with NaB3H4 (kindly provided by J. Baezinger).
Twelve drop fractions (approximately 250 pi) were collected
and counted using ScintiVerse LC (Fisher Scientific). All
runs were performed at 37. Recovery varied somewhat
between runs, but it was between 30-80% of the loaded
radioactivity. The void volume (Vo) was determined with
blue dextran (at a concentration of 0.2%), and the inclusion
volume (Vi) with either bromo phenol blue (0.2%) or
[3H]mannose (approximately 2,000 dpm). The relative elution
coefficient (Rev) for each component was determined from the
elution volume (Ve): Rev = (Ve-Vo)/(Vi-Vo).
Results
Analysis of Cellular Fucoproteins by SDS-PAGE
HL250 amoebae were grown in minimal defined media
supplemented with L-[ 3H]fucose. During logarithmic growth
phase, cells were harvested, washed, and resuspended in 0.25
M sucrose buffer supplemented with protease inhibitors.
Immediately, the cells were lysed and the lysate was
clarified from unbroken cells and nuclei by centrifugation
at 2k x g for 5 min. The resulting supernatant from this
spin was centrifuged at 100k x g for 1 hour. Both fractions
were analyzed by SDS-PAGE and the gels stained, cut into 2.2
mm slices, and counted. While total protein distributed in


153
Latallo et al., 1990). Since the fucosyltransferase
reported in my studies appears to be cytosolic, it would be
important to determine what is the relationship between
microsomal and cytosolic fucosyltransferases. All
fucosyltransferases utilize the same sugar nucleotide donor,
GDP-fucose, so it is likely that the GDP-fucose binding site
would be similar for all enzymes. In addition, comparisons
among the fucosyltransferases may reveal important
information regarding intracellular targeting and possible
evolutionary relationships. There is evidence that a
retaining sequence allows glycosyltransferases to remain in
the Golgi apparatus and endoplasmic reticulum (Paulson and
Colley, 1989). The fact that the fucosyltransferase
reported in these studies localizes to the cytosol raises
the possibility that the fucosyltransferase would lack the
targeting and retaining sequences.
In order to compare the cytosolic fucosyltransferase
with the sequenced fucosyltransferases (Kukowska-Latallo et
al., 1990; Goelz et al., 1990), it will be necessary to
sequence the cytosolic fucosyltransferase. The
aforementioned enzymes were cloned using a gene transfer
system in which cloned cDNAs determined the expression of
the enzyme in a recipient host that did not express such
activity. It could be possible to do the same for the
cytosolic fucosyltransferase, using the type I analog
synthetic acceptor to screen for activity of transfected


pmol fucose/ 30 min
o no co cn
time (min)
pmol fucose/ mg protein
101


154
clones. A suitable host to express the cytosolic
fucosyltransferase cDNA would be a mutant Dictyostelium
strain, although there are no such mutants available at the
present. On the other hand, yeast could be used, since it
has been shown yeast cells do not carry out fucosylation
(Kukuruzinska et al., 1987). One of the complications that
may arise in trying to screen for clones expressing the
cytosolic fucosyltransferase is the transfection of
microsomal fucosyltransferases. It remains to be determined
whether the P100 fucosyltransferase activity is capable of
fucosylating the type I analog. If the activity in the P100
does not utilize the type I analog [galfl( 1,3) GlcNAc/1-8-
methoxycarbonyloctyl] as acceptor, then clones can be
screened using the synthetic acceptor. However, if there
are fucosyltransferases in the P100 that utilize the type I
analog as acceptor, it will be necessary to differentiate
the activity in vitro before the transfection experiments.
P100 extracts will be assayed for the ability to fucosylate
galfil,3GlcNAcfi-8-methoxycarbonyloctyl. If the fraction
fucosylates the acceptor, the sensitivity to cations will be
examined for possible differences with the cytosolic
fucosyltransferase activity. If the fraction is active in
the presence of EDTA, fucosyltransferase positive clones may
be screened in the presence and absence of Mg++. Those that
express activity only in the presence of Mg++ may represent
positive clones. In the event that the activity in the P100


99
transfer [ 14C ] from GDP-[ 14C ] fucose into TCA precipitable
endogenous material. As shown in figure 4-1, the cytosolic
activity was dependent on time and protein content. The
cytosolic fucosyltransferase activity had the properties of
being enzyme-mediated. Table 4-1 shows the effect of
boiling, denaturants and temperature on the cytosolic
fucosyltransferase activity. While the non-ionic detergent
Triton X-100 inhibited all activity, Tween-20 was only
slightly inhibitory. 30 was the optimal temperature of
those tested (22, 30, and 37). Consistent with an
enzyme-mediated process, only unlabelled GDP-fi-fucose was
able to inhibit incorporation of radioactivity proportionate
to its relative concentration (GDP-a-fucose was without
effect), demonstrating the stereospecificity of the enzyme
(lower section of table 4-1). It also implies a
fucosyltransferase that catalyses an alpha-fucosyl linkage
is being assayed.
Earlier I explored the possibility of FP21 arising by
contamination from the vesicular fraction. The same
question was asked about the fucosyltransferase activity in
the cytosol, since known fucosyltransferases are Golgi
enzymes and activity was detectable in the P100 (see next
section). Hence, I tried to deplete the S100 of
fucosyltransferase activity by centrifuging at 170k x g for
2.5 hours (instead of 1 hr at 100k x g). This step was used
to sediment any population of small or low density vesicles


139
overall fucosylation process in the P100 fraction, including
transport into the intact vesicles. The apparent Michaelis
constants for the P100 activity in the presence of Tween-20
and in the intact fraction were 38.2 pM and 28.3 /jM,
respectively. The similarity of the apparent Km values
suggested that the GDP-fucose transporter in the P100
vesicles had a similar or lower K relative to that of the
m
bulk P100 fucosyltransferase activity, since if it had a
much higher apparent Km, GDP-fucose transport would have
been rate limiting. The GDP-fucose transporter from rat
liver Golgi-enriched vesicles has an apparent Km of 7.5 /jM
(Sommers and Hirschberg, 1982). The apparent Km for the
cytosolic fucosyltransferase is relatively low compared to
that of the bulk P100 activity. Though the relative
concentrations of GDP-fucose in the cytosol and vesicles are
not known, vesicles have the ability to concentrate GDP-
fucose relative to the outside (Perez and Hirschberg, 1986).
Thus it is not unreasonable to predict that a cytosolic
fucosyltransferase would have a higher affinity for GDP-
fucose since the concentration of GDP-fucose is probably
lower in the cytosol than in the vesicles.
The studies described in this chapter concerning the
P100, characterized the bulk activity in the fraction and
cannot differentiate among different fucosyltransferases
that may be present. The fucosyltransferase activity in the
P100 may be a product of different fucosyltransferases with


42
producing GDP-fucose via the salvage pathway and that the
rest of the fucosylation machinery is probably normal.
Earlier studies have reported the phenotypic reversion of
mammalian mutants with a defective GDP-mannose to GDP-fucose
conversion pathway when the cells were supplied with
extracellular fucose (Ripka and Stanley, 1986; Reitman et
al., 1980). However, Ripka and Stanley (1986) used the
recovery of lectin sensitivity as a marker for phenotypic
reversion, but did not report measuring the fucose content
of the cells or show the data for lectin binding compared to
the parental strain. Reitman et al. (1980) showed that a
mouse lymphoma cell line which has a defect in the
conversion of GDP-mannose to GDP-fucose was defective in pea
lectin binding compared to the parental cell line. The
ability to bind pea lectin was restored to wild type
parental cell line levels after culturing in 10 mM fucose.
Fucose is an important determinant in the carbohydrate
binding specificity of pea lectin (Kornfeld et al., 1981).
It is important to note that the mutant mouse lymphoma cell
line had approximately one fifth the amount of fucose and
one third the number of high affinity lectin-binding sites
as the parental line, indicating that mutant cells were
salvaging fucose from the medium, or that the mutation was
only partial. The evaluation for phenotypic reversion of
HL250 is more rigorous since it demands the expression of a


Figure 3-2. Proteinaceous nature of FP21.
Approximately 1,500 dpm of metabolically [3H]fucose-labelled
FP21 from HL250 was gel purified, as described in Materials
and Methods, and subjected to either a mock or pronase
digestion. Resulting digests were electrophoresed on a 15%
SDS polyacrylamide gel, which was then sliced into 0.5 cm
pieces and counted. Open circles, FP21; closed circles,
FP21 digested with pronase.


60
gel slice


43
carbohydrate epitope and measures total levels of fucose
from a cell that was grown in fucose-free media.
The specific activity of the medium was compared to the
specific activity of the intracellular macromolecular
fucose. Consistent with a lesion in the GDP-mannose to GDP-
fucose conversion pathway, the mutant cells relied on
extracellular fucose as their only fucose source. In
contrast, extracellular fucose only contributed to a small
fraction of the total Ax3 fucose pool. This is useful
because it means that radioactivity from cells grown in
[3H]fucose can be used as a direct measure of fucosylation
in HL250.
In conclusion, even though HL250 has a severe
glycosylation lesion which renders it unable to carry out
fucosylation when grown in the absence of fucose, the strain
is able to grow, develop, and form spores. To my knowledge,
there are no previous reports in the literature of
eukaryotic cells defective in the GDP-mannose to GDP-fucose
conversion pathway that can survive in fucose-free media.
The fucosylation mutants reported are cell lines that have a
functional salvage pathway maintained in culture in the
presence of animal serum, so they can synthesize GDP-fucose
from the fucose present in the cell culture media (Ripka and
Stanley, 1986; Reitman et al., 1980). For this reason,
these investigators were unable to totally deprive the
mutants of fucose, as I am able to do with Dictyostelium.


162
Palcic, M.M., L.D. Heerze, M. Pierce, and 0. Hindsgaul.
1988. The use of hydrophobic synthetic glycosides as
acceptors in glycosyltransferase assays. Glycoconjugate
J. 5:49-63.
Paulson, J.C. and K.J. Colley. 1989. Glycosyltransferases.
Structure, localization, and control of cell type-
specific glycosylation. J. Biol. Chem. 264:17615-17618.
Paulson, J.C., J.-P. Prieels, L.R. Glasgow, and R.L. Hill.
1978. Sialyl- and fucosyltransferases in the
biosynthesis of asparaginyl-linked oligosaccharides in
glycoproteins. Mutually exclusive glycosylation by fi-
galactoside a2-6 sialyltransferase and N-
acetylglucosaminide al-3 fucosyltransferase. J. Biol.
Chem. 253:5617-5624.
Perez, M. and C.B. Hirschberg. 1986. Transport of sugar
nucleotides and adenosine 3'-phosphate 5'-
phosphosulfate into vesicles derived from the Golgi
apparatus. Biochim. Biophys. Acta 864:213-222.
Poole, S., R.A. Firtel, and E. Lamar. 1981. Sequence and
expression of the discoidin I gene family in
Dictyostelium discoideum. J. Mol. Biol. 153:273-289.
Potvin, B., R. Kumar, D.R. Howard, and P. Stanley. 1990.
Transfection of a human a-(1,3)fucosyltransferase gene
into Chinese hamster ovary cells. Complication arise
from activion of endogenous a-(1,3)fucosyltransferases.
J. Biol. Chem. 265:1615-1622.
Rajan, V.P., R.D. Larsen, S. Ajmera, L.K. Ernst, and J.B.
Lowe. 1989. A cloned human DNA restriction fragment
determines expression of a GDP-L-fucose:B-D-galactoside
2-a-L-fucosyltransferase in transfected cells. J. Biol.
Chem. 264:11158-11167.
Reeves, R. and D. Chang. 1983. Investigations of the
possible functions for glycosylation in the high
mobility group proteins. Evidence for a role in nuclear
matrix association. J. Biol. Chem. 258:679-687.
Reeves, R., D. Chang, and S.-C. Chung. 1981. Carbohydrate
modifications of the high mobility group proteins.
Proc. Natl. Acad. Sci. USA 78:6704-6708.
Reitman, M.L., I.S. Trowbridge, and S. Kornfeld. 1980. Mouse
lymphoma cell lines resistant to pea lectin are
defective in fucose metabolism. J. Biol. Chem.
255:9900-9906.


145
Mutant extracts were incapable of carrying out the
conversion of GDP-mannose to GDP-fucose in vitro. In other
organisms, this pathway is the sole pathway of GDP-fucose
synthesis in the absence of extracellular fucose. The low
fucose biochemical phenotype can be explained by the model
that the conversion pathway is defective. HL250 cells and
extracts in vitro can still fucosylate, showing that GDP-
fucose transport and fucosyltransferase(s) are still active.
Although the possibility remains that there are other
genetic defects in this mutagenized strain, there is no
reason to suspect that other genes of the fucosylation
pathway have been affected.
After determining that the source of macromolecular
fucose in HL250 grown in normal medium was derived from
extracellular fucose, I examined the compartmentalization of
fucosylation. The results of the experiments described in
Chapter III show the existence of a fucosylated protein in
the cytosol and are summarized as follows: 1) The major
fucosylated species in the S100 is FP21. It is present in
both Ax3 and HL250. 2) Analysis of FP21 revealed that the
oligosaccharide in FP21 was O-linked with a size of 4.8
glucose units. 3) FP21 appears to be endogenous to the
cytosol, and not derived from a sedimentable compartment
during preparation of the extracts. 4) Glycopeptides
released from FP21 by pronase digestion differ from 21 kD MW


S100 (DPM)
58
600
500
400
300
200
100
0
P100 (DPM)


Table 4-1. Effect of different treatments on the cytosolic fucosyltransferase activity.
pmol fucose/mq/30 min
condition
GDP-r Clfucose (pM)
control
experimental
relative activity
no cation
0.36
17.3
0.97
0.06
5 mM EDTA
1.32
9.23
0.06
0.01
boiled
1.32
9.23
<0.01
<0.01
0.15% Triton X-100
0.36
18.5
0.22
0.02
0.1% Tween-20
0.36
16.6
10.5
0.63
10% ethanol
6
54.4
3.00
0.06
22
6
31.2
22.8
0.73
37
6
35.3
16.1
0.46
6 pM GDP-a-fucose
0.36
20.8
21.3
1.02
6 pM GDP-fi-fucose
0.36
20.8
1.09
0.05
HL250 vegetative cells were harvested, filter-lysed, and fractionated into an S100 and
P100 fractions. The S100 was assayed for fucosyltransferase activity in the presence
of 100-300 pg protein, the noted concentration of GDP-[ 14C ] fucose, 5 mM ATP, 0.25 mM NaF,
5 mM MgCl2, in 50 mM MES, pH 7.4, for 30 min at 30 (standard conditions), unless other
conditions are specified. Additives were added at the noted concentration; when EDTA
was present divalent cations were omitted from the reaction mixture; the experiments
performed at 22 contained 10 mM MgCl2 and 10 mM MnCl ; those at 37 were supplemented
with 10 mM MgCl2. The results are a compilation of different experiments carried out
at the given GDP-[ C]fucose concentrations. Relative activity refers to the activity
exhibited under experimental conditions compared to control conditions. Fucose
incorporation into endogenous acceptors was calculated from the amount of TCA-
precipitable [14C]radioactivity. Results are the average of two determinations.
102


Figure 4-8. Haworth projections of the structures of the
synthetic glycolipid acceptors.


113
fractions were assayed for fucosyltransferase activity at
various pH values as described in Materials and Methods.
The pH profiles of the S100 and P100 fucosyltransferase
activities in the presence of Tween-20 are shown in figure
4-4. Activity was maximal for the S100 from pH 6.8 to 7.8
and for the P100 from pH 6.4 to 7.8. At pH 9.6 the activity
in the S100 was inhibited more than 20-fold compared to
maximum (p<0.05), whereas the activity in the P100 was only
inhibited threefold (p<0.05). The pH-dependent activity
profiles were not affected by the exclusion of Tween-20 (not
shown). Thus, while the general profile is similar for the
activities from the S100 and P100, the cytosolic activity
was more sensitive to alkaline pH than the P100 activity.
The dependence of S100 and P100 fucosyltransferase
activities on the GDP-fucose concentration in the presence
of Tween-20 was also studied. S100 and P100 fractions were
prepared from HL250 vegetative cells and assayed for
fucosyltransferase activity under standard conditions, at
increasing concentrations of GDP-fucose in the presence of
0.1% Tween 20 (figure 4-5). The apparent Km for GDP-fucose
was 1.7 pM and 38.2 pM for the S100 and P100, respectively.
The apparent Vmax for the S100 was 42.7 pmol fuc/ mg protein/
30 min and 122 pmol fuc/ mg protein/ 30 min for the P100.
As evidenced by the lower apparent Km (22-fold lower), the
affinity of the S100 fucosyltransferase for GDP-fucose was
higher than the one from the P100 activity. This accounts


152
to be examined more carefully. It is possible that the
enzyme fucosylates the cytosolic acceptor, FP21, while being
membrane-associated, but facing the cytosol. There is a
membrane-associated glycosyltransferase that utilizes
cytosolic acceptors (Haltiwanger et al., 1990). If a
portion of the fucosyltransferase pool was to partition to
the outside of the vesicles, the activity on intact vesicles
should fucosylate the type I analog. Fucosylation of this
synthetic acceptor by intact P100 vesicles should be
dependent on added Mg++.
Another approach to study the cytosolic
fucosyltransferase is to clone and seguence the enzyme,
avoiding purification of the protein. To date, only two
fucosyltransferases have been sequenced, one encodes an
al,3/l,4fucosyltransferase and the other an
al,3fucosyltransferase (Kukowska-Latallo et al., 1990; Goelz
et al., 1990). There is 57% identity between the two
enzymes at the C-terminus, for a stretch of two-thirds the
length of the protein (Goelz et al., 1990). Both enzymes
appear to be type II transmembrane proteins, each composed
of a short amino-terminal cytoplasmic domain with no
discernible signal sequence, and a putative single
transmembrane signal/anchor domain (Kukowska-Latallo et al.,
1990; Goelz et al., 1990). The sequenced
fucosyltransferases possessed N-linked glycosylation sites,
and one of them was shown to be a glycoprotein (Kukowska-


121
Fucosyltransferase Activity Cannot be Detected In Vitro in
Ax3 S100 Extracts
Normal growing cells expressed [3H]fucose metabolically
labelled FP21 and the main glycopeptide and oligosaccharide
products released by pronase digestion and fl-elimination,
respectively, were indistinguishable from those derived from
the mutant HL250 (Chapter III). All the studies reported
above on the cytosolic fucosyltransferase activity were
carried out in HL250 because in vitro transfer of fucose
from GDP-fucose to endogenous acceptors cannot be detected
in Ax3 S100 fractions by TCA-precipitation, SDS-PAGE
analysis, or C18 SepPak fractionation (see Materials and
Methods for description of methods). A plausible
explanation for the lack of activity in the Ax3 would be
that the S100 from Ax3 contained an inhibitor for the
activity. To investigate this possibility, I performed
several mixing experiments. Vegetative Ax3 and HL250 cells
were harvested, filter-lysed and fractionated into an S100
and P100. The S100 fractions were assayed individually or
mixed in different ratios before assaying for
fucosyltransferase activity. The fractions were used intact
or desalted prior to assay (table 4-4). There was no
evidence for an inhibitor since experiments in which S100
fractions from Ax3 and HL250 were mixed in different ratios
showed activity commensurate to the HL250 contribution
(table 4-4). Dilution of labelled GDP-fucose with
endogenous unlabelled GDP-fucose is not an explanation


5
suggesting that there were fucose-containing membrane-bound
oligosaccharides on the outer nuclear membrane. Similar
results were obtained with the mannose- and glucose-specific
lectin concanavalin A (Con A), which was also found to
agglutinate purified nuclei (Nicolson et al., 1972).
However, subsequent work by another group revealed that, as
evidenced by electron microscopic examination, ferritin-Con
A appeared to stain only damaged nuclei (Virtanen and
Wartiovaara, 1976) raising some concerns about the studies
by Nicolson et al. (1972). Likewise, in the aforementioned
studies (Nicolson et al., 1972), integrity of the isolated
nuclei was not determined, allowing for the possibility that
the lectin was interacting with lumenal fucoproteins that
escaped organelles during nuclei isolation or with
contaminating fucoprotein-containing membranes, such as
plasma membrane.
Chromatin. The existence of chromatin associated
fucose-containing proteins has been suggested by several
laboratories. Early on, Stein et al. (1975) reported the
presence of [3H] labelled glycoconjugates in purified
chromatin from HeLa cells grown in the presence of
[3H]fucose. Although the label was not examined to
corroborate its presence as fucose, in the case of
radiolabelling macromolecules with radioactive sugar
precursors, fucose is an excellent candidate because it has


Figure 4-7. Fucosylation of mutant FP21 by Ax3 S100 fraction.
HL250 amoebae were harvested, homogenized, and fractionated
into S100 and P100 fractions. Unlabeled HL250 S100 was
mixed with an aliquot of in vitro [ 14C ] fucosylated HL250
S100. FP21 was purified from the S100 fraction by (NH4)S02
precipitation, QAE-ion exchange chromatography, and HPLC gel
filtration (described in detail in Materials and Methods).
[14C]FP21 eluted in fractions 20 and 21 of the HPLC gel
filtration chromatography, as confirmed by SDS-PAGE. Ax3
S100 was added to an aliquot of fraction 21 that was
previously dried down in the bottom of an assay tube in a
vacuum centrifuge to serve as acceptor in the in vitro Ax3
S100 fucosyltransferase reaction. The reaction mixture
contained 0.15 fjM GDP-[3H] fucose, 5 mM MgCl2, 349 pg of Ax3
S100 protein, was incubated for 60 min and the reaction
stopped by boiling in SDS/fi-mercaptoethanol electrophoresis
buffer. Sample was resolved on a 15% SDS-polyacrylamide
gel, which was cut into 0.5 cm slices and counted. No [ H]
radioactivity was incorporated into any MW species when the
wild type Ax3 S100 fraction was incubated in the absence of
added FP21 from mutant source (not shown). Open circles,
[14C] label derived from in vitro labelled purified FP21
from HL250; closed circles, [3H] radioactivity from in vitro
fucosylation by the Ax3 S100 fraction.


15
macromolecule may serve as a ligand for it or that another
lectin, possibly from the discoidin family, will be present
in the cytosol with fucose-binding capability.
Fucosyltransferases
Intracellular fucosyltransferases identified to date,
are localized in the lumen of the Golgi or possibly,
endoplasmic reticulum. Thus the presence of fucoproteins in
the nucleus and cytosolic compartments poses questions
pertaining to the mode of synthesis and/or intracellular
transport of such glycoproteins. Early on, Kawasaki and
Yamashina (1972) theorized, based on metabolic labelling,
that nuclear membrane glycoproteins were not synthesized in
and transported from microsomes, but were made in the
nuclear membrane or its vicinity. A few reports have
suggested the presence of glycosyltransferases in the
nucleus, nuclear membranes, or cytosol that may be involved
in the modification of several glycoproteins; however, none
of these glycoproteins was reported to contain fucose
(Richard et al., 1975; Galland et al., 1988, Haltiwanger et
al., 1990).
Though there is evidence for nuclear and cytosolic
fucosylated macromolecules, to my knowledge there are no
reports of nuclear or cytosolic fucosyltransferases.
Louisot and collaborators reported the purification and
separation of two soluble fucosyltransferase activities from


the figures contained herein. Particularly, I would like to
thank Drs. Kelly Selman and Gillian Small for the advice and
friendship they have provided during these years. I am also
grateful to have been in contact with other students, past
and present, which have made the time spent in the
department more enjoyable. I am grateful to Dan Tuttle for
reviewing my dissertation. I would also like to acknowledge
all the people that have worked in the laboratory of Dr.
West, including Scherwin Henry, for making the time spent in
the laboratory more rewarding and Mel Fields for his
assistance in some experiments.
I would like to acknowledge my professors at the
University of Puerto Rico, who encouraged me to pursue a
career in biological research.
I appreciate all the friends I have made in Gainesville
in the past years and would like to acknowledge their
importance in my education. Especially, I thank Hans van
Oostrom for his moral support, encouragement, and for
teaching me about the wonders of the electronic world and
those rare creatures called computers.
Finally, and most importantly, I would like to thank my
family. I thank my grandparents, especially Abuela Emma and
Abuela Yuya, for the faith they have showed in me and for
their prayers, which have helped me through my graduate
education. My brothers Germn, Omar, and Carlos have been
great sources of happiness and pride. Lastly, I thank my
iv


33
al., 1989), and vegetative cells grown in 1 mM L-fucose had
detectable amounts of macromolecular-associated fucose
(table 2-2). These results suggested that HL250 had a
normal salvage pathway, but a defect in the GDP-mannose to
GDP-fucose conversion pathway.
In vitro conversion of GDP-mannose to GDP-fucose.
Conversion of GDP-mannose to GDP-fucose has been reported in
bacteria, in higher plant cells, and mammalian cells
(Kornfeld and Ginsburg, 1966; Liao and Barber, 1971; Ripka
et al., 1986). It was assumed that Dictyostelium would
share this ability with other species, so I assayed if cell
extracts in vitro were able to convert GDP-mannose to GDP-
fucose. High speed supernatants from Ax3 and HL250 were
assayed in vitro for their ability to convert GDP-
[3H]mannose into GDP-[ 3H] fucose. Cells were homogenized and
a 100k x g soluble fraction (S100) assayed as described by
Ripka et al. (1986) and the effects of time, varying protein
and GDP-mannose concentrations were examined. Table 2-2
shows that the conversion activity in Ax3 was proportional
to the time of incubation. In contrast, mutant extracts
showed negligible activity. Ax3 and HL250 cytosols were
mixed to determine if a soluble inhibitor of GDP-mannose to
GDP-fucose conversion activity was present. When equal
amounts of extracts were mixed, activity was commensurate
with the Ax3 contribution (table 2-2) indicating HL250 does


71


39
Table 2-3. Specific activities of fucose.
cpm/nmol
fucose
strain
fucose concentration
medium
macromolecular
HL250
50 pM
1.lxlO5
9.5x10*
Ax3
0.1 pM
1.8xl07
4.7x10*
Cells were grown for 3 days in FM media in the presence of
6xl06 dpm/ml of [3H]fucose supplemented with non-radioactive
fucose to yield the noted concentration of fucose in the
media. Specific activity was determined as described in
Materials and Methods.


165
West, C.M. and W.F. Loomis. 1985. Absence of a carbohydrate
modification does not affect the level or subcellular
localization of three membrane glycoproteins in modB
mutants of Dictyostelium discoideum. J. Biol. Chem.
260:13803-13809.
Yamashita, K., T. Mizuochi, and A. Kobata. 1982. Analysis of
oligosaccharides by gel filtration. Meth. Enzymol.
83:105-126.
Yurchenko, P.D. and P.H. Atkinson. 1975. Fucosyl-
glycoprotein and precursor pools in HeLa cells.
Biochemistry 14:3107-3114.
Yurchenko, P.D., C. Ceccarini, and P.H. Atkinson. 1978.
Labeling complex carbohydrates of animal cells with
monosaccharides. Meth. Enzymol. 50:175-204.
Zatz, M. and S.H. Barondes. 1971. Particulate and
solubilized fucosyltransferases from mouse brain. J.
Neurochem. 18:1625-1637.
Zinn, A.B., J.S. Marshall, and D.M. Carlson. 1978.
Preparation of glycopeptides and oligosaccharides from
thyroxine-binding globulin. J. Biol. Chem. 253:6761-
6767.


65
Table 3-2. Radioactivity recovered in the second S100 after
different P100 treatments.
condition
% radioactivity recovered
untreated
sonicated
mixed*
2.2%
11.3%
1.4%
Cells were labelled in vivo by growing in 2 pCi/ml of
[3H]fucose in FM media for 3 days, lysed, and fractionated
into an S100 and P100. The P100 was then subjected to
different treatments and recentrifuged at 100k x g for 1 hr.
The S100 from this second centrifugation was analysed for
radioactivity. *P100 from metabolically labelled cells was
mixed with unlabelled post-nuclear supernatant from cells
grown in FM media and recentrifuged at 100k x g for 1 hr.


6
been shown to be incorporated as such, with minimal
metabolizing of the label (Yurchenko et al., 1978). In
these studies the authors examined a very pure chromatin
preparation, with essentially no contamination from nuclear
or plasma membranes. The authors argued against plasma
membrane contamination based on mixing experiments, in which
radiolabeled cell-surface trypsinates were combined with
unlabeled nuclear preparations prior to chromatin isolation.
The fucosylated chromatin-associated macromolecules were
deemed to be fucoproteins based on their sensitivity to
pronase digestion. Unfortunately, these studies have not
been pursued further, and many questions remain unanswered
with respect to their structure and biosynthesis.
Chromatin-associated fucoproteins were also detected in
normal rat liver and Novikoff hepatoma ascites cells
(Goldberg et al., 1978) using the L-fucose-specific lectin
asparagus pea agglutinin (APA). The authors reported a
strongly basic fucoprotein, that was sensitive to pronase
digestion. Based on the reactivity with the lectin, they
calculated that the protein was three times more
concentrated in tumor chromatin than in normal liver cells.
However, the method for chromatin purification did not
eliminate adequately nuclear inner membrane contamination,
and since binding to the lectin was assayed by
affinoelectrophoresis a molecular weight for the protein was
not determined (Goldberg et al., 1978).


82
have reported the existence of pronase-resistant material in
the particulate fraction of vegetative cells, so it seems
possible that the pronase-resistant material eluting in the
void volume is, or is related to, the smear previously
described, although SDS-PAGE analysis will be needed to
confirm this supposition.
Fucose is Covalently Bound to FP21
The fact that the radioactivity in FP21 was not
released by boiling in SDS/ fl-mercaptoethanol, nor after
boiling in SDS under reducing conditions followed by SDS-
PAGE, suggested that 3H was covalently bound to in vivo
labelled FP21. Nevertheless, there have been reports of
covalent-bound enzymatic intermediates in the literature
(Scrimgeour, 1977). Although ES (enzyme-substrate)
intermediates are very reactive and usually cannot be
isolated without some sort of stabilization or chemical
trapping technique, the possibility that FP21 is really a
cytosolic fucosyltransferase that binds GDP-fucose or other
fucose metabolites covalently was considered. The Rev of
the radioactivity released by mild alkaline hydrolysis
suggests that it is not related to GDP-fucose or fucose
because it elutes with a different Rev than GDP-fucose or
fucose. Additional evidence that 3H is present as fucose
was presented in Chapter II, where it was shown that more


85
P100 (DPM)


127
Cytosolic Fucosyltransferase Preferentially Fucosylated a
Type I Acceptor
The size of the fucose-containing oligosaccharide in
FP21 (determined in Chapter III to be 4.8 glucose units)
implies there is more than one sugar residue, so the
acceptor site on FP21 may be another sugar. As a first step
to determine whether the cytosolic fucosyltransferase could
fucosylate model acceptor analogs utilized by known
fucosyltransferases, the S100 from Ax3 and mutant origin
were screened for activity towards hydrophobic model
acceptors. The incorporation of radioactivity into
synthetic sugar acceptors that contained 8-methoxy
carbonyloctyl, or methyl nonanoate, [CH3(CH2)7COOCH3,
referred to as R throughout the text] as the hydrophobic
tail by the S100 was determined by the C18 Sep-Pak method,
which employs a hydrophobic interaction column. Unreacted
GDP-fucose does not interact with the column, eluting in the
water wash while the glycolipid acceptor is eluted from the
column with methanol. As shown in table 4-5, only the type
I acceptor analog (known as lacto-N-biose I or
gal/11,3GlcNAc/l-R) sustained activity in the Ax3 S100. In
contrast, type II (known as N-acetyllactosamine or gal/11,4-
GlcNAc-R) and /1-gal-R were not suitable acceptors (see
figure 4-8 for structures). The mutant S100 was also active
with gal/11,3GlcNAc/l-R, but only one tenth as active (on a
per protein basis) as Ax3.


40 60 80 1 00 -40 -20 0 20 40 60
GDP-fucose (pM) GDP-fucose (|iM)
pmol fuc/mg protein/30 min
pmol fuc/mg protein/30 min
II


Figure 2-1. Localization of SP96 in prespore and spore cells.
Slugs cells were dissociated, placed onto polylysine-coated
glass slides, fixed, permeabilized, and processed for
indirect immunofluorescence using mAbs 83.5 or A6.2.


10
factors and, at least in one case, glycosylation may have
influenced transcriptional activity (Lichtsteiner and
Schibler, 1989; Jackson and Tjian, 1989). As mentioned
earlier, in the case of fucosylated macromolecules, it was
found that in Novikoff hepatoma cells chromatin had three
times the amount of a fucoprotein of normal liver cells
based on APA binding (Goldberg et al., 1978). However, the
identity, size, or number of such proteins were not well
documented. Putative fucoproteins were found preferentially
associated with euchromatin (Kan and Pinto da Silva, 1986).
Fucosylated histones were found in the macronucleus of
Tetrahymena, where transcriptionally active chromatin is
compartmentalized (Levy-Wilson, 1983). Nevertheless,
histones from non-active chromatin were not studied, so no
comparisons can be made with heterochromatin. Levy-Wilson
(1983) suggested that the reason why other investigators
have failed to detect fucosylation in mammalian histones is
due to the low proportion of transcriptionally active genome
which would imply a low concentration of fucosylated
histones.
Although the data gathered in these reports are
interesting, they are the result of isolated studies from
diverse organisms and it is difficult to draw generalized
conclusions. In addition, virtually nothing is known about
the structure or biosynthesis of the fucose-containing
moieties. Even though the fractions in all these reports


140
different specificities. This may be the case in
Dictyostelium because, even though fucosyltransferases have
not been well characterized in this organism, various
fucosyltransferases have been localized to microsomes in
other eukaryotes (Hirschberg and Snider, 1987; Kornfeld and
Kornfeld, 1985). The fact that I was able to differentiate
the bulk activity in the P100 from the cytosolic
fucosyltransferase supported the idea that the
fucosyltransferase in the S100 is unrelated to the P100
activity and thus endogenous to the cytosol.
However, my observations do not rule out other
possibilities. For example, the cytosolic
fucosyltransferase could have derived from vesicles but was
preferentially lost during isolation and the remaining
enzyme, though with distinct properties from the majority of
the P100 activity, is in the minority. The inability of
EDTA to inhibit activity further when compared to no
addition of divalent cations to the P100, may mean that the
enzyme does not need cations at all. Conversely, since the
bulk activity in the P100 is stimulated by cations, it is
possible that the activity retains tightly bound cations
which EDTA cannot remove. Another possibility is that the
acceptor, FP21, is not present in the P100, either due to a
cytosolic compartmentalization, or to leakage from the
vesicles. A definitive confirmation that the cytosolic
fucosyltransferase is different from any fucosyltransferase


Table 4-6. Evaluation of the suitability of p-nitro-phenyl glycosides as acceptors for
cytosolic fucosyltransferase activity in HL250 and Ax3 S100.
treatment concentration
control
p-nitro-0-cr-D-glucose
p-nitro-0-fi-D-glucose
p-nitro-0-a-D-GlcNAc
p-nitro-0-i-D-GlcNAc
p-nitro-0-fi-D-galactose
p-nitro-0-fi-D-glc
tetraacetate
DPM HL250
DPM HL250
DPM Ax3
(TCA)
(SepPak C18)
(SepPak i
1003
<3
<3
996
21
4
1020
19
n.d.
1066
26
10
1032
31
18
1059
38
3
1134
24
10
10
mM
10
mM
10
mM
10
mM
10
mM
10
mM
CO
ho
S100 fractions from Ax3 and HL250 were assayed for fucosyltransferase activity and
fucose incorporation determined by TCA-precipitation or SepPak C method (see Materials
and Methods fordetaiis). n=l; n.d., not determined; 0, phenyl; 0.89 pM GDP-[ 14C ] f ucose;
0.1% Tween-20; 619 pg protein; 60 min assay.


52
down and resuspended in water to a volume of 200 pi
containing at least 5xl03 dpm. In other experiments, 200 pi
of the entire P100 fraction were left intact or made 0.1%
Triton X-100. 200 pi of freshly dissolved 1% pronase
(CalBiochem) in 0.1 M Tris-HCl (pH 8.0), 1 mM CaCl2, were
added at 0, 24, and 48 hours. Incubation was at 50 and a
few drops of toluene were added to prevent microbial growth.
At 72 hours the reaction was stopped by incubating in a
boiling water bath for 3 min.
Oligosaccharide Release by Alkaline-Borohydride Treatment
The oligosaccharide in FP21 was released by mild
alkaline hydrolysis under reducing conditions, also known as
fl-elimination. To approximately 5xl03 dpm of gel purified
protein, freshly made NaOH and NaBH4 concentrated solutions
were added in that order to yield a final concentration of
0.1 M and 1 M, respectively. Samples were incubated for 15
hours in a water bath at 45. The reaction was stopped by
the addition of acetic acid to a final concentration of 1 M.
The samples were dried down by vacuum centrifugation,
resuspended once in 1 ml 100 mM HAc, dried again, and
resuspended twice in methanol and stored dry at -80 until
ready to use.
The oligosaccharide was also released by strong
alkaline-borohydride treatment. The method is that of Zinn
et al. (1978) and very similar to the procedure for B-


-0.09 -0.00 0.09 0.18 0.27
1/[S]
1 /V
nmol fucose/mg protein/30 min
nmol fucose/30 min


24
NaOH. Quantitation of the conversion of GDP-mannose to GDP-
fucose was achieved by determining the amount of fucose
present after acid hydrolysis. Free mannose, released from
GDP-mannose, was phosphorylated by adding 4 units of
hexokinase in the presence of 5 mM ATP and 5 mM MgCl2 at 37
for 1 hour. Hexokinase catalyzes the transfer of one
phosphate to C-6 of any acceptor hexose. Since fucose lacks
a hydroxyl group at position C-6, it cannot be
phosphorylated. Parallel controls with no enzyme were used
to correct for losses. The mixture was passed over a Dowex-
1 (formate) column (0.6 x 5 cm) and eluted with water. 1 ml
fractions were collected an aliquot of 100 pi from the
eluate was counted using ScintiVerse LC (Fisher); fucose-
associated radioactivity usually eluted by the first 2 ml.
Calculation of K and V was done by the Lineweaver-Burk
double reciprocal plot (1/v vs. 1/[S]) as discussed by
Henderson (1985).
Results
Phenotypic Description of a Fucosylation Mutant
The normal strain Ax3 was mutagenized with
nitrosoguanidine and surviving clones were screened with
anti-SP96 antiserum (Loomis, 1987). One of the clones,
HL250, was selected due to its inability to bind anti-SP96
antiserum which recognizes carbohydrate and peptide
epitopes. HL250 has been found to have a more permeable


95
1982). After 2 hours of incubation less than 20% of the
initial GDP-[ 14C ] fucose had been hydrolyzed. Incorporation
of 14C did not exceed 30% of the initial radioactivity on
any given experiment. For analysis by gel electrophoresis
of the endogenous acceptors of the in vitro
fucosyltransferase activity, the reaction was terminated by
3 min boiling in sample buffer. For the determination of pH
optima experiments, assays were buffered using concentrated
solutions of MES previously adjusted to different pH values
with either HC1 or NaOH. The final pH value of each
reaction was determined on an equivalent reaction mixture
lOOx the volume, without GDP-[ 14C]fucose. Where indicated,
the S100 was desalted on a 0.85 x 13 cm on a BioRad BioGel
P-2 column (200-400 mesh) equilibrated with 50 mM MES, pH
7.4, titrated with NaOH, at 4 with a flow rate of 0.5
ml/min. If supplied, synthetic acceptors and/or FP21 were
previously dried down onto the bottom of the assay tubes in
a vacuum centrifuge. Acceptors were subsequently
resuspended in water or in the reaction mixture. Results
are expressed as average of two determinations (variations
in the duplicates did not exceed 15% of the average value)
or average of three measurements + standard error of the
mean (s.e.m.). Calculations of K and V were done by the
Hanes single reciprocal plot ([S]/v vs. [S]) as discussed by
Henderson (1985).


151
times, nor if its preferentially expressed in any cell type
during development, have been determined. Ideally, it would
be useful to produce antibodies with specificity for the
glycosylated protein, and specificity for the peptide moiety
of FP21 (similar to other mAb produced in the laboratory;
see West et al., 1986). Using these antibodies,
fucosylation of FP21 during development could be followed by
immunoprecipitation of FP21.
Studies on the Cytosolic Fucosyltransferase
Another aspect of my project was the evidence presented
for a novel fucosyltransferase that appears to be cytosolic
and seems to differ from the bulk sedimentable
fucosyltransferase activity. The first question to be
addressed will be the compartmentalization of the enzyme.
The cytosolic enzyme could be purified by conventional
methods (Beyer et al., 1980; Foster et al., 1991; Martin et
al., 1987). Once purified, antibodies could be raised
against the enzyme and used for immunolocalization of the
fucosyltransferase. Currently, a purification protocol is
being developed in the laboratory. If localization of the
enzyme is done by immunofluorescence and the enzyme is a
soluble cytosolic protein, it should be possible to observe
a cytosolic distribution of the enzyme and an absence from
intracellular vesicles. However, if the immunofluorescence
pattern shows labelling of vesicles, the results will need


147
to be an al,4fucosyltransferase lacking ctl,3 activity. 5)
Fucosylation of the type I acceptor analog (galfil,3GlcNAcfi-
8-methoxycarbonyloctyl) was inhibited by addition of
purified FP21, suggesting the same activity was responsible
for fucosylation of both molecules.
Based on the results obtained in my studies, I propose
a model for fucosylation in Dictyostelium, acknowledging the
existence of a fucosyltransferase in the cytosol that
fucosylates a cytosolic protein, FP21. There are
fucosyltransferases in vesicles and in the cytosol; the
preferential acceptor for the cytosolic fucosyltransferase
is FP21. This model is appealing because all of the
elements necessary for fucosylation, biosynthesis of GDP-
fucose, a fucosyltransferase, and the acceptor,
compartmentalize in the cytosol. The model also concurs
with emerging views of glycosylation in the cytosol (Hart et
al., 1989a; Hart et al., 1989b). Initially I showed that
Dictyostelium possesses a GDP-fucose conversion pathway
similar to that reported earlier for other organisms
(Kornfeld and Ginsburg, 1966; Liao and Barber, 1971; Ripka
et al., 1986). It was shown that Dictyostelium can convert
GDP-mannose into GDP-fucose, and that when this biosynthetic
pathway is defective, GDP-fucose is formed from fucose
supplied in the extracellular medium by the salvage pathway.
This is the first time evidence has been presented that
suggests Dictyostelium has GDP-fucose biosynthetic pathways


DPM DPM
63


DPM
74


131
8-methoxycarbonyloctyl synthetic acceptors are not
available commercially, so I investigated the possibility
that other hydrophobic glycosides, which can be readily
obtained from commercial sources, would serve as acceptors.
Some of these phenyl derivatives have been shown by others
to be suitable acceptors for a fucosyltransferase (Potvin,
et al., 1990; Palcic et al., 1988). S100 fractions from
HL250 and Ax3 were assayed for fucosyltransferase activity
in the presence of p-nitro-phenyl glycoside derivatives;
fucosylation of endogenous substrates was monitored by TCA-
precipitation, and fucosylation of p-nitro-phenyl glycosides
by the Sep-Pak method (table 4-6). Inhibition of
incorporation of radioactivity into FP21 in HL250 was
examined by TCA precipitation of endogenous acceptors.
Millimolar concentrations of these compounds failed to
inhibit significantly incorporation. Likewise, none of the
p-nitro-phenyl glycosides served as acceptor for the mutant
nor the Ax3 cytosolic fucosyltransferase activity when
assayed by the C Sep-Pak method. Thus it is concluded,
that from the acceptor candidates examined, only
galfil,3GlcNAcfi-R is a suitable acceptor under the conditions
used.
The activity responsible for fucosylation of the type I
acceptor analog in Ax3 was examined by varying the
concentration of acceptor or the concentration of nucleotide
sugar donor. The type I acceptor analog was fucosylated


143
described, which exhibit al,3 activity as well (Kukowska-
Latallo et al, 1990; Stroup et al., 1990). However, there
are some limitations to the studies employing synthetic
acceptors. To conserve synthetic acceptors, which were not
commercially available, the concentration of the acceptors
was well below the Km (0.145 mM, while the apparent Kn was
determined to be approximately 1 mM). The possibility still
exists that the enzyme is able to use 8-methoxycarbonyloctyl
galiil,4GlcNAcfi as acceptor, but will only be evident at
higher concentrations. Tentatively, an al,4 specificity is
being assigned to the cytosolic fucosyltransferase, but
definitive proof will require characterization of the enzyme
purified to homogeneity.
Finally, slug stage extracts were examined for
fucosyltransferase activity, because it was found by
metabolic labelling experiments in Chapter III that a
fucosylated protein of 21 kD fractionated with the S100.
The S100 and P100 fractions from HL250 had considerable
activity, but from the Ax3 fractions only the P100 showed
activity, consistent with the results from vegetative cells.
The detection of a cytosolic fucosyltransferase in slug-
stage cell extracts is consistent with their ability to
fucosylate FP21 in vivo as determined by metabolic
labelling. The apparent absence of activity in Ax3 cells
indicated that, as found for vegetative stage cells, FP21
was quantitatively fucosylated.


Figure 2-2. Biosynthesis of GDP-fucose.
Diagrams of the conversion pathway elucidated in bacteria
and mammalian cells (adapted from Flowers, 1981) and the
salvage pathway as described in HeLa cells (adapted from
Yurchenko et al., 1978).


Figure 3-8. Incorporation of [3H]fucose into macromolecular
species of slug stage cells.
HL250 cells were harvested, plated on filters for
development, and exposed to [3H]fucose for 5 h as described
in the text. Cells were then collected, disaggregated,
filter-lysed, and S100 and P100 fractions were prepared and
analyzed by 7-20% linear gradient SDS-PAGE. 70 jug and 24.6
fjq of protein were electrophoresed for the S100 and P100,
respectively. Open circles, P100; closed circles, S100;
arrow, migration position of trypsin.


41
the mutant HL250 may have a defect in (1) fucosyltransferase
activities, (2) endogenous acceptors for
fucosyltransferases, (3) transport of GDP-fucose into
microsomal vesicles, and/or (4) synthesis of GDP-fucose. In
vitro microsomal extracts of normal and mutant cells were
active in the transfer of [14C]fucose from GDP-[ 14C] fucose to
endogenous acceptors and the activity was latent (see
Chapter IV), so it was reasoned that fucosyltransferases may
be normal and probably the uptake of GDP-fucose by vesicles
was not impaired, so the lesion might be at another point in
the fucosylation pathway. Since fucose is normally added as
a terminal modification, the fucose minus phenotype could be
the result of a lack of formation of acceptors for the
fucosyltransferases (Stanley, 1984; Hirschberg et al.,
1982). HL250 expresses levels comparable with Ax3 of other
carbohydrate epitopes and has normal neutral monosaccharide
composition (Gonzlez-Yanes et al., 1989; West et al.,
1986), for these reasons it was speculated that the defect
was not in an earlier step of glycosylation but it involved
the fucosylation pathway directly.
The glycosylation defect in HL250 appears to result
from an inability to produce GDP-fucose. I have found that
the GDP-mannose to GDP-fucose conversion activity in vitro
is reduced to undetectable levels in mutant cell extracts.
The fact that there is partial rescue when the cells are
grown in the presence of fucose, suggests that the cells are


Figure 3-6. Gel filtration chromatography of FP21
oligosaccharides.
Vegetative cells were metabolically labelled with 2 ^Ci/ml
of [3H]fucose, an S100 was prepared, and FP21 was isolated
by SDS-PAGE and electroelution. Gel purified FP21 from Ax3
and HL250 were subjected to fi-elimination or strong alkaline
hydrolysis followed by fractionation by gel filtration
chromatography. Data obtained from one representative
experiment.
Panel A. ft-elimination of Ax3 FP21. Vo, 36; Vi, 141.
Panel
B.
fi-elimination of HL250
FP21. Vo,
30;
<
p.
140.
Panel
C.
Alkaline hydrolysis of
Ax3 FP21.
o
>
38;
Vi,


CHAPTER II
CHARACTERIZATION OF A FUCOSYLATION MUTANT
Introduction
Since fucosylation comprises numerous steps which are
potential sites for mutations, many fucosylation mutants
have been obtained. Fucosylation consists of the synthesis
of the sugar nucleotide donor GDP-fucose and the transfer of
fucose from the GDP-fucose to an acceptor (Kornfeld and
Kornfeld, 1985). The main source of GDP-fucose is the
conversion pathway of GDP-mannose to GDP-fucose (Yurchenco
et al., 1978; Flowers, 1985). Alternatively, synthesis of
GDP-fucose by the fucose salvage pathway can occur in the
presence of extracellular L-fucose, allowing cells defective
in the conversion pathway to phenotypically revert (Ripka
and Stanley, 1986; Reitman et al., 1980). Lesions may
affect the formation of GDP-fucose, the transport of GDP-
fucose to the fucosylation compartment, the transferases
responsible for fucosylation and/or the biosynthesis or
transport of the acceptor species to the fucosylation
compartment. Two mutants deficient in protein-associated
fucose were shown to be defective in the formation of GDP-
fucose (Reitman et al., 1980; Ripka et al., 1986). A
Chinese hamster ovary (CHO) cell line that showed a marked
18


118
for the higher activity of the S100 fraction in the presence
of detergent at the concentration of GDP-fucose used in most
assays (0.36 pM), despite the higher Vmay of the P100.
Since the above results were obtained in the presence
of detergent, I investigated the effect of GDP-fucose
concentration on the fucosyltransferase activities in the
intact fractions. It was reasoned that, in the case of the
P100, it would give us some insight into the overall
fucosylation process, including uptake of GDP-fucose into
the vesicles. S100 and P100 fractions were prepared from
HL250 vegetative cells and assayed for fucosyltransferase
activity under standard conditions at increasing
concentrations of GDP-fucose (figure 4-6). An apparent Km
of 0.44 pM and Vmax of 25.5 pmol fuc/ mg protein/ 30 min was
calculated for the S100. For the P100 an apparent Kn of
28.3 pM and Vmax of 233 pmol fuc/ mg protein/ 30 min was
determined.
The fucosyltransferase activities from the S100 and
P100 differed in the acceptor species that were fucosylated,
the sensitivity to high pH, divalent cation dependence, and
apparent affinity for GDP-fucose. These differences in
enzymatic behavior support a model for separate
compartmentalization of the two fucosyltransferase
activities; the S100 activity is free in the cytosol and the
P100 activity is in a membrane bound organelle.


7
In a more recent study on duodenal columnar cells, Kan
and Pinto da Silva (1986) used UEA-I conjugated to colloidal
gold in freeze-fracture electron microscopy of cross-
fractured nuclei. They compared binding of the conjugated
lectin to euchromatin, heterochromatin, and nucleolus;
compartments which can be differentiated ultrastructurally.
Binding of UEA-I showed that colloidal gold particles were
almost exclusively confined to cross-fractured areas where
euchromatin was exposed. Labelling was abolished by
pretreatment and incubation in the presence of L-fucose, as
expected for a specific label. Pre-digestion of the
fractions with trypsin also abolished labelling, suggesting
the receptors for UEA-I binding were glycoproteins. The
preferential binding to euchromatin may be of importance,
because replication and transcription take place at
euchromatin regions. Although the results reported are
intriguing, the authors did not identify the type of fucose-
containing proteins detected. DNA-associated proteins can
be classified as either histone or non-histone proteins, and
as summarized below, both types of proteins appear to be
fucosylated.
Histones. The histones are the most abundant proteins
associated with DNA. Histones are very basic proteins and
are found in all nuclei (Darnell et al., 1986). Based on
the specific binding to UEA-I, Levy-Wilson (1983) presented


13
Fucose-Bindinq Proteins
The presence of fucoproteins in the nucleus and cytosol
prompted the idea that there might be specific proteins
inside the cell that bind, and/or modify these fucoproteins,
as is the case with the fucoproteins in the secretory
pathway. This would include fucose-binding proteins,
fucosyltransferases responsible for fucosylation, and
fucosidases responsible for fucose removal. There are some
studies suggesting the existence of fucose-binding lectins
and fucosyltransferases. However, there is no evidence for
a nuclear or cytosolic fucosidase.
Endogenous Lectins
In light of evidence for glycoproteins in the cytosolic
and nuclear compartments, investigators sought the existence
of carbohydrate-binding proteins that would colocalize with
such glycoproteins. Endogenous lectins have been detected
in preparations of nucleoplasmic and/or cytosolic fractions
of a variety of cells (Hart et al., 1989a).
Aided by fluorescein-labeled neoglycoproteins, Sve et
al. (1986) have postulated the presence of endogenous
fucose-specific lectins. Baby hamster kidney cell nuclei
were isolated by two different procedures, cell lysis and
enucleation in Ficoll, to argue against contamination by
cytoplasmic or membrane-derived components. Using
fluorescein-labelled BSA conjugated to fucose in the order


35
not contain an inhibitor for the activity. Conversion was
linear with respect to protein through 800 pg (figure 2-3,
panel A). In agreement with the results from the time
dependence experiment, HL250 expressed less than 1% of the
activity possessed by Ax3 at all concentrations of protein
assayed (figure 2-3, panel A). Conversion by Ax3 S100 was
also dependent on the amount of GDP-mannose present while
the activity present in HL250 was insignificant (figure 2-3,
panel B). The Ax3 GDP-mannose to GDP-fucose conversion
activity showed an apparent Km of 14.1 pM and Vnwy of 18.3
nmol fucose/mg protein/30 min (fig. 2-3, panel C). Previous
reported values for the apparent Km of the conversion
pathway range from 2 pM in CHO cells (Ripka et al., 1986) to
160 pM in the higher plant Phaseolus vulgaris (Liao and
Barber, 1971).
The conversion activity is absent in mutant extracts in
vitro, if absent in vivo, this would explain the lack of
fucose in living cells and the correction by exogenous
fucose. Taken together, all these results point to the
conversion pathway as the site of the lesion in HL250.
Furthermore, the defect seems to be in the first step of the
conversion of GDP-mannose to GDP-fucose because no
radioactivity was recovered after hexokinase treatment. If
the 4,6-dehydratase was active, the product would have
eluted from the ion-exchange column after hydrolysis and
phosphorylation by hexokinase (see figure 2-2).


87
in the cytosol of developing cells, suggesting FP21 was not
limited to the vegetative stage in Dictyostelium, and was
fucosylated during development.
I believe that FP21 is recovered in the S100 because it
resides in the cytosol in living cells, and not as default
location from ruptured vesicles, for several reasons. The
S100 fraction was shown to be equivalent to the cytosol and
essentially devoid of organellar markers. S100 and P100
fractions from [3H]fucose metabolically labelled cells
exhibited a different radioactive profile by SDS-PAGE.
Sonication of the P100 fraction failed to release FP21 into
the supernatant. FP21 appeared to be endogenous to the
cytosol and not derived from organellar vesicles, because
control experiments suggested there was no generalized
breakage of vesicles during the preparation of the cytosolic
fraction. In addition, as a control for contamination from
P100 material, glycopeptides derived from comigrating 21 kD
MW species from the P100 fraction were compared with FP21
glycopeptides. The fucopeptide in FP21 (ca. 5.5 glucose
unit) does not seem to be a product of vesicular
fucosylation, since it is not shared by macromolecules of 21
kD MW in the P100 fraction which yielded a major peak with a
different Rev (ca. 4.3 glucose unit) than the one derived
from FP21. When the entire P100 fraction was subjected to
Pronase digestion, a heterogeneous mixture of fucosylated


78
oligosaccharides (Zinn et al., 1978; Biermann, 1988). The
results of this reaction are seen in figure 3-6, panel C.
More than 75% of the radioactivity was released, eluting
with a Rev of 0.49. Most of the remainder of the
radioactivity eluted in the void volume. I expected to see
a change in Rev by strong alkaline hydrolysis compared to
mild alkaline hydrolysis if mild alkaline hydrolysis did not
release the oligosaccharide, but hydrolysed the protein.
The fact that similar results are obtained by mild and
strong conditions suggested that fi-elimination occurred to
release the oligosaccharide. Since mild alkaline hydrolysis
had been shown earlier not to cleave N-linked sugars
(Biermann, 1988), I conclude that the oligosaccharide in
FP21 is linked via an O-linkage. The released
oligosaccharide eluted as an asymmetrical peak, both by mild
and strong alkaline hydrolysis. These results suggest that
there is more than one type of oligosaccharide in FP21 that
differ slightly in size. These results also suggest that
the slight heterogeneity seen in the glycopeptide size may
reflect amino acid heterogeneity.
Comparison of qlycopeptides derived from vesicular and
cytosolic material. To further address the possibility of
FP21 arising by contamination of the S100 from the P100
fraction, glycopeptides derived from the Ax3 P100 were
compared to gel purified pronase digested FP21 from Ax3.
P100 derived glycopeptides were obtained in three different


2
apparatus utilizing GDP-fucose as the sugar nucleotide donor
(Bennett et al., 1974; Hirschberg and Snider, 1987).
However, there are a few exceptions to these remarks, since
fucose has been found to be attached directly to serine and
threonine, although the site for this modification has not
been identified (Hallgreen et al., 1975), and to be present
in homopolymers, as in fucoidans (Flowers, 1981). It has
been suggested that fucosylation also occurs in the
endoplasmic reticulum as well as in the Golgi apparatus in
thyrotrophs under different physiological conditions (Magner
et al., 1986). There are two pathways for the biosynthesis
of GDP-fucose. The main source of GDP-fucose is the
conversion pathway of GDP-mannose to GDP-fucose (Yurchenco
et al., 1978; Flowers, 1985). Alternatively, synthesis of
GDP-fucose may occur by the fucose salvage pathway in the
presence of extracellular L-fucose (Yurchenco et al., 1978;
Flowers, 1985; Ripka and Stanley, 1986; Reitman et al.,
1980).
Glycosylation is carried out by diverse specific
glycosyltransferases which have been located in the
endoplasmic reticulum and Golgi apparatus. One underlying
assumption is that the acceptor macromolecule must
colocalize with the transferase enzyme in order to be
glycosylated. However, there have been occasional reports
of glycoproteins in compartments topologically discontinuous
with the lumen of the rough endoplasmic reticulum and Golgi


Abstract of Dissertation Presented to the Graduate School
of the University Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
A NOVEL FUCOSYLATION PATHWAY IN THE CYTOSOL
OF DICTYOSTELIUM DISCOIDEUM
By
Beatriz Gonzlez-Yanes
December 1991
Chairman: Dr. Christopher M. West
Major Department: Anatomy and Cell Biology
The existence of a fucosylation pathway in the cytosol
of Dictyostelium discoideum was investigated in the
glycosylation mutant strain, HL250, and the normal parental
strain, Ax3. HL250 was characterized as a conditional
mutant that cannot convert GDP(guanosine-5'-diphosphate)-
mannose to GDP-fucose, resulting in a lack of macromolecular
fucosylation unless grown in the presence of extracellular
fucose. HL250 or Ax3 cells were metabolically labelled with
[3H]fucose, filter-lysed, and fractionated by high speed
centrifugation into sedimentable (P100) and soluble (S100)
fractions. The fractions exhibited unique profiles of
fucoconjugates as analyzed by gel electrophoresis. The
major acceptor in the S100 was a 21 kilodalton molecular
weight protein, FP21. Analysis of FP21 oligosaccharide by
mild alkaline hydrolysis and gel filtration chromatography
xiv


91
for the addition of O-GlcNAc was in the same topological
compartment where translation takes place, the cytosol.
Thus it is possible that fucosylation may take place in the
cytosol, but it has escaped detection by previous
investigators for a variety of reasons. One of the
difficulties in assaying cytosolic enzymes is that the
endogenous acceptors for the enzymes may be present in low
quantities in the cell, complicating purification of large
amounts for use as substrates. If acceptors are already
fucosylated, in vitro assays that utilize endogenous
acceptors would not detect enzymatic activity. Hart and
coworkers have circumvented this problem with the use of
synthetic peptides with a sequence based on O-GlcNAc
glycosylation sites (Hart et al., 1989b). They have
identified an enzymatic activity capable of O-linked GlcNAc
transfer in rat hepatocytes that was recovered in both the
soluble and membrane fractions (Haltiwanger et al., 1990).
The membrane-associated activity was releasable by high salt
treatment and oriented towards the cytosol, not the lumen of
the vesicles (Haltiwanger et al., 1990).
With the help of the conditional fucosylation mutant
HL250, I have addressed the existence of a fucosylation
pathway in the cytosol. Total protein in HL250 is
underfucosylated relative to the normal strain, so it was
reasoned that it would be a useful strain to assay
fucosylation in vitro due to the availability of


86
the high speed supernatant, S100, being the major
fucosylated species in this fraction. The recoverability of
FP21 after TCA precipitation, HPLC anion exchange
chromatography, boiling in SDS/fi-mercaptoethanol, and
methanol/acetic acid fixation of the gels suggested a
covalent nature for the association of radioactivity. It is
unlikely that FP21 is a cytosolic fucosyltransferase that
binds GDP-[3H]fucose or another fucose metabolite
covalently, since FP21 does not copurify with the cytosolic
fucosyltransferase (C.M. West, unpublished results).
Employing enzymatic and chemical analysis, the
oligosaccharide in FP21 was examined and characterized as a
small (4.8 glucose unit) oligosaccharide. The
oligosaccharide appeared to be O-linked based on its
insensitivity to PNGase F and the releasability from FP21
under alkaline, reducing conditions. Ax3 and HL250 produced
FP21-derived glycopeptides and oligosaccharides of similar
size, suggesting both strains carry out similar
modifications in vivo, despite the starvation for fucose in
HL250. Thus, it appears there are no competing reactions
for the unfucosylated oligosaccharide, unlike the case for
outer fucose in N-linked glycans from mammals (Paulson et
al., 1978). Based on antiserum specificity, molecular
weight, and/or fractionation by HPLC anion exchange
chromatography, FP21 was shown to be a protein unrelated to
discoidin or gp24. Finally, a 21 kD fucoprotein was present


55
a ratio of almost 1:1 (P100:S100), the specific activity
(expressed as dpm/mg of protein) distributed roughly in a
6:1 ratio (P100:S100) (table 3-1). The majority of the
radioactivity fractionated with the P100. However, the
amounts of radioactivity recovered in the S100 were
unexpected; 36% of the radioactivity in the S100 (this value
varied in individual experiments from 30-68%) migrated as
one peak at the 21 kD position, slightly ahead of trypsin,
which was used as a molecular weight marker. On the other
hand, the P100 showed two main broad peaks, one at 10-25 kD
and another at 58-84 kD with 10% of the total radioactivity
in the P100 migrating at the 21 kD level (figure 3-1). The
proteinaceous nature of the S100 fucoprotein was confirmed
by digestion with the proteases trypsin and pronase with
quantitative recovery of radioactivity at lower MW positions
in unfixed gels (results from pronase digestion shown in
figure 3-2). The S100 fucosylated protein has been called
FP21 for fucoprotein of 21 kD molecular weight.
The abundance of FP21 was estimated based on the
specific activity of fucose and determined to be 103
molecules in FP21, assuming one fucose molecule per molecule
of FP21. Based on the dilution of fucose specific activity
in Ax3 (from table 2-3, Chapter II), there would be 4 x 105
molecules in Ax3. If there is one fucose per copy of FP21,
and all FP21 molecules are fucosylated (evidence for
quantitative FP21 fucosylation in Ax3 will be presented in